Timetree of life

The Timetree of Life This page intentionally left blank The Timetree of Life Edited by S. Blair Hedges and Sudhir K...

Author: S. Blair Hedges | Sudhir Kumar

189 downloads 1221 Views 8MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

The Timetree of Life

This page intentionally left blank

The Timetree of Life Edited by S. Blair Hedges and Sudhir Kumar Foreword by James D. Watson

1

3

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © S. Blair Hedges and Sudhir Kumar 2009 The moral rights of the authors have been asserted Database right Oxford University Press (maker) First published 2009 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by Newgen Imaging Systems (P) Ltd., Chennai, India Printed in China on acid-free paper by Asia Pacific Offset Limited ISBN 978–0–19–953503–3 (Hbk.) 10 9 8 7 6 5 4 3 2 1

Contents Foreword Preface List of Contributors

I. INTRODUCTION Discovering the timetree of life Timetrees: beyond cladograms, phenograms, and phylograms The geologic time scale Calibrating and constraining molecular clocks

II. TIMETREES Life

ix xi xv

1 3 19 26 35

87 89

Superkingdoms Archaebacteria Eubacteria Eukaryotes (Eukaryota)

99 101 106 116

Protists Haptophyte algae (Haptophyta) Diatoms (Bacillariophyta)

121 123 127

Plants

131 133 138 146 153 157 161

Land plants (Embryophyta) Mosses (Bryophyta) Liverworts (Marchantiophyta) Ferns Gymnosperms Flowering plants (Magnoliophyta)

vi

CONTENTS

Magnoliids Eudicots Asterids Eurosid I Eurosid II Monocots Fungi Fungi

166 169 177 188 197 203 213 215

Animals Animals (Metazoa)

221 223

Invertebrates Cnidarians (Cnidaria) Scaphopod mollusks (Scaphopoda) Cephalopod mollusks (Cephalopoda) Nematodes (Nematoda) Arthropods (Arthropoda) Spiders (Araneae) Holometabolous insects (Holometabola) Bees, ants, and stinging wasps (Aculeata) True flies (Diptera) Beetles (Coleoptera) Lacewings (Neuroptera) Crabs, shrimps, and lobsters (Decapoda) Stalked and acorn barnacles (Thoracica) Sea urchins (Echinoidea)

231 233 239 242 247 251 255 260 264 270 278 290 293 298 302

Vertebrates Vertebrates (Vertebrata)

307 309

Fishes

315 317 320 328 332 335 339 344 348

Jawless fishes (Cyclostomata) Cartilaginous fishes (Chondrichthyes) Ray-finned fishes (Actinopterygii) Sturgeons and paddlefishes (Acipenseriformes) Teleost fishes (Teleostei) Notothenioid fishes (Notothenioidei) Labyrinth fishes (Anabantoidei) Lungfishes (Dipnoi)

Contents

vii

Amphibians Amphibians (Lissamphibia) Frogs and toads (Anura) Salamanders (Caudata) Caecilians (Gymnophiona)

351 353 357 365 369

Amniotes

373 375

Amniotes (Amniota) Reptiles Lizards, snakes, and amphisbaenians (Squamata) Snakes (Serpentes) Turtles (Testudines) Crocodylians (Crocodylia) Birds Birds (Aves) Ratites and tinamous (Paleognathae) Waterfowl and gamefowl (Galloanserae) Advanced birds (Neoaves) Passerine birds (Passeriformes) Shorebirds (Charadriiformes) Diurnal birds of prey (Falconiformes) Cranes, rails, and allies (Gruiformes) Woodpeckers, toucans, barbets, and allies (Piciformes) Owls (Strigiformes) Swifts, treeswifts, and hummingbirds (Apodiformes) Mammals Mammals (Mammalia) Monotremes (Prototheria) Marsupials (Metatheria) Placental mammals (Eutheria) Armadillos, anteaters, and sloths (Xenarthra) Tenrecs and golden moles (Afrosoricida) Primates (Primates) Pikas, hares, and rabbits (Lagomorpha) Rodents (Rodentia)

381 383 390 398 402 407 409 412 415 419 423 432 436 440 445 451 454 457 459 462 466 471 475 479 482 487 490

viii

CONTENTS

Hedgehogs, shrews, moles, and solenodons (Eulipotyphla) Bats (Chiroptera) Carnivores (Carnivora) Rhinoceroses, tapirs, and horses (Perissodactyla) Whales and even-toed ungulates (Cetartiodactyla) Index

495 499 504 508 511 517

Foreword I first began to appreciate the variety of living organisms when I went bird watching with my father, in Jackson Park, Chicago. My father had been a devoted bird-watcher since his high school days, and it was not just to escape Mass that I began to join him on Sunday morning visits to the park. I much enjoyed learning to distinguish birds that to the uninitiated were merely differing shades of brown. My appreciation of the diversity of the living world was further enhanced by visits to the wonderful Field Museum of Natural History, named in honor of Marshall Field. (I have reason to like Marshall Field. His grandson, Marshall Field III, played an important role in the history of Cold Spring Harbor Laboratory, hosting a 1932 fund raising gala at his 1750 acre estate on Long Island.) While I knew that evolution was at the bottom of it all, it was still a source of wonder to me that naturalists were able to make sense of, and name, all of these creatures. Then, and for many years after, the naming of species and determining their relationships, was based on observable characters–the number of bones in a skull, the structure of gentitalia, and so on. This began to change in the 1950s with the use of paper, followed by starch and acrylamide gel, electrophoresis to examine proteins. But a key step was taken in 1962, when Emile Zuckerkandl and Linus Pauling introduced the “molecular clock,” bringing together the relationships between organisms and the times of their divergence. By the 1970s, analysis had moved away from proteins and to using DNA hybridization methods which provided genome-wide data. However, these methods were not without problems, and there were some celebrated (or infamous) controversies. It was the development of DNA sequencing in the 1970s, but especially the invention of automated DNA sequencing that heralded a new phase in the study of biological processes. For the first time, it was possible to produce large amounts of DNA sequence and so comparisons could be made between many genes across many organisms. The latest sequencing techniques can generate gigabases of nucleotide sequence, as Craig Venter and I know from seeing the complete sequences of our own genomes. One can only hope that the cost of sequencing will continue to decline so that all biologists can sequence their favorite organism. Now, I look in wonder at The Timetree of Life, at the breadth of life that it covers, and the extraordinary data presented in it. Darwin himself drew trees, most famously the sketch (Fig. 1) that appears in what is known as Notebook B, on “Transmutation of Species” (1837–1838). The branching pattern illustrates how Darwin thought

x

FOREWORD

species arose by descent with modifications from an ancestral species, and explained the relationships between existing and extinct species. How thrilled he would be to know how his insights continue to be the foundation of all that we do in biology. James D. Watson Cold Spring Harbor Laboratory

Fig. 1. A page written by Charles Darwin in a notebook in 1837, at age 28, showing a phylogenetic tree and collecting his early thoughts about the evolution of species. He writes “I think. Case must be that one generation then should be as many living as now. To do this & to have many species in same genus (as is) requires extinction. Thus between A & B immense gap of relation. C & B the finest gradation, B & D rather greater distinction. Thus genera would be formed—bearing relation [next page] to ancient types—with several extinct forms for if each species an ancient (1) is capable of making 13 recent forms, twelve of the contemporarys must have left no offspring at all, so as to keep number of species constant” (page 36 of “Notebook B,” reproduced by kind permission of the Syndics of Cambridge University Library).

Preface This book addresses one of the most basic of biological questions: the Tree of Life and its timescale. Our goal was to bring together experts on all of the major groups of organisms to produce a state-of-the-art synthesis of the molecular timescales of life. At the same time, we wanted to make this information on phylogenetic trees scaled to time—timetrees—accessible to everyone, including students and scientists of all disciplines, to facilitate interdisciplinary research and discovery. The result is essentially an encyclopedia of The Tree of Life. On the one hand it has a uniform style and minimum of jargon to make it useful for nonspecialists, while on the other hand it contains the data and literature references needed by active researchers. Early in the project, we made two decisions to ensure that a single, manageable volume was produced. First, we decided to include only divergence times estimated by molecular clocks. The integration of fossil and molecular timetrees is complex and the number of fossil taxa is quite large. Second, we limited coverage of taxa to the family level and above, because any finer resolution below the family level (e.g., genus or species) would have required multiple volumes. In the future, we look forward to relaxing both of those limitations. Chapters in this book correspond to evolutionary groups (taxa) and as such are arranged in a hierarchy. For example, the chapter on placental mammals covers only the orders whereas chapters on individual orders of placental mammals cover the families. Beyond this, our decisions as to the scope of each chapter were largely based on the published literature on molecular timescales, which is uneven in its taxonomic coverage. For example, there are more chapters on plants and tetrapods than on protists and invertebrates, reflecting the imbalance in the number of studies reporting time estimates pertaining to those groups. Also, if a group (e.g., class or order) was covered in a single published study we were more likely to treat that taxon as a single chapter rather than splitting or merging it with other taxa. We took great efforts to include all evolutionary groups—prokaryotes and eukaryotes—with published times estimates, but additionally needed to include new data and analyses to fill in taxonomic gaps in 13 chapters. The current gaps in coverage are certain to disappear in the near future as more timetrees become available. Each chapter begins by summarizing aspects of the diversity and distribution of the group in question, showing a color image of one or more representatives as the first figure. If fossils exist, highlights of the fossil record are mentioned. The remainder of the chapter then discusses the phylogenetic relationships of the included taxa

xii

PREFACE

and their times of divergence based on molecular data. The time estimates, whether published or new, are presented in a table with measures of statistical confidence. The first two columns of every table pertain to the nodes and times in the timetree, which is the second figure of each chapter. We asked the authors to be precise in discussing divergence times, referring to the split of two lineages rather than the “age” of a taxon. Authors were encouraged to conclude their chapters by relating the timetree to other aspects of evolutionary history and Earth history such as plate tectonics, climate change, and biogeography. Therefore, even in chapters that do not contain new data, this section may contain a new synthesis and new ideas. One of the most difficult tasks of each author was to construct a single summary timetree for their group. Multiple timetrees were not permitted, because of space limitations. As guidance, we urged authors to average time estimates across studies, for each node, and use those mean time estimates in the timetree. For groups in which multiple studies exist, and each study has estimates for every node (i.e., a complete table), this averaging approach usually worked well. However, in cases where one published study included all taxa and other studies estimated times for only one or two nodes (i.e., a sparse table), the averaging approach was often impractical, in which case the most complete study was used for the summary timetree by some authors. A few authors also preferred not to average time estimates if the outcome was inconsistent with currently accepted tree topology or if they disagreed with the conclusions of a particular study. In this sense, contributing authors made the best possible choices in coming up with the final timetrees. All timetree chapters received rigorous peer review. In the case of those authored or coauthored by one of the editors (S.B.H.), the other editor (S.K.) handled the review process independently. All reviewers were kept anonymous, even when they asked to be identified to the authors. We owe a great debt to this large body of reviewers who significantly improved the overall quality of the book. All of the timetrees in this book are drawn in the same format. Higher taxa names, if present, are shown in vertical text on the right. Nodes are numbered from oldest to youngest, and there is a timescale at the bottom showing two levels of geologic periods. The geologic timescale, colors, and names of periods follow a recent, widely used system (A Geologic Timescale 2004, edited by F. M. Gradstein et al., Cambridge University Press, 2004). Node numbers were used on the tree specifically so that users of the book will be able to determine the exact time estimates for each node from the table. This system also permits users to find confidence intervals for each node from different published studies. It was not practical to place confidence intervals directly on the timetrees because the timetree summarizes multiple studies, each with different confidence intervals. Although we contemplated this book for some years, a catalyst for going forward was the great enthusiasm of the participants in a symposium that we organized in 2006, “Discovering the Timetree of Life,” at the annual meeting of the Society for Molecular Biology and Evolution at the Biodesign Institute in Tempe, Arizona. The Astrobiology Institute of the U.S. National Aeronautics and Space Administration (NASA) provided funding for that symposium, and the speakers were John C. Avise, Fabia U. Battistuzzi, Debashish Bhattacharya, Michael J. Benton, Jaime E. Blair, Franky Bossuyt, Linda E. Graham, S. Blair Hedges, Sudhir Kumar, William J. Murphy, Juan

Preface

xiii

C. Opazo, Michael J. Sanderson, Jeff rey L. Thorne, Marcel Van Tuinen, and Shuhai Xiao. Melissa Kirven-Brooks (NASA) assisted with the logistics. A majority of our work on this book took place while both of us were on sabbatical, and therefore we thank our institutions (Pennsylvania State University and Arizona State University) for providing time away from other duties. Staff and students in our laboratories (Solny Adalsteinsson, Matthew P. Heinicke, Wayne Parkhurst, Omy Keyes, Jana McAlpin, and Michael Suleski) provided technical assistance. We especially thank Michael Suleski for redrawing all of the timetrees in the book— for consistency—based on originals submitted by the authors. We also thank the funding agencies supporting our laboratories and research on molecular clocks: the National Science Foundation (S.B.H. and S.K.), NASA Astrobiology Institute (S.B.H.), the National Institutes of Health (S.K.), Japan Society for the Promotion of Science (S.K.), and the Science Foundation of Arizona (S.K.). We are grateful to Ian Sherman and the staff of Oxford University Press (especially Helen Eaton) for supporting this project. Finally, we thank all the authors for their excellent contributions and for tolerating numerous text and graphical edits that enabled us to maintain a uniformity of style. S. Blair Hedges and Sudhir Kumar


List of Contributors A. Louise Allcock The Martin Ryan Marine Science Institute, National University of Ireland Galway, University Road, Galway, Ireland. Cajsa L. Anderson Department of Biodiversity and Conservation, Real Jardin Botanico, CSIC, Plaza de Murillo 2, 28014 Madrid, Spain. Robert J. Asher Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, UK. John C. Avise Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92697, USA. Nadia A. Ayoub Department of Biology, University of California, Riverside, CA 92521, USA. Allan J. Baker Department of Natural History, Royal Ontario Museum, 100 Queen’s Park Crescent, Toronto, ON, M5S 2C6, Canada; Department of Ecology and Evolutionary Biology, University of Toronto, Toronto, ON, M5S 1A1, Canada. F. Keith Barker Bell Museum of Natural History, University of Minnesota, 1987 Upper Buford Circle, St Paul, MN 55108, USA. Fabia U. Battistuzzi Center for Evolutionary Functional Genomics, The Biodesign Institute, Arizona State University, Tempe, AZ 85287-5301, USA. Michael J. Benton Department of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK. Matthew A. Bertone Department of Entomology, Campus Box 7613, North Carolina State University, Raleigh, NC 27695-7613, USA. Debashish Bhattacharya Department of Biological Sciences and Roy J. Carver Center for Comparative Genomics, 446 Biology Building, University of Iowa, Iowa City, IA 52242, USA. Jaime E. Blair Department of Biology, Franklin and Marshall College, Lancaster, PA, 17604 USA. Mark Blaxter Institute of Evolutionary Biology, School of Biological Sciences, University of Edinburgh, Ashworth Laboratories, King’s Buildings, Edinburgh EH9 3JT, UK.

xvi

LIST OF CONTRIBUTORS

Franky Bossuyt Biology Department, Unit of Ecology & Systematics, Amphibian Evolution Lab, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium. Seán G. Brady Department of Entomology and Laboratories of Analytical Biology, National Museum of Natural History, Smithsonian Institution, Washington, DC, 20560, USA. Birgitta Bremer Bergius Foundation, Royal Swedish Academy of Sciences and Botany Department, Stockholm University, SE-106 91 Stockholm, Sweden. Christopher A. Brochu Department of Geoscience, University of Iowa, Iowa City, IA 52242, USA. Joseph W. Brown Department of Ecology and Evolutionary Biology & University of Michigan Museum of Zoology, 1109 Geddes Road, University of Michigan, Ann Arbor, MI 48109-1079, USA. David C. Cannatella Section of Integrative Biology and Texas Memorial Museum, 1 University Station C0930, University of Texas, Austin, TX 78712, USA. Mark W. Chase Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS, UK. Arnaud Couloux Centre national de séquençage, Genoscope, 2 rue GastonCrémieux, CP5706, 91057 Evry cedex, France. Joel Cracraft Department of Ornithology, American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, USA. Keith A. Crandall Department of Biology, Brigham Young University, Provo, UT 84602, USA; Monte L. Bean Life Sciences Museum, Brigham Young University, Provo, UT 84602, USA. Bryan N. Danforth Department of Entomology, Comstock Hall, Cornell University, Ithaca, NY 14853, USA. Frédéric Delsuc Université Montpellier II, Place Eugène Bataillon, 34095 Montpellier Cedex 05, France; CNRS, Institut des Sciences de l'Evolution (UMR 5554), CC064-Place Eugène Bataillon, 34095 Montpellier Cedex 05, France. Rui Diogo Department of Anthropology, The George Washington University, Washington, DC, 20052, USA. Philip C. J. Donoghue Department of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK. Christophe J. Douady Université de Lyon, F-69622, Lyon, France; Université Lyon 1, F-69622 Villeurbanne, France; Laboratoire d'Ecologie des Hydrosystèmes Fluviaux (UMR CNRS 5023), F-69622 Villeurbanne, France. Emmanuel J. P. Douzery Université Montpellier II, Place Eugène Bataillon, 34095 Montpellier Cedex 05, France; CNRS, Institut des Sciences de l'Evolution (UMR 5554), CC064-Place Eugène Bataillon, 34095 Montpellier Cedex 05, France.

List of Contributors

xvii

Scott V. Edwards Department of Organismic and Evolutionary Biology, Museum of Comparative Zoology, 26 Oxford Street, Harvard University, Cambridge, MA 02138, USA. Eduardo Eizirik Faculdade de Biociências, PUCRS, Av. Ipiranga, 6681, Porto Alegre, RS 90619-900, Brazil; Instituto Pró-Carnívoros, Av. Horácio Neto,1030, Atibaia SP 12945-010, Brazil. Brian D. Farrell Department of Organismic and Evolutionary Biology, 26 Oxford Street, Harvard University, Cambridge, MA 02138, USA. Félix Forest Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS, UK. John Gatesy Department of Biology, University of California, Riverside, CA 92521, USA. David J. Gower Department of Zoology, The Natural History Museum, London SW7 5BD, UK. Felix M. Gradstein Museum of Natural History, University of Oslo, N-0318 Oslo, Norway. Jeremiah D. Hackett University of Arizona, Ecology and Evolutionary Biology Department, Biosciences West 336, Tucson, AZ 85721, USA. Cheryl Y. Hayashi Department of Biology, University of California, Riverside, CA 92521, USA. Shunping He Institute of Hydrobiology, The Chinese Academy of Sciences, Wuhan, 430072, China. S. Blair Hedges Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA. Matthew P. Heinicke Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA. Xiaolan He-Nygrén Botanical Museum, Finnish Museum of Natural History, University of Helsinki, P.O. Box 7, 00014 Helsinki, Finland. Khidir W. Hilu Department of Biological Sciences, Virginia Tech, Blacksburg, VA 24061, USA. Jens T. Høeg Comparative Zoology and Department of Biology, Institute of Biology, University of Copenhagen, Universitetsparken 15, DK-2100, Copenhagen, Denmark. Rodney L. Honeycutt Division of Natural Science, Pepperdine University, 24255 Pacific Coast Highway, Malibu, CA 90263-4321, USA. Peter Houde Department of Biology, New Mexico State University, Box 30001 MSC 3AF, Las Cruces, NM 88003-8001, USA. Thomas Janßen Research Institute Senckenberg, Department of Botany and Molecular Evolution, Senckenberganlage 25, 60325 Frankfurt, Germany.

xviii


Jungwook Kim Department of Entomology, North Carolina State University, Raleigh, NC 27695-7613, USA. Carey W. Krajewski Department of Zoology, Southern Illinois University, Carbondale, IL 92901, USA. Sudhir Kumar Center for Evolutionary Functional Genomics, The Biodesign Institute, and School of Life Sciences, Arizona State University, Tempe, AZ 85287-5301, USA. Shigehiro Kuraku Laboratory for Evolutionary Morphology, RIKEN Center for Developmental Biology, 2-2-3 Minatojima-minami, Chuo-ku, Kobe 650-0047, Japan; Lehrstuhl für Zoologie und Evolutionsbiologie, Department of Biology, University of Konstanz, Universitätsstrasse 10, D-78457 Konstanz, Germany. Shigeru Kuratani Laboratory for Evolutionary Morphology, RIKEN Center for Developmental Biology, 2-2-3 Minatojima-minami, Chuo-ku, Kobe 650-0047, Japan. Leah Larkin Department of Biological Sciences, University of the Pacific, Stockton, CA 95211, USA. Annie Lindgren Museum of Biological Diversity, The Ohio State University, 1315 Kinnear Road, Columbus, OH 43215, USA. Arne Ludwig Department of Evolutionary Genetics, Leibniz-Institute for Zoo and Wildlife Research, Alfred-Kowalke-Street 17, 10315 Berlin, Germany. Ole Madsen Department of Biomolecular Chemistry 271, Radboud University Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands; Animal Breeding and Genetics Group, Wageningen University, P. O. Box 338, 6700 AH, Wageningen, The Netherlands. Susana Magallón Departamento de Botánica, Instituto de Biología, Universidad Nacional Autónoma de México, 3er Circuito de Ciudad Universitaria, Del. Coyoacán, México D.F. 04510, Mexico. Conrad A. Matthee Evolutionary Genomics Group, Department of Zoology, University of Stellenbosch, Stellenbosch, 7602 South Africa. Duane D. McKenna Department of Organismic and Evolutionary Biology, 26 Oxford Street, Harvard University, Cambridge, MA 02138, USA. Linda K. Medlin Marine Biological Association of the UK, The Citadel, Plymouth PL1 2PB, UK. Robert W. Meredith Department of Biology, University of California, Riverside, CA 92521, USA. Kathleen J. Miglia Duke University Medical Center, Molecular Genetics & Microbiology, 210 Jones Building, Box 3020 Durham, NC 27710, USA. David P. Mindell California Academy of Sciences, 55 Concourse Drive, Golden Gate Park, San Francisco, CA 94118, USA.


xix

Masaki Miya Natural History Museum and Institute, Chiba, 955-2 Aoba-cho, Chuo-ku, Chiba 260-8682, Japan. William S. Moore Department of Biological Sciences, Wayne State University, Detroit, MI 48202, USA. William J. Murphy Department of Veterinary Integrative Biosciences, Texas A&M University, College Station, TX 77843-4458, USA. Gavin J. P. Naylor School of Computational Biology, Dirac Science Library, Florida State University, Tallahassee, FL 32306, USA. Thomas J. Near Department of Ecology and Evolutionary Biology & Peabody Museum of Natural History, Yale University, New Haven, CT 06520, USA. Angela E. Newton Department of Botany, Natural History Museum, London, SW7 5BD, UK. James G. Ogg Department of Earth and Atmospheric Sciences, Purdue University, West Lafayette, IN 47907, USA. Kinya G. Ota Laboratory for Evolutionary Morphology, RIKEN Center for Developmental Biology, 2-2-3 Minatojima-minami, Chuo-ku, Kobe 650-0047, Japan. Zuogang Peng School of Biology, Georgia Institute of Technology, Atlanta, GA 30332, USA. Sérgio L. Pereira Department of Ecology and Evolutionary Biology, University of Toronto, Toronto, ON, M5S 1A1, Canada. Marcos Pérez-Losada CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, Universidade do Porto, Campus Agrário de Vairão, 4485661 Vairão, Portugal; Department of Biology, Brigham Young University, Provo, UT 84602-5181, USA. Davide Pisani Laboratory of Evolutionary Biology, Department of Biology, The National University of Ireland, Maynooth, Co. Kildare, Ireland. Megan L. Porter Department of Biological Sciences, University of Maryland Baltimore County, Baltimore, MD 21250, USA. Céline Poux Vertebrate Department, Royal Belgian Institute of Natural Sciences, Vautierstraat 29, 1000 Brussels, Belgium; Department of Biomolecular Chemistry 271, Radboud University Nijmegen, PO Box 9101, 6500 HB Nijmegen, The Netherlands. Kathleen M. Pryer Department of Biology, Duke University, Durham, NC 27708, USA. Jean-Claude Rage UMR 5143, Paléobiodiversité & Paléoenvironnements, Département Histoire de la Terre, C. P. 38, Muséum National d’Histoire Naturelle, 8 rue Buffon, Paris 75005, France. Susanne Renner Department of Biology, Menzingerstr. 67, University of Munich, Munich, Germany.

xx


Kim Roelants Biology Department, Unit of Ecology and Systematics, Amphibian Evolution Lab, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium. Alex D. Rogers Institute of Zoology, Zoological Society of London, Regent’s Park, London NW1 4RY, UK. Lukas Rüber Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK. Oliver A. Ryder Conservation and Research for Endangered Species, Zoological Society of San Diego, 15600 San Pasqual Valley Road, Escondido, CA 920277000, USA. Jennifer M. Sander Department of Evolution, Ecology, and Organismal Biology, Ohio State University, Columbus, OH 43210-1293, USA. Eric Schuettpelz Department of Biology, Duke University, Durham, NC 27708, USA. H. Bradley Shaffer Department of Evolution and Ecology, University of California, Davis, CA 95616, USA. A. Jonathan Shaw Department of Biology, Duke University, Durham, NC 27703, USA. Andrew M. Shedlock Department of Organismic and Evolutionary Biology, Museum of Comparative Zoology, 26 Oxford Street, Harvard University, Cambridge, MA 02138, USA. Andrew B. Smith Department of Palaeontology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK. Mark S. Springer Department of Biology, University of California, Riverside, CA 92521, USA. Michael E. Steiper Department of Anthropology, Hunter College of the City University of New York, 695 Park Avenue, New York, NY 10065, USA; Programs in Anthropology and Biology, The Graduate Center of the City University of New York, 365 Fift h Avenue, New York, NY 10016, USA. Jan M. Strugnell Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, UK. Emma C. Teeling UCD School of Biology and Environmental Science, Science Center West, University College Dublin, Belfield, Dublin 4, Ireland. Michelle D. Trautwein Department of Entomology, North Carolina State University, Raleigh, NC 27695-7613, USA. Marcel van Tuinen Department of Biology and Marine Biology, 601 South College Road, University of North Carolina at Wilmington, Wilmington, NC 28403-5915, USA.


xxi

Nicolas Vidal UMR 7138, Systématique, Evolution, Adaptation, Département Systématique et Evolution, C.P. 26, Muséum National d’Histoire Naturelle, 43 Rue Cuvier, Paris 75005, France. David R. Vieites Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Sciences Building, University of California, Berkeley, CA 94720, USA. David B. Wake Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Sciences Building, University of California, Berkeley, CA 94720, USA. Marvalee H. Wake Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Sciences Building, University of California, Berkeley, CA 94720, USA. Brian M. Wiegmann Department of Entomology, North Carolina State University, Raleigh, NC 27695-7613, USA. Niklas WikstrÖm Department of Systematic Botany, Evolutionary Biology Centre, Uppsala University, Norbyvägen 18D 75236, Uppsala, Sweden. Mark Wilkinson Department of Zoology, The Natural History Museum, London SW7 5BD, UK. Shaun L. Winterton Entomology, Queensland Department of Primary Industries & Fisheries, 80 Meiers Road, Indooroopilly, Queensland, Australia 4068. Hwan Su Yoon Bigelow Laboratory for Ocean Sciences, West Boothbay Harbor, ME 04575, USA. Nathan M. Young University of California at San Francisco, 1001 Portrero Avenue, San Francisco, CA 94110, USA. Peng Zhang Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Sciences Building, University of California, Berkeley, CA 94720, USA.


INTRODUCTION


Discovering the Timetree of Life S. Blair Hedgesa,* and Sudhir Kumar b a

Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA; bCenter for Evolutionary Functional Genomics, The Biodesign Institute, Arizona State University, Tempe, AZ 85287-5301, USA *To whom correspondence should be addressed ([email protected])

Abstract The two primary components of evolutionary history are the relationships of organisms (phylogeny) and their times of divergence. Together they form a timetree: a phylogenetic tree scaled to time. The fossil record initially provided the timescale, but this has been supplemented in recent years with the application of molecular clocks. The Timetree of Life is now being discovered, largely through phylogenetic and chronological analyses of DNA and protein sequences. The addition of a temporal dimension to the tree of life is driving major advances in evolutionary biology, providing a better understanding of the mechanisms of evolution, and revealing the reciprocal interactions between life and the environment throughout Earth history.

The evolutionary tree of life describes how species are related and organized in the greatest of biological hierarchies. In a letter to Huxley, 2 years before publication of Origin of Species (1), Charles Darwin predicted: The time will come, I believe, though I shall not live to see it, when we shall have fairly true genealogical trees of each great kingdom of Nature.

With great delight, we can say that the time has come, and that it is now. Certainly, many important details remain to be worked out, such as deep branching patterns among major taxonomic groups and the interrelationships of many species, but much of the tree of life already has taken shape (2). This revolution in evolution has occurred largely through advances in molecular biology over the last half century, building on a foundation laid by paleontology and comparative biology. It would not have been possible without many discoveries, progressively building on previous work, such as the structure of DNA (3), methods to sequence proteins

and DNA (4–6), a technique to rapidly amplify DNA (7), and advances in statistical methods of data analysis. Some—perhaps most—of the resulting molecular phylogenies have corroborated trees based on morphology and cell biology, but many findings were unexpected including the discoveries of archaebacteria (8) and an African clade of mammals (9) to name just two. Our current understanding of the tree of life draws from fossils, morphology, and—especially in the last two decades— many molecular phylogenies. However, a phylogenetic tree provides only half of the picture. Evolutionary history has two primary components—relationships and timescale—and both are important. Together, they form something that does not have a specific name; hence we use the word “timetree” for any tree scaled to time. It is preferred over the more general term “chronogram” which does not indicate that a tree is involved, or “phylogeny.” (Phylogeny is the specific branching order—relationships—without the temporal component.) In past decades, the two words have been used separately (“time tree”), although rarely, with the compound form appearing only a few times in recent years (e.g., 10, 11). The word “phylochronology” has been applied recently to the study of populations through time using ancient DNA methods (12), but it is just as applicable to the study of timetrees in general. The dimension of time provides a direct connection with other fields of science, and the ability to relate biological evolution with climate change and Earth history in general. A timetree of life offers a more complete view of the framework of evolutionary history than the tree of life alone, and the utility of this perspective is broadly recognized (13). Timetrees were not invented with molecular data. In fact, Darwin’s only figure in the Origin of Species (1) is essentially a timetree—a hypothetical one—scaled to generations rather than years. Subsequently, paleontologists (e.g., 14) were the major producers of timetrees, because fossils initially provided the only information to establish the evolutionary timescale. This changed in the early 1960s. Enough protein sequence information became available to show that molecular change is more predictable and quantifiable than morphological change (15), an observation that has come to be known

S. B. Hedges and S. Kumar. Discovering the Timetree of Life. Pp. 3–18 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

4

THE TIMETREE OF LIFE

as the “molecular clock” (16). Subsequently, studies of viral DNA evolution, where dates of divergence are known, confirmed that time-dependent change occurs at the molecular level (17). The basic principle of molecular clocks is that there is a strong enough correlation between molecular change and geologic time, such that the rate can be used to measure time in parts of the tree lacking a calibration time. Correlations between fossil and molecular times are often strong (18), but large differences have been encountered as well (19–21), leading to intense scrutiny of both types of data and the generation of new hypotheses to explain the difference. The theoretical basis of molecular clocks is a separate issue from the application of clocks for the same reason that early civilizations counted days—quite accurately—by watching the sun move across the sky without knowing how that process occurred. From the beginning, it has been suggested that the explanation for the molecular clock—in part or in whole—lies in the selective neutrality of mutations and substitutions (15, 16, 22). Because mutations occur randomly, this would mean that molecular clocks are stochastic clocks, in the same class as geologic clocks that rely on random isotopic decays. This is in contrast to regular clocks, such as those based on pendulums or atomic resonance. While it is likely that molecular clocks are stochastic clocks, it is less certain whether they are driven by neutral mutations. Explanations involving selection have been proposed as well (23–25) and no consensus has been reached as to the basis for the clock-like change in molecular data (26). But lineage-specific variation should not be confused with the wide variation in rates of change among genes and proteins, which is the result of selection on function (27, 28). Moreover gene-specific variation based on selection is fully compatible with the hypothesis of neutrality and stochasticity in any given gene or protein. While the theoretical underpinning of molecular clocks and the dynamics of rate variation continue to be studied (26), the field has firmly entered an empirical phase. Divergence times are being estimated in many groups of organisms and this effort is growing at a rapid pace, driven by the greater accessibility of DNA sequence data (Fig. 1).

Advances in methodology Methodology for molecular time estimation has developed considerably since the 1960s, in parallel with improved methods of data acquisition. Initially, most researchers estimating divergence times used distance

A 3000

Publications

2000 1000 0 B 250 200 150 100 50 0 C

80 60

Taxa (thousands)

Sequences (millions)

40 20 0 1990

1995

2000

2005

Fig. 1 The rapid expansion of published molecular time estimates. (A) Number of research publications that have discussed molecular clocks or molecular time estimates. Data are from the Web of Science online resource, with results checked for appropriateness to molecular evolutionary clocks. The search included only titles, abstracts, and keywords and therefore represented only a fraction of the total number of articles, although not all identified articles present new time data. (B) Thousands of taxa in Genbank (NCBI). (C) Millions of DNA sequences in Genbank during the same period. All counts are cumulative.

data from immunological, allozyme, and DNA–DNA hybridization methods (e.g., 20, 29) until DNA sequencing became more accessible (30). Relative rate tests were used almost from the beginning to test the significance of clock-like behavior and to adjust time estimates based on lineage-specific rate differences (31–33). Methods for accommodating lineage-specific rate variation in a molecular clock analysis continued to be developed (33, 34) and were useful in some focused studies (e.g., 35) but were not mature for routine analyses. Instead, most molecular clock studies before the year 2000 took an approach that used statistical tests to identify and exclude genes and lineages that violated rate constancy before estimating time (36, 37). In the last decade, further improvements were made in methods for accommodating rate variation among lineages (38–42). These methods have reduced the need to exclude species or genes that show rate variation, facilitating molecular clock analysis. As a consequence, most molecular clock studies now use these “relaxed clock” methods.

Discovering the Timetree of Life

Any device for timing requires calibration, and this is perhaps the most critical aspect of molecular clocks and the one most often debated. If timing is done with a gene deemed to be evolving at a constant rate, only a single calibration is needed, by definition—the rate established at any one point in the tree can be propagated to all points in the tree. For this reason, emphasis was often placed on the quality of the calibration (18), whether it was from a single fossil or an average rate from multiple fossils. The closer the calibration is to the true evolutionary divergence, the more accurate the resulting time estimates. Relaxed clock methods can work with a single calibration point, but they are expected to do better with multiple calibrations because modeling of rate of evolution across a tree can only be done reliably with multiple reference points. It is better to use as many good calibration points as possible, but only one calibration is required to estimate time. The number of calibrations used in any study is dictated by practical considerations—what is available—and, most importantly, by the quality and distribution of calibrations (37). For groups lacking calibrations and showing rate constancy in a gene (or genes), there is nothing wrong with using a rate established in another group of organisms as long as the assumptions are explicitly stated. This is an approach used frequently in the past for mitochondrial DNA analyses (e.g., 43, 44), as it provided valuable information on the timing of events otherwise unavailable. This flexibility is necessary because calibration points of any kind are absent from a vast majority of nodes in the Tree of Life, requiring rates of change to be extrapolated from one group or node having calibrations to another lacking calibrations, if divergence times are to be estimated at all. In fact, rate extrapolation—between one part of a tree and another—has to be done in every molecular clock study, regardless of the number of calibration points. As mentioned earlier, the best calibration point is one that is closest to the true evolutionary divergence. Such calibrations are rare and usually restricted to divergences caused by the separation of land (e.g., separation of two continents) for terrestrial organisms, or water for aquatic organisms, and that have associated geologic dates. Fossil calibrations are always minimum times of divergence and therefore will result in time estimates that are minimums as well, more recent than the true divergence. A major change in the use of calibrations has come with development of relaxed clock methods that permit both maximum and minimum calibrations in estimating divergence times. If both types of

5

calibrations are used in a study, the resulting times can be considered as estimates of the true divergence time. The difficulty comes in identifying valid maximum times of divergence for calibration. The age of the Earth (4600 million years ago, Ma) is a global maximum for all nodes in the Tree of Life, but it is too old for most purposes. The best maximum calibrations are those that involve emergence of land areas (e.g., islands), providing a maximum time for diversification of terrestrial organisms restricted to that area (45). Establishing a maximum in the fossil record is more difficult. In 1996, we proposed a method for establishing a maximum calibration using transitional fossils, in that case involving the transition from fishes to tetrapods ~380–360 Ma (19). A series of fossils documents the morphological transition from lobe-finned fishes to stem tetrapods (46, 47), thus constraining the maximum time of any divergence within tetrapods, such as the split between mammals and birds. However, fossils recording such evolutionary transitions are rarely available. Another method of establishing a maximum calibration in a relaxed clock analysis is to use the age of the earliest fossil evidence for a lineage (48, 49). This approach is problematic. Its use is tantamount to interpreting the fossil record as the true record of evolutionary history, which is guaranteed to underestimate the true time of divergence. A related method uses the age of the oldest fossil of the most closely related group to the node in question as the maximum age for the node (50, 51). This would often result in a narrow interval between the minimum and maximum time of divergence of two lineages, and its use would force relaxed clock methods to produce estimates that are not older than the maximum calibration. Consequently, the power of sophisticated statistical methods would fail to be realized in molecular clock analysis and their outcomes will effectively not be different from reading the fossil record as a literal interpretation of the Timetree of Life (or, at the least, biased by this method). In general, the oldest fossil of a lineage can only establish a minimum for that lineage, not a maximum for another lineage (52, 53). Another method of determining the maximum constraint of a node from the fossil record again considers the earliest fossils of closest relatives, but it instead places emphasis on the absence of fossils of the clade in question from earlier deposits that otherwise should contain representatives (53). Such “soft maximum” constraints (54) have been proposed, along with minimum calibration constraints, for a moderate number of divergences among animals (53, 55, 56). However, soft maximum

6


calibration points tied to assumptions regarding the absence of earlier fossils may cause considerable underestimation of divergence time in the same way as mentioned earlier. A case in point concerns the continuing debate over the Cambrian Explosion model, which suggests that most animal phyla originated (diverged from one another) during a relatively short period of time in the latest Precambrian and early Cambrian, ~550 Ma (57–59). Instead, most studies using molecular clocks have found that divergences among phyla occurred hundreds of millions of years before their appearance in the fossil record (reviewed in 60–62); those studies claiming younger time estimates (63–68) have been shown to be flawed (60, 62, 69–71). If the soft maximum constraint of 581 Ma (53) is used in relaxed clock studies, it will cause the inferred time estimates to be close to fossil time estimates. In other words, if an investigator believes that the maximum date for protostome deuterostome divergence is close to 581 Ma, then the test of the Cambrian explosion hypothesis by molecular data is a pointless exercise. In addition to minimum constraints, we have proposed that probability distributions describing the likelihood of the true divergence time for a calibration node be considered in molecular clock analyses (18, 19, 37). For example, we have described how the divergence of birds and mammals is associated with a higher probability closer to the minimum (310 Ma) than to the maximum (370 Ma), which leads to a distribution with a similar probability rather than one with a central tendency (e.g., a normal distribution) or uniform distribution (19). Based on the birth–death model of diversification, a long-tailed (logistic) distribution for calibrations has been favored (53), although other distributions are possible (37). In fact, a distribution can be envisioned in which the highest density (likelihood) is near the oldest time when only the maximum constraint is known, and in which the tail of the distribution is toward the younger time. Some of these probability distributions of calibrations have already been used while estimating the confidence intervals of divergence times (72) and in Bayesian relaxed clock analyses (41). While probability distributions for calibrations may be useful, they should also be used with caution. For example, the use of nonuniform distributions (e.g., logistic, exponential) may agree with some patterns of diversification (53), but they may not be the best model for all divergences in a data set—and hence may impose a bias—if a nonbiological process affecting fossilization is responsible. For example, oxygen levels in the atmosphere

rose sharply in the late Precambrian, 580–542 Ma, from 1% to 10% of the present level to nearly 100% of the present level (73). Before 580 Ma, it is unlikely that there would have been sufficient oxygen for large, hard-bodied animals. Because the probability of fossilization apparently changed dramatically at 580 Ma, the logistic distribution would no longer be appropriate. Instead, a uniform calibration probability distribution might be more appropriate. Considering this, and the fact that we have a poor understanding of gaps in the fossil record in general (74), a uniform distribution (or, the assignment of a maximum or minimum calibration point without a distribution) may be the least biased approach for most or all nodes unless there is justification for using a particular distribution. Clearly, this is one area of research in great need of attention from both modelers of fossil preservation potential (74–77) and developers of relaxed clock methods (38, 39, 41, 42, 54).

The current state of knowledge The Timetree of Life (78) summarizes the current state of knowledge on the Timetree of Life, with some caveats. The first is that it includes only living organisms, sampled by molecular methods (the molecular record). A synthesis of living and fossil groups is an important future goal (see later). Secondly, there is unequal coverage among the kingdoms and phyla, directly related to the limited availability of molecular data for certain groups of organisms. Thus, plants, cartilaginous fishes, amphibians, reptiles, and mammals are relatively well covered, whereas protists, fungi, invertebrates, and ray-finned fishes are poorly represented. Coverage of prokaryotes is moderate: roughly two-thirds of the families of archaebacteria and onethird of the families of eubacteria are included in these timetrees. Thirdly, the book covers only the families and groups above that taxonomic level. To include genera and species would have been impossible in a single volume, although such expanded coverage is planned for the future in a similar format, and is available elsewhere in an online database (11). Another caveat is that taxonomic ranks above the species level are partly arbitrary with regards to the temporal depth of included divergences, resulting in disproportionate coverage of evolutionary lineages among the groups (79). Lastly, while it is an authority-based synthesis, opinions often differ among experts and therefore some biases will necessarily be present in how the data and conclusions are presented. Despite these caveats, the book chapters are windows into


Chondrichthyes, Dipnoi, Serpentes, Crocodylia, and Piciformes. All individual timetrees in The Timetree of Life were assembled here into a single timetree (Fig. 2). It includes all three superkingdoms and 1610 families (or family-level

the literature of an area and mostly (if not all) all available divergence times relevant to the scope of the chapter. New divergence time data or analyses are included in 13 chapters: Archaebacteria, Eubacteria, Eukaryota, Monocots, Metazoa, Scaphopoda, Aculeata, Coleoptera,

5 Millions of years

5

Eub

acte

ria ha

e.

als

mm

Pro

Arc

Ma

tis

ts

25 M

os

50

se

75 100

Bir

ds

s

s ort erw Liv

250

n Fer

500

y sG

ates

1000

Verteb r

2000

Plants

Reptiles

4000

ng pla n

ibian Amph

s

Floweri

Ec

hi

n.

F is

h

ts

i

ng

Fu

C M

oll

. N ds

Arthropo

Invertebrates

Fig. 2 The Timetree of Life from an assembly of individual timetrees. Each of the 1610 terminal branches represents a family or family-level taxon, although only major clades are labeled. This global timetree was assembled from all of the timetrees in The Timetree of Life (78), with very little additional manipulation or editing. For three nodes lacking molecular time estimates (Scaphopoda/Cephalopoda, Decapodiformes/ Octapodiformes, and Anabantoidei/Notothenoidei), the ages

7

of the earliest pertinent fossils were considered in setting the time. Also, Nematoda was joined with Priapulida and Arthropoda, and set to the same time of divergence. A total of 36 nodes in 13 chapters (0.2% of all nodes) were slightly older than parental nodes in other chapters and were adjusted (average, 6.7%) to eliminate the conflict. Abbreviations: C (Cnidaria), Gy (Gymnosperms), Echin. (Echinodermata), and Moll. (Mollusks).

8


taxa). It is unlikely that a complete timetree of all life will be available in the foreseeable future (or ever) given the uncertainties in some parts of the tree, problems of horizontal gene transfer, incompleteness of the fossil record, ongoing extinctions through human impact, and inability to collect and sample—with molecular methods—all living species for practical reasons. Nonetheless, this first major synthesis could be called the Timetree of Life, while acknowledging that it is a work in progress by many people, and that it is not too far from the true version. The next decade or two will see the Timetree of Life come into much better focus, but it is unlikely there will be any one day where the community can proclaim that the Tree of Life or Timetree of Life has been achieved. In the future, history will probably record the discovery of the Timetree of Life as having happened during a short period—perhaps in the first two or three decades of the twenty-first century.

Table 1. Ages of family lineages of organisms as measured by molecular clocks, based on an analysis of data in The Timetree of Life (78). Group

Average age (Myr)

No. families

SE

Archaebacteria

2567

13

300

Eubacteria

1388

89

80.6

Eukaryotes

102.4

1378

1.9

Land plants

103.1

463

3.1

Mosses

156.3

59

10.1

Liverworts

167.3

42

12.5

Ferns

209.9

21

18.9

Gymnosperms

229.2

12

19.4

Flowering plants

73.9

329

1.5

Eudicots

69.3

235

1.8

Monocots Animals

85.4

77

2.2

102.1

915

2.3

Invertebrates

143.7

401

3.2

Families through time

Mollusks

183.9

26

20.2

This synthesis and analysis of the timetrees and divergence times in Timetree of Life provides the opportunity to make broad comparisons across all of life and reveals some new patterns. One useful comparison is the average age of a family lineage (the elapsed time since the divergence with its closest relative). Taxonomic ranks are often used in comparative studies, even though their biological meaning is unclear (80). Therefore, it is useful to know how family lineages differ in age among groups. Sparse coverage of families will bias lineages towards older ages, and therefore some caution must be used in evaluating the results (Table 1), especially in the poorly sampled groups such as invertebrates and ray-finned fishes. There is great disparity among groups, as is already known from fossil evidence (81). However, molecular time estimates should provide a better quantification of the difference because they correspond to lineage originations rather than the earliest occurrence as recorded by fossils. Typical families within the three superkingdoms differ greatly in age (Table 1). Families of Archaebacteria and Eubacteria, on average, are 25 and 14 times, respectively, as old as those of eukaryotes. However, plant and animal families are nearly identical in age, on average (~100 Myr old), despite their long and separate histories of taxonomic practice. Among animals, the average age of an invertebrate family (144 Myr) is about twice that of a vertebrate family (69.7 Myr). Within vertebrates, the

Nematodes

249.8

5

34.3

Spiders

158.6

26

10.3

Bees, ants, & stingless wasps

117.1

22

6.2

Beetles

127.6

183

3.4

Lacewings

206.5

17

7.3

True flies

142.9

51

7.5

Crustaceans

185.1

44

11.6

Sea urchins

95.9

27

13.2

69.7

514

2.5

Vertebrates Jawless fishes

482.3

2

0.0

Cartilaginous fishes

143.0

57

7.8

Ray-finned fishes

53.4

15

13.7

Lungfishes

172.3

3

52.3

Frogs and toads

92.2

59

6.0

Salamanders

146.6

10

8.8

Caecilians

118.9

7

25.0

Turtles

91.3

14

11.3

Lizards, snakes, and amphisbaenians

84.2

54

5.6

Crocodilians

76.7

3

12.9

Birds

42.8

149

1.7

Mammals

37.1

141

1.3

Note: Age was measured as the time of divergence between that family and its closest relative. Myr = million years and SE = standard error.


fishes show wide variation in the average age of families, ranging from 482 Myr in jawless fishes (two families) to 53 Myr in a small selection of ray-finned fishes. The nonavian reptiles (turtles, lizards, snakes, amphisbaenians, and crocodilians) have similar mean ages, dating to the late Cretaceous (91–77 Myr). The three orders of amphibians also have average ages in the Cretaceous, although a bit earlier (146–92 Myr). Families of birds and mammals typically are younger (Cenozoic) and are similar in average age (43 and 37 Myr, respectively). Fossils and molecules

A comparison between 46 fossil minimum divergence times (55) and the corresponding molecular times (Fig. 3) shows a high correlation (r = 0.96), as has been observed previously in broad surveys (18, 36). Only two fossil dates were older than molecular dates: Ochotonidae vs. Leporidae and Aves vs. Crocodylia. Both are controversial and have been discussed in the literature (82, 83). In the other comparisons, fossil times are 27.5% (4–57%) younger than molecular times on average, with the largest differences involving the earliest animal divergences (the Cambrian Explosion). A maximum calibration (e.g., from the fossil record) should be older than the (unknown) true divergence

time. Likewise, if the molecular divergence time is an unbiased estimate of the true divergence time, and its variance is not excessive, it should be younger than maximum calibrations in a majority of cases. However, while the correlation coefficient of molecular vs. soft maximum times was high (0.96), nearly half (48%) of the soft maximum divergence times (55) were younger than the corresponding molecular time estimates (Fig. 2). They ranged from 99% older to 44% younger. This indicates that either (i) the molecular time estimates are overestimates of the true divergence, (ii) these soft maximum times are underestimates of the true divergence, or (iii) the variance in the molecular estimates is so large that a large fraction of estimates—by chance—are older than the fossil maximum estimates. The results of the diversification analyses in the next section give some independent support for the accuracy of molecular time estimates, which would suggest that the soft maximum times (55) are underestimates of the true divergence. The diversification of life

If we were to know the total number of species that existed at all times in the past, we would have a complete view of the rate of evolutionary diversification (speciation) through time. This information would be valuable

900 800 700

Fossil time (Ma)

600 500 400 300 200 100 0 0

200

400

600

9

800

1000

1200

1400

Molecular time (Ma)

Fig. 3 Relationship between molecular time estimates (x-axis), based on data from The Timetree of Life (78) and fossil time estimates (y-axis) (55). Open circles = fossil minimum times; closed diamonds = fossil soft maximum times; dashed line = 1:1 relationship; Ma = million years ago.

10


for understanding not only how diversification in some groups (e.g., predators) was affected by diversification in other groups (e.g., prey), but also the relationship between biological diversification and Earth history. Syntheses of the fossil record, usually focusing on genera or families, already have alluded to some of these patterns (75, 84), although evidence from molecular clocks has not always reproduced them (19, 21), leading to debates over evolutionary mechanisms. But the two estimates are not expected to agree all the time. Origination events in the fossil record measure the first occurrence of a taxon, recognizable by morphology, which must postdate the phylogenetic divergence of that lineage and its closest relative, measured by molecular clocks. In theory, the time difference between the phylogenetic divergence and the fossil origination could be slight (~1%), but that would require an extraordinarily complete fossil record and the evolution of diagnostic morphological characters over a very short period of time. In practice, however, a substantial difference between the fossil and molecular time estimates is expected, and such a difference (28%, average) was found, as discussed in the previous section. Also, at least some of the pattern of diversification recorded in the fossil record can be attributed to known biases such as peaks in originations ascribed to rare sites of fossilization (75, 85). Despite the significant taxonomic gaps, the data assembled in The Timetree of Life provide an opportunity, for the first time, to compare global, fossil-based patterns of diversification (75, 84, 86, 87) with those measured by molecular clocks. The two data sets are based on familylevel (and above) taxa, with 7186 families in the fossil record database (81) and 1610 families in the synthesis here of molecular times. The molecular data set does not include extinct lineages whereas the fossil data set lacks many extant lineages. The fossil record database also contains many more marine and invertebrate taxa than are currently available in the molecular database. On the other hand, the early history of life is much better represented by the molecular data set. There are essentially no fossils of any groups of extant taxa—aside from cyanobacteria and a handful of uncontested eukaryote fossils (88, 89)—before about 600 Ma (81). In this sense, the molecular data provide the first glimpse of diversification patterns in the Precambrian. Also, divergences can be sampled continuously and evenly throughout the molecular timetree, whereas fossils are necessarily assigned to geologic periods which vary in length. In the analyses detailed here, we follow previous authors in presenting diversification curves as lineage originations,

both cumulatively and as originations or rates binned to time intervals (75, 84). Because of the relatively small fraction of exclusively marine families in the molecular data set, a distinction was not made between marine and continental taxa. The global diversification curve (Fig. 4A and B) shows a relatively smooth exponential increase (linear on a log scale) except for the last 300 Myr where the rate trajectory shifts steeply upward. (The slight plateau seen in the most recent intervals, the last ~30–40 Myr, is probably an artifact of the taxonomic scope of the data being restricted to the family level and above). This is a much different and smoother curve than has been observed in past analyses of diversification based on the fossil record. The results of those fossil analyses have been debated as to whether they conform to a dampened exponential curve (75, 86, 87, 90, 91) or a logistic model (84, 92–95). The explanation for the exponential curve is that diversification proceeds continuously at about the same rate without being limited by competition. In contrast, the second school of thought contends that competition among lineages for ecological niches causes diversification to follow a logistic curve, with an early rapid rate and a late slow rate, eventually reaching a plateau. These analyses (Fig. 4) support the first, expansionist model, and show a good fit to a standard exponential curve for most (93%) of the history of life. The rate curve (Fig. 4C) also shows the sharp increase in origination rate in the last ~300 Myr. The rate between 4000 and 400 Ma averages 17% per 200 Myr whereas the rate for the last 200 Myr is 64%. A biological explanation for this dramatic rate increase is not immediately obvious. The timing could suggest that it is related to the colonization of land and diversification of terrestrial organisms in the late Paleozoic Era, especially in the Carboniferous and Permian (359–251 Ma). If so, it could also be associated with a major pulse in atmospheric oxygen ~340–250 Ma (96, 97). However, there is not a complete concordance between oxygen levels and diversification, because oxygen declined in the early Mesozoic at the same time that the lineage origination rate continued to increase. Yet another, and perhaps, simpler explanation is that the rate spike of the last ~300 Myr is an artifact of the extinction process and the fact that only living lineages are being examined. This well known “Pull of the Present” bias results when an excess of recent lineages are sampled that will soon become extinct and removed from the surviving curve of continuous lineages. In other words, If we traveled back through time and sampled

Discovering the Timetree of Life A

11

1800

Lineages

1200

B

600

0 10000

Lineages

1000 100 10

C

1

Originations (%)

80 60 40 20 0

4000

Ea

Pa Ma ARCHEAN

Pp

Na

Np

Mp

3000

2000

Pz PH

PROTEROZOIC 1000

0

Million years ago

Fig. 4 Diversification curves of life based on analyses presented here. (A) Cumulative curve, showing the total number of lineages at any point in time, based on all timetrees and lineages (1610) in The Timetree of Life (78). The curve samples data in 1-Myr intervals and is not smoothed. (B) Cumulative curve as in A, in 1-Myr intervals, plotted on a log axis (also not smoothed). (C) Rate curve, showing the number of

lineage originations as a percentage of standing diversity (originations/total lineages at that point in time), in 200-Myr intervals. Abbreviations: Ea (Eoarchean), Ma (Mesoarchean), Mp (Mesoproterozoic), Na (Neoarchean), Np (Neoproterozoic), Pa (Paleoarchean), PH (Phanerozoic), Pp (Paleoproterozoic), and Pz (Paleozoic).

lineages living at 300 Ma, we might see the same spike in diversification, but occurring in the preceding 300 Myr (600–300 Ma). This is similar—but not identical—to the “Pull of the Recent” bias in the fossil record thought to be caused by the greater number of fossil sites in more recent times but still debated vigorously (87, 98, 99). Turning now to the last billion years, finer sampling intervals reveal more details in the diversification rate curve (Fig. 5A–C). Surprisingly, it shows a noticeable depression at 250 Ma, which is the time of the

Permian–Triassic extinctions—the largest in the fossil record. The depression corresponds roughly to a 50% decrease in rate of lineage origination. In the finer sample of 10-Myr intervals, it is a low of 2.7% surrounded by peaks of ~7% (≅20 lineages) (Fig. 5B) while in the 20-Myr-interval plot it is a low of 6.7% bordered by peaks of 11–13% (= 32–47 lineages) (Fig. 5C). This Permian– Triassic rate depression is unlikely the result of a bias from fossil calibrations, because molecular time estimates are often considerably older than fossil estimates

12

THE TIMETREE OF LIFE 200

Originations

A

150 P/T

100 50 0

B

Originations (%)

12

8

4

0 20

Originations (%)

C

15 10 5 0 20

Originations (%)

D

Simulated Cambrian Explosion (not observed)

15 10 5 0

Tonian

Cryogenian

Ed

NEOPROTEROZOIC 1000

800

600

C Cm O S D PALEOZOIC 400

P

J Tr K MESOZOIC 200

Pg CZ 0

Million years ago

Fig. 5 Diversification curves of life, for the last 1 billion years, based on analyses presented here. (A) Rate curve, showing the total number of lineage originations in 20-Myr intervals. (B) Rate curve, plotted as a percent of standing diversity (originations/ total lineages) in 10-million-year intervals. (C) Rate curve, plotted as a percent of standing diversity (originations/total lineages) in 20-Myr intervals. (D) Same rate curve as in C, but

showing expected distribution if all animal lineages originated no earlier than the late Precambrian (560 Ma), corresponding to the Cambrian Explosion hypothesis. This distribution was not observed. Abbreviations: C (Carboniferous), Cm (Cambrian), CZ (Cenozoic), D (Devonian), Ed (Edicaran), J (Jurassic), K (Cretaceous), O (Ordovician), P (Permian), Pg (Paleogene), P/T (Permian–Triassic extinctions), S (Silurian), and Tr (Triassic).


(or calibrations), as noted earlier. A detailed analysis of diversification in mammals (100) did not reveal a rate effect corresponding to the Mesozoic–Cenozoic extinction event (66 Ma). Our analysis also did not find any clear rate effect at 66 Ma, although the taxonomic limitations of this data set (families and above) reduced resolution of events during the Cenozoic. If the rate depression at ~250 Ma is an effect of the Permian–Triassic extinctions, it adds increased confidence in the accuracy of molecular time estimates, and hence The Timetree of Life (Fig. 2). The Cretaceous peaks in the rate curves, ~140–100 Ma (Fig. 5A–C) correspond to a time when the supercontinents were breaking up, possibly explaining an increased rate of diversification, as has been suggested for bird and mammal ordinal lineages (19). Lineages adapting to a great diversity of niches within a continent are more likely to survive the vagaries of the extinction process, in the long term, than those diversifying in more localized settings. Alternatively, the Cretaceous peaks may be the product of two artifacts in combination: the Pull of the Present bias causing an increased rate of recent diversification (since ~300 Ma) combined with the family-level taxonomic bias in the data set, causing a decline in rates in the most recent sampled intervals (3000 Ma), before the first peak, represent mostly hyperthermophiles (both superkingdoms) and methanogens (Archaebacteria). All peaks correspond to diversification of Eubacteria. The first is in the late Archean (~2500–2800 Ma), followed by a notable depression (~2500–2000 Ma), and then a cluster of peaks in the late Paleoproterozoic through end of the Precambrian (~1600–600 Ma). The depression corresponds to the Great Oxidation Event (GOE), the first conclusive evidence of a rise in atmospheric oxygen (101, 102). Because oxygen was likely toxic to anoxygenic organisms, we may speculate that the GOE caused a mass extinction event, which led to depressed origination rates. The curve shows that, following the depression, a large amount of diversification occurred during the middle to late Proterozoic, corresponding to the time when eukaryotes were beginning their diversification. Diversification curves for other groups (Fig. 7) largely reflect aspects of their evolution already well established, and discussed in individual chapters (78). This includes differences in the time of onset of diversification and the average age of family lineages. All of these distributions have limitations for drawing conclusions about the evolutionary history of life. Determining whether the variation in diversification rate discussed above is significant may require many more lineages and finer taxonomic sampling. Future syntheses that integrate the fossil record, and include groups that are missing here, will provide a more complete view of diversification. However, the addition of taxa below the family level (genera and species) may not change broad patterns in eukaryotes, except for the last ~100 Myr.

A global repository of divergence times A global consortium maintains public sequence data and alignments but until recently there was no public database for molecular divergence times. We created TimeTree (http://www.timetree.org) (11) to fill this gap. Although the hierarchical nature of the data required complexity

14

THE TIMETREE OF LIFE 10

A

Originations

8 6 4 2 0 B

12

Originations

9 GOE

6 3 0

4000

Ea

Pa Ma ARCHEAN

Pp

Na

Np

Mp

3000

2000

Pz PH

PROTEROZOIC 1000

0

Million years ago

Fig. 6 Diversification rate curve for prokaryotes. (A) Lineage originations (family level and above) sampled in 100-Myr intervals. (B) Lineage originations sampled in 200-Myr intervals. Abbreviations: Ea (Eoarchean), GOE (Great Oxidation Event),

Ma (Mesoarchean), Mp (Mesoproterozoic), Na (Neoarchean), Np (Neoproterozoic), Pa (Paleoarchean), PH (Phanerozoic), Pp (Paleoproterozoic), and Pz (Paleozoic).

in the database design, it also permitted broader utility in the data presented. TimeTree makes use of a single, large, and conservative guide tree (a version of the Tree of Life) and maps divergence times from a diverse array of published timetrees and divergence time estimates. A query to the database consists of asking for the divergence of species (or taxon) A from species B. The results show all of the pertinent published studies and time estimates bearing on that species divergence, and time estimates are summarized in different ways for the user. A good illustration of how the system works is the divergence of cat and dog. After the user searches for those two species, TimeTree identifies the two most-inclusive groups containing those taxa, the suborders Feliformia (cat) and Caniformia (dog) of the mammalian order Carnivora. All published times of divergence between Feliformia and Caniformia are then assembled, because every one traces through the same node. For example, the true divergence time of mongoose (Feliformia) and raccoon (Caniformia) is identical to that of cat and dog. Therefore, any single time estimate between two species might be used for hundreds

or thousands of other pairwise comparisons and, conversely, many time estimates might pertain to the divergence of two species never actually sequenced. This means that every database query involves tree-based computation and analysis rather than simple retrieval of tabular data. In the results, additional information is provided such as error estimates, links to abstracts and sequence data, and summary statistics (unweighted, and weighted by number of genes). This gives the user the opportunity to evaluate the published results and determine, visually, if there is a consensus in the field regarding the time estimate for a node in question. That approach differs from the one taken in The Timetree of Life (78), where experts for each group evaluate the evidence and draw conclusions, sometimes favoring one result over another. Both approaches have their advantages and disadvantages. The number of genes in a study is only one variable of many, some of which are hard to represent in a database. An expert familiar with the literature of a particular group would be the best person to evaluate the different timetrees and time estimates for that group. On the other

Discovering the Timetree of Life Lineages

Originations

600 400 200 0

Plants

150 100 50 0

Plants (excl. angiosperms)

400

Angiosperms

100 50 0 40 20 0 100

200

50

0

0

1000

Animals

150 100 50 0

300 200 100 0

Insects

40

600 400 200 0

Vertebrates

150 100 50 0

100

Amphibians

20

500 0

20 0

50

10

0

0

100

15 10 5 0

Reptiles (excl. birds)

50 0 200

100

Birds

100

50

0

0

200

60 40 20 0

Mammals

100 0

15

O

D

C

PALEOZOIC

500

400

P Tr

J

K

MESOZOIC

300 200 100 Million years ago

O

CZ

D

C

PALEOZOIC

0

500

400

P Tr

J

K

MESOZOIC


CZ

0

Fig. 7 Diversification curves of selected grou ps. Cumulative curves are on left, showing the total number of lineages (family level and above) sampled in 1-Myr intervals; rate curves are on right, showing lineage originations sampled in 20-Myr intervals. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Tr (Triassic).

16


hand, many nodes in the Timetree of Life do not have a relevant expert. For those nodes, and for others where the volume of data has made separate evaluation difficult, a database approach is preferred.

Future prospects Extinct species and groups of taxa need to be incorporated with living taxa to provide a more complete picture of the Timetree of Life. Also, efforts should be made to fi ll in the major gaps in taxonomic coverage. Protists, fungi, invertebrates, and ray-finned fishes are all poorly represented, even though many molecular phylogenies exist for these groups and calibrations are available. The most difficult problems to be solved will be the earliest divergences in prokaryotes and eukaryotes and placement of the root of the tree, both phylogenetically and temporally. Resolution of those questions may occur with the increased number of prokaryote and eukaryote genomes that will be available in coming years along with development of new methods and approaches, but they are difficult problems to solve. Methodology for calibration and time estimation is still in its early stage of development and will likely see major improvements during the next decade. Besides new algorithms and approaches, there is also a great need for the design of user-friendly soft ware (103) to make methods more accessible to the community. The immense value of having a robust Timetree of Life—for all fields of science—cannot be overstated. It will provide a means of estimating rates of change for almost anything biological—for example, morphological structures, behaviors, genes, proteins, non-coding regions of genomes—in any group of organisms. In that sense it will catalyze a Renaissance in comparative biology. For paleontologists, geologists, geochemists, and climatologists, it will provide a biological timeline for comparison, prediction, and synchronization with Earth history. In turn this will help formulate better hypotheses for how the biosphere has evolved on Earth and provide insights into evolutionary mechanisms in the Universe.

Acknowledgment We thank all of the contributors to The Timetree of Life for making such a synthesis possible. This work was supported by grants to S.B.H. from the U.S. National Science Foundation (NSF) and National Aeronautics and Space Administration (Astrobiology Institute) and to S.K. from the NSF and National Institutes of Health.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

16.

17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

C. Darwin, On the Origin of Species by Means of Natural Selection (John Murray, London, 1859). J. Cracraft, M. J. Donoghue, Eds., Assembling the Tree of Life (Oxford University Press, Oxford, 2004). J. D. Watson, F. H. C. Crick, Nature 171, 737 (1953). F. Sanger, Science 129, 1340 (1959). F. Sanger, S. Nicklen, A. R. Coulson, Proc. Natl. Acad. Sci. U.S.A. 74, 5463 (1977). A. M. Maxam, W. Gilbert, Proc. Natl. Acad. Sci. U.S.A. 74, 560 (1977). K. Mullis et al., Cold Spring Harbor Symp. Quant. Biol. 51, 263 (1986). C. R. Woese, G. E. Fox, Proc. Natl. Acad. Sci. U.S.A. 74, 5088 (1977). M. S. Springer et al., Nature 388, 61 (1997). L. M. Hower, S. B. Hedges, Caribb. J. Sci. 39, 298 (2003). S. B. Hedges, J. Dudley, S. Kumar, Bioinformatics 22, 2971 (2006). E. A. Hadly et al., PLOS Biol. 2, 1600 (2004). R. Dawkins, The Ancestor’s Tale: A Pilgrimage to the Dawn of Evolution (Houghton Mifflin, New York, 2004). A. S. Romer, Vertebrate Paleontology (The University of Chicago Press, Chicago, 1966). E. Zuckerkandl, L. Pauling, in Horizons in Biochemistry, M. Marsha, B. Pullman, Eds. (Academic Press, New York, 1962), pp. 189–225. E. Zuckerkandl, L. Pauling, in Evolving Genes and Proteins, V. Bryson, H. J. Vogel, Eds. (Academic Press, New York, 1965), pp. 97–165. H. Hayashida, H. Toh, R. Kikuno, T. Miyata, Mol. Biol. Evol. 2, 289 (1985). S. Kumar, S. B. Hedges, Nature 392, 917 (1998). S. B. Hedges, P. H. Parker, C. G. Sibley, S. Kumar, Nature 381, 226 (1996). V. M. Sarich, A. C. Wilson, Science 158, 1200 (1967). G. A. Wray, J. S. Levinton, L. H. Shapiro, Science 274, 568 (1996). M. Kimura, Nature 217, 624 (1968). E. Margoliash, Proc. Natl. Acad. Sci. U.S.A. 50, 672 (1963). D. Hartl, D. Dykhuizen, Nature 281, 230 (1979). M. Goodman, G. Braunitzer, A. Stangl, B. Schrank, Nature 303, 546 (1983). S. Kumar, Nat. Rev. Genet. 6, 654 (2005). W.-H. Li, Molecular Evolution (Sinauer Associates, Sunderland, Massachusetts, 1997). M. Nei, Molecular Evolutionary Genetics (Columbia University Press, New York, 1987). C. Sibley, J. Ahlquist, J. Mol. Evol. 20, 2 (1984). T. D. Kocher et al., Proc. Natl. Acad. Sci. U.S.A. 86, 6196 (1989). V. M. Sarich, A. C. Wilson, Proc. Natl. Acad. Sci. U.S.A. 58, 142 (1967).


32. F. Tajima, Genetics 135, 599 (1993). 33. N. Takezaki, A. Rzhetsky, M. Nei, Mol. Biol. Evol. 12, 823 (1995). 34. M. Hasegawa, H. Kishino, T. Yano, J. Hum. Evol. 18, 461 (1989). 35. C. D. Schubart, R. Diesel, S. B. Hedges, Nature 393, 363 (1998). 36. S. B. Hedges, S. Kumar, Trends Genet. 19, 200 (2003). 37. S. B. Hedges, S. Kumar, Trends Genet. 20, 242 (2004). 38. M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997). 39. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 40. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 41. A. J. Drummond, S. Y. W. Ho, M. J. Phillips, A. Rambaut, PLOS Biol. 4, 699 (2006). 42. T. Britton, C. L. Anderson, D. Jacquet, S. Lundqvist, K. Bremer, Syst. Biol. 56, 741 (2007). 43. W. M. Brown, M. George, A. C. Wilson, Proc. Natl. Acad. Sci. U.S.A. 76, 1967 (1979). 44. G. F. Shields, A. C. Wilson, J. Mol. Evol. 24, 212 (1987). 45. M. P. Heinicke, W. E. Duellman, S. B. Hedges, Proc. Natl. Acad. Sci. U.S.A. 104, 10092 (2007). 46. P. E. Ahlberg, J. A. Clack, Nature 440, 747 (2006). 47. E. B. Daeschler, N. H. Shubin, F. A. Jenkins, Nature 440, 757 (2006). 48. S. A. Magallon, M. J. Sanderson, Evolution 59, 1653 (2005). 49. M. J. Sanderson, Am. J. Bot. 90, 954 (2003). 50. J. Müller, R. R. Reisz, BioEssays 27, 1069 (2005). 51. R. R. Reisz, J. Müller, Trends Genet. 20, 237 (2004). 52. S. B. Hedges, S. Kumar, M. van Tuinen, BioEssays 28, 770 (2006). 53. M. J. Benton, P. C. J. Donoghue, Mol. Biol. Evol. 24, 26 (2007). 54. Z. H. Yang, B. Rannala, Mol. Biol. Evol. 23, 212 (2006). 55. M. J. Benton, C. J. Donoghue, R. J. Asher, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 35–86. 56. P. C. J. Donoghue, M. J. Benton, Trends Ecol. Evol. 22, 424 (2007). 57. S. Conway Morris, Nature 361, 219 (1993). 58. J. W. Valentine, D. Jablonski, D. H. Erwin, Development 126, 851 (1999). 59. G. E. Budd, N. J. Butterfield, S. Jensen, Science 294, U1 (2001). 60. S. B. Hedges, F. U. Battistuzzi, J. E. Blair, in Neoproterozoic Geobiology and Paleobiology, S. Xiao, A. J. Kaufman, Eds. (Springer, New York, 2006), pp. 199–229. 61. G. A. Wray, Genome Biol. 3, 1 (2001). 62. J. E. Blair, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 223–230. 63. S. Aris-Brosou, Z. Yang, Syst. Biol. 51, 703 (2002). 64. S. Aris-Brosou, Z. Yang, Mol. Biol. Evol. 20, 1947 (2003).

65.

66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79.

80. 81. 82. 83.

84. 85. 86. 87. 88. 89.

90. 91.

92. 93. 94.

17

E. J. P. Douzery, E. A. Snell, E. Bapteste, F. Delsuc, H. Philippe, Proc. Natl. Acad. Sci. U.S.A. 101, 15386 (2004). K. J. Peterson et al., Proc. Natl. Acad. Sci. U.S.A. 101, 6536 (2004). K. J. Peterson, N. J. Butterfield, Proc. Natl. Acad. Sci. U.S.A. 102, 9547 (2005). K. J. Peterson, J. A. Cotton, J. G. Gehling, D. Pisani, Philos. Trans. Roy. Soc. Lond. B 363, 1435 (2008). J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 387 (2005). L. A. Hug, A. J. Roger, Mol. Biol. Evol. 24, 1889 (2007). A. J. Roger, L. A. Hug, Philos. Trans. Roy. Soc. Lond. B 361, 1039 (2006). S. Kumar, A. Filipski, V. Swarna, A. Walker, S. B. Hedges, Proc. Natl. Acad. Sci. U.S.A. 102, 18842 (2005). D. E. Canfield, S. W. Poulton, G. M. Narbonne, Science 315, 92 (2007). A. B. Smith, J Geol Soc 164, 731 (2007). M. J. Benton, Science 268, 52 (1995). M. Foote, J. P. Hunter, C. M. Janis, J. J. J. Sepkoski, Science 283, 1310 (1999). S. Tavaré, C. R. Marshall, O. Will, C. Soligo, R. D. Martin, Nature 416, 726 (2002). S. B. Hedges, S. Kumar, Eds., The Timetree of Life (Oxford University Press, New York, 2009). J. C. Avise, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 19–25. J. C. Avise, G. C. Johns, Proc. Natl. Acad. Sci. U.S.A. 96, 7358 (1999). M. J. Benton, The Fossil Record 2 (Chapman & Hall, London, 1993). A. F. Hugall, R. Foster, M. S. Y. Lee, Syst. Biol. 56, 543 (2007). C. A. Matthee, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 487–489. J. J. Sepkoski, Phil. Trans. Roy. Soc. Lond. B 353, 315 (1998). A. B. Smith, Systematics and the Fossil Record (Blackwell Scientific Publications, London, 1994). M. J. Benton, Trends Ecol. Evol. 12, 490 (1997). M. J. Benton, B. C. Emerson, Palaeontology 50, 23 (2007). S. M. Porter, Paleon. Soc. Pap. 10, 35 (2004). S. M. Porter, in Neoproterozoic Geobiology and Paleobiology, S. Xiao, A. J. Kaufman, Eds. (Springer, Dordrecht, The Netherlands, 2006), pp. 1–21. J. W. Valentine, Palaeontology (Oxford) 12, 684 (1969). P. W. Signor, in Phanerozoic Diversity Patterns: Profiles in Macroevolution, J. W. Valentine, Ed. (Princeton University Press, Princeton, 1985). D. M. Raup, Science 177, 1065 (1972). J. J. Sepkoski, Paleobiology 4, 223 (1978). J. J. Sepkoski, Paleobiology 5, 222 (1979).

18


95. J. J. Sepkoski, Paleobiology 10, 246 (1984). 96. R. A. Berner, D. J. Beerling, R. Dudley, J. M. Robinson, R. A. Wildman, Ann. Rev. Earth Planet. Sci. 31, 105 (2003). 97. R. A. Berner, Geochim. Cosmochim. Acta 70, 5653 (2006). 98. J. Alroy et al., Proc. Natl. Acad. Sci. U.S.A. 98, 6261 (2001).

99. 100. 101. 102. 103.

M. J. Benton, M. A. Wills, R. Hitchin, Nature 403, 534 (2000). O. R. P. Bininda-Emonds et al., Nature 446, 507 (2007). D. Canfield, Ann. Rev. Earth Planet. Sci. 33, 1 (2005). H. D. Holland, Geochim. Cosmochim. Acta 21, 3811 (2002). S. Kumar, J. Dudley, Bioinformatics 23, 1713 (2007).

Timetrees: beyond cladograms, phenograms, and phylograms John C. Avise Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92697, USA ([email protected])

Abstract For several historical reasons discussed herein, until recently the absolute temporal dimension of many phylogenetic trees has been relatively ignored whereas the branching (cladistic) aspect typically has been the focus of most phylogeny-reconstruction efforts. This unfortunate neglect of “timetrees” is now being remedied, as this book will attest. Many scientific benefits can emerge from superimposing robust estimates of geological time on cladograms, including opportunities to: improve phylogenetic reconstructions of phenotypic evolution; illuminate causal geological or other events and processes in the history of life; and develop a universal time-standardized framework for biological classification that will facilitate studies in comparative evolution.

. . . the extent of variation of the primary structure [of proteins] . . . may give rough approximations of the time elapsed since the lines of evolution leading to any two species diverged. —Emanuel Margoliash (1963).

The notion that genetic differences between species tend to increase with time has been prevalent since the inception of the field of molecular evolution. Margoliash’s comment above was in reference to early empirical data for cytochrome c showing that horse and pig differ at only three amino acid sites, whereas horse and tuna display 19 amino acid substitutions and horse and yeast display at least 44 such changes. From such observations, Margoliash (1) concluded: “relatively closely related species show few [genetic] differences . . . phylogenetically distant species exhibit wider dissimilarities.” Zuckerkandl and Pauling (2) had noted similar kinds of outcomes in four members of the hemoglobin protein family in 1962, and in 1965 they coined the term “molecular clock” to encapsulate the notion that protein

sequences appear to diverge with some regularity across evolutionary time (3). During the ensuing four decades, the general time-dependent nature of molecular evolution (but not always any great precision for particular molecular clocks) gained voluminous support from studies on a wide variety of proteins and subsequently of DNA sequences (reviewed in 4).

Stepped cladogram Pig Horse Tuna Yeast

Distance phenogram Pig Horse Tuna Yeast 20

10

0

Genetic distance (AA differences)

Evolutionary timetree

1200

600

Pig Horse Tuna Yeast 0

Evolutionary time (millions of years)

Fig. 1 Alternative phylogenetic depictions for pig, horse, tuna, and yeast based on the cytochrome c sequence data considered by Margoliash (1; see text). Top panel: stepped cladogram showing only the cladistic order of phylogenetic nodes (branch lengths have no meaning). Middle panel: distance phenogram showing branch lengths (amino acid changes) in addition to branching topology (thereby making this depiction a phylogram also; see text). Bottom panel: an evolutionary timetree showing estimates of nodal dates in addition to the branching topology.

J. C. Avise. Timetrees: beyond cladograms, phenograms, and phylograms. Pp. 19–25 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

20


The word phylogeny, from the Greek roots “phyl” meaning tribe and “geny” meaning origin, refers to the genealogical history of life. Given the early recognition of molecular clocks and the vast popularity of molecular systematics since the mid-1960s (5), it might seem that phylogeneticists would have been preoccupied with dating evolutionary trees, that is, in capitalizing upon the wealth of temporal (as well as cladistic information) seemingly inherent in the primary structures of proteins and nucleic acids. Ironically, however, phylogenetic chronograms (“timetrees,” for short) have been relatively neglected until recently in most molecular studies in favor of efforts instead to reconstruct merely the branching topologies (cladistic structures) of phylogenetic trees. In support of this contention, a recent compendium on the “tree of life” (6) included nearly 200 phylogenetic branching diagrams, fewer than 10 of which (ca. 5%) provided explicit indications of the absolute dates of evolutionary nodes. If tallies likewise were to be conducted on the phylogenetic representations published during the last four decades in leading journals such as Systematic Biology, Molecular Phylogenetics and Evolution, or Evolution, a similar neglect of evolutionary dates in phylogenetic trees would undoubtedly be evidenced. This state of affairs is highly ironic, because timetrees can, in principle, encapsulate far more information about evolutionary history than do cladograms or phenograms (Fig. 1). Here, I will speculate on why timetrees have been relatively neglected in biological systematics (at least until very recently), and, more importantly, why this situation can and should be remedied.

Past neglect of timetrees Historical reasons for the relative disinterest in timetrees (compared to the enthusiasm for cladograms) probably entail both scientific and sociological factors. Here I conjecture on why timetree reconstructions have not always been pursued with vigor. Cladistic–phenetic distractions

The rise of molecular biology that gave birth to molecular phylogenetics in the early 1960s happened to coincide with two seminal developments in traditional (i.e., nonmolecular) systematics: the publication in 1963 of Principles of Numerical Taxonomy by Sokal and Sneath (7), and the 1966 translation into English of Hennig’s

Phylogenetic Systematics (8). These books gave rise, respectively, to the polarized schools of phenetics and cladistics that were to dominate philosophical discourse in systematics for more than two decades (see chapter 4 in 9). Pheneticists argued that organisms should be grouped and classified according to their overall similarity as measured by defined rules using as many quantifiable traits as possible. Results were summarized as phenograms depicting phenetic similarities (not necessarily phylogenetic relationships) among taxa. Cladists countered that organisms should be grouped and classified according explicitly to their evolutionary relationships as evidenced by (even a few) shared-derived traits, that is, synapomorphies. Results were summarized as cladograms depicting cladistic topologies in phylogenetic trees. Especially in the 1960s through the 1980s, the young field of molecular evolution found itself distracted by (and sometimes immersed in) the cladistic–phenetic wars. On one battlefront, some molecular systematists were forced to defend their approaches against hard-core cladists who automatically discredited any “phenetic” method (e.g., DNA–DNA hybridization) that merely yielded genetic similarity or distance estimates between taxa. Many cladists also impugned any statistical clustering algorithms for phylogenetic inference (e.g., UPGMA or neighbor-joining, 10) that employed composite genetic distance estimates. On entirely another battlefront, molecular systematists were sometimes forced to counter the notion by a few hard-core pheneticists that phylogenetic reconstruction itself was not one of the primary achievable goals of systematics. Interestingly, neither the pheneticists nor the cladists devoted much attention to how absolute time of evolutionary separation between taxa might be extracted from empirical information (molecular or otherwise). Thus, the relatively few timetrees estimated from molecular or other data were not readily accepted into either the traditional phenetic or cladistic camps. This is ironic for the following reasons. In one important sense, various molecular approaches approximate a cladistic ideal, because the data are genetic and the volume of molecular information can be so vast as to provide a strong collective signal regarding branching topologies (as well as branch lengths) in phylogenetic trees. In another sense, however, various molecular approaches also closely approximate a phenetic ideal, because the number of assayed traits can be huge (up to millions of nucleotide positions in some current DNA sequence comparisons) and the data are inherently

Beyond Cladograms, Phenograms, and Phylograms

quantifiable in terms of overall similarities or distances among taxa. Furthermore, given that evolutionary convergences or reversals at individual nucleotide positions can be rather common (in part because only four interconvertible character states exist per site), any cladistic analysis that focuses unduly on any few presumptive shared-derived characters in molecular data could be inappropriate. Thus, molecular phylogenetic approaches are neither cladistic nor phenetic exclusively, but rather they can encompass some of the best of both worlds. Irrespective of the method of analysis (e.g., via distance-based, parsimony, maximum likelihood, or Bayesian methods), large molecular data sets can yield phylograms that portray both the cladogenetic and anagenetic components of phylogeny. However, the main thesis of this chapter is that phylogeneticists should now strive for even more information by estimating absolute dates for evolutionary nodes. To the extent that this task is successfully accomplished, the resulting timetrees will convey much more information than traditional cladograms, phenograms, or phylograms. Reservations about molecular clocks

A second set of historical obstacles to greater enthusiasm for timetrees involved some misunderstandings about molecular clocks. On the conceptual side, the notion of molecular clocks meshed well with an emerging neutrality theory predicting that molecular evolutionary rates were driven by (indeed were equitable with) rates of neutral mutation (11). Accordingly, praise or condemnation of molecular clocks often hinged on a researcher’s philosophical stance in the broader neutralist–selectionist debate. But this situation was inappropriate because molecular clocks are also compatible with most selectionist scenarios. For example, if large numbers of genes are acted upon by multifarious selection pressures over long periods of time, then any short-term fluctuations in selection intensity might average out such that the overall genetic distances among taxa correlate well with elapsed times since common ancestry. On the empirical side, any hopes for a universal molecular timepiece of great precision were dashed as empirical data accumulated showing significant variation in molecular rates at several levels: among nucleotide positions within codons, among nonhomologous genes within a lineage, among various classes of DNA such as coding and noncoding, and between full genomes such as nuclear and mitochondrial (review in 12). Seemingly

21

most damning to universal clock arguments were rate differences also reported at homologous loci among different taxonomic lineages (e.g., 13–16). Some systematists interpreted such findings to doom all efforts to estimate timetrees from molecular data, but this reaction was overly negative because, in principle and often in practice, rate variation (as well as complicating phenomena such as saturation effects in the nucleotide substitution process) can be recognized and accommodated in timetree reconstructions (e.g., 17–19). Thus, contrary to a widespread sentiment, attempts to date phylogenetic nodes do not necessarily hinge critically on the precise ticking of molecular clocks. Furthermore, many interesting questions in evolutionary biology can be answered (at least to a first approximation) with even ballpark temporal estimates. In addition, pronounced rate variation is likely to be greater on local than on global genomic scales. A useful analogy might be to the gas laws in physics. Much as the individual molecules within a gas have idiosyncratic and unpredictable movements in submicroscopic space, individual nucleotide sites and loci within a genome can display idiosyncratic and unpredictable changes over evolutionary time. But large collections of gas molecules also have consistent composite properties (summarized by the gas laws of physics), and likewise large collections of DNA nucleotides may have fairly consistent conglomerate behaviors such as mean rates of sequence evolution. In other words, with regard to temporal signal, DNA sequences are detail-noisy yet composite-rich. A final and oft-overlooked point is that estimating timetrees is not an endeavor exclusively for the field of molecular evolution. To the contrary, nodes in phylogenetic trees are most securely dated when multiple lines of evidence (from paleontology, historical geography, comparative biology, etc.) are also thoroughly integrated into the analysis. Indeed, every initial calibration of a molecular clock requires at least one absolute temporal landmark (e.g., from fossil or biogeographic evidence) independent of the molecular data. Thus, if objections are to be raised against the promotion of timetrees, they should not stem from a (misplaced) sentiment that traditional nonmolecular approaches to systematics are thereby somehow being excluded.

Rationales and prospects for timetrees Having speculated on why time–time approaches have been relatively neglected, I want to suggest why this situation can and should change. First, molecular data

22


are now being gathered and phylogenetically analyzed for thousands of species, and major structural features plus many finer details of the Tree of Life are quickly emerging. Second, because these newest phylogenetic appraisals are typically based on unprecedented volumes of sequence information (sometimes of entire genomes), the resulting phylogenetic estimates should be almost as secure as might ever become possible. Third, cladogenetic topologies and anagenetic branch lengths (alone or together) paint incomplete phylogenetic pictures unless secure estimates of nodal dates are included in the representations as well. Thus, the time is right for biologists to begin exploring much more fully the absolute temporal dimensions of the phylogenetic trees they reconstruct. Secure timetrees will then offer many potential added benefits to biology, including the following. Character mappings will be enriched

All branches of comparative biology could profit from better knowledge about the evolutionary histories of morphological, behavioral, physiological, biochemical, or other phenotypic traits. In recent years, phylogenetic character mapping (PCM) has become a wildly popular exercise wherein scientists use molecular trees as historical backdrops for deciphering the evolutionary pathways traversed by all sorts of phenotypic traits in a wide variety of plant, animal, and microbial taxa (review in 20). Well-supported phylograms are especially useful in PCM reconstructions via maximum likelihood analyses, often yielding insights about ancestral character states that would not be apparent from maximum parsimony reconstructions based on cladogram structures alone (e.g., 21). Furthermore, time is the common denominator in all rate estimates, so assessments of both absolute and relative evolutionary rates in various phenotypic (and molecular) traits are obviously facilitated when well-dated nodes in the relevant timetrees are available. PCM exercises have illustrated two broader ironies about traditional systematics. First, for many taxa, extensive discussions and debates have often centered on fine details of alternative branching orders within particular cladograms, while the no-less-important temporal contexts of the phylogenies were often virtually neglected. Second, whereas phylogenetic treatments have traditionally focused on particular taxa one at a time, much of interest can emerge from comparative assessments (for which molecular data are uniquely well suited) across even disparate taxonomic groups. In short, the regular

presentation of explicit timetrees will open many conceptual worlds for fruitful discussions in comparative biology. Consider, for example, how the identical-stepped cladograms in Fig. 2 take on new meaning and raise novel questions when properly stretched to reflect their relative temporal dimensions. Causal historical events will be illuminated

Another benefit of timetrees is that they will inevitably help to focus scientific attention on causal processes underlying phylogenetic diversity. Several examples are evident in this book, but to introduce the broader argument I will briefly mention two additional cases here. Interest has long centered on whether ancient vicariant events (related to plate tectonic movements) or subsequent over-water dispersals (e.g., by rafting) account for the presence of various terrestrial vertebrates in the West Indies. Geologic evidence indicates that the Greater Antilles formed in close proximity to North and South America during the mid-Cretaceous and that these islands began drifting from their continental partners at least 80 million years ago (Ma). The vicariance scenario thus predicts that sister clades on the islands vs. mainlands separated more than 80 Ma, whereas dispersal

Hominidae

Human Chimpanzee Gibbon

Hylobatidae Trionychidae

Soft-shelled Flap-shelled Pig-nosed turtle

Carettochelyidae

180

120

60

0

Millions of years ago

Fig. 2 Example of widely differing temporal frameworks for cladograms with identical branching sequence. Above: stepped cladograms for four species representing two taxonomic families of primates and for four species representing two families of marine turtles. Below: timetrees for these respective assemblages [estimates of nodal evolutionary dates are provisional].


scenarios predict that the separations were more recent and probably variable in time. From molecular phylogenetic appraisals of nodal dates for more than 35 relevant pairs of vertebrate taxa, Hedges (22) effectively falsified the ancient vicariance hypothesis for these faunas. Another illustrative example of an informative timetree is reproduced in Fig. 3. By comparing extensive sequence data from several nuclear and mitochondrial genes in species representing 139 genera and nearly all higher taxa of extant ants, and by incorporating dates of relevant fossils into the phylogenetic analyses, Moreau et al. (23) concluded that an early evolutionary radiation of ant lineages coincided with the great proliferation of angiosperm plants (as well as coleopteran and hemipteran insects) during the Late Cretaceous (100–66 Ma). These findings, which indicate that ants diversified much earlier than previously supposed, are generating exciting

Time calibration using minimum fossil ages 0 Ma

50 Ma

100 Ma

Myrmicinae

K-P boundary

100 Ma

50 Ma

0 Ma

Time calibration using maximum fossil ages

Fig. 3 Timetree for 52 genera (each right-terminal node is a different genus) in the ant subfamily Myrmicinae after Moreau et al. (23). This diagram includes only a representative subset of the total of 139 ant genera (in 19 subfamilies) included in the broader phylogenetic study by Moreau and colleagues. Note the evident proliferation of lineages before the Cretaceous– Paleogene (K-P) boundary (see text).

23

new hypotheses about what biological factors might have been responsible. For example, ground litter favorable for ants is highly diverse in angiosperm forests and may have provided many novel habitat opportunities; and/or, the expansion of herbivorous insects at about that same time may have offered diverse food resources ripe for exploitation by the newly diversifying ants. Biological classifications can be universally standardized

If and when secure timetrees become widely available, opportunities will also arise to develop the first ever universally standardized scheme of biological classification. Unfortunately, current classifications provide no assurance that one taxonomic genus or family of mammals, for example, is comparable to another, much less to a genus or family of fishes, insects, or beetles. Indeed, no standards have been adopted by which such assurances might even be attempted. Another aspect of inconsistency in current classifications is that whereas many taxa are valid clades, others are polyphyletic or paraphyletic grades (see, e.g., the top of Fig. 1), and the nomenclature gives no indication which is which. Hennig (8) bemoaned these states of affairs when he wrote, “If systematics is to be a science it must bow to the self-evident requirement that objects to which the same label is given must be comparable in some way.” Others have echoed similar thoughts: “No scientific enterprise, least of all one that considers the promotion of nomenclatural universality as one of its primary objectives, can accept the inconsistencies and ambiguities current in biological taxonomy” (24). Although systematists readily admit that the biological classifications in use today are wildly nonuniform across disparate taxonomic groups, little has been accomplished to rectify this huge flaw. This unfortunate situation stems in part from the lack of biological universality in the morphological, physiological, or other phenotypes that systematists traditionally use to classify organisms. For example, traits conventionally employed in fish systematics (number of lateral-line scales, fins placements, etc.) often have no useful counterparts in mammals or insects, one net result being that few systematists seem to have given much thought to how a universally standardized taxonomy might be erected or used. By contrast, many DNA and protein molecules are more or less universal and could in principle serve as common yardsticks for biological classifications. However, different lineages show rather variable evolutionary rates and patterns (see earlier), so

24


molecular sequences per se still do not provide an ideal standard for cross-taxon comparisons. Instead, evolutionary time itself is the ultimate gold standard for developing any universally consistent framework for biological classification. Avise and Johns (25) detail the rationale for this sentiment and also suggest a logical means for implementing it (at least in principle). Their approach, following Hennig (8), involves the notion of “temporal banding.” For any securely dated timetree, temporal windows or bands of defined but arbitrary width (ideally sanctioned by a consensus of systematists) would be superimposed. By definition, each window would correspond to a particular rank in the conventional Linnaean hierarchy (such as genus, family, or order), and extant members of each clade would automatically be assigned a taxonomic rank empirically defined by the evolutionary window that contains the stem node. Under this temporal-banding proposal, all named taxa would be clades, and all clades of a given taxonomic rank would have originated within the same window of evolutionary time. Such clades could be of mammals, fishes, insects, corals, or any other kinds of creature for which timetrees become securely estimated. Assuming that timetrees for many or most forms of life will eventually emerge from systematists’ collective efforts, an implementation of the temporal-banding protocol could yield biology’s first universally standardized classification. However, a major difficulty with this proposal (in its original formulation) concerns the large number of rank shifts and nomenclatural changes that likely would be entailed. Taxonomic stability is also very important in systematics (26), so any wholesale transformation of classification in the name of global standardization could be counterproductive if it complicated more so than facilitated the communication and retrieval of biological information. To overcome this problem, we recently introduced an addendum to the original temporal-banding proposal (20). Rather than transmogrify existing classifications to conform to temporal windows, the temporal window for each clade could simply be indicated, in shorthand, by an appropriate time-clip attached to the existing taxon name. Each time-clip would refer to the specific temporal window (such as a geological period or epoch) within which the clade is thought to have arisen. Thus, the timeclip proposal would provide a practical way to retain the familiar Linnaean system and traditional taxon names yet simultaneously incorporate important temporal information into existing classifications.

Conclusion: Time is of the essence For all of these reasons, the time seems right to focus much greater attention on the temporal component (in addition to cladistic component) of phylogenetic trees. The routine estimation and utilization of timetrees could add exciting new dimensions to biology, including enhanced opportunities to integrate large molecular data sets with fossil and biogeographic evidence (and thereby foster greater communication between molecular and traditional systematists); estimate not only ancestral character states but also evolutionary rates in numerous categories of organismal phenotype; help establish more reliable associations between causal historical processes and biological outcomes; develop a universally standardized scheme for biological classifications; and, in general, promote novel avenues of thought in many arenas of comparative evolutionary biology.

References 1. E. Margoliash, Proc. Natl. Acad. Sci. U.S.A. 50, 672 (1963). 2. E. Zuckerkandl, L. Pauling, in Horizons in Biochemistry, M. Kasha, B. Pullman, Eds. (Academic Press, New York, 1962), pp. 189–225. 3. E. Zuckerkandl, L. Pauling, in Evolving Genes and Proteins, V. Bryson, H. J. Vogel, Eds. (Academic Press, New York, 1965), pp. 97–166. 4. S. Kumar, Nat. Rev. Genet. 6, 654 (August 2005). 5. D. M. Hillis, C. Moritz, B. K. Mable, Molecular Systematics (Sinauer Associates, Sunderland, Mass., 1996). 6. J. Cracraft, M. J. Donoghue, Assembling the Tree of Life (Oxford University Press, New York, 2004). 7. R. R. Sokal, P. H. A. Sneath, Principles of Numerical Taxonomy (W. H. Freeman, San Francisco, 1963). 8. W. Hennig, Phylogenetic Systematics (University of Illinois Press, Urbana, 1966). 9. J. C. Avise, Molecular Markers, Natural History, and Evolution (Sinauer Associates, Sunderland, Mass., 2004). 10. M. Nei, S. Kumar, Molecular Evolution and Phylogenetics (Oxford University Press, New York, 2000). 11. M. Kimura, Nature 217, 624 (1968). 12. W.-H. Li, Molecular Evolution (Sinauer Associates, Sunderland, Mass., 1997). 13. F. J. Ayala, Proc. Natl. Acad. Sci. U.S.A. 94, 7776 (1997). 14. X. Gu, W.-H. Li, Mol. Phylo. Evol. 1, 211 (1992). 15. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 16. D. M. Rand, Trends Ecol. Evol. 9, 125 (1994). 17. M. Hasegawa, J. L. Thorne, H. Kishino, Genes Genet. Syst. 78, 267 (2003). 18. M. S. Springer, W. J. Murphy, E. Eizirik, S. J. O’Brien, Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003).


19. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 20. J. C. Avise, Evolutionary Pathways in Nature: A Phylogenetic Approach (Cambridge University Press, Cambridge, 2006). 21. D. Schluter, T. Price, A. O. Mooers, D. Ludwig, Evolution 51, 1699 (1997). 22. S. B. Hedges, Ann. Rev. Ecol. Syst. 27, 163 (1996).

23.

25

C. S. Moreau, C. D. Bell, R. Vila, S. B. Archibald, N. E. Pierce, Science 312, 101 (2006). 24. K. deQueiroz, J. Gauthier, Ann. Rev. Ecol. Syst. 23, 449 (1992). 25. J. C. Avise, G. C. Johns, Proc. Natl. Acad. Sci. U.S.A. 96, 7358 (1999). 26. E. Mayr, The Growth of Biological Thought: Diversity, Evolution, and Inheritance (Belknap Press, Cambridge, Mass., 1982).

The geologic time scale Felix M. Gradsteina,* and James G. Oggb a

Museum of Natural History, University of Oslo, N-0318 Oslo, Norway; bDepartment of Earth and Atmospheric Sciences, Purdue University, West Lafayette, IN 47907, USA *To whom correspondence should be addressed (felix.gradstein@ nhm.uio.no)

Abstract Construction and assembly of the Geologic Time Scale involves: (a) constructing a relative (chronostratigraphic) standard scale for key periods in the Earth’s rock record; (b) identifying high-resolution linear age dates to calibrate this relative scale in linear time; (c) astronomically tuning intervals with cyclic sediments or stable isotope sequences which have sufficient fossil or geomagnetic ties to be merged in the standard scale, and increase its resolution; (d) interpolating for those relative time intervals where direct linear age information is insufficient; and (e) estimating error bars on the age of boundaries and on unit durations.

Time is an indispensable tool for all of us. The time kept by innumerable watches and a great variety of clocks regulates our everyday life, while the familiar calendar governs our weekly, monthly, and yearly doings. These eventually condense into the historical record of the events over centuries. The standard unit of modern time keeping is the second, defined by a precise number of vibrations of the cesium atomic clock. The atomic second is defined as the duration of 9.192.631.770 periods of the radiation corresponding to the transition between two hyperfine levels of the ground state of the Cesium 133 atom. This value was established to agree as closely as possible with the ephemeris second based on the Earth’s motion. The advantage of having the atomic second as the unit of time in the International System of Units is the relative ease, in theory, for anyone to build and calibrate an atomic clock with a precision of 1 part per 1011 (or better). In practice, clocks are calibrated against broadcast time signals, with frequency oscillations in Hertz being the “pendulum” of the atomic time-keeping device. The tick of the second paces the quick heart beat, and traditionally was the 60th part of the 60th part of the 24th part of the 24-h day, with the minute and the hour

being convenient multiples to organize our daily life and productivity. The day carries the record of light and dark, the month the regularly returning shapes of the moon, and the year the cycle of the seasons and the apparent path of the sun. All is clear, and we have grown up with the notion that time is a vector, pointing from the present to the future. Events along its path mark the arrow of time, and the arrow is graded either in relative “natural” units, or in units of duration—the standard second and its multiples, like hours and years, and millions of years.

Geologic time and the sediment record A majority of geologists consider time as a vector pointing from the distant past to the present. Instead of “distant past,” the term “deep time” has been coined in the vernacular. What is exactly the concept of geologic time, what are its natural units, how are they defined, and how do we use these units properly? A good understanding of geologic time is vital for every scientist who deals with events in the Earth sediment and rock record, or with the genetic record of evolution in living organisms, especially those who strive to understand past processes and determine rates of change. This understanding takes place in a framework called Earth Geological History, a super calendar of local and global events. The challenge to this understanding is reading, organizing, and sorting the Earth’s stone calendar pages. In the process, we often have to reconstruct the content of missing pages. Correlation of the rock record between regions is a vital part of the reconstruction process. One of the earliest reconstructions is by Nicolas Steno (1631–1687) who made careful and original stratigraphic observations. Based on these observations, Steno concluded that the Earth’s strata contain the superimposed records of a chronological sequence of events that can be correlated worldwide. Geological correlation formally is expressed in terms of five consecutive operations (each is followed by one or more examples): (a) Rock units, like formations or well log intervals = lithostratigraphic correlation Kimmeridge Clay Formation of England

F. M. Gradstein and J. G. Ogg. The geologic time scale. Pp. 26–34 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

The Geologic Time Scale

(b) Fossil units, like zones = biostratigraphic correlation Turrilina alsatica benthic foraminifer zone (c) Relative time units = geochronologic (“Earth time”) correlations Jurassic Period, Eocene Epoch, Oxfordian Age, polarity chron C29r (d) Rocks deposited during these time units = chronostratigraphic (time–rock) correlation Jurassic System, Eocene Series, Oxfordian Stage, polarity zone C29r (e) Linear time units or ages = geochronologic correlation 150 million years ago (Ma), 10,000 years ago (ka) Without correlation to a global reference scale, successions of strata or events in time derived in one area are unique and contribute nothing to the understanding of Earth history elsewhere. The rules of hierarchy in geological correlation, from rocks and fossils to relative and linear time, are carefully laid down in the International Stratigraphic Guide. An abbreviated copy of this “rule book” with further references may be found on the Web site of the International Commission on Stratigraphy (ICS) under www.stratigraphy.org. Before we deal with linear geological time, a few words are necessary about the common geological calendar built from relative age units. This chronostratigraphic scheme is not unlike a historical calendar in which societal periods, for example, the Minoan Period, the reign of Louis XIV, the American Civil War, are used as building blocks, devoid of a linear scale. Archeological relics deposited during these intervals (e.g., the Palace of Minos on Crete, Versailles or spent cannon balls at Gettysburg, respectively) comprise the associated physical and chronostratigraphic record. A chronostratigraphic scale is assembled from rock sequences stacked and segmented in relative units based on their unique fossil and physical content. When unique local fossil and physical records are matched with those of other rock sequences across the globe—in a process known as stratigraphic correlation—a relative scale can be assembled that, when calibrated to stage type sections, becomes a chronostratigraphic scale. The standard chronostratigraphic scale, in downloadable graphics format, is available from the ICS Web site. This time scale is made of successive stages in the rock record, like Cenomanian, Turonian, Coniacian, and so on, within the Cretaceous system. Originally, each stage unit was a well-defined body of rocks at a specific location of an assigned and agreed upon relative age span, younger than typical rocks of the

27

underlying stage and older than the typical rocks of the next higher stage. This is the concept of defining stage units with type sections, commonly referred to as stratotype sections. The principles and building blocks of this chronostratigraphy were slowly established during centuries of study in many discontinuous and incomplete outcrop sections. Inevitably, lateral changes in lithology between regions and lack of agreement on criteria, particularly in which fossils were characteristics of a relative unit of rock, have always resulted in a considerable amount of confusion and disagreement on stage nomenclature and stage use. Almost invariably classical stage stratotypes turned out to only represent part of stages. Hence, a suite of global subdivisions with precise correlation horizons was required.

Global stratotype section and point Now, relatively rapid progress is being made with definition of Global Stratotype Sections and Points (GSSPs) to fi x the lower boundary of all geologic stages, using discrete fossil and physical events that correlate well in the rock record. For the ladder of chronostratigraphy, this GSSP concept switches the emphasis from marking the spaces between steps (stage stratotypes) to fi xing the rungs (boundaries of stages). Each progressive pair of GSSPs in the rock record also precisely defines the associated subdivision of geologic time. It is now 25 years ago that a “golden spike” struck the first GSSP. This event of historic proportions for the geologic time scale involved the boundary between the Silurian and Devonian Periods, or rather the lower limit of the Devonian, at a locality called Klonk in Czechoslovakia. The problem of the Silurian–Devonian boundary and its consensus settlement in the Klonk section hinged on a century-old debate known as the “Hercynian Question” that touched many outstanding geoscientists of the nineteenth century. The issue came to the forefront after 1877, when Kaiser stated that the youngest stages (étages) of Barrande’s “Silurian System” in Bohemia correspond to the Devonian System in the Harz Mountains of Germany and other regions. Kaiser’s findings contrasted with the conventional nineteenth century wisdom that graptolite fossils became extinct at the end of the Silurian. Eventually, it became clear that so-called Silurian graptolites in some sections occur together with so-called Devonian fossils in other sections, leading to the modern consensus that graptolites are not limited to Silurian strata. A bronze plaque in the Klonk outcrop shows the exact position of the modern Silurian–Devonian Boundary,

28


At present over 55 GSSPs have been defined (Fig. 1; see www.stratigraphy.org for details), but there are more stages in the Phanerozoic Eon in need of base definition. Fortunately, a majority of those now have target definitions, and are awaiting consensus on the best outcrop or borehole section to place a “golden spike.” Thus, with the definitions in place, we can proceed to scale the “deep time” stage units linearly. This brings us to Geochronology, referring to the geochronologic calendar of Earth events called the Geologic Time Scale. While the chronostratigraphic scale is a convention to be agreed upon rather than discovered, calibration of the scale in seconds and (mega-) years is a matter for discovery and estimation rather than agreement. Like human time, linear geological time is expressed in units of standard duration—the second and hence (thousands or millions of) years.

Building a geological time scale The ideal time scale is built from accurate radiometric ages, taken precisely at stage boundaries throughout the stratigraphic column in the Phanerozoic Eon. For more detailed resolution, the exact number of orbitally tuned sedimentary cycles is counted within each stage, such that calibrations and correlation may be achieved within

orb

ital

tun ing sea f spr loor ead ing dire ct d atin g det aile dat d di rec ing t pro por sub tion al zon sca e scali led ng stan com pos dar ite cub d ic ma splin e x. erro likelih fitting r es ood & tim atio n

which is taken at the base of the Lochkovian Stage, the lowest stage in the Devonian. The base of the Lochkovian Stage is defined by the first occurrence of the Devonian graptolite Monograptus uniformis in bed #20 of the Klonk Section, northeast of the village of Suchomasty. The lower Lochkovian index trilobites with representatives of the Warburgella rugulosa group occur in the next younger limestone bed #21 of that section. The concept of the GSSP has gained acceptance among those stratigraphers who consider it a pragmatic and practical solution to the common problem that conventional stage type sections inevitably leave gaps, or lead to overlap between successive stages. The boundary stratotype very much relies on the notion that it is possible to arrive at accuracy in correlation through the use of events, like a geomagnetic reversal, a global change in a stable isotope value, or the evolutionary appearance of one or more prominent and widespread fossil taxa. Thus, the limits of a stage can now be defined with multiple event criteria that to the best of our current knowledge are synchronous over the world. Delimiting successive stages in a clear and practical manner enhances their value as standard units in chronostratigraphy and ultimately in geochronology. Without standardized units neither the (relative) stratigraphic scale nor the (linear) time scale can exist.

Ma 0 Cenozoic 90

180

Cretaceous Jurassic Triassic

270

Permian Carboniferous

360

450

540

Devonian Silurian Ordovician Cambrian Methods used to construct Geologic Time Scale 2004 (GTS 2004)

Fig. 1 Methods used to construct Geologic Time Scale 2004 (GTS2004) (1).


20 thousand years for the last 540 million years or so . . . . If this sounds too good to be true, let it rest. Back to reality.

Geologic reality is schematically illustrated in Fig. 1, providing a quick overview of the actual methodology applied to construct Geologic Time Scale 2004 (GTS2004) (1), the most recent standard time scale. Before the Cambrian, the first period of the Phanerozoic Era, the geologic time scale is less sophisticated, and based only on sparse radiometric dates. The steps involved in Phanerozoic time scale construction may be summarized as follows: (a) Construct a relative (chronostratigraphic) standard scale for the key periods in the Earth’s rock record (b) Identify high-resolution linear age dates to calibrate this relative scale in linear time (c) Astronomically tune (see later) intervals with cyclic sediments or stable isotope sequences which have sufficient fossil or geomagnetic correlation ties to be merged in the standard scale, and increase its resolution (d) Interpolate for those relative time intervals where direct linear age information is insufficient (e) Estimate error bars on the age of boundaries and on unit durations. The first step, integrating multiple types of stratigraphic information to construct the standard chronostratigraphic scale, is the most time consuming; it summarizes and synthesizes centuries of detailed geological research and tries to understand all relative correlations and calibration to the standard. The second and third steps, identifying which radiometric and cycle-stratigraphic studies to use as the primary constraints for assigning linear ages, are the ones that have much evolved. Historically, Phanerozoic time scale building went from an exercise with very few and relatively inaccurate radiometric dates, as available to the pioneer of the geologic time scale Arthur Holmes, to one with many dates with greatly varying analytical precision, as in the mid-1980s. Next, time scale studies started to appear of selected intervals, like Paleogene, Late Cretaceous, or Ordovician, that selected a small suite of radiometric dates with high analytical precision and relatively precise stratigraphic position. At the same time, a high-resolution Neogene time scale started to take shape, using orbital tuning of long sequences of sedimentary and/or oxygen isotope cycles in the Mediterranean region and in Atlantic and Pacific

29

pelagic sediments. The present trend for the pre-Neogene is to incorporate radiometric dates that have very small analytical and stratigraphic uncertainties, and pass the most stringent tests. The fourth step, interpolating the stratigraphic and radiometric information, has much evolved. An early method already constructed the basic two-way graph, used until now. It plotted the cumulative sum of maximum global thickness of strata per stratigraphic unit along the vertical axis and selected radiometric dates from volcanic tuffs and other suitable layers along the horizontal linear axis. This best fit line method interpolated ages to the stages, but is a far cry from methods used today that scale stages along the vertical axis with composite standards of fossil zones. In the mid-1990s, Frits Agterberg and Felix Gradstein started to apply mathematical/statistical error analysis to the time scale ages, which, for the first time, allowed them to assign fairly realistic error bars to ages of Mesozoic stage boundaries, a trend that persists today for the whole of Phanerozoic below the Neogene. The following is a simplified introduction to the modern building tools depicted in Fig. 1.

Music of the spheres Let us start with a brief outline of the principle of the sedimentary cycles approach to time scale building, as is now standard for the last 23 Ma (Neogene), and provides superior resolution and precision. Gravitational interactions of the Earth with the Sun, Moon, and other planets cause systematic changes in the Earth’s orbital and rotational system. These interactions give rise to cyclic oscillations in the eccentricity of the Earth’s orbit, and in the tilt and precession of the Earth’s axis, with mean dominant periods of 100,000, 41,000, and 21,000 years, respectively. The associated cyclic variations in annual and seasonal solar radiation onto different latitudes alter long-term climate in colder vs. warmer and wetter vs. dryer periods that lead to easily recognizable sedimentary cycles, such as regular interbeds of limy and shaly facies. Massive outcrops of hundreds or thousands of such cycles are observed in numerous geological basins, for example around the Mediterranean, and in sediment cores from ocean-drilling sites. Counting of this centimeter to meter thick cycles in great detail over land outcrops and in ocean-drilling wells, combined with the additional correlation aids provided by magnetostratigraphy, oxygen isotope stratigraphy, and biostratigraphy, produced a very detailed

30


Neogene cycle pattern. The critical step is the direct linkage of each cycle to the theoretical computed astronomical scale of the 21,000, 41,000, and 100,000-year paleoclimatic cycles. This astronomical tuning of the geological cycle record from the Mediterranean and Atlantic by earth scientists at Utrecht and Cambridge Universities such as Luc Lourens, Frits Hilgen, and Nick Shackleton led to unprecedented accuracy and resolution for the last 23 million years (2). In New Zealand, Tim Naish and colleagues have calibrated the upper Neogene record to the standard Neogene time scale. Using the high-resolution land-based cycle, isotope and magnetic record in the Wanganui Basin, these authors thereby transferred precise absolute ages to local shallow marine sediments and demonstrated the link between sequence and cycle stratigraphy. Efforts are underway to extend the continuous astrochronologic scale back into Oligocene and Eocene by applying a combination of cycle stratigraphy, improved astronomical projections, oxygen isotope stratigraphy, and magnetostratigraphy to the deep sea record. A special application of orbitally tuned cyclic sediment sequences is to “rubber-band” stratigraphically floating units, like parts of Paleocene, Albian, and parts of Lower Jurassic, skilfully executed by specialists like Ursula Rohl, Tim Herbert, and Graham Weedon. A quantitative estimation of the duration of all cycles within a stratigraphic unit allows estimates of their duration.

Decay of atoms For rocks older than Neogene, the derivation of a numerical time scale depends on the availability of suitable radiometric ages. Radiometric dating generally involves measuring the ratio of the original element in a mineral, like sanidine feldspar or zircon, to its isotopic daughter products. The age of a mineral may then be calculated by means of the isotopic decay constant. Depending on the half-life of the element, several radiometric clocks are available; 40Ar/39Ar and the family of U/Pb isotopes are the most common suites nowadays applied to the Phanerozoic, because of analytical precision and utility with tuffaceous beds in marine or non-marine sequences. Radiometric dating of sedimentary rocks follows several geological strategies: (a) Dating of igneous intrusions within sediments records the time of primary cooling, when the igneous rocks were emplaced and had cooled sufficiently (to a few hundreds of degrees centigrade) to set the radiometric decay clock in

action. Because of uncertainty in the relation of the intrusion to the host sediment, such dates may be of limited stratigraphic use. (b) Dating of volcanic flows and tuffs as part of the stratified sedimentary succession. (c) Dating of authigenic sedimentary minerals, mainly involving glauconite, found widespread in many marine sediments. Mild heating or overburden pressure after burial may lead to loss of argon, the daughter product measured in the 40K/40Ar clock in glauconite. Another problem is that glauconite also contains an abundance of tiny flakes that allow diff usion of Ar at low temperatures. The result is that glauconite dates may be too young. Because of such problems which may be difficult to detect, modern geologic time scales avoid dates based on glauconite. Calibration of the decay constants or measurement standards can be enhanced by intercalibration to other radiometric methods, or by dating rocks of a known age, for example a volcanic ash within an astronomically tuned succession. Astrochronologic and interlaboratory recalibration of the 40Ar/39Ar monitor standard indicates that many of the 40Ar/39Ar ages used in previous Phanerozoic time scales are too young by about 0.5% to 1.0%. For example, the 65.0 Ma age, assigned 10 years ago to the top Cretaceous, is now 65.5 Ma. Radiometric dating techniques with less than 1% analytical error are providing suites of high-precision U/Pb and Ar/Ar dates for the Paleozoic and Mesozoic. Surprisingly, perhaps, there are only seven direct age dates on period or stage boundaries (Fig. 1), with a majority of the 200+ radiometric age dates used for GTS2004 “floating” at some level within a stage. The integration of this level of chronometric precision with high-resolution biostratigraphy, magnetostratigraphy, or cyclic scales is a major challenge to time scale studies. Even the most detailed biostratigraphic scheme probably has no biozonal units of less than 0.5–1.0 million year (my) duration, not to speak of the actual precision in dating a particular “stratigraphic piercing” point, for which an U/Pb age estimate would be available with an analytical uncertainty of 0.1 to 0.5 my. Similarly, combination of analytically less precise K/Ar dates with much more precise Ar/Ar or U/Pb dates in statistical interpolations creates a strong bias toward the latter, despite the fact that both may have equal litho-, bio-, and chronostratigraphic precision. Nevertheless, the combination of precise stratigraphic definitions through GSSPs and accurate


radiometric dates near these levels is paving the way for a substantial increase in the precision and accuracy of the Geologic Time Scale. The bases of Paleozoic, Mesozoic, and Cenozoic Eras are bracketed by analytically precise ages at their GSSP or primary correlation markers—542.0 ± 1.0 Ma, 251.0 ± 0.4 Ma, and 65.5 ± 0.3 Ma respectively—and there are direct age dates for the base Carboniferous, base Permian, base Jurassic, base Aptian, base Cenomanian, and the base Oligocene. Most other period or stage boundaries lack direct age control. Therefore, the third step, linear interpolation, also plays a key role for the time scale.

Interpolation and statistics Despite the progress in standardization and dating, parts of the Mesozoic and Paleozoic Eras have sparse radiometric records (see Fig. 2). Ideally, each of the 90+ stage boundaries that comprise the Paleozoic, Mesozoic, and Cenozoic Eras of the Phanerozoic Eon should coincide with an accurate radiometric date from volcanic ash. However, this coincidence is rare in the geological record. The combined number of fossil events and magnetic reversals far exceeds the total number of radiometrically datable horizons in the Phanerozoic. Therefore, a framework of bio-, magneto-, and chronostratigraphy provides the principal fabric for stretching of the relative time scale between dated tiepoints on the loom of linear time. For such stretching, interpolation methods that are employed are both geological and statistical in nature. Earlier, we mentioned the outdated method of plotting the cumulative global thickness of periods against selected linear age dates. Among the modern geological scaling methods, an assumption of relative constancy of seafloor spreading over limited periods of time is a common tool for interpolating the Latest Cretaceous through Paleogene relative scale. Magnetic polarity chrons, the units of magnetochronology, can be recognized both on the ocean floor as magnetic anomalies measured in kilometers from the mid-ocean spreading center, and in marine sediments as polarity zones that contain biostratigraphic events and can be linked to linear time. Knowing the linear age of a few ocean crust magnetic anomalies (earth magnetic reversals or magnetochrons) allows interpolation of the ages of the intervening magnetic pattern, which in turn can be correlated to the fossil record and geological stage boundaries. The subduction of pre-late Jurassic oceanic crust precludes such an interpolation approach for older Mesozoic and the Paleozoic strata.

31

A second geological method involves building a zonal composite to scale stages. Several outstanding examples are documented in GTS2004 built by a large international team of scientists under the direction of Felix Gradstein and James Ogg. For this scale, Roger Cooper and colleagues have built a very detailed composite standard of graptolite zones from 200+ sections in oceanic and slope environment basins for the uppermost Cambrian, Ordovician, and Silurian intervals. With zone thickness taken as directly proportional to zone duration, the detailed composite sequence was scaled using selected, high-precision age dates. For the Carboniferous through Permian, a composite standard of conodont, fusulinid, and ammonoid events from many classical sections can now be calibrated to a combination of U/Pb and 40Ar/39Ar dates. A composite standard of conodont zones was used for early Triassic. This procedure directly scales all stage boundaries and biostratigraphic horizons. The two-way graph of linear age vs. scaled stages requires a best fitting method, and that is where statistics comes into play, with cubic spline fitting and maximum likelihood interpolation most suitable. On the time scale chart (late 2008 edition; Fig. 3), a majority of Phanerozoic stage boundaries for the first time show error bars; an exception is the Neogene Period where analytical errors are negligible. The error bars reflect both radiometric and stratigraphic uncertainty; in addition, error bars were calculated on stage duration. Uncertainty in the duration of the age units is less than the error in age of their boundaries.

TS-Creator© Now, Adam Lugowski, Ogg, and Gradstein are producing an electronic version of the Geologic Time Scale with the international standard bio-magneto-sequence time scale charts. There are charts for Paleozoic, Mesozoic, and Cenozoic Eras and for each period. This JAVA language package, called TS-Creator©, can be freely downloaded from the ICS website (www.stratigraphy.org). It contains tables of Cambrian through Holocene stratigraphic events calibrated to GTS2004 ages. There are nearly 15,000 biostratigraphic, sea-level, and magnetic zones and levels, plus a suite of geochemical curves. Documentation of zonal definitions, relative age assignments, and how these events were recalibrated to GTS2004 was also compiled. This included updating cross correlations and enhancing detail for selected stratigraphic methods using trilobites, conodonts, graptolites, ammonoids, fusulinids, chitinozoans, megaspores,

Stage/Age boundaries

Epoch Pleistocene

Holocene Pliocene EL L

Miocene Oligocene

Eocene Paleocene

M E L E L M

Piacenzian/Gelasian Zanclean Messinian

Tortonian

Serravallian Langhian

Burdigalian Aquitanian Chattian Rupelian Priabonian Bartonian

Lutetian

E

Ypresian

L M

Thanetian Selandian

E

UP Ar/ b A Rb r, K-A -Sr r

Cenozoic

Quatern.

Paleogene Neogene

Era Period

0

Danian Maastrichtian Campanian

Cretaceous

Late

Albian Aptian

Early

Barremian Hauterivian Valanginian Berriasian Tithonian Kimmeridgian Oxfordian Callovian

Late

Jurassic

Mesozoic

100

Santonian Coniacian Turonian Cenomanian

Middle

Bathonian

Early

Pliensbachian

Bajocian Aalenian Toarcian

Sinemurian Hettangian Rhaetian

Triassic

200

Norian

Late

Carnian

Middle

Ladinian Anisian Olenekian

Early

Carboniferous

300

Guadalupian

Cisuralian

Induan Changhsingian

Wuchiapingian Capitanian Wordian/Roadian Kungurian Artinskian Sakmarian

PennMississippian sylvanian

Permian

Lopingian

Late Middle Early

Asselian Gzhelian Kasimovian Moscovian Bashkirian

Late

Serpukhovian

Middle

Visean

Early

Tournaisian

Devonian

Frasnian

Middle

Cambrian

Eifelian Pragian Lochkovian

Pridoli

Ludlow Wenlock

Ludfordian/Gorstian Homerian/Sheinwoodian Telychian

Llandovery

Aeronian/Rhuddanian Hirnantian

Late

Katian Sandbian

Middle

Darriwilian Floian

Early Furongian

500

Givetian Emsian

Early

Ordovician Silurian

400

Paleozoic

Famennian

Late

Series 3 Series 2 Series 1

Tremadocian Stage 10 Stage 9 Paibian Stage 7 Drumian Stage 5 Stage 4 Stage 3 Stage 2

Legend GSSP Direct radiometric constraint on GSSP rocks Other radiometric date

Stage 1

20 0 10 # radiometric dates Fig. 2 Geologic Time Scale 2004 showing which stage and period boundaries have a Global Boundary Stratotype Section and Point (GSSP), which ones are dated directly and which other age dates were used. Intervals with sparse or no dates required interpolation.

INTERNATIONAL STRATIGRAPHIC CHART

Callovian

3.600 5.332

Messinian 7.246 11.608 13.82

Langhian 15.97

Aquitanian Oligocene

Chattian Rupelian

23.03 28.4 ±0.1 33.9 ±0.1

Eocene

40.4 ±0.2

Lutetian Ypresian Thanetian

Paleocene

37.2 ±0.1

Selandian

48.6 ±0.2 55.8 ±0.2

Danian 65.5 ±0.3

Maastrichtian

70.6 ±0.6

Campanian

Mesozoic Cretaceous

Upper

Santonian Coniacian

83.5 ±0.7 85.8 ±0.7 ~ 88.6

Turonian 93.6 ±0.8

Cenomanian Albian

99.6 ±0.9 112.0 ±1.0

Aptian Barremian Lower

Hauterivian Valanginian Berriasian

125.0 ±1.0 130.0 ±1.5 ~ 133.9 140.2 ±3.0 145.5 ±4.0

Bajocian

Middle

Pliensbachian

Hettangian Rhaetian Upper

203.6 ±1.5 216.5 ±2.0

Carnian

Lower

199.6 ±0.6

Norian

Ladinian Middle

196.5 ±1.0

Anisian Olenekian

~ 228.7 237.0 ±2.0 ~ 245.9 ~ 249.5

Induan 251.0 ±0.4

Changhsingian Lopingian

Wuchiapingian

253.8 ±0.7 260.4 ±0.7

Capitanian Guadalupian

Wordian

265.8 ±0.7 268.0 ±0.7

Roadian 270.6 ±0.7

Kungurian Artinskian Cisuralian

Sakmarian Asselian

275.6 ±0.7 284.4 ±0.7 294.6 ±0.8 299.0 ±0.8

Gzhelian Upper

303.4 ±0.9

Kasimovian 307.2 ±1.0

Middle

Moscovian

Lower

Bashkirian

Upper Serpukhovian Middle

311.7 ±1.1 318.1 ±1.3 328.3 ±1.6

Visean 345.3 ±2.1

Lower

Tournaisian 359.2 ±2.5

* Definition of the Quaternary and revision of the Pleistocene are under discussion. Base of the Pleistocene is at 1.81 Ma (base of Calabrian), but may be extended to 2.59 Ma (base of Gelasian). The historic “Tertiary” comprises the Paleogene and Neogene, and has no official rank.

Ludlow

411.2 ±2.8

418.7 ±2.7

Ludfordian

421.3 ±2.6

Gorstian 422.9 ±2.5

Homerian Wenlock

426.2 ±2.4

Sheinwoodian Telychian

Llandovery

Aeronian

428.2 ±2.3 436.0 ±1.9 439.0 ±1.8

Rhuddanian 443.7 ±1.5

Hirnantian Upper

Katian

Middle

850 1000

Stenian Mesoproterozoic

1200

Ectasian

1600

Statherian 1800

Paleoproterozoic

Orosirian 2050

Rhyacian Siderian

471.8 ±1.6

2300 2500

Neoarchean 2800

Mesoarchean 3200

Paleoarchean 3600

Eoarchean 4000

460.9 ±1.6 468.1 ±1.6

1400

Calymmian

445.6 ±1.5 455.8 ±1.6

Dapingian Floian

GSSP GSSA

Cryogenian

Sandbian Darriwilian

Age Ma

System Period

Erathem Era

Eonothem Eon

GSSP

Age Ma

Pragian 416.0 ±2.8

189.6 ±1.5

Sinemurian

397.5 ±2.7 407.0 ±2.8

Pridoli

542 ~635

Tonian 391.8 ±2.7

Lochkovian

183.0 ±1.5

Neoproterozoic

Givetian Eifelian

Lower

Toarcian

Ediacaran

374.5 ±2.6

Frasnian

Emsian

175.6 ±2.0

Lower

Stage Age

Series Epoch

Eonothem Eon Erathem Era System Period

GSSP

Age Ma 167.7 ±3.5

359.2 ±2.5

385.3 ±2.6

Aalenian

58.7 ±0.2 ~ 61.1

164.7 ±4.0

171.6 ±3.0

20.43

Priabonian Bartonian

Middle

Pennsylvanian

Burdigalian

161.2 ±4.0

Bathonian

Mississippian

Serravallian

Triassic

Miocene

~ 155.6

Upper

Hadean (informal) ~4600

Subdivisions of the global geologic record are Tremadocian formally defined by their lower boundary. Each unit 488.3 ±1.7 of the Phanerozoic (~542 Ma to Present) and the Stage 10 base of Ediacaran are defined by a basal Global ~ 492 * Standard Section and Point (GSSP ), whereas Stage 9 Furongian ~ 496 * Precambrian units are formally subdivided by Paibian absolute age (Global Standard Stratigraphic Age, ~ 499 GSSA). Details of each GSSP are posted on the Guzhangian ICS website (www.stratigraphy.org). ~ 503 Numerical ages of the unit boundaries in the Series 3 Drumian ~ 506.5 Phanerozoic are subject to revision. Some stages Stage 5 within the Cambrian will be formally named upon ~ 510 * international agreement on their GSSP limits. Most Stage 4 sub-Series boundaries (e.g., Middle and Upper ~ 515 * Series 2 Aptian) are not formally defined. Stage 3 ~ 521 * Colors are according to the Commission for the Stage 2 Geological Map of the World (www.cgmw.org). ~ 528 * Terreneuvian The listed numerical ages are from 'A Geologic Fortunian 542.0 ±1.0 Time Scale 2004', by F.M. Gradstein, J.G. Ogg, This chart was drafted by Gabi Ogg. Intra Cambrian unit ages A.G. Smith, et al. (2004; Cambridge University Press) with * are informal, and awaiting ratified definitions. and “The Concise Geologic Time Scale” by J.G. Ogg, Copyright © 2008 International Commission on Stratigraphy G. Ogg and F.M. Gradstein (2008). Lower

478.6 ±1.7

Cambrian

Meso zoic Jurassic

2.588

Famennian

150.8 ±4.0

Precambrian Archean Proterozoic

0.781

Piacenzian

Tortonian

Stage Age

Oxfordian

1.806

Zanclean

Kimmeridgian

145.5 ±4.0

Devonian

0.126

Paleo zoic Carboniferous Permian

“Ionian” Calabrian

Pliocene

Series Epoch


GSSP

Upper

Upper Pleistocene

Gelasian

Phanerozoic Cenozoic Paleogene Neogene

Tithonian

0.0117

Phanerozoic Paleo zoic Silurian Ordovician

Holocene

Age Ma

Stage Age

Series Epoch

International Commission on Stratigraphy

Phanerozoic

Quaternary *


ICS

December 2008

Fig. 3 The International Stratigraphic Chart summarizes the set of chronostratigraphic units (geologic stages, periods) and their computed ages, which are the main framework for Geologic Time Scale 2004. Uncertainties on ages are expressed as two-sigma (95% confidence). This version incorporates changes made by the International Commission up to December, 2008.

34


nannofossils, foraminifers, dinoflagellates, radiolarians, diatoms, strontium isotope, and C-org curves. Numerical ages are calculated within the database using the calibrations; therefore, all ages can be automatically recomputed when control ages are improved in future time scales. Regional scales of selected areas (e.g., Russia, China, North America, and New Zealand) are also included. TS-Creator© automatically takes the reference database, gets instructions from the user on the stratigraphic interval and stratigraphic information to be displayed, and then generates both on-screen and scalable-vector graphic (SVG) renditions that directly input into drafting programs. Next, the user can click on a value, zone, or boundary in the charts on the computer screen, and a window opens with an explanation of the calibration, definition, and interpolated age. This “hot-linked” chart suite is currently a back-looking reference to information in the source tables, but in the future will also provide links to other tables and text from the GTS2004 book, images of stage-boundary outcrops and fossil taxa, and the additional enhancements anticipated during the major update for “Geologic Time Scale 2010.”

Additional information on geologic time

The goal of this brief synopsis was to introduce the basic concepts involved in the construction of the geologic time scale. Further details can be found elsewhere (1–18).

References 1. F. M. Gradstein, J. G. Ogg, A. G. Smith, A Geologic Time Scale 2004 (Cambridge University Press, New York, 2004).

2.

3. 4. 5.

6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17. 18.

L. Lourens, F. Hilgen, N. J. Shackleton, L. Laskar, D. Wilson, in Geologic Time Scale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, New York, 2004), pp. 409–440. S. C. Cande, D. V. Kent, J. Geophys. Res. Solid Earth 100, 6093 (1995). S. A. Bowring, D. H. Erwin, Y. Isozaki, Proc. Natl. Acad. Sci. U.S.A. 96, 8827 (1998). R. M. Carter, T. R. Naish, Eds., The High-Resolution Chronostratigraphic and Sequence Stratigraphic Record of the Plio-Pleistocene Wanganui Basin (Institute of Geological and Nuclear Sciences, New Zealand, 1999). F. M. Gradstein, et al., SEPM Spec. Publ. 54, 95 (1995). W. B. Harland et al., A Geologic Time Scale 1989 (Cambridge University Press, Cambridge, 1990). F. J. Hilgen, W. Krijgsman, C. G. Langereis, L. J. Lourens, EOS Trans. Am. Geophys. Union 78, 285 (1997). A. Holmes, Trans. Geol. Soc. Glasgow 21, 117 (1947). A. Holmes, Trans. Edinburgh Geol. Soc. 17, 183 (1960). S. L. Kamo et al., Earth and Planetary Science Letters 214, 75 (2003). A. Martinsson, The Siluro-Devonian Boundary. International Union of Geological Sciences Series A, 5 (Schweizerbart’sche Verlagsbuchhandlung, Stuttgart, 1977). P. R. Renne et al., Geology 22, 783 (1994). N. J. Shackleton, S. J. Crowhurst, G. P. Weedon, J. Laskar, Phil. Trans. Roy. Soc. Lond. A 357, 1907 (1999). G. P. Weedon, H. C. Jenkyns, A. L. Coe, S. P. Hesselbo, Phil. Trans. Roy. Soc. Lond. A 357, 1787 (1999). T. D. Herbert, S. L. D’Hondt, I. Premoli-Silva, E. Erba, A. G. Fischer, SEPM Spec. Vol. 54, 81 (1995). J. D. Obradovich, Geol. Assoc. Canada Spec. Pap. 39, 379 (1993). U. O. Röhl, J.G., T. L. Geib, G. Wefer, Geol. Soc., Spec. Publ. 183, 163 (2001).

Calibrating and constraining molecular clocks Michael J. Bentona,*, Philip C. J. Donoghuea, and Robert J. Asher b a

Department of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK; bDepartment of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, UK *To whom correspondence should be addressed (Mike.Benton@ bristol.ac.uk)

Abstract In dating phylogenetic trees, it is important to work to the strengths of paleontology and molecular phylogeny estimation. Minimum constraints on calibrations (i.e., oldest fossils in a crown clade) may be calculated with some precision and may be treated as hard bounds, while maximum constraints are soft bounds that may be represented most honestly by probability distributions that reflect the distribution of fossiliferous rocks around the time in question, but allow a small probability of truly ancient dates as well. We present detailed documentation of 63 key calibration dates, with thorough evidence and error expressions, for a wide range of organisms.

For well over 200 years, natural scientists have used fossils with varying degrees of confidence to date the evolution of life. The field has advanced dramatically in the last few years, and it would be useful now to review some of the key issues and to suggest an outline of a modus operandi for the future. In explaining the role of fossils in establishing timescales, it is useful to review the historical sequence in which key observations were made. Much of this early history predates the 1960s concept of the molecular clock; but the way in which fossils should be used today depends crucially on those earlier geological and paleontological observations. In presenting these observations in a logical sequence, we highlight what can and cannot be done with fossils, and link this to the relative strengths and weaknesses of molecular data. The value of carrying out this survey now is that it is not framed in the old and rather worn narrative of a “conflict” between fossils and

molecules (e.g., 1, 2). It is no longer a question of which is better than the other, or how far one can go with one or the other source of data. It is evident that both fossils and molecules have great strengths, but in recognizing the weaknesses of each source of data, a realistic plan for collaboration between paleontologists and molecular biologists can be proposed (3, 4). In this chapter we review the key qualities of the fossil record in a semi-historical account that provides explanations and key references. We then outline a framework for collaboration in dating the tree of life. Finally, we present further, documented, evidence of key fossil-based data for dating.

The key attributes of the fossil record Fossils occur in temporal order

It seems self-evident that older rocks lie at the bottom of the pile, and younger rocks in successive layers on top. And yet, it was only when Nicholas Steno (1638–1686) in 1669 enunciated this as the law of superposition of strata (5) that observers took the point. Fossils were known in Steno’s day, but they were seen as rather random in occurrence and not linked to any pattern reflecting the history of life. By the 1790s, when scientists had finally accepted the idea of extinction, Georges Cuvier (1769–1832) in France and William Smith (1769–1839) in England showed that assemblages of fossils occur in successive rock strata and that the sequence of rocks and of fossils reflects some aspects of the history of the Earth and of life. In more recent times, this conclusion has been confirmed by the observation that temporally older vertebrate fossils tend to occupy cladistically more basal nodes (cf. 6). Fossils and fossil assemblages are characteristic of units of past time

Cuvier and Smith also noted the predictability of some aspects of the fossil record, in particular the totality of fossils that had been collected. Specifically, they noted that particular fossils, or assemblages of fossils, appeared

M. Benton, P. C. J. Donoghue, and R. J. Asher. Calibrating and constraining molecular clocks. Pp. 35–86 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

36


to characterize particular rock units and to occur in the same order every time they were seen. This observation was used practically by geologists to correlate rock units from place to place, to name them (e.g., “Carboniferous,” “Lias,” “Old Red Sandstone”) and to match stratigraphic units on the first geological maps. The principle is fundamental to modern biostratigraphy and lies behind the practical search for new sources of oil and gas today. However, the practices of biostratigraphy and correlation have evolved substantially, in particular with the development of graphic correlation which facilitates the integration of stratigraphic phenomena, including isotope anomalies and geochronologically dated ash layers (7), to derive a global composite standard stratigraphy against which local sections may be compared and correlated (8). Fossils may be included in phylogenies of modern organisms

Cuvier had no sympathy with then-current ideas of evolution, but he gave the world the science of comparative anatomy. Whatever the causes, he recognized the prevalence of anatomical similarities in often widely different organisms, and he realized these indicated closeness of relationship in some sense. He famously demonstrated his skills in public demonstrations in Paris in the early nineteenth century, by taking a single fossil bone, and reconstructing the whole animal and its biology and habits in some detail, before an assistant revealed the whole skeleton. Cuvier never plotted an evolutionary tree, but key concepts such as homology arose from his pre-Darwinian comparative anatomy. There is a single geological timescale that may be used as a yardstick of time

William Smith around 1800 mapped his correlatable geological units across England, and guessed they extended over Europe. Roderick Murchison (1792–1871) and others in the 1830s drove these ideas forward, naming the major divisions of geological time, the Paleozoic, Mesozoic, and Cenozoic eras, and the various geological periods (Cambrian, Silurian, Devonian, etc.). Murchison hurtled across Europe and Russia in his coach and mapped his British units across the Ural Mountains. He declared that the new geological eras and periods, established in Western Europe (mainly in England and Wales), provided a yardstick of deep time that would work worldwide. Initial studies in Africa, North America, and

Australia about 1840 showed he was right. Since 1840, the international geologic timescale has not been substantially revised, but some major time units have been added, most notably the Ediacaran in 2004 (9). The order of fossils matches the pattern of the evolution of life

Although faintly discerned by Smith, Cuvier, and others, the link between deep time and the phylogeny of life could not be made without an understanding that life had evolved. Cuvier’s great rival at the Muséum Nationale d’Histoire Naturelle in the 1790s, Jean-Baptiste Lamarck (1744–1829), was the first serious proponent of evolution, the idea that species change through time. His model of evolution was more an escalator than a tree, the great scala naturae, where every species is arrayed along one or more moving walkways from rocks to angels: the humans of today were once apes, and the apes of today may some day be humans. Many people in the 1830s accepted the idea of progress, or directionality, in the order of fossils in the rocks, but others, notably Charles Lyell (1797–1875), saw time as a series of cycles rather than a unidirectional arrow, and so sought to deny the idea of progress from simple to complex through the succession of fossils in the rocks. There is a tree of life

Charles Darwin (1809–1882) was the first to understand that the evolution of life was not an escalator, nor any other kind of linear progression, but a branching tree that links all living and fossil species, and the lines unite backwards in time to the single common ancestor of all life. His first branching tree appeared in manuscript notes in 1838; and the sole illustration in the Origin of Species was a branching tree (10). The fossil record is incomplete

Charles Darwin also spent some time in the Origin of Species discussing the “imperfection of the geological record.” He pointed out that many living organisms have only soft parts, and so are unlikely to be preserved. Others live in environments such as mountainsides or beaches, where erosion dominates, and sediment does not accumulate. He noted the patchiness of geological strata, and the fact that intermediate fossil forms are rare. Darwin did, however, predict that intense efforts by paleontologists would fill many of the gaps in the

Calibrating the Molecular Clocks

record and allow the deeper parts of the tree of life to be disentangled. In many cases (e.g., basal tetrapods, synapsids, and diapsids) his prediction has been fantastically confirmed, whereas others remain more difficult to document paleontologically. Raup (11) summarized the issue clearly, arguing that there are biases in the fossil record; for example, the quality of the record must diminish as one goes further back in time. It has been shown, however, that although there is a diminution in quality back in time, this does not erase large-scale evolutionary patterns (12). The fossil record can only be as complete as the rock record in which it is preserved; and since attempts have been made to compile large-scale databases of organismal diversity though geological time, there has been a worry that it is biased by inconsistencies in the rock record (11). These include the consequence of plate tectonics, like the fact that open ocean sediments are invariably destroyed at destructive plate margins, along with the oceanic crust on which they rest. Thus, the only open ocean sediments from which we may sample past diversity are of Triassic age or younger; older sediments from these environments are very rare and are represented only by tectonically and thermally abused mélanges scraped from the surface of oceanic crust as it subducted into the Earth’s mantle. Continental interiors are invariably regions of net erosion, rather than sedimentation, and so it is only at the continental margins that we can hope to maintain a record of relatively continuous sedimentation, in which fossil remains may be preserved over long geological timescales. However, because environments shift in position with respect to the rise and fall of sea level over geological timescales, there are concomitant secular variations in the environments—and their hosted organisms—that are preserved (13). Most worrying of all, variance in the availability of rock for sampling and the numbers of species found in those rocks are closely correlated (14–16). This may indicate either that biases in the rock record dominate the paleontological diversity signal, or that the rise and fall in sea level controls rock volume and species diversity. These alternative hypotheses are difficult to reconcile (17). However, the general congruence between phylogenetic branching order and the temporal sequence in which fossils are found (6, 12) indicates that even if the rock record is heterogeneous, the primary paleontological signal is not overwhelmed. Thus, we may place some faith in the fossil record, our only direct record of evolutionary history, but it must nevertheless be interpreted with considerable care.

37

Deep time may be dated

Throughout the nineteenth century, most scientists accepted that the Earth was very ancient, but there was no meaningful way to determine exact dates for any rocks. Calculations based on estimates of the rate of cooling of the Earth from an initial supposedly molten state gave rise to rather short histories of the Earth, measured in tens or at best hundreds of millions of years, whereas estimates based on rates of sedimentation or the rate of solution of salts in the sea, yielded rather higher estimates, in the hundreds to thousands of millions of years. The discovery and application of radiometric dating by Arthur Holmes (1890–1965) and others in the early years of the twentieth century, showed that the Earth was some 4.5 billion years old, and that the fossiliferous Phanerozoic represented the past 0.5 billion years or so (18). Further details of radiometric dating are given in The Geologic Time Scale. Radiometric dating has evolved substantially and, although the geologic timescale is in constant revision, there is an ever-diminishing error, and the calibration of geologic time is being extended into the Proterozoic (19). A dated phylogenetic tree offers valuable information on evolution

In the early to mid-twentieth century, various paleontologists used trees of one sort or another to calculate rates of evolution, whether rates of change of individual characters, of all the characters of a group of organisms (the transformational approach), or rates of appearance and turnover of species or genera (the taxic approach). Even with uncertain timescales, such rates could be established in a relative way, as shown by George Gaylord Simpson (1902–1984) in his classic work, Tempo and Mode in Evolution (20). Since the 1940s, there has been a renaissance in paleobiology, with extensive work on rates and patterns of evolution, origination and extinction. Effective molecular clock methodologies afford a new opportunity for both molecular and morphological evolution, allowing us to approach fundamental questions such as the relationship between morphological and molecular evolution (21, 22). Phylogenies are cladistic

Until the 1960s, phylogenies were put together in a somewhat impressionistic way, whereby systematists used their judgment to determine propinquity of relationship. There was a concept of phylogenetically useful morphological

38


characters, as opposed to less useful characters, but there were no clear rules or protocols that allowed analyses to be repeated or challenged. Hennig (23) famously distinguished plesiomorphic (“ancestral”) from apomorphic (“derived”) characters, but it is well known also that his message was not widely appreciated until his book was translated into English in 1966 (24). Fossils may belong to stems or crowns

There has been much confusion in discussions about the tree of life because of sloppy use of taxic terms. Hennig (24, 25) and Jefferies (26) distinguished crowns and stems: a crown clade consists of all living members of a group, their common ancestor, and everything in between, regardless of whether it is living or extinct. A stem is paraphyletic and composed of all those extinct lineages more closely related to the crown group in question, than to another (Fig. 1). So Archaeopteryx is a stem bird but not a crown bird. This confusion between stems and crowns led to some difficulty in the early debates about the timing of origin of major clades (27, 28) but while clarification of these issues has led to the resolution of dispute in some debates (29), it has actually increased confusion in others (30, 31). Fossil taxa need not exhibit all derived characters of the crown clade in question because, invariably, we seek a date to constrain divergence between one crown group and another. Thus, it is the fossils assigned to the stem that, by definition, lack the full complement of crown-group characters, in which we have most interest. However, fossils may be fragmentary and it may be difficult to distinguish whether the absence of characters in a fossil taxon reflects an aspect of evolution or simply incomplete preservation.

There are many practitioners of tree dating, and each has a more-or-less unique approach to the problem. The emphasis on fossils as useful (or not), the number of calibration points recommended, the ways in which fossil-based dates should be determined and cited, the ways in which such calibration dates should be assessed for congruence or not, and how they should be combined with molecular trees, are all under active debate. We consider these points in turn, and seek to make broadly defensible recommendations that may contribute to the development of a new protocol that will be generally acceptable. We are optimistic that a fair number of previously debated points may now be resolved, and that methods now exist that play to the strengths of paleontological data on the one hand, and molecular data on the other (3, 4). One calibration date or many?

It was commonplace through the 1960s to 1990s to use a single calibration point in molecular clock analyses, generally because data sets were small and algorithms simpler. Most analysts of tetrapod phylogeny, for example, used the mammal–bird date of 305–315 Mya (32) as their sole reference point. This stance has been criticized (33–37) because of the risk that the whole enterprise will be skewed by possible errors in the sole calibration point. Others (e.g., 35, 38–40) have argued forcefully that multiple calibration dates should be used, suggesting that greater numbers will reduce uncertainty and improve statistical robustness. It is perhaps generally correct that several dates are better than one, but it would be facile to argue that more is always better. There is no benefit in simply increasing the number of paleontological dates used in a calibration exercise without thought about their quality: numerous erroneous dates will give a meaningless result (41–43).

p

Total group

Strategies for using fossils to date trees

St

em

gr ou

Time

Cr ow n

gr ou p

Present day

With an understanding of the strengths and limitations of fossils, it will now be possible to consider current assumptions about how fossils may be used to contribute to the grand enterprise of dating the tree of life.

Fig. 1 Stem, crown and total-group definitions, following Hennig (25) and Jefferies (26).

Cross validation of potential calibration points

One problem with using multiple fossil calibration dates is that there is a risk of mixing useful and erroneous calibrations. Near and Sanderson (42) and Near et al. (41)


have suggested that it is better to cross validate different potential calibration points across a single phylogeny and determine which are consistent with each other, and which are inconsistent. The consistent calibration dates, which all point to the same solutions for unknown dates are assumed to be a more-or-less correct set of fossil dates, close to their relevant nodes. The inconsistent dates may be too young or too old, indicating either unusually poorly sampled lineages (too young) or incorrectly assigned fossils (too old). Although it is likely that this method will indeed identify a consistent set of dates in most cases, and these dates will presumably be close to the relevant nodes, this need not be the case. We note that cross validation methods may not always work: Hugall et al. (44) note a case where two calibration points can be correct but appear incongruent because rate smoothing failed to give the correct relative branch lengths. In an instructive application of cross validation of paleontological calibration dates, Douzery et al. (45) applied seven minimum divergence estimates from the mammalian fossil record to their molecular clock study focusing on rodents. Four were within crown Rodentia, one based on a close relative of crown Lagomorpha, one from crown Cetartiodactyla, and one based on a 1996 report of a “Paleocene” proboscidean taxon (Phosphatherium) from Morocco (46). When analyzed alone, each of these calibrations predicted at least one of the six other dates within a 95% confidence interval except for one: their Paleocene calibration for Paenungulata. Importantly, it was exactly this date that had been geologically misinterpreted in its original publication (46). According to more recent analyses of the Moroccan Ouled Abdoun Basin localities from which Phosphatherium is known (47), the original report of a Paleocene age was in error; these fossils are not Paleocene but Eocene, ca. 5–7 Ma younger than the 60 Ma value given in Douzery et al. (45). The early Eocene remains of proboscideans from Morocco remain among the oldest known fossils of crown Afrotheria. Although somewhat by accident, Douzery et al. (45) recognized this very instructive error by identifying their 60 Ma paenungulate date as incompatible with any of their other six calibration points. Importantly, the resulting clock estimates from Douzery et al. (45) for placental mammals yielded dates within the Tertiary for intraordinal divergences, and did not exceed 80 Ma for the common ancestor of Placentalia. These values are generally consistent with estimates based on the mammalian fossil record (48) and with some other, independent molecular clock estimates (49), but not others (e.g., 50–52).

39

It is possible that cross validating paleontological minima may make the resulting estimates “too young.” Such minimum estimates from the fossil record are, after all, “minimum” and, barring misidentification, they will underestimate actual divergence times. However, when several such calibrations are mutually compatible (which is distinct from biasing dates in a single direction), this indicates either accurate identification of a genuine divergence date or systematic error affecting all calibrations, as may be the case when preservational factors affect multiple fossil lineages. Such factors usually act in limited geographic areas, and fossil finds from other regions can plug gaps. Some geologic events, such as major sealevel changes, can create a global-scale hiatus, such as the nonpreservation of coastal habitats. Such global hiatuses are generally brief, geologically speaking, and the habitat and its fossil record reappear, and so can be identified and measured. In any event, all such cases must be evaluated individually and on their own merits. In general, we regard cross validation as a valuable tool to help improve the accuracy of fossil calibrations as applied in molecular clock studies. Choice of dates and the quality of the fossil record

Fossil date estimates for divergence events have errors associated with them that arise because of all the various imperfections of the fossil record, as well as the often tortuous means by which a numerical date can be assigned. However, these errors are rarely if ever acknowledged—an astonishing fact given that calibration is, by definition, the rate-determining step in molecular clock analyses. Reisz and Müller (37, 53–55) have argued that this may be overcome by quoting errors on the dates, where the errors are indicated by the age of more and less derived relatives of the fossil organism that provides the main calibration date. Thus, an age span, rather than a point date can be used to calibrate a clock, faithfully reflecting the error associated with the paleontological estimate. However, Reisz and Müller go further and argue that some paleontological estimates have much broader errors than others, providing a measure of their relative quality. In particular, they single out the bird–mammal split, which is the most widely adopted of all fossil calibrations, as a less-than-ideal example of paleontological calibration. This is because of a dearth of more primitive relatives that are close in age; indeed, there is a dearth of sites from which such fossils might be found. Thus, the errors on the paleontological estimate, particularly for its

40


maximum bound, are very broad indeed. Many alternative calibrations are available with much smaller attendant errors associated with them. For some questions, it may still be desirable to use the bird–mammal split as a calibration, because of its applicability to a breadth of sequence data in public databases and relevance to certain high-profile scientific questions. Nevertheless, Reisz and Müller have made a valid point with which we fundamentally agree: not all calibrations are of the same quality and when possible, those with more paleontological data pertinent to constraining a soft maximum estimate (as when fossil-bearing strata older than the minimum estimate are well sampled) should be preferred. The debate about accuracy of paleontological calibration dates conflates two issues: the relative accuracy of minimum and maximum constraints. For the minimum constraint, the dating error on a securely identified fossil is simply the error in dating the rock formation in which it is contained. For a maximum constraint, the error encompasses this error but, much more significantly, it is also a measure of uncertainty that the oldest possible age estimate really lies below the branching point in question. In the next sections, we argue first that error on a fossil calibration cannot be symmetrical and cannot be generalized about a single fossil point. We then show that fossils can act as relatively secure “hard” minimum constraints on a particular branching point, and a soft maximum constraint can also be estimated. Magnitude and symmetry of error on fossil calibration dates

Error bars on paleontological calibration points, if used at all, have generally been assumed to be symmetrical. This seems logical, because uncertainties about dating rock layers and uncertainties about the identity of the fossil might be assumed to be equal in both directions, up and down. However, errors were usually not indicated on fossil calibration dates because there has not been an obvious or reasonable way to calculate them (3). One “quick fi x” for this problem was proposed by Douzery et al. (40), who used the whole span of the geological period in which the calibration fossil was found as their error range. So, for example, they used the span of the Devonian period (354–417 myr) as the error on timing of the split of mammals and actinopterygian fishes (i.e., the base of clade Osteichthyes). As Hedges et al. (56) pointed out, this is a far wider age range than is necessary when compared to either the date of the oldest osteichthyan (418.7 ± 2.6 myr), or the range from soft

maximum to minimum constraint for that branching point (421.75–416 myr). The debate about error bars on fossil dates, whether to use them or not, how to calculate them, and whether they are symmetrical or not, is circumvented, we suggest, by a recognition of the use of fossils to determine minimum and soft maximum constraints. Fossil dates as estimates of origin or as minimum constraints

Until recently, paleontologists and molecular clock practitioners have been perhaps a little unclear about just what the fossil dates represent. The tenor of many of the to-and-fro debates between defenders and opponents of the merits of the fossil record in dating the tree of life (e.g., 1, 2) suggested that both sides were treating their dates as pointing at the same thing, namely the actual time of origin of a clade. It is clear, however, that the two dates are different. Paleontologists are limited in recognizing the origin of a clade because clades may start as rare, founding taxa, located in only one small part of the world and, by definition, they will lack many or all diagnostic crown-group characters. So, the oldest fossil X will always be younger than the origin of clade X, whether by a few thousand years (geologically negligible) or many million. Molecular clocks, of course, attempt to date clade divergence. Once it is accepted that molecular dates are dates of origin, and that paleontological dates always postdate them, then the arguments about error bars and “acceptable” and “unacceptable” dates become less acute. Paleontologists estimate minimum constraints on the ages of clades (37, 57, 58). Providing the fossil is correctly assigned to a clade, and providing its provenance is known, the date reflects simply current best knowledge on the age of the geological formation that contains the oldest phylogenetically secure fossil. We emphasize this point because compendia of fossil dates for the inception and demise of clades (e.g., 59) are littered with records that are optimistically interpreted. The oldest fossil records of a clade will always be phylogenetically uncertain because, by definition, they will exhibit the fewest of all characters to justify membership. This problem is further exacerbated by the fact that the oldest possible records are also very often extremely poorly preserved fragmentary fossils; indeed, they may be little more than a fragment (28). For instance, the oldest records of the shark lineage are a series of isolated scales (60), not the complete articulated skeleton that we might prefer.


In such circumstances, it can be difficult to distinguish between fossils that fail to exhibit diagnostic characters because they are primitive, and fossils that preserve few characters at all for reasons of fossil preservation (25). Thus, it is important that molecular clocks are calibrated using phylogenetically secure fossil records, demonstrating the primitive absence of a number of crown-group diagnostic characters, even at the expense of alternative dates, based on poorer data, that are maybe tens of millions of years older. If fossil dates are accepted as minimum constraints (3, 4), their attendant errors must still be considered; but it is the youngest limit of the resulting age span that should be adopted as the minimum age constraint for a lineage split. As such, no fossil date provides a poor minimum constraint—unless it is actually older than the lineage divergence that it purports to date. Returning to the infamous mammal–bird split, corresponding to the origin of the clade Amniota, the minimum constraint is the minimum age of the Joggins Formation (in which occurs the oldest diapsid and synapsid, respectively, Paleothyris and Protoclepsydrops). The age span of this geological formation is 314.5–313.4 Ma ± 1.1 myr, a date based on biostratigraphy (palynology) and exact dating from elsewhere, resulting in a minimum constraint of 312.3 Ma (3). Again this does not mean that the divergence could not have happened earlier, only that it could not have happened later. Calculating soft maximum constraints and codifying probability densities

A consequence of accepting that fossils provide only minimum constraints, rather than direct calibrations on molecular clock analyses, is that it is desirable to somehow capture a soft maximum constraint on the calibration of the clock. Various authors have suggested that this can be achieved by codifying an arbitrarily diminishing probability density extending back in time from the minimum constraint (57, 61–64). The manner in which these constraints are established and the nature of the variance in probability density can be better informed by paleontological data. Benton and Donoghue (3, 4) have suggested that soft maximum constraints can be established broadly in the manner that Reisz and Müller (37) suggested to establish the older error on calibrations. Note, however, that one of the recommendations by these authors was to use the date of the oldest fossil belonging to the nearest relative (sister group) as a guide to the maximum constraint;

41

as noted elsewhere (65), this does not provide a maximum constraint, merely a minimum constraint on that nearest relative’s lineage. Probability densities may be estimated using models of diversification and preservation probability below the oldest known fossil in a clade (66, 67) and/or informed by older, more tenuous records of the existence of clades, otherwise rejected because they are phylogenetically less secure (Fig. 2). Crucially, it is now possible to implement such constraints without precluding the possibility that the timing of divergence predates the “soft” maximum constraint, but with the assumption that it is increasingly improbable with increasing departure from this constraint (62).

Tree making and the molecular clock Variable rates

In the early years of dating trees, the molecular clock was assumed to be constant (68, 69). It soon became clear, however, that the clock ran at different rates both within and between lineages (70, 71) and this at first led some to doubt the possibility of using the molecular clock to calibrate dates on a tree. However, estimating divergence times need not depend on a constant clock, and techniques were developed to test for variable rates (72–74). These techniques allowed analysts to determine typical rates within the clade of interest, and to eliminate those that deviated from the norm. Unusually fastevolving genes in particular could give anomalously ancient divergence dates if they were not recognized, and so it was often recommended that such genes should be eliminated (e.g., 75, 76). The clock tests, however, lack power for shorter sequences and for genes with low rates of change, and they detect only a small proportion of cases of rate variations in the kinds of genes commonly used for molecular clock studies (77–79). For this reason, current tree reconstruction and dating techniques are “relaxed” in that they allow analysts to assume any number of local molecular clocks within a phylogeny (80). Some of these techniques assume that evolutionary rates among closely related lineages are similar, or “autocorrelated” (81, 82), but this assumption has been rejected by others (e.g., 63) who recommend methods that allow independent rates as a first pass, and then evaluate whether the reconstructed rates are autocorrelated (49, 63, 83). Further advances in methods of tree reconstruction and dating that allow for the vagaries of evolution and

42

THE TIMETREE OF LIFE Young

(A)

(B) Minimum constraint

Oldest certain

Probability density

Probability density

Oldest possible

5% 5%

Maximum constraint

Maximum constraint

0

0

Old

Fig. 2 Two hypothesized patterns for the distribution of probabilities between the soft maximum and minimum constraints on the date of origin of a clade. In (A) the curve is a logistic, corresponding to a standard birth–death model of

diversification and an equilibrium at “normal” diversity, when fossils become abundant. In (B) an assumption is added that there might be some older possible fossils, corresponding to an expansion of the probability distribution.

multiple uncertainties include explicit modeling of the evolution of the rate of evolution using Bayesian methods (82–84), the use of nonparametric or semiparametric models of rate evolution (81, 85), and methods that allow for phylogenetic uncertainty as well as rate uncertainty (63, 86). These methods often lead to better concordance overall between molecular and paleontological dates (e.g., 40, 87, 88), but this is not always uniformly the case. For example, Hedges and Kumar (89) argue that the majority of dates in vertebrate evolution agree well between their molecular analyses and the fossils, but certain dates, such as those for the origin of modern mammals and birds, are more ancient than the oldest fossils. On the other hand, Kitazoe et al. (49) argue that when abrupt changes in mutation rate and convergent evolution are taken into account in models of molecular evolution over time, divergence estimates for mammals correspond more closely to the fossil record than previously reported. For example, Kitazoe et al. (49) estimate the common ancestor of placental mammals at ca. 84 Ma, slightly older than the molecular clock estimate reported by Douzery et al. (45), and much younger than other clock estimates of 105 Ma (51) or 129 Ma (50). Concordance between largely independent data sets,

such as here, might be a guide to the quality of the result, but need not be of course. Finding the correct answer

How are we to judge the current state of tree making and tree dating? One view is that the new methods, notably Bayesian approaches and relaxed clocks, must be superior to other techniques because they generate better concordance between paleontological and molecular evidence (e.g., 90, 91), an assumption that the traditional evidence is a yardstick against which new methods and new evidence may be assessed. Of course, in tree making and tree dating there can only be a single answer, a single correct tree, and a single date for any branching point. Others argue that the new methods are so accommodating, one might even claim that they are so relaxed as to be laid back, that they can produce any desired outcome (e.g., 92). A key point about relaxed clock methods is that although there are more inherent assumptions, they are statistically weaker, but in assuming so much about the evolutionary process when the diversity of competing models betrays how little we understand of it, the concern is that our multifarious assumptions will preclude us from better knowing it (93).


Counterintuitively perhaps, relaxation of assumptions need not produce answers that are so broad as to be meaningless. For example, theoretical and empirical studies into the use of wide confidence intervals on multiple calibration dates by Yang and Rannala (62, 94) have shown that these have self-correcting properties. They show that hard calibration dates may conflict and produce unsatisfactory results, but dates expressed with at least one soft bound (the soft maximum constraint) interact so as to correct poor calibrations. Poor calibration data expressed with unjustifiable hard confidence intervals may, on the other hand, produce results with misleadingly high precision. These results are encouraging because they show that flexible calibration data, expressed in line with paleontological reality, can interact to cancel out a great deal of the uncertainty. This can then feed back information to the paleontologists that certain calibration dates appear to work better than others, and that the poor dates require closer study to see why they do not work so well. The example of cross validation of calibrations from Douzery et al. (45) illustrates this point nicely. The search for the single answer must also apply treetesting techniques with equal rigor to both morphological/paleontological and molecular trees and dating evidence. There are issues of quality and quantity of data. Statistically speaking, the best data matrix is the largest—more data can lead to better statistical measures. In particular, estimates of divergence times are improved by dense packing of calibration points around the node of interest. But, are more genes and more species always best? Yes: but more progress may be made through the addition of further paleontological constraints, or by further constraining established constraints (94). Quantity and quality

The focus of future work must pursue quantity and quality of data equally. It is self-evident that data sets must increase in size, with the inclusion of ever more species and genes, and larger numbers of calibration dates. But, data quality ought to be determined based on biological and geological criteria, not on statistical expediency. On the whole, taxa should not be deleted from analyses unless there are serious doubts about the accuracy of data, for example from an incomplete or damaged fossil specimen. A statistically “rogue” taxon that has an unusually bad effect on a cladistic analysis (e.g., 95, 96) is worthy of attention: it must fit in the tree somewhere, and cannot simply be regarded as awkward and

43

so eliminated. Certain genes, like certain morphological characters, can be shown to contain no phylogenetic signal, and so should be omitted on the basis of such auxiliary evidence. Ornithologists have realized for decades that certain kinds of characters, such as plumage color, are generally not helpful in determining deep phylogenetic relationships, and these are rarely included in morphological phylogenetic analyses. Certain individual genes, or even classes of genes, may similarly be accepted as phylogenetically uninformative at particular levels, and so can only confound a statistical analysis. We have argued, as have others (e.g., 3, 36–38, 41, 57) that more calibration dates are generally better than few. But the quality of those dates, or the data they are based on, is critical (37). Quality of calibration constraints may be assessed subjectively, based on accuracy of identification of fossils, distinction of crown clades from stems, cross-checking of geological dates, and evaluation of the extent of a well-documented underlying fossil record. We do not believe there is an objective way to determine that one constraint is good, and another is bad; the only bad constraint predates (in the case of a minimum constraint) or postdates (in the case of a soft maximum constraint) the evolutionary event in question. The use of hard minimum constraints and soft maximum constraints, as recommended here, is a more meaningful representation of the nature of paleontological evidence than the attempt to present single fossil-based dates as the holy grail, whether with or without error bars. At present, tree making and tree dating is often a two-step process, where a single model tree is generated, and dates are then calculated against that tree. Better approaches for the future will probably be to combine the two steps, to allow uncertainty in the tree topology and in the calibrations and calculated dates (63, 97). The best solution must maximize the fit of topology and dates against external evidence. Fixing the tree first, and the dates second could well mean that we miss an even better resolution of the data. The reasons for the two-step process are (a) tractability and (b) distinguishing rates from time. Typically, analysts use one program, or set of programs, to calculate the tree that best fits their data, and then a separate program or programs to apply calibration dates and calculate unknown divergence times. There is no fundamental reason, however, that the two steps could not run in parallel within a single process of calculation, except when the class of data from which rates of evolution are typically calculated (i.e., DNA sequences) are missing entirely for elements of the tree to be built (i.e., fossils).


55.6

61

62.

5

og eh it bb

8 .5

ka

Pi

33

.7

rel

uir

124

61

Sq

.6

.5

61

16

n

55

2.9

cke

Gu

.5

ch

Rat

5

66

52.

2.3

66

61.5

rew

h Tree S

330 .4

Lizard

225.9

239

31

dile

se

Mou

10.4

Emu

Croco

ig

aP

ine

61

fin

Chi

dg

Sh 39.6 62.5 62.5

.5

He

62.5 62.5

Ra

52.4

rew

Bat

llo

ant

m

us

bra

Sheep

adi

ph

.4

typ

Ze

52.4

48

su

Pla

Cow

Arm

Ele

os

18

rec

n Te oo ar

ng

Ka

Op

Dolphin

A

Pig

Benton and Donoghue (3) presented a set of 30 calibration dates for key genomic animals, from nematode to human, increased to 40 by Donoghue and Benton (4).

Dog

Key calibration dates for molecular clock practitioner

Cat

These have been corrected, revised, and updated in light of new data, and augmented by 25 further calibration dates required because of genome sequencing projects that are approaching maturity, giving a current total of 65 dating arguments (Fig. 3A and B). Further revisions will be provided at http://www.fossilrecord.net. It is striking that in the short time since the publication of Benton and Donoghue (3), revisions have had to be made to some of their paleontological date estimates. These revisions arise more from clarifications of previously incompletely determined materials rather than from outright errors. In particular, two groups of early mammals, the zhelestids and zalambdalestids, had been attributed to many locations in the tree of basal placental mammals; in some analyses they have been interpreted as evidence for an early date of the minimum constraint on several fundamental branching

Distinguishing rates from time is harder. A long branch in a tree could represent an ancient divergence time or a fast-evolving genome. These alternatives can be tested against different genes: if the branch is always long, then an ancient divergence time is the more likely explanation. Fossil dates can also help to distinguish rates from time in some cases. Theoretical studies (62, 94) suggest that, with an infinite amount of sequence data, uncertainties in time estimates usually reflect uncertainties in fossil calibrations rather than in branch lengths.

Horse

Bushbaby

Toad

33.7

61.5

416.0

Lemur

460.6

96.9

back

le

Stick

o

a Tetr

gu

kifu

2

pr

ey

Inv ert e

m La

0.

S

br ate s

nt

ha

rk ha

Fig. 3 Continues

Macaq

ue

Gibb

on

Ora

ngu

Ch

im

p Ne an de rth Hu al m an

6.5

Ta

p Ele

Marmoset

33.7

.25 32

don

11. 2 11. 2 2 3.5

a Medak

149.85 96.9

Zebra Fish

421.75

55.6

44

tan


points within Eutheria. Zhelestids and zalambdalestids are each represented by several species from mid- to late Cretaceous localities in central Asia, with an oldest occurrence around 95.3 Ma. Archibald (98) and Archibald et al. (99) associated these taxa with crownplacental lineages, specifically zhelestids with “ungulatomorphs,” a grade that includes hoofed artiodactyls and perissodactyls, and zalambdalestids with Glires (i.e., rodents and lagomorphs). The latter hypothesis dates to the 1960s (100, 101). However, some older assessments of Cretaceous mammal affinities (e.g., 102 on zalambdalestids) as well as more recent phylogenetic analyses sampling an adequate number of both recent and fossil clades (e.g., 48, 103–106) have indicated instead that no Cretaceous taxon has an exclusive, sister taxon relationship to any single crown-placental clade. The revised interpretation of zhelestid affinities reduces four of the hard minimum age estimates presented in Benton and Donoghue (Table 1 in ref. 3), namely for “cow–dog,” “human–cow,” and “human– armadillo,” from Late Cretaceous to Paleocene, and for “human–tenrec” from Late Cretaceous to early Eocene. Zalambdalestids had not been used as minimum constraints in Benton and Donoghue (3, 4).

Dates for calibrating and constraining molecular clocks

The paleontological constraints provided here were selected because they represent the divergence events between metazoans whose genomes have been sequenced. The reasoning underpinning this is that molecular clock analyses are usually codified on the basis of available data and, thus, these constraints will be of the greatest utility to the greatest number of analysts. The constraints do not attempt to date the individual fossil on which they are based. Rather, they establish firm minimum and soft maximum constraints on the timing of the component divergence events. Thus, in contrast to common practice, minimum constraints are not established on the precise age, or the maximum possible age of the oldest phylogenetically secure member of the two lineages that result from the divergence event that is to be constrained. Rather, we attempt to derive the minimum possible date of this fossil. If the only stratigraphic constraint available for a fossil is for instance Danian, the minimum constraint would be provided by the geochronological age of the top of the Danian Stage. This is 61.70 Ma ± 0.2 myr (107) and, thus, the quoted minimum

aria

ella

hal

a

.6

.0

n hi rc

s

xu io

ph

Am

238.5

tes

ica

n Tu

307.2

3.7

28

aU

ni

168 238.5

Se

ph

Le

e

152

199

h ec

Ca

hid

pit

mp

Sea

Bio

Hare

t Limpe

Honey bee

Fruit Fly

ito

Mosqu

le

Beet

us

dni

Rho Ap us

Lo Da

Hy dra

Wasp

B

Ne m at od e An em on e

.2

307

30

5.5

Spo

nge

470.2 0

52

0.5

51

es

brat

e Vert

0.5

52

7

4.9

45

531.5

63

518.5

Fig. 3 Summary of paleontological constraints (in myr) on metazoan phylogeny. A. Vertebrates. B. Invertebrates. The dates are minimum constraints, and soft maximum constraints are given in the text, and at http://www.fossilrecord.net/.

46


constraint is 61.5 Ma. This principle is explained graphically in Fig. 4. Soft maximum constraints are established on the basis of well-preserved assemblages of more plesiomorphic (ancestral) relatives of the clade that lack members of the clade itself. The maximum possible age of that assemblage would be quoted. Thus, if the only stratigraphic constraint on the age of such an assemblage was again Danian, the soft maximum constraint would be would be provided by the geochronological age of the base of the Danian Stage. This is 65.5 Ma ± 0.3 myr (107) and, thus, the quoted soft maximum constraint is 65.8 Ma. It should, but often does not, go without saying that the trees on which paleontological constraints were determined should be compatible with the molecular phylogenies in which they are employed. In the paleontological constraints presented below, most were established within an uncontroversial phylogenetic framework. However, certain nodes have recently undergone revision and so we emphasize here the phylogenetic schemes followed in deriving a constraint in such instances. The high-level mammalian clades identified herein are based on the topology figured by Springer and Murphy (108; Fig. 1); interrelationships of deuterostomes and jawless vertebrates follow Bourlat et al. (109); other clade relations are not contentious. Terminology for high-level mammalian clades is based on Simpson (110), Waddell et al. (111), and Springer and Murphy (108). Note that there is no convention for the conversion of traditional taxonomic concepts, based on the classification of living constituents alone, to crown or total-group-based definitions with the inclusion of fossil taxa, especially stemtaxa. Hennig (25) argued that traditional taxa should be converted to total-group definitions but, in practice, different approaches have been adopted, without justification, in different scions in the Tree of Life. For instance, among tetrapods and plants, traditional taxa have generally been converted to crown-based definitions while, elsewhere, total-group definitions have been adopted (28). In what follows, we have followed convention and there is no consistency with regard to the adoption of total or crown-based definitions of taxa. However, in each instance we are explicit with regard to the composition of the clades for which we provide constraints on divergence timing. Homo sapiens: human–neanderthal (minimum = 0.2 Ma; soft maximum = 1 Ma) H. sapiens neanderthalensis represents an anatomically distinctive European, antedating the European arrival of anatomically modern humans by many thousands of years.

The oldest neanderthal fossils date to just under 0.2 Ma from France (Biache St. Vaast) and Germany (Ehringsdorf) (112). The oldest anatomically modern members of our own subspecies are probably remains from the Levant, from the Skhul cave in Israel, dating to ca. 0.100–0.135 Ma (113). Because these ages postdate the Geologic Time Scale 2004 (GTS2004) marine timescale (114), we tentatively rely on the radiometric and faunal dates provided in the primary literature. Hence, the paleontologically minimum date for this split may be estimated based on the older, undisputed neanderthal sites in Europe at 0.2 Ma. For the soft maximum split within H. sapiens, we would suggest the widespread occurrences of H. erectus outside Africa over 1 Ma (112). In the case of central Asia, there appear to have been populations of H. erectus nearly 1.8 Ma (115). Despite occurrences of the genus Homo throughout Asia and Africa by around 1 Ma, no evidence for neanderthals or anatomically modern humans from this time is yet known (112). Hominoidea: chimp–human, neanderthal (minimum = 5.7 Ma; soft maximum = 10 Ma) The dating of the chimp–human split has been discussed for nearly a century. Early paleontological estimates, up to the 1970s, placed the branching point deep in the Miocene, at perhaps 20–15 Ma, but this was revised dramatically upward to about 5 Ma by early molecular studies (116), and estimates as low as 2.7 Ma have been quoted (117). Paleontological evidence for the branching point was distinctly one-sided until recently, since the only fossils fell on the human line, and so the question of the date of divergence of humans and chimps became synonymous, for paleontologists, with the date of the oldest certain hominin (species on the human, not chimp, line). The oldest chimpanzee fossils are, at ca. 0.5 Ma, comparatively young (118). The date of the oldest hominin has extended backward rapidly in the last 25 years. Until 1980, the oldest fossils were gracile and robust australopithecines from 3 Ma. The discovery of “Lucy,” now termed Praeanthropus afarensis in Ethiopia (119) extended the age back to 3.2 Ma at most. Then, two further hominin species pushed the age back to over 4 myr: Ardipithecus ramidus from rocks dated as 4.4 Ma from Ethiopia (120) and Praeanthropus anamensis from rocks dated as 4.1–3.9 Ma from Kenya (121). More recent finds, remarkably, have pushed the dates back to 6 myr: A. ramidus kadabba from Ethiopia (5.8–5.2 Ma; 122), Ororrin tugenenis from Kenya (c. 6 Ma; 123), and Sahelanthropus tchadensis from Chad (6–7 Ma; 124). The last two taxa have proved highly controversial, with claims that one or other, or both, are not

2. Lithostratigraphy of formation

Me da ka Sti ckl eb ack Tak ifu go Tet rao do n

Ze b

raf i

sh

1. Oldest certain fossil in lineage

T.

Mšrnsheim Beds Mörnsheim Beds

ti3

L., O. ti2b

Ostariophysi

Fish fossils

Euteleostei Clupeocephala

ti2a

Lower Solnhofen Plattenkalk

ti1

Ršgling Beds Rögling Beds

Tischlingerichthys

3. Biostratigraphy

4. Chronostratigraphy

AGE (Ma)

145

Stage CRETACEOUS (Berriasian) 145.5 ± 4.0

Polarity Ammonite zones Chron Sub-Mediterranean M18 M19 M20 M21

149.9 ± 0.05

M22

Glochiceras lithographicum

Taramelliceras prolithographicum

150.8 ± 4.0

Gravesia

Hybonoticeras hybonotum

150

Tithonian

Fig. 4 Graphical summary of the process by which a time constraint on lineage divergence is derived, using the split between zebrafish and euteleost models as an example. 1. The oldest phylogenetically secure fossil is identified that falls within the overall clade circumscribed by the lineage divergence. In this instance it is Tischlingerichthys, among others; there is equivocation over its membership of any component of the crown euteleost lineage, but it is clearly a member of the euteleost total-group. 2. The stratigraphic age of the fossil taxon is established: Tischlingerichthys has been recovered from zone ti2b of the Solnholfen Limestones. 3. A means of correlating the stratigraphic age of the fossil is established: the strata bearing Tischlingerichthys also bear the ammonite Hybonoticeras hybonotum, which is a diagnostic zone fossil for the lower Tithonian. A minimum age for Tischlingerichthys

M22A

M23

Kimmeridgian

155

154.55 ± 4.0

S. semiforme S. darwini Hybonoticeras hybonotum Hybonoticeras beckeri

M24 M24A M24B M25 M25A

can be obtained from the base of the succeeding S. darwini ammonite biozone since the Solnholfen Limestones fall fully within the H. hybonotum Biozone. 4. The minimum age for the oldest record of Tischlingerichthys is established using the 2004 Geologic Timescale in which the base of the S. darwini Biozone has been determined to correlate to the base of the M22n magnetostratigraphic polarity chron, which has itself been dated at 149.9 Ma ± 0.05 myr. Thus, the minimum constraint on the divergence of Ostariophysi and Euteleostei is the minimum age interpretation of the errors on this date, equating to 149.85 Ma. As tortuous as this argumentation appears, the majority of splits require many more steps of lithostratigraphic, biostratigraphic, chemostratigraphic, and/or magnetostratigraphic correlation before they can be tied to a geochronological age (4).

48


hominin, but ape-like (125). Recently, evidence has been presented to bolster the case that Orrorin, at least, is in fact a hominin (126). Dating of the Sahelanthropus beds in Chad is not direct. Biostratigraphic evidence from mammals in particular, but with cross-checking from fish and reptile specimens, indicates that the unit is definitely late Miocene (i.e., older than 5.33 Ma), and that it may be older than the Lukeino Formation of Kenya, the source of Orrorin (dated at 6.56–5.73 Ma from Ar/Ar dates on volcanic layers; 127), equivalent to the lower fossiliferous units of the Nawata Formation at Lothagam (dated as 7.4–6.5 Ma; 128). This would suggest a date for the sediments containing Sahelanthropus of 7.5–6.5 Ma, based on biostratigraphy and external dating. However, since the status of Sahelanthropus remains contentious, we based our minimum constraint on Orrorin, which has minimally an age of 5.73 Ma (129). Thus, we determine a 5.7 Ma age for the minimum constraint on the human–chimp split. Kumar et al. (130) have recently calculated a range of ages for the human–chimp divergence of 4.98–7.02 Ma; their minimum constraint (4.98 Ma) is younger than the oldest fossils (Orrorin, Sahelanthropus). The report by Suwa et al. (131) on a late Miocene fossil gorilla hints at a similar age for the gorilla and orangutan lineages. Some late Miocene ape fossils, such as Gigantopithecus and Sivapithecus may be stem-orangs. Nonetheless, a range of such apes, Ankarapithecus from Turkey (10 Ma), Gigantopithecus from China (8–0.3 Ma), Lufengopithecus from China (10 Ma), Ouranopithecus from Greece (10–9 Ma), and Sivapithecus from Pakistan (10–7 Ma) give maximum ages of 10 Ma, early in the late Miocene, and these deposits have yielded no fossils attributable to either chimps or humans. This is taken as the soft maximum constraint on the human–chimp divergence. Pongidae: orangutan–chimp, human, neanderthal (minimum = 11.2 Ma; soft maximum = 33.7 Ma) While there are numerous taxa of fossil hominoids from Africa and Eurasia, few can be unambiguously attributed to an extant ape lineage; and most of these are hominids. Numerous fossils attributable to Pongo are known from Pleistocene sites in east Asia; and another fossil orangutan relative, Gigantopithecus, is known from the late Miocene and Pliocene (132). Sivapithecus from the Miocene of southern Asia is the oldest ape fossil interpreted with some confidence as a close relative to orangutans (132). There is a possible, but tentative, record of a fossil orangutan relative dating to the late Oligocene in

northern Kenya (133). The oldest definitive occurrence of an orangutan relative is Sivapithecus from the Chinji Formation of Pakistan, corresponding to magnetic polarity chron 5Ar, estimated to be ca. 12.5 Ma before present (134). This correlates to the Serravallian stage, the top of which is at 11.2 Ma, our minimum estimate for the divergence of the orangutan from other great apes. As a soft maximum we suggest the first diverse occurrence of anthropoids from the earliest Oligocene of the Fayûm, Egypt. These primates comprise a diverse radiation just on the Oligocene side of the Eocene–Oligocene boundary at 33.7 Ma (135), and lack derived features of the extant great ape lineages. Hylobatidae: gibbon–orangutan, chimp, neanderthal, human (minimum = 11.2 Ma; soft maximum = 33.7 Ma) Numerous taxa of Miocene apes may share a close relation with extant gibbons to the exclusion of other hominoids. European taxa such as Dryopithecus and Oreopithecus, with a record dating to the middle Miocene, have over the years occasionally been linked to hylobatids (132). However, recent cladistic analyses do not place any of these taxa with hylobatids to the exclusion of other catarrhines (136). Hence, gibbons cannot be said to have a definitive fossil record before Pleistocene occurrences of the extant genus in east Asia; and previously named Miocene species of Pongo are now recognized under other Miocene hominoid genera (137). The cladistic divergence of hylobatids from other catarrhines is therefore constrained by the record of great apes from South Asia (see Sivapithecus discussed earlier), minimally dated at the top of the Serravallian at 11.2 Ma. Divergences within hylobatids themselves would of course be much younger. As a soft maximum we would again suggest the first diverse occurrence of anthropoids from the earliest Oligocene of the Fayûm, Egypt. These primates comprise a diverse radiation just on the Oligocene side of the Eocene–Oligocene boundary at 33.7 Ma (135), and lack derived features of extant hominoids. Catarrhini: macaque–gibbon, orangutan, chimp, neanderthal, human (minimum = 23.5 Ma; soft maximum = 34.0 Ma) The human–macaque split is equivalent to the branching of Old World monkeys (Cercopithecoidea) and apes (Hominoidea), which together form the clade Catarrhini. The oldest cercopithecoids are Victoriapithecus macinnesi from Kenya, and two species of Prohylobates from Libya and Egypt. Miller (138) surveyed all fossils of these


two genera, and compared ages of their respective deposits. The oldest cercopithecoid fossil is a tooth identified as Victoriapithecus sp. from Napak V, Uganda (c. 19 Ma), followed by Prohylobates tandyi from Moghara, Egypt (18–17 Ma) and Prohylobates sp. from Buluk, Kenya (>17.2 Ma), Prohylobates simonsi from Gebel Zelten, Libya (c. 17–15 Ma), and V. macinnesi from Maboko, Kenya (ca. 16–14.7 Ma). MacLatchy et al. (139) report an even older cercopithecoid, a fragment of a maxilla from the Moroto II locality in Uganda, which has been radiometrically dated to be older than 20.6 Ma ± 0.05 myr (140). The oldest hominoids include Morotopithecus, also from the Moroto II locality in Uganda (140). Young and MacLatchy (141) determined that this taxon is a hominoid, located in the cladogram above the gibbons, and so not the most basal member of the group. Because of incompleteness of the material, Finarelli and Clyde (142) are less certain of its phylogenetic position, but Morotopithecus is certainly a catarrhine. Even older is the first record of the long-ranging hominoid genus Proconsul from Meswa Bridge in Kenya, biostratigraphically constrained to ~23.5 Ma (143, 144). Even older still is the purported hominoid Kamoyapithecus from the Eragaliet Beds of the Lothidok Formation of Kenya, dated at 24.3–27.5 Ma (145), but the material is insufficient to determine whether it is a hominoid or a catarrhine, possibly lying below the human–macaque split (142). So, the minimum constraint on the human–macaque split is 23.5 Ma, based on the oldest record of Proconsul, biostratigraphy and external dating. The soft maximum constraint is based on members of the stem of Catarrhini, namely the Families Propliopithecidae (Propliopithecus, Aegyptopithecus) and Oligopithecidae (Oligopithecus, Catopithecus) that are basal to the cercopithecoid–hominoid split (146). These are represented in particular from the rich Fayûm beds in Egypt, which possess a diverse anthropoid primate fauna, including stem platyrrhines and catarrhines (147) during the early Oligocene ((135); 33.9–28.4 Ma ± 0.1 myr). Hence, at the base of the Oligocene at 33.9 Ma ± 0.1 myr, the Fayûm shows a diversity of primates and other mammals, but no members of crown-group hominoids or cercopithecoids. Anthropoidea: marmoset–macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 33.7 Ma; soft maximum = 65.8 Ma) The oldest South American primate is Branisella from the late Oligocene of Bolivia; this taxon cannot be more explicitly linked to any of the modern platyrrhine groups

49

(132). The middle Miocene locality of La Venta, Columbia, has produced the oldest remains of essentially modern platyrrhines, including the possible marmosets Micodon, Patasola, and Lagonimico (148). The oldest crown anthropoid, that is, the oldest primate reconstructed within the clade of living New and Old World monkeys, consists of the Fayûm taxon Catopithecus from the Fayûm Quarry L-41, now dated at the end of the Priabonian (135) with an upper bound of 33.7 Ma. Older taxa such as African Altiatlasius and Biretia, and Asian eosimiids and amphipithecids, may in fact be anthropoids, but appear to fall outside crown Anthropoidea (147). Given the fact that the oldest known euprimate (Altiatlasius) has been regarded tentatively as an anthropoid sister taxon (147), the soft maximum for anthropoids must predate this occurrence in the late Paleocene. Early Paleocene strata have yielded fossils of several groups (plesiadapids, paromomyids, carpolestids) reconstructed closer to crown primates than to Scandentia or Dermoptera, but have not yielded any definitive crown primates. Hence, the paleontological soft maximum can be defined by the base of the Paleocene, at 65.8 Ma. Primates: bushbaby, lemur–marmoset, macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 55.6 Ma; soft maximum = 65.8 Ma) Crown-group Primates, or Euprimates, encompass living forms plus the extinct adapoids and omomyoids, as the latter are more closely related to extant lemuriforms than to anthropoids (149, 150). The oldest fossil often attributed to primates is Purgatorius known from the basal Paleocene (64 Ma ± 1 Ma), reputed from time to time to have been latest Cretaceous in age. However, the latest Cretaceous specimen is an unidentifiable tooth (150). The oldest euprimate is Altiatlasius from the late Paleocene of Morocco (151), and is recognized as the most basal stem anthropoid (147, 152) or, alternatively, as the most basal euprimate (150). Broadly speaking, this specimen is currently regarded as the oldest euprimate fossil. Altiatlasius comes from the Adrar Mgorn 1 locality in the Ouarzazate Basin of Morocco, dated generally as late Paleocene (Thanetian stage). Magnetostratigraphic study (153) narrows the age range of the locality to “late or latest Thanetian,” and so the age of the top of the Thanetian Stage provides the minimum constraint. This is 55.80 Ma ± 0.2 myr (107), and so the minimum constraint is 55.6 Ma. The soft maximum constraint may be marked by older possible primate fossils. McKenna and Bell (154) implied in their classification that carpolestids, with a record in

50


the Danian (early Paleocene) are euprimates, but this has not been substantiated elsewhere. They also attribute the basal Paleocene Decoredon from China to Primates, although this has been regarded as a hyopsodontid by others (e.g., 155). In general, early Paleocene strata have yielded fossils of several groups (plesiadapids, paromomyids, carpolestids) reconstructed closer to crown primates than to Scandentia or Dermoptera (150), but have not yielded any definitive crown primates. Hence, the paleontological soft maximum constraint can be defined by the base of the Paleocene, at 65.8 Ma. Strepsirhini: mouse lemur–bushbaby (minimum = 33.7 Ma; soft maximum = 55.6 Ma) Malagasy primates are extraordinarily diverse, but nevertheless comprise an extant radiation that shares a single common ancestor to the exclusion of other primates such as galagos, lorises, and monkeys. There is an extraordinary diversity of subfossil primates from Madagascar, but neither these nor other lemuriforms have a fossil record demonstrably older than the Holocene (152). In contrast, the lorisiform sister-radiation of Malagasy lemurids (including galagos and bushbabies) shows a less ambiguous fossil record through the late Eocene, including the oldest known records of the toothcombed prosimians from the Birket Quarun Formation of the Fayûm, Egypt [e.g., Karanisia, Seiffert et al. (156)]. Following Seiffert (135, 152) this unit corresponds to the Priabonian with an upper bound of 33.7 Myr before present. Fossil primates are relatively common mammalian fossils at many localities in North America and Eurasia throughout the Eocene; yet toothcombed prosimians remain conspicuously absent before the end of the middle Eocene. Hence, we suggest the first appearance of euprimates, represented by Altiatlasius from the late Paleocene of Morocco at 55.80 Myr ± 0.2 myr (102) (or 55.6 Myr) as the soft maximum for Strepsirhini. Archonta: tree shrew–bushbaby, lemur, marmoset, macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 61.5 Ma; soft maximum = 131.5 Ma) Tree shrews are members of Scandentia, an order within Archonta (= Euarchonta) that has long been seen as a close relative of Primates. Current trees (e.g., 51, 150) place Scandentia as sister group of Dermoptera, and those two as sister group of Primates. The minimum constraint on dating the split between tree shrews and any of the primates is set then by determination of the oldest member of Orders Scandentia, Dermoptera, or Primates or their respective stem relatives.

Scandentia (tree shrews) and Dermoptera (flying lemurs) are small, non-speciose orders of mammals, with sparse fossil records. The oldest scandentian is Eodendrogale from the middle Miocene of China (149), and the oldest dermopteran is a single specimen from the late Eocene of Thailand, Dermotherium (149). The oldest Primates are more diverse, and clearly extend the date of origin of the order to the beginning of the Paleocene, as already discussed. Hence, the minimum constraint on the Scandentia–Primates split (i.e., origin of Archonta) is set in the early Paleocene based on Torrejonian occurrences of extinct primate sister taxa such as carpolestids and plesiadapids (150). The upper bound of the Torrejonian North American Land Mammal Age (NALMA) correlates to the top of the Danian, 61.5 Ma. The soft maximum constraint could correspond to the date of origin of Archontoglires, the larger clade including Archonta (primates + flying lemurs + tree shrews) and Glires (rodents + rabbits). However, the problem is that no definitive records of Archontoglires, or any other crown-group placental mammal, exist before the K-T boundary. A tenuous link between ca. 95 Ma zalambdalestids and lagomorphs and/or rodents has been suggested on several occasions (e.g., 99, 101, 157); but this link has been disputed (e.g., 158) and is not supported by well-sampled phylogenetic studies (48, 104–106). As there is a relatively good fossil record that documents numerous mammalian groups during the Cretaceous, including the stem-Metatheria and Eutheria, but lacking any crown placental, using the criteria outlined earlier we are left with a similar value for both the paleontological minimum and soft maximum in the early Paleocene. Because we do not want to rule out a priori the possibility suggested by some molecular clock analyses of an older radiation of crown placental groups deep in the Cretaceous, we have relaxed the criteria for identifying a paleontological soft maximum for supraordinal placental mammal clades. Hence, the next, best-documented therian node lower than Archontoglires that is paleontologically well documented is the divergence between Metatheria and Eutheria, in the Early Cretaceous, represented by the Liaoning fossils of Eomaia (159) and Sinodelphys (160). Liaoning fossils of the Jehol biota have been estimated to be between Barremian and Aptian. Importantly, the association of Jehol fossils and dated horizons is not without ambiguity (161); hence we conservatively use the age of the Barremian to define the ages of Jehol specimens (162), which indicates a lower bound of 130 Ma ± 1.5 myr, or 131.5 Ma.


Archontoglires: rabbit, pika, squirrel, guinea pig, mouse, rat–tree shrew, bushbaby, lemur, marmoset, macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 61.5 Ma; soft maximum = 131.5 Ma) The human–mouse split is synonymous with the latest branching point between the mammalian Orders Primates and Rodentia. Both orders are members of the clade Archontoglires, sometimes called Euarchontoglires. Archontoglires is composed of two clades, the Archonta and the Glires, and Primates belongs to the former, Rodentia to the latter. Thus, the human–mouse split becomes synonymous with the origin of Archontoglires. As stated earlier, fossil evidence for this branching point does not exceed 65.2 Ma, the beginning of the Paleogene (base of Cenozoic, base of Tertiary), and corresponding to the extinction of the dinosaurs and the beginning of the radiation of placental mammal orders. Several molecular analyses have suggested that crownplacental orders might have their origin at some point much deeper in the Cretaceous, ranging from over 125 Ma (74, 163) to much younger dates that are in line with fossil evidence (45, 49). As stated earlier, the oldest confirmed primates are from the Paleocene–Eocene transition, 55.5 Ma (164). The oldest “plesiadapiform”-grade mammals include early Paleocene Palaechthon and carpolestids. The oldest undisputed fossil rodents are known with confidence from the Thanetian (late Paleocene, 58.7–55.8 Ma), including members of the Family Ischyromidae from North America and Europe (101). Some or all eurymyloids from Asia may fall on the stem to Rodentia and/ or Lagomorpha (48, 165), which would provide a minimum record for Glires in the early Paleocene, corresponding to the record of Heomys from the Shanghuan Asian Land Mammal Age (166). The upper bound for the Shanghuan ALMA, as for the Torrejonian NALMA, is the top of the Danian, 61.5 Ma. As stated earlier for the common ancestor of Scandentia, Dermoptera, and Primates, given the lack of crown placental fossils during the Cretaceous, we suggest a soft maximum paleontological estimate for Archontoglires at 131.5 Ma, based on the record of Eomaia and Sinodelphys from China (see Archonta). Glires: pika, rabbit–squirrel, guinea pig, mouse, rat (minimum = 61.5 Ma; soft maximum = 131.5 Ma) The rabbit–mouse split is synonymous with the clade Glires, comprising Orders Rodentia plus Lagomorpha). The date would have been assumed traditionally to lie at 65 Ma, or younger, marking the time of purported

51

placental mammal radiation after the extinction of the dinosaurs (see earlier under Archontoglires). The oldest crown lagomorphs are somewhat younger. Stucky and McKenna (101) indicate several Eocene rabbits from the Lutetian: Lushilagus from China, Procoprolagus from Canada, and Mytonolagus from the United States. Meng and Wyss (167) note an older possible stem lagomorph, Mimotona, from the early to late Paleocene (Doumu Formation, Nonshangian, Qianshan Basin, China), the same unit that yielded the putative earliest stem rodent Heomys. The minimum constraint on the age of clade Glires, and so for the rabbit–mouse split, based on Heomys, is the same as that for Archontoglires, described earlier at 61.5 Ma. Again we suggest a paleontological soft maximum of 131.5 Ma. Lagomorpha: pika–rabbit (minimum = 48.6 Ma; soft maximum = 65.8 Ma) The modern order Lagomorpha consists of two groups: leporids (rabbits and hares) and ochotonids (pikas). There are just under a dozen species of pikas (Ochotonidae, Ochotona), the oldest relative of which (to the exclusion of leporids) has been reported to be the late Eocene Asian form Desmatolagus (154). A yet older taxon, Decipomys from the early Eocene of central Asia, shows a pattern of enamel microstructure that could be a “structural predecessor” to that of modern ochotonids (168). The status of Decipomys as an ochotonid, or of Eocene palaeolagids as close relatives of rabbits and hares, would indicate a divergence within crown Lagomorpha by the early or middle Eocene, respectively. A recent analysis of isolated hindlimb elements from China and India (169) also indicates that leporids and ochotonids were distinct by the early Eocene. Although fragmentary, these elements are surprisingly diagnostic for the Leporidae. These identifications are consistent with the interpretation of the early Eocene Strenulagus and Gobiolagus from central Asia as leporids (170), although Lopatin and Averianov (171) have more cautiously assigned them to “Lagomorpha, Family Strengulidae” without specifying a crown affi liation. Hence, our minimum estimate for crown Lagomorpha is based on the Indian leporid fossils described by Rose and colleagues (169) from the middle Ypresian-equivalent Cambay Shale in West-Central India. The top of the Ypresian is dated at 48.6 Ma. There are many fossil Glires on the stem to Lagomorpha that long predate the first unambiguous occurrence of a leporid or ochotonid (48, 172). For the soft maximum divergence of crown lagomorphs we choose the K-T

52


boundary at 65.8 Ma, based on the occurrence of basal Glires such as Mimotona and Heomys, reconstructed near the base of modern Rodentia and Lagomorpha. None of these Paleocene Glires can be defended as a member of Leporidae or Ochotonidae. Rodentia: squirrel–guinea pig, mouse, rat (minimum = 55.6 Ma; soft maximum = 65.8 Ma) Recent phylogenies place squirrels (sciuromorphs) external to guinea pigs and murid rodents (i.e., Caviomorpha + Myomorpha) (48, 173, 174). Ischyromyids dating to the late Paleocene from North America and Europe have been linked to modern sciurids (154), which would give a minimum constraint for crown Rodentia of 55.6 Ma. Recent phylogenies do not consistently place ischyromyids such as Paramys and extant sciurids in the same clade, but they generally do fall within crown-group Rodentia (48, 103). Early Paleocene eurymylids may not be rodents proper, but members of a larger clade including Simplicidentata, or they may fall outside Simplicidentata, but within Glires, as outgroup to rodents and rabbits (48). Therefore, a reasonable soft maximum constraint on the base of crown Rodentia could be set by these early rodents at the base of the Paleocene at 65.8 Ma. Rodentia (minus sciurids): guinea pig–mouse, rat (minimum = 52.5 Ma; soft maximum = 58.9 Ma) The guinea pig (Cavia), a member of Caviomorpha and Ctenohystrica (173, 175), is the closest relative to the Muridae, the family containing mouse (Mus) and rat (Rattus). Caviomorpha is a member of the larger clade Hystricognatha, and Muridae is a part of the larger clade Myomorpha, which in turn falls in the major clade Sciurognatha, according to traditional classifications. If so, this puts the guinea pig–mouse divergence as equivalent to the origin of crown-group Rodentia. A newer molecular phylogeny makes Myomorpha closest to Hystricognatha, and Sciuromorpha closest to those two (173). The oldest member of the Caviomorpha stem group is Tsaganomys from the mid-Oligocene Hsanda Gol Formation of Mongolia (154, 167), while the oldest members of the Myomorpha stem group are early Eocene dipodids such as Ulkenulastomys, Blentosomys, and Aksyiromys from the Obayla Svita of the Zaysan Basin, Kazakhstan (176). Lucas (177) assigns an Irdinmanhan age to this site, based on comparisons of the contained mammals, moving it from early Eocene to the base of the middle Eocene, and with a soft maximum age of K/Ar age of 55–56 Ma (= 52.5–54.7 Ma when corrected to new

IUGS constants) measured from extrusive volcanic rocks that underlie the Toruaygyr mammalian Assemblage, also Idinmanhan in age, in Kirgyzstan. We take the upper end of this range, 52.5 Ma, as our minimum constraint on the guinea pig–mouse split date. The soft maximum constraint might be taken as equivalent to the age of the ischyromyids and other entirely extinct rodent groups from the late Paleocene (Thanetian) of North America and Europe (154), 58.9 Ma. Muridae: mouse–rat (minimum = 10.4 Ma; soft maximum = 14.0 Ma) The mouse (Mus musculus) and rat (Rattus norvegicus) are both members of the Subfamily Murinae within the Family Muridae, members of the larger clade of muroid rodents. The Old World rats and mice are hugely diverse, with over 500 species, and they appear to have radiated relatively rapidly in Europe, Africa, Asia, and Australia. The phylogeny of all genera within Murinae has not been determined, so the location of the split between Mus and Rattus is somewhat speculative at present. However, all current morphological and molecular phylogenies (178–181), indicate that Mus and Rattus diverged early in the evolution of Murinae, but not at the base of the divergence of that clade. A lower limit to the mouse– rat divergence is indicated by the oldest known murine fossil, Antemus chinjiensis from the middle Miocene Chinji Formation of Pakistan, dated at about 14.0–12.7 Ma on the basis of magnetostratigraphy and radiometric dating (182). The oldest fossil example of Mus dates from 7.3 Ma, a specimen of Mus sp. from locality Y457 in the Siwaliks (182). Fossils of Rattus are not known until the latest Pliocene and the Pleistocene of Thailand (183) and China (184), no more than 3 Ma. The divergence of the two lineages leading to Mus and Rattus was stated to be 14–8 Ma by Jacobs and Pilbeam (185), in a first review of the fossil evidence. This range was narrowed down at its older end to 12 Ma in subsequent studies (186, 187), based on the first appearance of the fossil genus Progonomys, early members of which were assumed to include the common ancestor of Mus and Rattus. The 12 Ma figure has most commonly been selected as the mouse–rat calibration point, but dates in the range from 16–8.8 Ma have been used in recent molecular studies. In a thorough review of the fossil evidence, Jacobs and Flynn (182) show that records of Progonomys in the Siwalik succession extend from 12.3 to 8.1 Ma, with the later forms (10.4–8.1 Ma) assumed to lie on the Mus lineage. The extinct genus Karnimata (11.1–6.4 Ma) is interpreted as a member of the lineage leading to Rattus. The


oldest record (11.1 Ma) is uncertain, but the next (at 10.4 Ma) is unquestionable. The early species, Progonomys hussaini (11.5–11.1 Ma) is interpreted as an undifferentiated basal murine antedating the common ancestor of Mus and Rattus by Jacobs and Flynn (182), and so they place the Progonomys–Karnimata split (equivalent to the Mus–Rattus split) at not much beyond 11 Ma, “although it may be younger.” The dating is based on detailed field stratigraphic study of the long Siwaliks sedimentary sequence, with dating from magnetostratigraphy and radiometric dating (188, 189). The soft maximum constraint on this date is taken as the oldest record of Antemurus at 14.0 Ma. Boreoeutheria: hedgehog, European shrew, bat, cow, sheep, dolphin, pig, horse, dog, cat–rabbit, pika, squirrel, guinea pig, mouse, rat, tree shrew, bushbaby, lemur, marmoset, macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 61.5 Ma; soft maximum = 131.5 Ma) The human–cow divergence is synonymous with the origin of Boreoeutheria. This clade is composed of the clades Archontoglires (human) and Laurasiatheria (cow). A number of Late Cretaceous putative laurasiatherians have been cited. The oldest supposed laurasiatherians have been said to be the zhelestids, from the Bissekty Formation of Dzharakuduk, Kyzylkum Desert, Uzbekistan, and the even older Khodzhakul Formation at Sheikhdzhili, which would provide a very ancient minimum age constraint on the clade (early Cenomanian, 95.3 Ma) if the assignment is correct. Some authors (99, 190, 191) place the zhelestids in Laurasiatheria, basal to the hoofed artiodactyls and perissodactyls. This has been challenged, however, and more comprehensive cladistic analyses of basal Eutheria (104, 105) place zhelestids outside of the crown clade of extant orders. As for the other superordinal placental clades indicated earlier, the minimum paleontological constraint for Boreoeutheria is again the early Paleocene, constrained here by the To1 record of Protictis (61.5 Ma; see later). The paleontological soft maximum cannot be better constrained than the Early Cretaceous records of Eomaia and Sinodelphys from the Liaoning beds of China (131.5 Ma; see Archonta). Laurasiatheria: shrew, hedgehog–cow, sheep, dolphin, pig, horse, cat, dog, bat (minimum = 62.5 Ma; soft maximum = 131.5 Ma) The split between the lipotyphlans (shrew, hedgehog, mole, and Solenodon), on the one hand, and the cow, sheep, whale, pig, horse, cat, dog, pangolin, and

53

bat, on the other, is equivalent to the base of the clade Laurasiatheria. Arguably, the crown-placental order most commonly regarded among paleontologists to have representatives in the Cretaceous is Lipotyphla. McKenna and Bell (154) reported the oldest lipotyphlan as Otlestes from the Cenomanian (99.6–93.5 Ma) of Uzbekistan, but Archibald (190) regarded it as a basal eutherian, lacking derived characters of Lipotyphla, or any other modern order. More recently, Averianov and Archibald (191) synonymized it with Bobolestes (from the same local fauna) and regarded it as a questionable zalambdalestoid. Next in time is Paranyctoides from the Turonian (93.4–89.3 Ma) of Asia and the Campanian (83.5–70.6 Ma) of North America, and Batodon from the Maastrichtian (70.6–65.5 Ma) of North America, both regarded as lipotyphlans by McKenna and Bell (154). Archibald (190) is uncertain, but retains these records pending discovery of further specimens. The oldest, relatively uncontroversial records of Laurasiatheria are Paleocene carnivorans (cf. Protictis). As above for Carnivora, we tentatively assign an early Paleocene minimum age constraint to this node, corresponding to To1 at 62.5 Ma. The soft maximum constraint is, as for the other supraordinal divergences, 131.5 Ma. Lipotyphla: European shrew–hedgehog (minimum = 61.5 Ma; soft maximum = 131.5 Ma) The shrew and hedgehog are members of clade Lipotyphla/Insectivora. They represent, respectively, the larger clades Soricomorpha and Erinaceomorpha. The oldest erinaceomorphs include Adunator from the early and late Paleocene of North America (154). The fossil record of the other extant lipotyphlans (or “eulipotyphlans”), Talpidae and Solenodontidae, is much younger. McKenna and Bell (154) noted a number of putative Late Cretaceous soricomorphs: Otlestes, Paranyctoides, and Batodon, but these have all been reinterpreted (see above) as basal Lipotyphla, or as basal to extant orders. Archibald (190) interpreted Otlestes as a basal eutherian, lacking derived characters of Lipotyphla, or any other modern order. Later, Averianov and Archibald (191) synonymized it with Bobolestes (from the same local fauna) and regarded it as a questionable zalambdalestoid. Archibald (190) was uncertain about Paranyctoides and Batodon, but retained these records as putative lipotyphlans pending discovery of further specimens, but we cannot confidently assert they are soricomorphs. Micropternodontids such as Carnilestes and Prosarcodon from the early Paleocene of Asia may be soricid relatives (154), but this has not been rigorously

54


tested. We tentatively regard Paleocene erinaceomorphs (192; see below) as the minimum constraint for a soricid– erinaceid clade. With records from the North American Torrejonian, the minimum estimate of Adunator can be tied to the top of the Danian at 61.5 Ma. For the soft maximum paleontological bound we suggest 131.5 Ma (see Archonta). Laurasiatheria (minus Lipotyphla): bat–cow, sheep, dolphin, pig, horse, dog, cat (minimum = 62.5 Ma; soft maximum = 131.5 Ma) The split between Chiroptera (bats) and its sister clade Ferungulata (Cetartiodactyla + Perissodactyla + Carnivora) depends on the oldest members of these included clades. The oldest artiodactyl is Diacodexis from the early Eocene of North America (c. 55 Ma). Artiodactyls are part of the larger clade Cetartiodactyla, with the Cetacea, whales and relatives, and these date back to the early Eocene as well, at about 53.5 Ma (193). The clade may also include the extinct mesonychids, which are known first from the Danian/Thanetian, some 62 Ma (101). The oldest carnivoramorphans are Protictis (above) and the miacoid Ravenictis from the Danian (Puercan, early Paleocene) of North America. Several carnivoran families radiated in the mid to late Paleocene of that continent (194). Undisputed perissodactyls do not appear until the early Eocene. The oldest bats are Archaeonycteris, Palaeochiropteryx, and Icaronycteris (101, 195). Icaronycteris is reported first from the late Paleocene Clarkforkian Mammal Age, substage 3 (CF3), the Phenacodus/Ectocion acme zone, dated at 55.8–55.0 Ma (196). Archaeonycteris and Palaeochiropteryx, are marginally younger, coming from the MP7 level at Dormaal in Belgium and Rians, and possibly Meudon, in France. The European land mammal age MP7 is dated on the Paleocene/Eocene boundary, so 55.8 Ma ± 0.2 myr (107). Hence, neither bats nor cetartiodactyls have as old a fossil record as Carnivora, which dates to the early Torrejonian occurrence of Protictis (discussed earlier), for which we use the date 62.5 Ma. The soft maximum constraint is again set at 131.5 Ma (see Archonta), given the lack of undisputed crown placentals throughout the Cretaceous. Ferungulata: cow, sheep, dolphin, pig–horse, dog, cat (minimum = 62.5 Ma; soft maximum = 131.5 Ma) The cow–dog split is equivalent to the branching point between the clades containing the Orders Cetartiodactyla (even-toed ungulates and whales) and Carnivora (flesheating placental mammals). This is synonymous with

the clade Ferungulata, a clade within Laurasiatheria [as noted below this also includes Pholidota, following Waddell et al. (111)]. The oldest artiodactyl is Diacodexis from the early Eocene of North America (c. 55 Ma). Artiodactyls are part of the larger clade Cetartiodactyla, with the Cetacea, whales and relatives, and these date back to the early Eocene as well, at about 53.5 Ma (197). Mesonychids may be extinct relatives of cetartiodactyls, as they show conspicuous craniodental similarities with the latter, but lack the crucial pedal derived characters of cetartiodactyls (198). Mesonychids are known first from the Danian/ Thanetian, some 62 Ma (101). The oldest carnivoramorphan is the miacoid Ravenictis from the Danian (Puercan, early Paleocene) of North America, and several carnivoran families radiated in the mid- to late Paleocene of that continent (194). The clade Ferungulata includes also the Orders Perissodactyla and Pholidota, but neither of these dates back before the early Eocene. The oldest ferungulates are then Danian (Torrejonian, early Paleocene) in age, so the minimum age constraint for this clade is under 62.5 Ma. The soft maximum constraint is conservatively set at 131.5 Ma (see Archonta). Zooamata: horse–cat, dog (minimum = 62.5 Ma; soft maximum = 131.5 Ma) The dog–horse split is equivalent to the branching point between the Orders Carnivora and Perissodactyla [a clade that also encompasses pangolins in the Order Pholidota and was named Zooamata by Waddell and colleagues (111)]. The minimum age is determined from the oldest members of the carnivoran and perissodactyl lineages, as they predate records of Eocene pangolins (cf. 199). Flynn et al. (200) and others have modified the meaning of Carnivora so that it is restricted by them to the crown clade consisting of Caniformia + Feliformia. The wider clade traditionally called Carnivora, they term Carnivoramorpha. The oldest carnivoramorphans are the viverravids. The oldest generally accepted viverravid is Protictis from the Fort Union/Polecat Bench Formation, assigned to the basal Torrejonian (To1) NALMA, and dated as 63.6–62.5 Ma (196). If Ravenictis from Canada is also a carnivoramorphan (201), and that is debated (202), it extends this date back to at least the Puercan (Pu2), 65.4–64.3 Ma ± 0.3 myr. Most authors also agree that the extinct group Creodonta is the closest relative to Carnivoramorpha (202), and these date back to the Thanetian, 58.7–55.8 Ma ± 0.2 myr, younger than the oldest carnivormorphans.


The oldest perissodactyl is represented by fragmentary teeth that resemble the brontotheriid Lambdotherium from the late Paleocene site of Bayan Ulan in China (203), but the perissodacyl lineage may be extended further back in time. This places the dog–horse split minimally at the basal Torrejonian (62.5 Ma). As for the other supraordinal clades for which no fossil evidence exists up until the Paleocene, a paleontological soft maximum of 131.5 Ma (see Archonta) is tentatively suggested. Carnivora: dog–cat (minimum = 39.68 Ma; soft maximum = 65.8 Ma) The dog–cat split is equivalent to the branching point between the clades Caniformia (dogs, bears, raccoons, seals) and Feliformia (cats, mongooses, hyaenas), the major subdivisions of the Order Carnivora (202). The oldest carnivorans are members of the Families “Miacidae” (paraphyletic) and Viverravidae, known from the early Paleocene onwards (101). However, these have recently been reconstructed outside the Caniformia– Feliformia clade (202), and so cannot provide a minimum date for the dog–cat split. The oldest caniforms are amphicyonids such as Daphoenus and canids such as Hesperocyon, known first from the earliest Duchesnean North American Land Mammal Age (NALMA), which corresponds to magnetochron 18N, and is dated as 39.74 Ma ± 0.07 myr, based on radiometric dating of the LaPoint Tuff (204). Tapocyon may be an even older caniform; it comes from the middle Eocene, Uintan, dated as 46–43 Ma (205), although Flynn and WesleyHunt (202) place this taxon outside the Carnivora. The oldest feliforms may be the nimravids, also known first from the White River carnivore fauna of the Chadronian NALMA, with uncertain records extending to the base of that unit (206). The earliest Chadronian corresponds to the top of magnetochron 17N, and an age of 37.2–36.7 Ma (206, 207). Flynn et al. (200) suggest a caniform–feliform split around 50 Ma. Based strictly on the undisputed occurrence of caniforms from the Duchesnean, the minimum constraint from this event is at 39.68 Ma. The soft maximum constraint is based on the occurrence of the oldest stem carnivorans (miacids, viverravids) in the Torrejonian NALMA of the early Paleocene (see dog–horse), rounded down to the base of the Paleocene at 65.8 Ma. Cetartiodactyla: pig–dolphin, cow, sheep (minimum = 52.4 Ma; soft maximum = 65.8 Ma) The cow–pig split is equivalent to the major division in Artiodactyla between Ruminantia–Tylopoda, and

55

“Suiformes,” containing the controversial, but now fairly well-established hippo-whale clade (208, 209). The oldest terrestrial artiodactyls (e.g., Diacodexis) fall outside this clade. However, the oldest cetaceans, for example, early Eocene Himalayacetus, are not much younger. Because cetaceans comprise closer relatives to ruminant artiodactyls than do either suids or camels (108), their fossil record constrains the minimum divergence of crown artiodactyls. Hence, the cow–pig division is dated minimally by the record of Himalayacetus from the base of the Subathu Formation in Pakistan (210) where it co-occurs with Nummulites atacicus, whose range correlates with nannoplankton zones 11–12, providing a minimum age of 52.4 Ma (211). The absence of any crown cetartiodactyls during the Paleocene may point to the soft maximum constraint at 65.8 Ma. Whippomorpha–Ruminantia: dolphin–cow, sheep (minimum = 52.4 Ma; soft maximum = 65.8 Ma) Within extant Cetartiodactyla, suiforms and camelids fall outside the whippomorph + ruminant clade [(209); “whippomorph” for cetaceans + hippopotamids was coined by Waddell et al. (111) and this name unfortunately appeared before more palatable alternatives such as Cetancodonta of Arnason et al. (212)]. However, the fossil cetacean Himalayacetus is both the oldest known whippomorph and crown cetartiodactyl, as older terrestrial artiodactyls (Diacodexis) cannot be unambiguously reconstructed within the crown clade (208). Hence, the minimum and soft maximum estimates for the last common ancestor of whippomorphs plus ruminants are the same as those for Cetartiodactyla already described. Bovidae: cow–sheep (minimum = 18 Ma; soft maximum = 28.55 Ma) The branching between the cow (Bos) and sheep (Ovis) is an intrafamilial split within the Family Bovidae. Bos is a member of the Tribe Bovini, and Ovis is a member of the Tribe Caprini, which belong respectively to the Subfamilies Bovinae and Antilopinae (213), although the monophyly of Antilopinae is questioned (214). These two subfamilies comprise the Family Bovidae, so the cow–sheep split corresponds to the point of origin of the crown Bovidae. Fernandez and Vrba (214) point to a series of splits within Bovidae that gave rise to the major subfamilies 25.4–22.3 Ma, and they link this to a climatic change at the Oligocene/Miocene boundary. This date is, however,

56


not based directly on fossil evidence, but on a number of best-fitting dates from published morphological and molecular phylogenies. A number of putative late Oligocene bovids (101) have since been rejected. The oldest putative bovid was Palaeohypsodontus zinensis from the Oligocene of the Bugti Hills, Balochistan, Pakistan, and the early Oligocene of Mongolia and China. This is a ruminant, but it lacks unequivocal anatomical features of Bovidae, and is currently excluded from that family (215, 216). Fossil bovids may be identified in the fossil record by the presence of horn cores. The oldest such records, ascribed to Eotragus, come from the early Miocene of Western Europe and Pakistan. For example, Eotragus noyi from the base of the terrestrial sequence on the Potwar Plateau is dated at ~18.3 Ma (217). Eotragus is attributed to Boselaphini, a tribe within the Subfamily Bovinae consisting of the nilgai and other four-horned antelopes. The oldest members of Antilopinae appear to come from the middle Miocene of three continents: Caprotragoides from Asia (India and Pakistan), Tethytragus from Europe (Spain and Turkey), and Gentrytragus from Africa (Kenya and Saudi Arabia), all dated at ~14 Ma (218). The oldest firmly dated bovid then places the minimum constraint on the origin of the family at 18.3 Ma, and we set the soft maximum constraint at the base of the late Oligocene (28.45 Ma ± 0.1 myr), encompassing the many equivocal stem bovids (101), so 28.55 Ma. Placentalia: tenrec, elephant, armadillo–hedgehog, European shrew, bat, cow, sheep, dolphin, pig, horse, dog, cat, rabbit, pika, squirrel, guinea pig, mouse, rat, tree shrew, bushbaby, lemur, marmoset, macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 62.5 Ma; soft maximum = 131.5 Ma) The human–tenrec split is equivalent to the origin of the clade comprising Boreoeutheria, Xenarthra, and Afrotheria. The oldest boreoeutherians are, as already noted (see, e.g., Zooamata), early Paleocene carnivorans, glires, or carpolestids, with the carnivorans at least 62.5 Ma. The oldest reported afrotherians are much younger, dating from the Eocene. The oldest are Phosphatherium and Daouitherium from Ypresian (lower Eocene) phosphorites of the Ouled Abdoun Basin of Morocco (219). The minimum constraint for the Boreoeutheria/ Xenarthra–Afrotheria split is then 62.5 Ma, and the soft maximum constraint is 131.5 Ma (see Archonta).

Atlantogenata: armadillo–tenrec, elephant (minimum = 55.6 Ma; soft maximum = 131.5 Ma) As indicated later, the oldest undisputed afrotherians are Ypresian (basal Eocene) proboscideans from Morocco. The oldest xenarthrans (armadillos, anteaters, sloths) are slightly more ancient, known from the late Paleocene locality of Itaboraí in Brazil (220). Hence, the paleontological minimum divergence for this clade is in the mid- to upper Paleocene, corresponding to the upper Thanetian, 55.6 Ma. The soft maximum is provided on the same basis as for Archonta, hence 131.5 Ma. Afrotheria: tenrec–elephant (minimum = 48.4 Ma; soft maximum = 131.5 Ma) The tenrec–elephant split represents a deep division within Afrotheria. According to current phylogenies, the tenrec, golden moles (Chrysochloridae), elephant shrews (Macroscelidea), and aardvark (Tubulidentata) may form one clade within Afrotheria, and the elephants, hyraxes and sirenians form the other, termed Paenungulata. Paenungulata is widely accepted as a valid clade, having been established on morphological characters, and now confirmed by molecular analyses. In any case, the last common ancestor of tenrec and elephant corresponds to the base of crown-clade Afrotheria. The oldest fossil aardvarks, tenrecs, golden moles, and elephant shrews are generally stated to be Miocene (154), with a possible older elephant shrew, Metoldobotes from the late Eocene Jebel Qatrani Formation of Egypt. Tabuce et al. (221) report a much older golden mole, Chambius, from the Chambi locality in Tunisia (Ypresian). Seiffert and Simons (222) tentatively suggested that Widanelfarasia from near the Eocene/Oligocene boundary in Egypt may be the closest relative to a tenrec–golden moles clade. These records are equaled or predated by the oldest paenungulates. Zack et al. (223) argue that Paleocene apheliscines from North America share a close evolutionary relationship with elephant shrews. The oldest hyraxes are known from the Eocene of North Africa (219, 221). The oldest sirenians are Prorastomus and Pezosiren from the early middle Eocene of Jamaica (219). The oldest proboscidean fossils are Phosphatherium and Daouitherium from Ypresian (lower Eocene) phosphorites of the Ouled Abdoun Basin of Morocco (219). This basal age is confirmed by reports of early Eocene (Ypresian) hyrax and proboscidean fossils from the Tamaguélelt Formation of Mali (224). Extinct putative outgroups of crown-group Paenungulata, such as Desmostylia and Embrithopoda (Arsinoitherium) are younger, being Oligocene in age, while the Anthracobunidae date back to the early Eocene.


At present, no extant clade within Afrotheria, nor any confirmed extinct afrothere clade (with the apheliscine/“condylarth” possibility deserving further scrutiny), predates the Ypresian (early Eocene), dated at ca. 54 Ma, with a minimal age corresponding to the upper limit of the Ypresian at 48.4 Ma. Further study might reveal that certain Paleocene groups belong within one or other afrothere branch, and that could increase the minimum age constraint. The soft maximum constraint is 131.5 Ma (see Archonta). Theria: opossum, kangaroo–tenrec, elephant, armadillo, hedgehog, European shrew, bat, cow, sheep, dolphin, pig, horse, dog, cat, rabbit, pika, squirrel, guinea pig, mouse, rat, tree shrew, bushbaby, lemur, marmoset, macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 124 Ma; soft maximum = 171.2 Ma) The human–opossum branching point is of course synonymous with the split of marsupials and placentals. The earliest alleged “marsupial” dental fossils come from the mid-Cretaceous of North America, including Kokopellia juddi reported (225) from the Mussentuchit Member, in the upper part of the Cedar Mountain Formation, Utah, which is dated as middle to late Albian on the basis of bivalves and palynomorphs. A date of 98.37 Ma ± 0.07 myr was obtained from radiometric dating of zircons in a bentonitic clay layer. However Kokopellia has not been demonstrably placed within crown Marsupialia by any phylogenetic analysis, and indeed lies outside the marsupial crown (but still on its stem) in the few cases in which it has been tested (e.g., 226). Even older is the boreosphenidan Sinodelphys szalayi from the Yixian Formation, Liaoning Province, China, which is placed phylogenetically closer to marsupials than to placentals by Luo et al. (160). This then has taken the stem of the marsupial clade back to the Barremian with an age of at least 124 Ma. The oldest eutherians (on the stem to Placentalia) were also, until recently, restricted to the mid- and late Cretaceous (101), but subsequent finds have pushed the age back step-by-step deeper into the Early Cretaceous. First were Prokennalestes trofimovi and Prokennalestes minor, from the Höövör beds of Mongolia (227), dated as either Aptian or Albian. Then came Montanalestes keeblerorum (228) from the Cloverly Formation (late Aptian–early Albian, ~100 Ma). Then, Murtoilestes abramovi was named (229) from the Murtoi Svita, Buryatia, Transbaikalia, Russia, being dated as late Barremian to middle Aptian (say, 128–120 Ma). These three taxa were based on isolated jaws and teeth. These

57

were all topped by the spectacular find of Eomaia scansoria in the Yixian Formation of Liaoning Province, China (159), a complete skeleton with hair and soft parts preserved. Dating of the Jehol Group of China has been contentious, with early suggestions of a Late Jurassic age for some or all of the fossiliferous beds. Biostratigraphic evidence now confirms an Early Cretaceous (Barremian) age, with several radiometric dates, using different techniques, on three tuff layers that occur among the fossil beds of 124.6 Ma ± 0.01 myr, 125.06 Ma ± 0.18 myr, 125.2 Ma ± 0.9 myr (162). This gives an encompassing age designation of 125.0 Ma ± 0.4 myr for the span of the three tuff layers, and for the fossiliferous beds of the Yixian Formation, based on direct dating. Thus, we conclude a minimum constraint of ca. 124.6 Ma for the divergence of marsupial and placental mammals. An alternative view (230) places southern tribosphenic taxa (see later) on the stem to Theria, pushing the minimum age for Theria to the Jurassic in order to include such taxa as Ambondro. Given the proposal of Woodburne et al. (230) that southern, tribosphenic mammals such as Ambondro are therian (even eutherian), we would set the soft maximum age constraint for Theria within the Jurassic (Bathonian) at 167.7 Ma ± 3.5 myr, so 171.2 Ma. Marsupialia: opossum–kangaroo (minimum = 61.5 Ma; soft maximum = 131.5 Ma) The opossum–kangaroo split is equivalent to the divergence of the two main crown marsupial clades: Ameridelphia and Australidelphia. There are older metatherians from the Cretaceous, such as Kokopellia and Sinodelphys, but these fall phylogenetically outside of the marsupial crown radiation (e.g., 226). Indeed, nonmarsupial metatherians persist well into the Tertiary and, despite close dental similarities with crown didelphids (e.g., Herpetotherium), they fall outside Marsupialia when their full anatomical diversity is examined in a cladistic context (231). Hence, the oldest marsupials are from the Tiupampa fauna of Bolivia (232), dated close to 63 Ma and minimally corresponding to the top of the Danian at 61.7 Ma (107). Specifically, the Tiupampan genus Khasia has been reconstructed as having a particularly close relationship to microbiotheres (233), a radiation for which the only living representatives are two species of Dromiciops. Importantly, microbiotheres have been phylogenetically linked not with other South American marsupials, but with the Australidelphia, based on both morphological and molecular data (e.g., 231, 234). The oldest australidelphians from Australia

58


include possible bandicoots and/or microbiotheres from the early Eocene locality of Murgon (235). Other elements of the Tiupampa fauna, including the skeletally well-known Pucadelphys and Mayulestes, are not demonstrably part of crown Marsupialia (231), nor is the early Paleocene taxon Cocatherium from Chubut province, Argentina (233). Hence, the paleontological minimum constraint for Marsupialia is late Paleocene, 61.5 Ma. As noted earlier, numerous metatherian remains are known from the Cretaceous of both North America (Kokopellia; cf. 225, 228) and Mongolia (Asiatherium; cf. 236), with the Early Cretaceous S. szalayi being the oldest undisputed metatherian (160). Hence, we place the soft maximum constraint for Marsupialia at the Liaoning beds that produced Sinodelphys, at 131.5 Ma. Mammalia: platypus–opossum, kangaroo, tenrec, elephant, armadillo, hedgehog, European shrew, bat, cow, sheep, dolphin, pig, horse, dog, cat, rabbit, pika, squirrel, guinea pig, mouse, rat, tree shrew, bushbaby, lemur, marmoset, macaque, gibbon, orangutan, chimp, neanderthal, human (minimum = 162.9 Ma; soft maximum = 191.1 Ma) The base of the crown clade of modern mammals, marking the split between Monotremata, represented by the platypus, and Theria, represented by the human, might have a number of positions, depending on how many of the extinct Mesozoic mammal groups are included in the clade. As noted earlier, the oldest marsupial, Sinodelphys, and the oldest placental, Eomaia, take the age of Theria back to about 125 Ma. Vincelestes from the La Amarga Formation of Argentina is dated as Hauterivian, and shows the existence of stem Theria at least ca. 136 Ma. According to a widely accepted cladogram of Mesozoic mammals (160, 237, 238), the Theria are part of a larger clade Theriimorpha that includes further extinct clades: Triconodonta, Multituberculata, Symmetrodonta, and Dryolestoidea. Most of these originated in the Late Jurassic, but triconodonts and dryolestoids began earlier, in the Middle Jurassic. Basal triconodonts include Amphilestes and Phascolotherium from the Stonesfield Slate, referred to the Procerites progracilis Zone of the lower part of the middle Bathonian stage on the basis of ammonites (239), and so dated as 166.9–166.5 Ma ± 4.0 myr (240). Tooth-based mammal taxa from the Early Jurassic of India (Kotatherium, Nakundon) and North America (Amphidon) that have been ascribed to Symmetrodonta (238), are not convincingly members of the clade (241), and so are ignored here. The oldest

dryolestoid appears to be Amphitherium, also from the Stonesfield Slate. The oldest monotremes are Steropodon and Kollikodon from the Griman Creek Formation, Lightning Ridge, South Australia, and dated as middle to late Albian, 109–100 Ma. Teinolophos is from the Wonthaggi Formation, Flat Rocks, Victoria, and is dated as early Aptian, 125–121 Ma. In the new cladistic view (160, 226, 237, 238), the Ausktribosphenida from Gondwana are the closest relatives of Monotremata, forming together the Australosphenida. Oldest are Asfaltomylos from the late Middle Jurassic (Callovian) Cañadon Asfalto Formation of Cerro Condor, Argentina (242) and Ambondro from the upper part of the Isalo “Group” (Middle Jurassic, Bathonian) of Madagascar (243). The position of the Madagascar fi nd in the Bathonian is uncertain, so the age range is 167.7 Ma ± 3.5 myr–164.6 Ma ± 4.0 myr. The human–platypus split is then dated on the oldest theriimorph from 166.9–166.5 Ma ± 4.0 myr, similar in age to the less well-dated oldest australosphenidan. On the basis of the available evidence, we follow Luo and colleagues and accept a minimum constraint of 162.9 Ma. The closest relative of Australosphenida + Theriimorpha is Docodonta, and the oldest docodonts are from the Bathonian of Europe, with a possible earlier form from the Kota Formation of India. Further outgroups, Morganucodontidae, Sinoconodon, and Adelobasileus, are known from the Late Triassic and Early Jurassic. The Kota Formation, and several other units from other parts of the world that have yielded fossil mammals, but nothing assignable to the Australosphenida or Theriimorpha, date to the later half of the Early Jurassic, equivalent to the Pliensbachian and Toarcian stages (189.6 Ma ± 1.5 myr–175.6 Ma ± 2.0 myr), and so 191.1 Ma should be used as a soft maximum constraint. Amniota: bird–mammal (minimum = 312.3 Ma; soft maximum = 330.4 Ma) The ultimate divergence date between birds and mammals has been quoted many times as 310 Ma, generally tracing back to Benton (32). Van Tuinen and Hadly (244) trace the history of the use of this date in molecular analyses, and they quote a range of estimates from 338 to 247 Ma, with a preference for the 310 Ma date on the basis of reassessment of the Late Carboniferous timescale. This estimate has been criticized for being used without error bars (36, 244), for being based on uncertain fossils and hence too old (34), for being misdated (37, 244), and


for being poorly bracketed by outgroups above and below (37). Reisz and Müller (37, 54) indeed argue that this calibration point should no longer be used largely because its soft maximum bound is so poorly constrained. The ultimate ancestor of birds and mammals has to be tracked back to the base of the Synapsida and Sauropsida, the larger clades that include mammals and birds, respectively. These two clades together make up Amniota, the clade containing all tetrapods other than amphibians, and the relationships of major groups is agreed by most (e.g., 244–247). The question of the ultimate bird–mammal split becomes synonymous then with dating the origin of the clade Amniota. The oldest identified synapsid is Protoclepsydrops from the Joggins Formation of Joggins, Nova Scotia. The age of the Joggins Formation has been much debated, and figures in the range from 320 to 305 Ma have been cited recently. Reisz and Müller (37) indicate an age of 316–313 Ma, while Van Tuinen and Hadly (244) settle for 310.7 Ma ± 8.5 myr. Detailed field logging and biostratigraphy (248–250) confirm that the Joggins Formation falls entirely within the Langsettian European time unit, equivalent to the Westphalian A, and roughly matching the Russian Cheremshanian, in the later part of the Bashkirian Stage. Earlier dates for these units were equivocal (251), but the Langsettian is given as 314.5– 313.4 Ma ± 1.1 myr in GTS2004 (252). Protoclepsydrops has been classed as an ophiacodontid, not a member of the basalmost synapsid families— Eothyrididae, Caseidae, or Varanopseidae—whose basal members, if ever found, might be of the same age or older. Protoclepsydrops haplous is known from one incomplete partial skeleton and skull (253), but the remains are fragmentary; even the identification of these remains as a synapsid has been questioned (254, 255). Lee (34) used this uncertainty to reject Protoclepsydrops as informative in this discussion, and to look at the next oldest synapsids, such as Echinerpeton and Archaeothyris from the Morien Group of Florence, Nova Scotia (Myachkovian, upper Moscovian, 307.2–306.5 Ma). Because each retained only one derived character of Synapsida, Lee (34) rejected them, and moved up to more complete material of basal synapsids from some 288 Ma. Van Tuinen and Hadly (244) rejected Protoclepsydrops as a useful marker of the bird–mammal split, but accepted Archeothyris as reasonable, with a date of 306.1 Ma ± 8.5 myr. Phylogenetically, the basalmost member of the amniote branch, the Sauropsida (sometimes called Eureptilia, or Reptilia) is Coelostegus from the Upper Carboniferous

59

(Moscovian, Myachkovskian; 306.5 Ma ± 1 myr) of the Czech Republic (256). This is not, however, the oldest sauropsid. Basal sauropsids include a number of genera formerly assigned to the paraphyletic “Protorothyrididae,” outgroups to Diapsida, and the oldest of these in Hylonomus, also from the Joggins Formation at Joggins, Nova Scotia (257–259). Within Sauropsida, the sister clade to Diapsida + “Protorothyrididae” is the clade Captorhinidae, but the oldest captorhinid is younger— Romeria primus from the Moran Formation of Texas (Early Permian, Sakmarian, ~294–284 Ma; 260). Lee (34) cast doubt on the assignment on Hylonomus to the sauropsid clade, and preferred to redate that branch also to some 288 Ma. Lee’s (34) proposal would move the mammal–bird split date from somewhere around 310 to 290 Ma, whereas Van Tuinen and Hadly (244) settled for 305 Ma as a minimal date. However, Reisz and Müller (37) and Van Tuinen and Hadly (244) suggested that Lee was wrong to cast doubt on nearly all the Carboniferous synapsids and sauropsids—many are diagnostic of one or other group. More importantly though, Reisz and Müller (37) pointed out that the question of dating the ultimate bird–mammal split is synonymous with dating the origin of Amniota. So, it may be uncertain whether Protoclepsydrops is a synapsid, and Hylonomus is not a diapsid, and the “Protorothyrididae” are clearly paraphyletic (256), but all these taxa are diagnostically members of Amniota, so the origin of Amniota happened before the age of the Joggins Formation of Nova Scotia. Older evidence of amniotes has been reported by Falcon-Lang et al. (261), footprints with a number of amniote derived characters (pentadactyl manus and pes, slender digits whose relative lengths approximate a phalangeal formula of 23453 (manus) and 23454 (pes), narrow digit splay (40–63°), putative transverse scale impressions on digit pads, and straight tail drag) that come from the Grand Anse Formation of Nova Scotia. This unit lies 1 km below the Joggins Formation, and is dated at ~1 myr older than the Joggins. However, although we are convinced the footprint evidence represents amniotes, such a fossil is less reliable than a skeletal body fossil, and can be taken as an indication of a future increase in the minimum constraint on the bird– mammal split. The minimum constraint on the mammal–bird split, equivalent to the minimum age of the origin of clade Amniota corresponds to the age of the Joggins Formation. This is 314.5–313.4 Ma ± 1.1 myr, a date based on

60


biostratigraphy (palynology) and exact dating from elsewhere, conferring a minimum constraint of 312.3 Ma. The soft maximum constraint on the bird–mammal split is based on the next richly fossiliferous units lying below these horizons. The first is the East Kirkton locality, source of a diverse fauna of batrachomorphs and reptiliomorphs (see human–toad split later), but that has hitherto not yielded anything that could be called either a diapsid or a synapsid. Further fossiliferous sites of similar facies lie below the East Kirkton level, and they have not yielded reptile remains. We take the age of the fossiliferous Little Cliff Shale of the East Kirkton locality (Brigantian; 328.8–326.4 Ma ± 1.6 myr) as the basis for the soft maximum age constraint of 330.4 Ma.

Diapsida: lizard–crocodile, emu, chicken, zebrafinch (minimum = 255.9 Ma; soft maximum = 299.8 Ma) The clades Crocodylia (modern crocodilians and extinct relatives) and Squamata (modern lizards and snakes and their extinct relatives) are members, respectively, of the larger clades Archosauromorpha and Lepidosauromorpha. The ultimate split between crocodilians and lizards then is marked by the split between those two, and they, together with a number of basal outgroups, form the major clade Diapsida. Through a series of cladistic analyses (245, 257, 258, 262–265), the topology of the basal region of the cladogram around the split of Archosauromorpha and Lepidosauromorpha has been agreed (although some higher parts of the cladogram are still much debated, especially the placement of Sauropterygia and Ichthyosauria). The most ancient archosauromorph is the “prolacertiform” Protorosaurus speneri from the Kupferschiefer of Germany and the Marl Slate of NE England (266). Both geological units are correlated with each other on independent geological evidence, and defined as the basal unit of the Zechstein 1 (EZ1; Werra Folge) depositional cycle. The two units were generally assigned to the Kazanian (e.g., 260, p. 695), but subsequent stratigraphic revisions have shown that the Zechstein falls above the Illawarra Reversal, which is at the Wordian–Capitanian boundary, and the Zechstein I contains fossils characteristic of the Capitanian (267). It is unclear how much of the Capitanian is represented by the Zechstein, but it probably represents the upper part, so 263.8–260.4 Ma ± 0.7 myr. Roscher and Schneider (268) estimate a minimum age for the Kupferschiefer as 255.9 Ma, so we accept that here, as a younger estimate, based on new dating

evidence, than the date of 259.7 Ma given by Benton and Donoghue (3). The most ancient lepidosauromorph is debated— Benton (260, p. 688) indicated that Saurosternon bainii, sole representative of the Saurosternidae, may be the oldest, but he was uncertain. Other authors (246, 257, 269) were more convinced that this is a true lepidosauromorph. The doubt arises because the taxon is based on a single partial skeleton lacking the skull. Saurosternon is from the Cistecephalus or Dicynodon Assemblage Zone of South Africa (270) equivalent to the uppermost Wuchiapingian or Changhsingian, respectively, perhaps some 257–251 Ma. However, numerical cladistic analyses (271, 272) have shown unequivocally that Saurosternon and other supposed Permo-Triassic “lizards” are not lepidosauromorphs or even neodiapsids. If Saurosternon is not a lepidosauromorph, the next possibility would be a sauropterygian. The uncertainty here is whether sauropterygians are lepidosauromorphs—the group was unplaced phylogenetically for some time, but deBraga and Rieppel (245) and others have made a strong case that these marine reptiles are unequivocal lepidosauromorphs. Benton (260, p. 70) listed two Lower Triassic (Scythian) sauropterygians, Corosaurus and Placodus, but the dating of both is uncertain. Corosaurus is from the Alcova Limestone Member of the Chugwater Formation in Wyoming, formerly assigned to the Middle or Upper Triassic, but noted as Lower Triassic by Storrs (273). The precise age is hard to pin down. The Lower Triassic Placodus is from the Obere Buntsandstein of Pfalz, Germany, a unit dated as spanning the Olenekian–Anisian boundary, and ranging in age from 246–244 Ma ± 1.5 myr. Based on the oldest neodiapsid, Protorosaurus, the minimum constraint on the divergence of crocodilians and lizards is 255.9 Ma. In order to establish the soft maximum constraint on this divergence, outgroups to Neodiapsida are considered. Ichthyosauria are known first in the Early Triassic, younger than the minimum age constraint. Younginiformes, Weigeltisauridae (Coelurosauravus), and Claudiosaurus are of similar age to Protorosaurus, or younger. Next oldest is Apsisaurus from the Archer City Formation of Texas, dated as Asselian (299–294.6 Ma ± 0.8 myr) (260), and so 299.8 Ma, although its diapsid affinities have been questioned (274). This is a long way below the minimum age constraint, but there is a well-known “gap” in suitable fossiliferous formations through the mid-Permian, and we retain this possibly exaggerated soft maximum constraint.


Archosauria: crocodile–emu, chicken, zebrafinch (minimum = 239 Ma; soft maximum = 250.4 Ma) The most recent common ancestor of crocodilians and birds was an archosaur that lay at the deep junction of the two major clades within Archosauria: Avemetatarsalia/ Ornithodira, the “bird” line and Crurotarsi, the “crocodile” line (275–277). These two clades together form the Avesuchia (= “crown-group Archosauria”). The basal crurotarsans are the poposaurid Bromsgroveia from the Bromsgrove Sandstone Formation of England, and the “rauisuchians” Wangisuchus and Fenhosuchus from the Er-Ma-Ying Series of China, Vjushkovisaurus from the Donguz Svita of Russia, the ctenosauriscid Arizonasaurus from the Moenkopi Formation, and Stagonosuchus and “Mandasuchus” from the Manda Formation of Tanzania (260). All these records are dated as Anisian, but most cannot be dated more precisely; Arizonasaurus though can be assigned to the lower Anisian (278). This gives an age range of 245 Ma ± 1.5 myr–241 Ma ± 2.0 myr. The basal avemetatarsalian is Scleromochlus from the Carnian of Scotland, but older relatives are Marasuchus, Lagerpeton, and Pseudolagosuchus from the Chañares Formation of Argentina, dated as Ladinian, so 237 Ma ± 2.0 myr–228.0 Ma ± 2.0 myr. The minimum constraint on the divergence date for birds and crocodiles then falls at the top of the lower Anisian (245 Ma ± 1.5 myr–241 Ma ± 2.0 myr), and so 239 Ma, an increase of 4 myr over the date given by Benton and Donoghue (3) as a result of the closer dating of Arizonasaurus. The soft maximum constraint may be assessed from the age distribution of immediate outgroups to Avesuchia, the Proterochampsidae, Euparkeriidae, Erythrosuchidae, and Proterosuchidae (275–277). Numerous fossil sites from around the world in the Olenekian, the stage below the Anisian, have produced representatives of these outgroups, but not of avesuchians, and so the Olenekian (249.7 Ma ± 0.7 myr–245 Ma ± 1.5 myr) marks the soft maximum age constraint, and so 250.4 Ma. Neornithes: emu–chicken, zebrafinch (minimum = 66 Ma; soft maximum = 86.5 Ma) The divergence of emu and chicken is synonymous with the deep divergence between the Palaeognathae (ratites, or flightless birds) and the Neognathae (all other, flying, birds). The oldest palaeognaths are the lithornithids, a family known from the Paleocene and Eocene of North America. A putative latest Cretaceous lithornithid was reported by

61

Parris and Hope (279) from the New Jersey greensands. The age of these deposits has been much debated (280), and they fall either below or above the K-T boundary (65.5 Ma ± 0.3 myr). An older specimen might be mistakenly assigned here: the pelvis of a large flightless bird, Gargantuavis philoinos, reported (281) from the base of the Marnes de la Maurines Formation, in association with dinosaurs of late Campanian to early Maastrichtian aspect. These authors were clear that Gargantuavis was not a palaeognath, and suggested it might be related to the non-neornithine Patagopteryx. The oldest confirmed neognath fossil is the anseriform Vegavis from 66 Ma, and this has to be the minimum constraint on the divergence date for palaeognaths and neognaths. The soft maximum constraint is currently the same as for the chicken–zebrafinch split below, namely the clades Ichthyornithiformes and Hesperornithiformes of the Niobrara Chalk Formation, dated as Santonian (85.8–83.5 Ma ± 0.7 myr), and so 86.5 Ma. Neognathae: chicken–zebrafinch (minimum = 66 Ma; soft maximum = 86.5 Ma) The phylogeny of major groups of modern flying birds (clade Neognathae) has been hard to resolve. Recent morphological and molecular analyses now agree on a deep divergence between the clade Galloanserae, comprising Galliformes (chickens and game birds) and Anseriformes (ducks) on the one hand, and Neoaves (all other flying birds) on the other (280, 282). The oldest purported galloanserine is Teviornis gobiensis, a presbyornithid anseriform from the Gurilyn Tsav locality of Mongolia (283). Sediments here come from the lower portion of the Nemegt Horizon, at the base of the Nemegt Formation. The Nemegt Formation is assigned to the early Maastrichtian (284), dated as 70.6 Ma ± 0.6 myr to 69.6 Ma ± 0.6 myr. Doubt has been cast, however (285), on whether Teviornis is a neognath, let alone a galloanserine, so the next youngest purported neognath should be selected until this issue is clarified. A further latest Cretaceous anseriform is Vegavis iaai from lithostratigraphic unit K3 of Vega Island, Antarctica, dated as mid- to late Maastrichtian, ~68–66 Ma (286). The oldest galliform fossil that can be identified with confidence is much younger, early Eocene (282). The oldest neoavian is debated, with dozens of records of gaviiforms, pelecaniforms, charadriiforms, procellariiforms, and psittaciforms from the latest Cretaceous (most are close to the Cretaceous–Tertiary boundary, 65.5 Ma; 287, 288). The most complete fossil is Polarornis gregorii, described as a loon (gaviiform) from the Lopez

62


de Bertodano Formation of Seymour Island, Antarctica (289). This stratigraphic unit is dated as mid- to late Maastrichtian on the basis of microplankton (290), so 69.6–65.5 Ma ± 0.3 myr. Dyke and Van Tuinen (280) indicate some doubt about the taxonomic assignment of the specimen and about its geological provenance. Even if the various neoavian specimens fall close to the Maastrichtian–Danian boundary, and if there is some doubt about Polarornis and Teviornis, the galloanserine record of Vegavis is older, and dates the minimum constraint on chicken–zebrafinch divergence at 66 Ma, on the basis of biostratigraphy and indirect dating. The soft maximum constraint is based on older birdbearing deposits that match some at least of the facies represented in the late Maastrichtian, which are broadly from the shallow marine to coastal belt. Fossil birds, most notably, hundreds of specimens of Hesperornis, Baptornis, and Ichthyornis (members of the clades Ichthyornithiformes and Hesperornithiformes, both outgroups to Neornithes), but no Neornithes have been found in abundance from the Niobrara Chalk Formation of Kansas and neighboring states, dated as Santonian (85.8–83.5 Ma ± 0.7 myr), and so 86.5 Ma. Tetrapoda: toad–bird, mammal (minimum = 330.4 Ma; soft maximum = 350.1 Ma) The African clawed toad (Xenopus laevis) is a representative of modern Amphibia (the clade Lissamphibia, including frogs and toads, salamanders, and caecilians), and the human–toad split is equivalent to the deep branching point between Amphibia and Amniota. Within crown Tetrapoda, this is the split of Batrachomorpha (extant lissamphibians and extinct relatives) and Reptiliomorpha (extant amniotes and their extinct relatives). The oldest batrachomorph is Balanerpeton woodi, a basal temnospondyl from the East Kirkton locality in Scotland. Another putative basal batrachomorph is Eucritta melanolimnetes, from the same location, described as a possible baphetid (291), but possibly a batrachomorph (292). The fossils come from the Little Cliff Shale, a unit within the East Kirkton Limestone, a subdivision of the upper Oilshale Group of the Midland Valley of Scotland. The fossil beds are ascribed to the Brigantian (D2; lower portion) of the Viséan stage, based on biostratigraphic comparisons of the fossil plants, pollen, and bivalves with the rich records of Lower Carboniferous sites throughout Europe (293). The Brigantian regional stage is dated 328.8–326.4 Ma ± 1.6 myr. The oldest reptiliomorphs are the basal lepospondyl Westlothiana lizziae, and the aïstopod Lethiscus

stocki (292). Westlothiana and Lethiscus are both from the Viséan. Westlothiana comes from the East Kirkton locality, and is dated at 327.6 Ma ± 2.8 myr (see earlier). L. stocki is from the Wardie Shales, part of the Lower Oil Shale Group, near Edinburgh, and dated as older than the East Kirkton locality (293). The Wardie Shales are assigned to the Holkerian regional stage on the basis of fossil fishes and palynomorphs (294), dated as 339.2–332.4 Ma ± 2.0 myr. Van Tuinen and Hadly (244) reviewed the amphibian–amniote divergence date in detail, but assigned the Wardie Shales to the Asbian, the stage above the Holkerian, and so came to an age of 332.3 Ma. Further, they used radiometric dates from Menning et al. (251) which have been revised in GTS2004 (267). Our minimum constraint on the human–toad divergence is 330.4 Ma, based on Lethiscus, and biostratigraphic placement of the Wardie Shales Formation, with radiometric dating of the Holkerian from elsewhere. The soft maximum constraint is harder to determine because most of the close outgroups to the batrachomorph–reptiliomorph clade are known only from younger deposits: the oldest baphetids and crassigyrinids are from the Brigantian (260), the oldest colosteids from the Asbian (260). The whatcheeriids Whatcheeria and Pederpes, from North America and Europe, respectively, are older, however, and dated to the Ivorean regional North American stage, and so 348–345.3 Ma ± 2.1 myr. These horizons are underlain by further units of Famennian age, dated as 374.5 Ma ± 2.6 myr–359.2 Ma ± 2.5 myr, with basal tetrapods known from several continents, but no batrachomorphs or reptiliomorphs. We choose the whatcheeriids as marking the soft maximum constraint, even though they are phylogenetically more distant from crown Tetrapoda than baphetids and colosteids—but the latter two are younger than Lethiscus. Thus, we propose a date of 350.1 Ma as a soft maximum constraint. Osteichthyes: zebrafish, Medaka, stickleback, Takifugu, Tetraodon–toad, bird, mammal (minimum = 416.0 Ma; soft maximum = 421.75 Ma) This divergence event represents the origin of crown Osteichthyes and the splitting of Actinopterygii and Sarcopterygii. Thus, the minimum constraint depends on determining the oldest member of either clade. The earliest representative of total-group Actinopterygii may be Andreolepis hedei, known from microfragments from Gotland, Sweden (295–298), and elsewhere (299). It has been assigned to total-group Actinopterygii


on the following derived characters: rhomboid scale shape, ganoine-covered scales. The oldest occurrence that is readily constrained is from the lower part of division C of the Hemse Marl at Västlaus, Gotland, Sweden (296). Although there are no direct radiometric dates from the Ludlow of Gotland, these sections have been incorporated into a graphic correlation composite standard that includes radiometric dates (300, 301). Thus, a date for this occurrence can be established from the composite standard through the line of correlation, which equates to 421.75 Ma. The certainty with which A. hedei is assigned to Actinopterygii is obviously less than it might be were it known from articulated remains. It is known from a number of skeletal elements (295, 302–304), rather than mere scales, as are the other, slightly younger, early records of Actinopterygii (305, 306), but these have led some researchers to conclude a stem-Osteichthyes, rather than a stem-Actinopterygii affinity (304). Naxilepis, although known only from scales (306), possesses a further derived character of total-group Actinopterygii, in addition to those exhibited by A. hedei, namely a narrow-based dorsal peg and discrete rows of ganoine. The earliest occurrence is from the Miaogao Member of the Cuifengshan Formation of Quijing District, Yunnan, China, where it co-occurs with the conodont Ozarkodina crispa (306, 307), although this has not been substantiated. As later, this constrains the age of the first occurrence of Naxilepis between the middle Ludlow and the Ludlow–Pridoli boundary (418.7 Ma). The earliest macroremains assignable to total group Actinopterygii are of Dialipina markae from the Lochkovian of Siberia (308), which is also known from fully articulated remains from the Lower Devonian (Emsian) of the Canadian Arctic (309). Justification for the Lochkovian age assignment is not clear (310). The earliest record of the sarcopterygian total group is Psarolepis romeri, known (in stratigraphic order) from the Yulongsi (311), Xishancun (311), and Xitun (312) members of the Cuifengshan Formation, Quijing District, eastern Yunnan, China (the recently described Meemania eos is apparently a more basal member of the sarcopterygian stem but it is known only from the Xitun Member; 313). The dating of these occurrences relies primarily upon biostratigraphic dating of a lithostratigraphic correlation of the Yulongsi Member in neighboring Guangxi, where the conodont O. crispa has been found in the middle of the Yulongsi Member (314). The lower limit of the stratigraphic range of O. crispa is constrained by the Ludlow–Pridoli Boundary (315), which

63

has been dated as 418.7 Ma ± 2.7 myr (316). In the type Ludlow Series the upper range limit on O. crispa is just a few meters below its lower limit (315) (the latest Ludlow and earliest Pridoli are probably unrepresented in the Ludlow type area; 317). Although it is difficult to provide a direct date on this horizon, zircons from a bentonite 12 m deeper in the Ludlow type section have provided a U-Pb Zircon age of 420.2 Ma ± 3.9 myr (318). There is a report of O. crispa as low as “middle Ludlow” (319), although this is just one of a number of possible interpretations of the conflicting biostratigraphic data. Attempts to directly date the Quijing succession biostratigraphically have yielded the conodont Oulodus elegans detorta from the upper part of the Yulongsi Member (320). The stratigraphic range of O. elegans detorta is confined to its zone, which is the ultimate conodont zone of the Silurian (321). Thus, direct and indirect biostratigraphic dating is in agreement concerning the age of the middle and upper parts of the Yulongsi Member, indicating that the earliest record of Psarolepis is no younger than latest Ludlow (418.7 Ma ± 2.7 myr) and possibly older than 420.2 Ma ± 3.9 myr. Although originally described as a sarcopterygian (311, 312), Psarolepis has also been interpreted as stemOsteichthyes (322, 323). However, more recent and universal analyses have confirmed its assignment to the sarcopterygian stem-lineage (313, 324). After Psarolepis, the next oldest representatives of total-group Sarcopterygii, Diabolepis, Youngolepis, and Achoania are approximately coeval. They come from the Xishancun Member of the Cuifengshan Formation of Qujing District. The Xishancun Member is clearly younger than the underlying Yulongsi Formation, the upper part of which is dated as latest Silurian age on the occurrence of O. elegans detorta (see earlier), and it has been directly dated as Lochkovian on the basis of ostracode biostratigraphy (325). Outgroups of the Actinopterygii + Sarcopterygii clade may provide evidence for a soft maximum age constraint. Lophosteus superbus, described on the basis of a wide variety of microremains (326, 327) has been considered stem-Osteichthyes (328), although this is poorly substantiated (299, 304, 329). The earliest occurrence of L. superbus from the Pridoli of Gotland (326, 327), Estonia (298), and Latvia (298) is later than the first record of A. hedei which, despite concerns over assignment to Actinopterygii (329), has not been disputed membership of total-group Osteichthyes. Indeed, some of the evidence on which Andreolepis has been assigned to Actinopterygii can be called into question on the basis

64


of the discovery and phylogenetic position of Meemania, in which a ganoine-like tissue appears to be present (313). Thus, it is possible that Andreolepis presents only osteichthyan ancestral characters and that, on the basis of the available evidence, it is better assigned to stemOsteichthyes (304). Dating the earliest record of successive sister taxa is complicated by long-standing debate over the relative phylogenetic position and monophyly of the various groups. Acanthodii is generally considered the sister group of Osteichthyes and its earliest record is from the Ashgill of Siberia (330). Chondrichthyes is generally accepted as the succeeding sister taxon, the oldest record for which is Caradoc (60), although precious few characters bind these remains to the stem of Chondrichthyes (331). The oldest placoderms are undescribed forms from the Wenlock of China (329) and Vietnam (332). Conservative assessments of the age of the earliest remains readily assignable to the actinopterygian and sarcopterygian total groups are in close approximation (421.75 Ma ± 0 myr vs. 418.7 Ma ± 2.6 myr, respectively). However, phylogenetic assignment of these microremains rests on one or two equivocal derived characters, and this is insufficient evidence on which to justify constraining molecular clock analyses. Thus, we argue that it is best to rely on the evidence of better known and better phylogenetically constrained Psarolepis to provide a minimum constraint on the divergence of sarcopterygian and actinopterygian lineages. The firmest age dating on the earliest record of Psarolepis (based on biostratigraphic correlation) is 418.7 Ma ± 2.7 myr. Thus, a minimum constraint on the divergence of crown Osteichthyes should be quoted as 416.0 Ma. A soft maximum constraint could be provided by the age of the earliest record of A. hedei, dated at 421.75 Ma. Clupeocephala: zebrafish–Medaka, stickleback, Takifugu, Tetraodon (minimum = 149.85 Ma; soft maximum = 165.2 Ma) This divergence event represents the splitting of the ostariophysean and euteleost lineages. The earliest ostariophysean is Tischlingerichthys viohli from the Tithonian upper Solnholfen Limestone Formation of southern Germany (333). It is recognized on the basis of derived characters including the absence of a basisphenoid, and dorsomedial portions of the anterior neural arches expanding and abutting against each other and the posterior margin of the exoccipital. From the same deposit, Arratia (333) also described a number of additional taxa (Leptolepides,

Orthogonikleithrus) that qualify as the earliest record of the euteleost lineage. These were assigned to Salmoniformes. The security of their assignments to these higher-level clades within Euteleostei is questionable, although their assignment to the euteleost totalgroup is not, based not least on the presence of enlarged neural arches/spines. Thus, earliest representatives of both lineages are in precise agreement. However, this should come as no surprise given that they were found in the same deposit. Therefore, the fossil date is likely to be a considerable underestimate, subject to lagerstätten effect. There are no earlier records. The dating of the upper Solnhofen Limestone Formation has been based on ammonite zonation and the Formation is assigned to the ti2 division of the middle Tithonian, Late Jurassic. The Tithonian is dated as 150.8 Ma ± 4.0 myr–145.5 Ma ± 4.0 myr (240), but the upper Solnhofen Limestone Formation represents just the middle biohorizon of the lowest ammonite zone of the Tithonian (334), its base intercalated by the first (local) appearances of the ammonites H. hybonotum (and Gravesia) and Glochiceras lithographicum (335). In proposed stratotype sections, the base of the Tithonian is represented by the simultaneous first appearance of these two taxa plus the immediately subsequent appearance of Gravesia species (240). The base of the H. hybonotum Zone coincides with the base of the normal polarity Chron M22An which is dated at 150.8 Ma ± 0.1 myr (240). Given that the Solnhofen Formation falls fully within the H. hybonotum Zone, it is possible to derive a lower bound on its age from the base of the succeeding, S. darwini ammonite zone which coincides approximately with the M22n Chronozone, dated at 149.9 Ma ± 0.05 myr (240). Thus, the earliest paleontological evidence and, therefore, a lower bound on the split of Danio rerio–Takifugu rubripes, Tetraodon nigris can be considered to be 150.8 Ma ± 0.1 myr–149.9 Ma ± 0.05 myr, giving a minimum date of 149.85 Ma. However, note should be taken of the fact that the co-occurrence of the earliest records of these two lineages is an artefact of their presence in a Konservat-lagerstatten. A soft maximum constraint on the divergence of the ostariophysean and euteleost lineages is provided by the census of teleost–total group diversity provided by the assemblages found in the many Oxfordian localities in the Cordillera de Domeyko (336). Many species are known in conditions of exceptional preservation and these are stem teleosts; no otophysans or euteleosts are known from here or from older deposits.


The base of the Oxfordian (161.2 Ma ± 4.0 myr; 240) can be taken as the soft maximum constraint: 165.2 Ma. Medaka–stickleback, Takifugu, Tetraodon (minimum = 96.9 Ma; soft maximum = 150.9 Ma) This divergence event represents the split between Atherinomorpha and Percomorpha within Acanthopterygii. The oldest member of Atherinomorpha, based on otoliths of “Atherinidarum,” from Argile de Gan, Gan, Pyrénées-Atlantiques, France, has been assigned an early Eocene (Ypresian) age (337). The earliest skeletal records are late Eocene (Priabonian) (338). The oldest percomorph is the stem-tetraodontoform Plectocretacicus clarae, from the Cenomanian (Upper Cretaceous) of Hakel, Lebanon (339, 340). The age of the lithographic limestones at Hakel is derived from the occurrence of Mantelliceras mantelli and the benthic foraminifer Orbitulina concava (341). The stratigraphic range of O. concava is late Albian to early Cenomanian (342), while M. mantelli defines the basal ammonite zone of the Cenomanian. The base of the M. mantelli Zone is well dated on the basis of Ar–Ar and cyclostratigraphy at 99.1 Ma ± 0.4 Ma (343). Ogg et al. (344) provide a date of 97.8 Ma for the top of the M. mantelli Zone; errors on the timescale on surrounding zonal boundaries are 0.9 myr. Thus, the minimum age of the divergence of Atherinomorpha and Percomorpha can be based on the minimum age of the lithographic limestones of Hakel, which would be 96.9 Ma. The most appropriate soft maximum bound on the divergence of Gasterosteiformes and Tetraodontiformes would be the earliest euteleost records, provided by taxa such as T. viohli and associated crown euteleosts from the Tithonian of Solnhofen (333). Acanthopterygians (as are convincing members of any elopocephalan superorders or orders) are entirely absent. The soft maximum age of the Solnholfen lithographic limestones (justified above in connection with the ostariophysean–euteleost split) is 150.8 Ma ± 0.1 myr. Thus a soft maximum constraint for divergence of the gasterosteiform and tetraodontiform lineages is 150.9 Ma. Stickleback–Takifugu, Tetraodon (minimum = 96.9 Ma; soft maximum = 150.9 Ma) This divergence event represents the split between Gasterosteiformes and Tetraodontiformes within Percomorpha. The oldest member of Gasterosteiformes is Gasterorhamphosus zuppichinii from the Calcare di Mellissano, near Nardò, Lecce, Apulia, southeastern Italy (345), which is believed to be Campanian (Late Cretaceous) in age (338). This is younger than the oldest

65

known member of the tetraodontiform lineage, P. clarae, the earliest stem-tetraodontiform, from the Cenomanian (Upper Cretaceous) of Hakel, Lebanon (339, 340). The age of the lithographic limestones at Hakel is derived from the occurrence of M. mantelli and the benthic foraminifer O. concava (341). The stratigraphic range of O. concava is late Albian to early Cenomanian (342), while M. mantelli is more restricted temporally, and falls fully within the range of M. mantelli, defining the basal ammonite zone of the Cenomanian. The base of the M. mantelli Zone is well dated on the basis of Ar–Ar and cycle stratigraphy at 99.1 Ma ± 0.4 Ma (343). Ogg et al. (344) provide a date of 97.8 Ma for the top of the M. mantelli Zone; errors on the timescale on surrounding zonal boundaries are 0.9 myr. Thus, the minimum age of the divergence of Atherinomorpha and Percomorpha can be based on the age on the minimum age of the lithographic limestones of Hakel, which would be 96.9 Ma. Given that P. clarae is also the oldest known percomorph (338), the most appropriate soft maximum bound on the divergence of Gasterosteiformes and Tetraodontiformes would be the earliest euteleost records, provided by taxa such as T. viohli and associated crown euteleosts from the Tithonian of Solnholfen (333). Acanthopterygians (as are convincing members of any elopocephalan superorders or orders) are entirely absent. The soft maximum age of the Solnholfen lithographic limestones (justified above in connection with the ostariophysean– euteleost split) is 150.8 Ma ± 0.1 myr. Thus a soft maximum constraint for divergence of the gasterosteiform and tetraodontiform lineages is 150.9 Ma. Tetraodontidae: Takifugu–Tetraodon (minimum = 32.25 Ma; soft maximum = 56.0 Ma) Following the phylogenetic scheme of Holcroft (346) this divergence event represents the origin of crown-group Tetraodontidae. Archaeotetraodon winterbottomi has been identified as a member of this clade on the presence of numerous tetraodontid derived characters, including 11 caudal fin rays, 18 vertebrae, broadened neural and haemal spines and an absence of ribs (347). It has been recorded from the Pshekhsky Horizon, in the lower part of the Maikop Formation of the north Caucasus, Russia (348), making it the earliest known member of Tetraodontidae (347). The lower part of the Maikop Formation has been widely quoted as lower Oligocene (348, 349), although evidence is rarely presented in support of this. Leonov et al. (350) provide evidence on the age of the Pshekhsky Horizon on the basis of planktic and benthic

66


foram, nannoplankton and dinocyst biostratigraphy. The base of the Pshekhsky Horizon coincides with the base of the range of Globigerina tapuriensis, which belongs to Zone P18 of the Paleogene planktic foram zonation scheme (351). The base of P18 equates to the base of the Oligocene, which has been dated at 33.90 Ma ± 0.1 myr (107). The top of the Pshekhsky Horizon coincides approximately with the first appearance of the nannoplankton species Sphenolithus predistentus, and the base of NP23, a paleogene nannoplankton zone (350). The latter has been dated at 32.25 Ma (107), the errors on which are negligible, though there will be an inherent uncalculated error on the biostratigraphic correlation to the Caucasus. Thus, paleontological evidence on the divergence of the lineages leading to T. rubripes and Tetraodon nigroviridis provides a minimum constraint of 32.25 Ma. Relationships within Tetraodontiformes have been approached from anatomy and molecular phylogenetics, but remain poorly constrained. Nonetheless, the oldest records for the potential sister clades are all of Eocene age and among them, the oldest record is provided by the balistid Moclaybalistes danekrus, known from the lower Eocene Mo-Clay (Fur/Ølst) Formation, which has been dated using magnetostratigraphy and biostratigraphy using nannoplankton, dinoflagellate and pollen zones (352). The base of the Ølst Formation coincides with base of Dinoflagellate Zone 6 and the base of the Apectodinium augustum Zone, which coincides with the base of the Eocene. A soft maximum constraint on the split of T. rubripes and T. nigroviridus can thus be obtained from the age of the base of the Eocene which has been dated at 55.8 Ma ± 0.2 myr (107), thus 56.0 Ma. Gnathostomata: shark–fish, tetrapod (minimum = 421.75 Ma; soft maximum = 462.5 Ma) This divergence represents the origin of crown Gnathostomata and the splitting of Chondrichthyes and Osteichthyes. The oldest possible record of Chondrichthyes is based on isolated scales from the Late Ordovician Harding Sandstone of Colorado (60). These scales exhibit a single chondrichthyan derived character, the presence of a neck canal. There is a sequence of younger records, all based on isolated or fragmentary material and attributed to Chondrichthyes on one or, at most, a couple of derived characters (353–356). None of these is sufficient to establish the existence of total-group Chondrichthyes, at least to the degree of certainty necessary to calibrate or even constrain a molecular clock analysis. In this

regard, the earliest records that provide adequate evidence of chondrichthyan affinity are of Early to Middle Devonian age: Doliodus problematicus from the Emsian of Canada (357), and Pucapampella from the Emsian of South Africa (358, 359). These records are considerably younger than the oldest record of stem-Osteichthyes. The oldest possible record of Osteichthyes is based on isolated scales, attributed to acanthodians, from the Late Ordovician of Siberia (330). Further records of isolated acanthodian scales and spines are known from the Wenlock onwards but, given that the oldest articulated acanthodians are of Devonian age and younger, the degree to which acanthodian-like scales correlate with what is otherwise known of acanthodian anatomy is extremely uncertain. Thus, the oldest phylogenetically secure record of the divergence of crown gnathostomes is established on the basis of Andreolepis hedei, which is at least a stemOsteichthyan, if not a stem-Actinopterygian (see crown Osteichthyes). The oldest record of A. hedei is established on the basis of a graphic correlation composite standard, at 421.75 Ma (see crown Osteichthyes). A soft maximum constraint can be established on the basis of the oldest phylogenetically secure stemgnathostome, Sacabambaspis janvieri, dated at 462.5 Ma (see crown Vertebrata). Vertebrata: lamprey–shark, fish, tetrapod (minimum = 460.6 Ma; soft maximum = 581 Ma) Establishing a date on this divergence is complicated by debate over the interrelationships of hagfishes, lampreys, and gnathostomes. Hagfishes and lampreys were long united as cyclostomes to the exclusion of gnathostomes (360) until in the 1970s a number of authors independently recognized that lampreys and gnathostomes shared a number of morphological characters lacking in hagfishes (361–363). Morphology-based cladistic analyses continue to recognize a long and convincing inventory of features supporting this hypothesis of relationships even in the face of universal support for cyclostome monophyly from molecular datasets (364–366). In our view, the evidence from molecular data is now so compelling that we accept cyclostome monophyly and the likelihood that many characters hitherto considered derived features of lampreys + gnathostomes are more appropriately interpreted as ancestral vertebrate characters. Given the above, the divergence of lampreys and gnathostomes equates to the origin of crown vertebrates. A number of truly ancient fossil vertebrates have been recognized, extending establishment of crown vertebrates


into the Cambrian. These include numerous soft-bodied organisms from the Early Cambrian Chengjiang fauna, including Yunnanozoon and Haikouella, thought by some to represent early craniates (367, 368), and Zhongjianichthys, Myllokunmingia, and Haikouichthys, which exhibit convincing vertebrate derived characters (369–372). However, the evidence supporting their crown rather than stem-vertebrate affinity is not sufficiently convincing to justify their use in calibrating or constraining a molecular clock analysis. Similarly, conodonts have been widely touted as crown vertebrates, even stem-gnathostomes (373) but, while debate over the affinity of this group continues with its characteristic vigor, it would be inappropriate to use conodonts as evidence for calibrating or constraining the date of divergence of crown vertebrates. Although there are a number of records of armored stem-gnathostomes from the Early Ordovician (374– 376), the earliest phylogenetically secure records are Arandaspis prionotolepis (376) and S. janvieri (377), the oldest records of which are of Darriwilian age (375, 378). The best constraint on these earliest records is provided by Albanesi et al. (378), who identify co-occurring conodonts as indicative of the Pygodus anserinus Zone. Cooper and Sadler (379) interpolate a date of 462.2 Ma for the top of this zone; errors on the adjacent boundaries (top of the Darriwilian) are in the order of ±1.6 myr. Thus, this minimum constraint on the divergence of crown vertebrates is 460.6 Ma. Providing soft maximum bounds on the timing of crown-vertebrate divergence is contentious because of the possibility that some of the Early Cambrian Chengjiang vertebrates can be attributed to the vertebrate crown. Thus, we use as a soft maximum constraint the same evidence we use to constrain the divergence of bilaterian phyla (see crown Bilateria). Thus, a soft maximum constraint on the divergence of the echinoderm and chordate lineages may be taken as 581 Ma. Olfactores: tunicate–lamprey, shark, fish, tetrapod (minimum = 518.5 Ma; soft maximum = 581 Ma) This represents the origin of crown Olfactores, the clade comprised of tunicates and vertebrates (366). There are two putative fossil tunicates from the Early Cambrian Chengjiang biota (380, 381), though neither is sufficiently convincing to justify its use in calibrating or constraining a molecular clock analysis. Nevertheless, Zhongjianichthys, Myllokunmingia, and Haikouichthys (369–372), which are from the same deposit and can be attributed to the vertebrate total group at the very least,

67

serve as a minimum constraint on the divergence of tunicates and vertebrates. Thus, the minimum age of divergence of tunicates and vertebrates can be derived from the minimum age of the Yu’anshan Member of the Heilinpu Formation, in which the Chengjiang biota has been found. Unfortunately, this is equivocal because although its local stratigraphic assignment to the Eoredlichia wutingaspis Biozone is well constrained and long established (382), how this correlates to better-dated sections is not clear, not least because the fauna is largely endemic. The Heilinpu Formation belongs to the Qiongzhu Stage, which is considered to be Atdabanian in age. Thus, a minimum constraint may be provided by the age of the top of the Atdabanian, for which a date of 518.5 Ma is provided in the latest timescale (383). It should be noted, however, that this estimate is stratigraphically relatively remote from the nearest geochronological-derived date, and contingent upon the questionable conclusion that the Qiongzhu and Atdabanian are time equivalent. To provide a soft maximum bound on the timing of the crown Olfactores divergence, we follow the soft maximum bound on divergence of Bilateria (see crown Bilateria). Thus, a soft maximum constraint on the divergence of the echinoderm and chordate lineages may be taken as 581 Ma. Chordata: cephalochordate–tunicate, lamprey, shark, fish, tetrapod (minimum = 518.5 Ma; soft maximum = 581 Ma) Given the recent recognition that cephalochordates are the closest relatives to tunicates plus vertebrates, the split between cephalochordates, tunicates, and vertebrates equates to the origin of crown chordates. There are a number of putative fossil cephalochordates from the Cambrian including Pikaia graciliens from the Middle Cambrian Burgess Shale (384) and Cathaymyrus diadexus from the Early Cambrian Chengjiang fauna (385). At best, however, these fossils exhibit only chordate ancestral characters and, therefore, they provide no constraint over the timing of divergence of crown chordates. However, Zhongjianichthys, Myllokunmingia, and Haikouichthys (369–372) are from the same deposit and are attributable to the vertebrate total group at the very least. The minimum age of the Yu’anshan Member of the Heilinpu Formation, from which the Chengjiang biota has been found, is equivocal. This is because, although its local stratigraphic assignment to the E. wutingaspis Biozone is well constrained and long established (382),

68


how this correlates to better-dated sections is not clear, not least because the fauna is largely endemic. The Heilinpu Formation belongs to the Qiongzhu Stage, which is considered to be Atdabanian in age. Thus, a minimum constraint may be provided by the age of the top of the Atdabanian, for which a date of 518.5 Ma is provided in the latest timescale (383). It should be noted, however that this estimate is stratigraphically, relatively remote from the nearest geochronological-derived date, and contingent upon the questionable conclusion that the Qiongzhu and Atdabanian are time equivalent. For the soft maximum bound on the timing of crown chordate divergence, we follow the constraints on the divergence of Bilateria (see crown Bilateria). Thus, a soft maximum constraint on the divergence of the echinoderm and chordate lineages may be taken as 581 Ma. Deuterostomia: sea urchin–cephalochordate, tunicate, lamprey, shark, fish, tetrapod (minimum = 518.5 Ma; soft maximum = 581 Ma) This divergence event represents the splitting of crown deuterostomes into the chordate and ambulacrarian lineages, the latter clade composed of echinoderms and hemichordates. The oldest possible record of chordates dates from the Lower Cambrian Yu’anshan Member of the Heilinpu Formation (Chengjiang Biota) of Yunnan Province, South China, from which the remains of putative tunicates (380, 381), cephalochordates (385, 386), and even vertebrates (367–372) have been described. The problem with many of these records is that the characters defining clades at this deep level within phylogeny are largely cytological or embryological—not the kinds of characters that are preserved under even the most exceptional circumstances (387). Furthermore, both the living and fossil organisms are entirely soft-bodied and so precious few characters are preserved. And of these, many have been resolved to be deuterostome ancestral characters, rather than chordate or vertebrate-derived characters, with the recognition that echinoderms and hemichordates are sister taxa (387, 388). Thus, Yunnanozoon and Haikouella, thought by some to represent early craniates (367, 368), are interpreted by others as basal (perhaps even stem-) deuterostomes (331, 371, 389–393). Records of early tunicates (380, 381) have been questioned and the earliest unequivocal remains are Triassic in age (394). The putative vertebrates Zhongjianichthys, Myllokunmingia, and Haikouichthys (369–372) exhibit convincing vertebrate-derived characters, and these provide the best constraint on the minimum date of divergence of vertebrates

and chordates. There are contemporaneous records of more primitive deuterostomes, with the identification of vetulicystids as stem-echinoderms (389) and vetulicolians as stem-deuterostomes (371, 389, 391), although the veracity of the phylogenetic assignments of these taxa is a matter of some controversy (395–397). Earlier records of possible deuterostomes include Arkarua from among the enigmatic ediaracan biota (398). Although support for the identification of Arkarua as an echinoderm has found support from embryological homologies (399), all rests ultimately upon the presence of pentameral symmetry, which is not enough to rest an extension of tens of millions of years to a minimum date for divergence of deuterostomes and Bilateria upon. Thus, the vertebrates Zhongjianichthys, Myllokunmingia, and Haikouichthys (369–372) provide the best evidence for the minimum date of divergence of deuterostomes. Thus, a minimum constraint on the divergence of crown deuterostomes is based on the vertebrates Zhongjianichthys, Myllokunmingia, and Haikouichthys and the minimum age of the Yu’anshan Member of the Heilinpu Formation, from which the Chengjiang biota has been found. The age of the Chengjiang biota remains equivocal because, although its local stratigraphic assignment to the E. wutingaspis Biozone is well constrained and long established (382), how this correlates to betterdated sections is not clear, not least because the fauna is largely endemic. The Heilinpu Formation belongs to the Qiongzhu Stage, which is considered to be Atdabanian in age. Thus, a minimum constraint may be provided by the age of the top of the Atdabanian, for which a 518.5 Ma is provided in the latest timescale (383). It should be noted, however that this estimate is stratigraphically, relatively remote from the nearest geochronological-derived date, and contingent upon the questionable conclusion that the Qiongzhu and Atdabanian are time equivalent. The soft maximum bound on the timing of crown deuterostome divergence is based on the constraints on the divergence of bilaterians (see crown Bilateria). Thus, a soft maximum constraint on the divergence of the echinoderm and chordate lineages may be taken as 581 Ma. Bilateria/Nephrozoa: arthropod, nematode, annelid, mollusc–echinoderm, chordate (minimum = 531.5 Ma; soft maximum = 581 Ma) This divergence event represents the splitting of crown Bilateria, and the divergence of deuterostome and protostome lineages. The are numerous convincing chordates, among other putative deuterostomes, from the Lower Cambrian


Yu’anshan Member of the Heilinpu Formation (Chengjiang Biota) of Yunnan Province, South China (see Deuterostomia, Chordata, Olfactores, and Vertebrata, above), providing a minimum constraint of 518.5 Ma. The earliest evidence for the origin of arthropods are Rusophycus-like trace fossils from the upper NemakitDaldynian (early Tommotian) of Mongolia (400, 401) (520.5 Ma; see Arthropoda–Nematoda). However, there are still older representatives of the protostome lineage, further constraining the time of divergence of the human and fruitfly genomes, as well as the genomes of all integral taxa. The oldest of these is probably the mollusc Latouchella from the middle Purella Biozone, NemakitDaldynian, of Siberia (401, 402). There are a number of candidate crown bilaterians among the Ediacaran biota, among which a molluscan affinity for Kimberella has been most cogently argued (403). However, the evidence has not withstood scrutiny (393) and it is certainly insufficient to justify its use as a calibration for, or constraint on molecular clock analyses of metazoan evolution. The boundary between the Nemakit-Daldynian and the succeeding Tommotian Stage remains equivocal and so a more reliable minimum constraint might be provided by the current best estimate for the base of the Tommotian, which is 531.5 Ma (383). Thus, on the basis of the available paleontological, stratigraphic, and chronological data, the best minimum constraint for the divergence of crown Bilateria is 531.5 Ma. Providing soft maximum bounds on the timing of crown bilaterian divergence is extremely contentious. Nevertheless, following the same criteria used to provide constraints on other divergence events, it is possible to constrain the timing of divergence of bilaterian phyla on the occurrence of older lagerstätten that preserve records of earlier branching lineages. Inevitably, these records are represented by the Ediacaran faunas, the interpretation of which is extremely contentious, though there is increasing agreement that crown bilaterians are not represented among them (393, 404). Thus, the youngest, most completely sampled Ediacaran assemblage may be used to provide the soft maximum constraint on the divergence of bilaterian phyla including the chordate and echinoderm lineages. This is the Doushantuo Formation, which provides a sampling of Ediacaran diversity in a number of facies and through a number of modes of exceptional preservation (405, 406); although a number of candidate bilaterians have been described from this deposit (407–411), these have not withstood scrutiny (412–415). The most exacting test of the existence of bilaterians such as deuterostomes is

69

provided by the phosphates that occur early within the sequence and include exquisitely preserved embryos and hatchlings of metazoan affinity (416–418) and adults of possible cnidarians (412). No uncontested bilaterians are present (417). Condon et al. (419) indicate that the phosphorites are younger than the tillites of Gaskiers glaciogenic event which has been dated at 580 Ma ± 1 myr. This date is older than all Ediacarans that have been proposed as bilaterians (420, 421). The date provided differs from Benton and Donoghue (3) [but not Donoghue and Benton (4)], who provided the minimum not the soft maximum age on the embryo-bearing horizons in the Doushantuo Formation. Protostomia: arthropod, nematode–annelid, mollusk (minimum = 531.5 Ma; soft maximum = 581 Ma) The constraints on the divergence of Bilateria/Nephrozoa apply as equally to the internal split within Protostomia, between Ecdysozoa and Lophotrochozoa. The principal record providing minimum constraint on the divergence of these lineages is that of the mollusk Latouchella (401, 402), providing a date of 531.5 Ma (see Bilateria/Nephrozoa). For the soft maximum bound we use evidence presented for the soft maximum bound on divergence of Bilateria (see crown Bilateria). Thus, a soft maximum constraint on the divergence of the ecdysozoan and lophotrochozoan lineages may be taken as 581 Ma. Annelida–mollusca: leech, polychaete–limpet, sea hare, Biomphalaria (minimum = 531.5 Ma; soft maximum = 581 Ma) The evidence marshaled to constrain the divergence of Bilateria/Nephrozoa applies equally to the divergence of annelida and mollusca because it is based on the earliest mollusk. The principal record is Latouchella (401, 402), providing a date of 531.5 Ma (see Bilateria/Nephrozoa); as discussed with regard to Bilateria, we do not consider the evidence supporting a molluscan affinity of the edicaran organism Kimberella as sufficient to justify its use as a minimum constraint on the establishment of the molluscan phylum. For the soft maximum bound we evidence presented for the soft maximum bound on divergence of Bilateria (see crown Bilateria). Thus, a soft maximum constraint on the divergence of the annelida– mollusca may be taken as 581 Ma. Gastropoda: limpet–Biomphalaria, sea hare (minimum = 470.2 Ma; soft maximum = 531.5 Ma) The divergence of the limpet Lottia from the euthyneurans Biomphalaria and Aplysia represents the divergence

70


of the two major living subclasses of Gastropoda, Eogastropoda, and Orthogastropoda, respectively, as well as the base of crown Gastropoda. There are many candidates for the earliest crown gastropod but the precise affinity of these early mollusks remains questionable. The earliest secure record of crown Gastropoda may be sought in the oldest vetigastropods, which are conventionally accepted as members of the living gastropod clade. Thus, a minimum constraint would rest on the sinuopeid Sinuopea sweeti from the Jordan Sandstone of Wisconsin, and the raphistomatid Schizopea typical from the Eminence Dolomite of Missouri (59). Both the Eminence Dolomite and Jordan Sandstone are generally quoted to be Trempealeauan in age, but the basis of this age justification is unclear for the Eminence Dolomite (422), while the Jordan Sandstone has been assigned to the Sunwaptan on the basis of its trilobite fauna (423). Thus, using S. sweeti, the minimum constraint on the divergence of crown gastropods would be 490 Ma (383). However, Wagner has questioned the reliability with which any early Palaeozoic gastropods may be assigned to each of the three main extant lineages of gastropods, with the exception of the eotomarioids, which he identifies as candidates for the ancestry of extant vetigastropods (424). The oldest member of Eotomarioidea is Turritoma acrea from the Catoche Formation of Western Newfoundland (425). The Catoche Formation falls fully within the Oepikodus communis conodont biozone (425) the top of which, in the sense that it is employed, coincides with the Ibexian/Whiterockian boundary; this would be the O. communis and Reutterodus andinus Biozones of Ross and colleagues (426). This coincides with the Early/Middle Ordovician Boundary in the 2004 Geologic Timescale, and a date of 471.8 Ma ± 1.6 myr (379). Thus, on this conservative view which we follow, the minimum constraint on the divergence of crown Gastropoda is 470.2 Ma. A soft maximum constraint may be provided by the oldest mollusk, Latouchella, dated at 531.5 Ma (see earlier). Euthyneura: sea hare–Biomphalaria (minimum = 168.6 Ma; soft maximum = 473.4 Ma) The divergence of the sea hare Aplysia and the airbreathing freshwater snail Biomphalaria reflects the divergence of Pulmonata and Opisthobranchia, and the base of crown Euthyneura. Molecular phylogenetic analyses indicate that although pulmonates and opisthobranchs are each other’s closest relatives, their monophyly is questionable (427, 428). On evidence of mitochondrial

synteny, pulmonates have been identified as paraphyletic with respect to a monophyletic Opisthobranchia and among the pulmonates considered, Biomphalaria has been identified as more closely related to opisthobranchs than to the other pulmonates included in the analysis (429). This indicates that the split between Biomphalaria and Aplysia does not coincide with the base of crown Euthyneura. The oldest records of Euthyneura, both opisthobranchs and pulmonates are Tournaisian (59) but, because of the uncertainty concerning the interrelationships of Euthyneura it would be safer to rely upon the earliest records of lower rank taxa to which Biomphalaria and Aplysia have been assigned. The oldest record of the Order Aplysiomorpha to which Aplysia has been assigned is Tertiary. The earliest record of Basommatophora, the order to which Biomphalaria is assigned, is also Tournaisian, but the monophyly of Basommatophora remains to be established and, thus, a more reliable minimum constraint may instead be provided by the oldest record of the Superfamily Planorboidea and Family Planorbidae to which Biomphalaria is assigned. This earliest record is Anisopsis calculus from Cajac, France, which is reported to be of Aalenian (Jurassic) age (59). Without further constraint, the age of the Aalenian– Bajocian boundary may be used which, following the 2004 Geologic Timescale is 171.6 Ma ± 3.0 myr (240). Thus, the minimum constraint on the divergence of Aplysia and Biomphalaria is 168.6 Ma. A soft maximum constraint on the divergence of these heterobranch orthogastropods may be provided by evidence for the establishment of Orthogastropoda, dated at 471.8 Ma ± 1.6 myr (see earlier) and, thus, 473.4 Ma. Capitellid polychaete–leech (minimum = 305.5 Ma; soft maximum = 581 Ma) The intrarelationships of annelids are in a state of flux, with the phylogenetic signal from competing molecular data sets conflicting with one another, and with morphology-based data sets. Some general conclusions are that clitellates, the clade to which leeches are assigned, are monophyletic, but nest within polychaetes, which are grossly paraphyletic (430, 431). The oldest possible clitellate is a putative leech described from the Middle Silurian of Wisconsin (432, 433), and a much younger form from the Jurassic of Solnholfen (434). However, the evidence presented in support of their assignment to the clitellates amounts to little more than their vaguely leech-like round mouth and segmented body. Pronaidites carbonarius was described as a


Carboniferous oligochaete, but this record requires careful redescription and reconsideration (435). The gross paraphyly of polychaetes renders the significance of Cambrian polychaetes moot; they have been assigned to the extant clade Phyllodocida (436) but their assignment to any extant clade within Annelida has recently been challenged (437). Polychaetes are well represented by their jaw elements in the fossil record, from the Early Ordovician onward (438), but although they are often considered eunicids, there is no real evidence to support this. This is unfortunate because although the precise affinity of eunicids is unclear in recent molecular phylogenies, in one manner or another, eunicids, along with amphinomid and flabelligerid polychaetes, intercalate the clade circumscribed by clitellates and capitellid polychaetes (430, 431). Thus, we may derive a minimum constraint from the oldest securely identified fossil eunicid, amphinomid, or flabelligerid, all of which (e.g., Esconites zelus, Rhaphidiophorus hystrix, and Flabelligeridae sp., respectively) are from the Pennsylvanian Mazon Creek fauna of Illinois (439, 440). The Mazon Creek fauna is derived from the Francis Creek Member of the Carbondale Formation in NE Illinois. The Francis Creek Shale Member has been dated as middle Desmoinesian and middle Westphalian D age on the basis of both palynological and paleobotanical data (441–443). This equates to the upper part of the Moscovian Stage, the top of which has been dated at 306.5 Ma ± 1.0 myr on the basis of a graphically correlated composite standard calibrated using radiometric dates (252). The top of the Westphalian D is slightly older at 307.2 Ma. (252). Thus, the minimum constraint on the divergence of Capitella from Helobdella is 305.5 Ma. To provide a soft maximum bound on the timing of capitellid polychaete–leech divergence, we follow the soft maximum bound on divergence of Bilateria (see crown Bilateria). Thus, a soft maximum constraint on the divergence of the capitellid polychaete and clitellate lineages may be taken as 581 Ma. Nematode–arthropod (minimum = 520.5 Ma; soft maximum = 581 Ma) This divergence event represents the splitting of the nematode and arthropod lineages. The earliest evidence for the origin of arthropods are Rusophycus-like trace fossils from the upper NemakitDaldynian (early Tommotian) of Mongolia (400, 401). Dating of the Early Cambrian is not well advanced, not least because a global scheme of stratigraphic zonation for the Early Cambrian has yet to be established.

71

The most appropriate date for constraining the age of these trace fossils is the top of the Tommotian, the Tommotian–Atdabanian boundary. The best available date for this is 522.5 Ma, provided by Shergold and Cooper (383), though it is an estimate based on younger and older geochronological dates, errors on which are reported in the order of 2 myr. Thus, the minimum constraint on the divergence of arthropod and nematode lineages is 520.5 Ma. There are older representatives of mollusks, and we may consider these in establishing soft maximum bounds on the timing of arthropod–nematode divergence. The oldest mollusk is Latouchella, from the middle Purella Biozone, Nemakit-Daldynian, of Siberia (401, 402). This indicates the existence of the ecdysozoan total group to which nematodes and arthropods belong, but this isolated record provides no confidence on which to judge whether or not the arthropods and nematode lineages had yet diverged. There are a number of candidate crown bilaterians among the Ediacaran biota, among which a molluscan affinity for Kimberella has been most cogently argued (403). However, the evidence has not withstood scrutiny (393) and it is certainly insufficient to justify its use as a constraint on molecular clock analyses of metazoan evolution. Thus, we follow the soft maximum constraint on the date of divergence of crown Bilateria, 581 Ma. Mandibulata: Daphnia–louse, Rhodnius, aphid, beetle, wasp, honeybee, mosquito, fruitfly (minimum = 510 Ma; soft maximum = 543 Ma) This represents the establishment of crown Mandibulata and the divergence of Crustaceomorpha from Atelocerata. The fossil record of crustaceans is by far the more extensive of the two lineages comprising Atelocerata. The earliest possible crustaceans have been reported from the Mount Cap Formation of northwestern Canada (444). However, these remains are fragmentary and their interpretation as crustaceans is based on the special similarity between individual fragments and the fi lter-feeding apparatus of modern branchiopod crustaceans, rather than on the basis of a suite of mutually corroborative phylogenetically informative characters. The earliest convincing evidence for the divergence of Atelocerata and Crustaceamorpha is the phosphatocopid Klausmuelleria salopiensis from the Lower Cambrian Comley Limestones of Shropshire, UK (445–447). Siveter et al. (447) indicate that the lower Comley Limestones can be assigned to the Protolenid-Strenuellid Biozone which correlates to the Botomian-Toyonian age within the

72


Siberian stratigraphic framework (447). Within the 2004 Geologic Timescale, this provides a minimum constraint of 510 Ma (383). A soft maximum constraint may be provided by the earliest evidence of arthropods, based upon Rusophycuslike trace fossils (see “Nematoda–Arthropoda” later) from the Nemakit-Daldynian (early Tommotian) of Mongolia (400, 401). A soft maximum constraint may therefore be derived from the base of the Nemakit-Daldynian which equates to the base of the Cambrian and, thus, 542 Ma ± 1.0 myr (383). Our soft maximum constraint is therefore 543 Ma. Eumetabola: louse, Rhodnius, aphid–beetle, wasp, honeybee, mosquito, fruitfly (minimum = 307.2 Ma; soft maximum = 414 Ma) The divergence of Paraneoptera from Holometabola. Providing a minimum constraint on the divergence of crown Eumetabola is complicated by the lack of resolution concerning the affinity of Palaeodictyopterida, which has been variably considered a member of the clade. Grimaldi and Engel exclude palaeodictyopterids from the clade, leaving Miomoptera as the oldest members of Eumetabola (448). These authors discuss the various possible affinities of Mimptera among Paraneoptera or Holometabola, but there appears no equivocation of their membership of Eumetabola. The oldest known record of Miomoptera is an undescribed specimen (Field Museum PE 293590 from the Pennsylvanian Mazon Creek Lagerstatte (449)). The Mazon Creek fauna is derived from the Francis Creek Member of the Carbondale Formation in NE Illinois. The Francis Creek Shale Member has been dated as middle Desmoinesian and middle Westphalian D age on the basis of both palynological and paleobotanical data (441–443). This equates to the upper part of the Moscovian Stage, the top of which has been dated at 306.5 Ma ± 1.0 myr on the basis of a graphically correlated composite standard calibrated using radiometric dates (252). The top of the Westphalian D is slightly older at 307.2 Ma. (252). Thus, the minimum constraint on the divergence of crown Eumetabola is 305.5 Ma. A soft maximum constraint may be provided by the age of the oldest insect Rhyniognatha hirtsi from the Early Devonian Rhynie Chert of northeast Scotland (450). The age of the Rhynie Chert has been best established on the basis of the composition of its spore assemblages which indicate an early Pragian to the earliest Emsian age span (451). Thus, we may establish a soft maximum constraint on the base of the Pragian which is 411.2 Ma ± 2.8 myr (452), equating to 414 Ma.

Paraneoptera: louse–Rhodnius, aphid (minimum = 283.7 Ma; soft maximum = 414 Ma) Though the assignment of Archescytinidae to the hemipteran crown group may be questioned, there is no question of its membership of Paraneoptera. There are older records of Paraneoptera, including Permopsocidae, but these are likely stem-Paraneoptera (25, 448). Thus, the best minimum constraint on the divergence of Paraneoptera is provided by an undescribed archescytinid from the middle Bacov Beds of Boscovice Furrow, Obora Czech Republic (453, 454). These rocks were described as Artinskian by Kukalová-Peck and Willmann (454), without justification, but they have subsequently been attributed to the Sakmarian using vertebrate microremains for biostratigraphic correlation (455, 456). On this basis we may use the top of the Sakmarian as our basis for a minimum constraint on the divergence of Paraneoptera and Holometabola which is as given as 284.4 ± 0.7 myr (457), providing the minimum constraint of 283.7 Ma. The most approximate soft maximum constraint on the divergence of Paraneoptera is provided by the earliest records of Neoptera, which are a paraphyletic assemblage of Late Carboniferous roach-like dictyopterans, sometimes grouped as the grade Blattodea or Blatoptera. The oldest such record is probably Ctenopilus elongatus (previously Eoblattina complexa) from the Stephanian B-C of the Commentary Basin, France (458). The Stephanian B of western Europe correlates to the upper Kasimovian of the 2004 Geologic Timescale, the base of which has been dated at 306.5 Ma ± 1.0 myr (252) and, thus, a soft maximum constraint of 307.5 Ma. However, given the reliance on temporally isolated lagerstatten for constraining the temporal diversification of insects, this envelope is perhaps too strict. Instead, a more appropriate soft maximum constraint may be provided by the oldest member of Pterygota, the oldest possible record of which is also the oldest known insect, Rhyniognatha hirsti (450), providing a constraint of 414 Ma (see Eumetabola earlier). Hemiptera: Rhodnius–aphids (minimum = 199.0 Ma; soft maximum = 307.5 Ma) The oldest known hemipterans are members of the Archescytinidae, the oldest record of which remains undescribed but has been recorded from the early Artinskian locality of Obora (453). Archescytinidae is identified by Shcherbakov and Popov as more closely related to aphids than to Cimicina and, hence, providing a minimum constraint on the split between Rhodnius and aphids (453). However, Engel and Grimaldi question this interpretation of the affinity of Archescytinidae


within Hemiptera because the necessary characters are not preserved. Engel and Grimaldi (p. 321) describe three unnamed heteropterans from the Triassic of Virginia (USA), but the oldest described taxon is the Lufengnacta (Corixidae, Nepomorpha, Panheteroptera, Heteroptera) from the Yipinglang Coal Series of Yunnan Province, China. The age of the Yipinglang Coal Series is widely agreed to be of Late Triassic age and has been used to justify the correlation of overlying units across South China. Its precise age may be constrained by the palynoflora (459) which provides a Rhaetian-Norian age. Thus, the minimum constraint on the divergence of crown Hemiptera is provided by the date for the end Rhaetian (end Triassic), which is 199.6 Ma ± 0.6 myr (267) and, thus, 199.0 Ma. A suitable soft maximum constraint may be provided by the earliest Neopteran, which is C. elongatus (458), providing a date of 307.5 Ma (see Paraneoptera earlier). Holometabola: beetle–wasp, honeybee, mosquito, fruitfly (minimum = 307.2 Ma; soft maximum = 414 Ma) Divergence of Coleoptera and Hymenoptera–Panorpida, and the establishment of crown Holometabola. The oldest recorded member of this clade appears to be an undescribed member of Coleopteroidea from the middle Carboniferous Mazon Creek fauna of Illinois, USA (449). The Mazon Creek fauna is derived from the Francis Creek Member of the Carbondale Formation in NE Illinois. The Francis Creek Shale Member has been dated as middle Desmoinesian and middle Westphalian D age on the basis of both palynological and paleobotanical data (441–443). This equates to the upper part of the Moscovian Stage, the top of which has been dated at 306.5 Ma ± 1.0 myr on the basis of a graphically correlated composite standard calibrated using radiometric dates (252). The top of the Westphalian D is slightly older at 307.2 Ma. (252) Thus, the minimum date on the divergence of these two clades is 307.2 Ma. A suitable soft maximum constraint may be provided by the earliest member of Pterygota, which is R. hirsti (450), providing a constraint of 414 Ma (see Eumetabola earlier). Hymenoptera–Panorpida: wasp, honeybee–fruitfly, mosquito (minimum = 238.5 Ma; soft maximum = 307.2 Ma) This divergence event represents the splitting of the Hymenoptera and Panorpoidea lineages. The earliest recognized Panorpoidea are the mecopteroids that are interpreted as stem-Panorpoidea (or

73

panorpoideans) and are known from records as early as the Permian, the very oldest of which are members of Kaltanidae, interpreted as stem-panorpoideans (460). The earliest recognized Hymenoptera are from the Middle Triassic of Central Asia (461, 462), and the Upper Triassic of Australia (463) and Africa (464), all of which are referred to the Archexyelinae within Xyelidae. This difference in first records of Hymenoptera and Panorpoidea has led to the suggestion that putative stempanorpoideans from the Permian are unified on shared ancestral characters of Panorpoidea + Hymenoptera (448). Thus, the minimum date for the divergence of Hymenoptera and Panorpoidea would be based on the earliest records from the Middle Triassic Madygen Formation of Central Asia (461, 462), which is dated as Ladinian and/or Carnian on the basis of palynological data (465, 466). In the absence of greater biostratigraphic control it is possible only to derive a minimum date from the base of the Norian (base Norian 216.5 Ma ± 2.0 myr; 267). Thus, a minimum constraining date would be 214.5 Ma. However, this inconsistency is predicated upon the assumption that Hymenoptera and Panorpoidea are sister taxa, a view that is not universally accepted. Rasnitsyn (467), for instance, maintains that Hymenoptera and Panorpoidea are more remotely related, the closest relatives of Panorpoidea being Neuropteroidea and Coleopteroidea (united on modified ovipositor (gonapophyses 9 (= dorsal valvula) lost, and the intromittant function transferred to gonocoxa 9 + gonostylus 9 (= valvula 3). In this view, Panorpoidea + Neuropteroidea + Coleopteroidea diverged from the lineage leading to Hymenoptera within the paraphyletic Order Palaeomanteida, at a time approximating to the Carboniferous/Permian boundary. Unfortunately, the systematics of this group are poorly resolved and it is unclear which represent the earliest members of the lineages ultimately leading to Panorpiodea and Hymenoptera. The best estimate must be provided by the earliest member of the clade Panorpoidea + Neuropteroidea + Coleopteroidea, but note should be taken of the fact that this date is likely to be extended in light of systematic revision of Palaeomanteida. The oldest known member of Coleoptera is Pseudomerope gallei, from the Asselian (299–294.6 Ma ± 0.8 myr) (Lower Permian) of Rícany, Czech Republic (454), though the basis of this age assignment is not clear. The oldest recorded member of this clade appears to be an undescribed member of Coleopteroidea from the middle Carboniferous Mazon Creek fauna of Illinois,

74


USA (449), providing a date of 307.2 Ma (see Eumetabola, earlier). Thus, within the phylogenetic milieu which posits that Hymenoptera are not immediate sister taxa (467), the minimum date on the divergence of these two clades is 307.2 Ma. In conclusion, however, it must be emphasized that Hymenoptera and Panorpoidea are conventionally viewed as sister taxa. Nevertheless, a minimum date for divergence of 214.5 Ma postdates the minimum date of is 238.5 Ma for the divergence of the lineages leading to Drosophila melanogaster and Anopheles gambiae. Apis mellifera falls outside this clade and so in the absence of better constraint over the interrelationships of Diptera and Hymenoptera, a minimum date for their divergence can be taken as 238.5 Ma. A soft maximum constraint can be provided by the less likely hypothesis that Panorpoidea are more closely related to Neuropteroidea and Coleopteroidea, using the oldest record of this clade, described earlier as 307.2 Ma. Apocrita: honeybee–wasp (minimum = 152 Ma; soft maximum = 243 Ma) The divergence of the honeybee Apis from the parasitic wasp Nasonia corresponds to the crown-group concept of the hymenopteran suborder Apocrita, and the divergence of Proctotrupoidea and Chalcidoidea, respectively. The oldest records of both lineages are at minimum, Late Jurassic in age, but the earliest records of Proctotrupoidea are the best dated. These records belong to Mesoserphidae, such as Mesoserphus and Karatoserphus, from the Early Jurassic Daohugou Beds of Inner Mongolia, China (468, 469). The age of these beds has been constrained radiometrically using U-Pb series dating to the interval 168–152 Ma (470, 471) and, thus, we take 152 Ma as the minimum constraint on the divergence of honeybee and wasp. A soft maximum constraint can be provided by the earliest record of Hymentoptera, the earliest recognized members of which are from the Middle Triassic Madygen Formation of Central Asia (461, 462), that is dated as Ladinian and/or Carnian on the basis of palynological data (465, 466). Thus, the constraint may be derived from the base of the Ladinian, which may be as much as 241 Ma ± 2.0 myr (267), equating to a soft maximum constraint of 243 Ma. Diptera: fruitfly–mosquito (minimum = 238.5 Ma; soft maximum = 295.4 Ma) This divergence event represents the splitting of Brachycera and Culicomorpha lineages. The oldest representative

of Culicomorpha is Aenne triassica from the Late Triassic (Rhaetic) Cotham Member of the Lilstock Formation, Penarth Group at Aust Cliff, near Bristol, UK (472). Although this displays chironomid derived characters, only the distal half of a wing is preserved. The base of the Cotham Member coincides with the base of SA5n.3r which equates to the E23r reverse polarity magnetozone of the Newark Supergroup (473), the base of which is estimated at 202 Ma ± 1 myr on the basis of volcanics in the upper part of the underlying E23 normal polarity magnetozone (267). Hounslow et al. (473) argue that the whole of the Cotham Member equates to the E23r magnetozone, the duration of which is beyond stratigraphic resolution in the current timescale (267). Thus, we conclude the age of the first possible representative of Culicomorpha to be 202 Ma ± 1 myr. The next oldest record is Aenne liasina from the lower Toarcian (Lower Jurassic) of Grimmen, NE Germany (474), followed by an abundance of other Culicomorpha records in the Lower and Middle Jurassic (448). The oldest documented representatives of Brachycera are from the Upper Triassic Dan River Group of Virginia (475, 476), although their assignment rests upon precious few and largely inconsistent venation characters (448). There remains an older record of Brachycera, Gallia alsatica, from the Grès-à-Voltzia Formation of Arzviller, northeast France (recognized on the basis of the following derived characters: cell m3 narrowed distally and Cu and A1 terminating in one point at the wing margin) (476, 477). The Grès à Meules facies of the Grès-a-Voltzia Formation, from which these remains are derived, has been dated as lower Anisian (478, 479), although the evidence on which this is based was not presented. The top of the lower Anisian is dated as 240.5 Ma, based on proportional scaling of major conodont zones (267) from a graphic correlation global composite standard (480), from which an error of ±2.0 myr is derived. Otherwise, there are convincing records from the Early Jurassic, including the Black Ven Marls (Sinemurian) at the cliff of Stonebarrow Hill near Charmouth, Dorest, UK (turneri-obtusum Zone) 194.1– 192.0 Ma (481), and the lower Toarcian (Harpoceras falciferum Zone) of Dobbertin, Mecklenburg, Germany 182.7–181.2 Ma (482). The oldest representatives of the clade comprising Culicomorpha and Brachycera are members of grauvogeliid Psychodomorpha, specifically, Grauvogelia arzvilleriana from the Middle Triassic Grès-a-Voltzia Formation of France (483). Crucially, this is neither the most primitive crown dipteran, nor the oldest known total-group


dipteran, but the oldest record that falls within the clade circumscribed by Anopheles and Drosophila, following the phylogenetic scheme presented in (448). Thus, on the record of G. arzvilleriana (483), its coincidence with the earliest (albeit undocumented) record of Brachycera (476, 477), and the phylogenetic hypothesis of Grimaldi and Engel (448), the minimum date for the divergence of the lineages leading to D. melanogaster and A. gambiae is 238.5 Ma. A soft maximum constraint is provided by the insect fauna of Boskovice Furrow, Oboro, Moravia, Czech Republic. A huge diversity of insects has been described from this deposit which is the single most important Paleozoic insect locality in the world (448). No members of the clade circumscribed by Brachycera and Culicomorpha have been described from here or from older deposits. The Oboro fauna has been dated at early Artinskian (454) and Sakmarian (456), although only the latter has been substantiated. The base of the Sakmarian has been dated at 294.6 Ma ± 0.8 myr (457). Thus, the soft maximum constraint on the divergence of Brachycera and Culicomorpha can be taken as 295.4 Ma. Eumetazoa: Cnidaria–Bilateria (minimum = 531.5 Ma; soft maximum = 581 Ma) The split between Cnidaria and Bilateria represents the origin of crown Eumetazoa. The oldest unequivocal record of Bilateria is the mollusc Latouchella from the middle Purella Biozone, Nemakit-Daldynian, of Siberia (401, 402). In the absence of better constraint, a numerical date may be derived from the boundary between the Nemakit-Daldynian and the succeeding Tommotian Stage. However, this remains equivocal and so a more reliable minimum constraint might be provided by the current best estimate for the base of the Tommotian, which is 531.5 Ma (383). Thus, on the basis of the available paleontological, stratigraphic, and chronological data, the oldest record of Bilateria is 531.5 Ma. The oldest possible record of a cnidarian is provided by Sinocylcocyclicus guizhouensis from the Ediacaran Doushantuo phosphorites (412), although the evidence in favor of a cnidarian affinity does not amount to more than its structural resemblance to tabulate corals. Innumerable putative medusoid cnidarians have been described among the Ediacaran biota, but these have been reinterpreted as microbial communities (484) or trace fossils (485). Frond-like Ediacarans such as Charnia, have traditionally been interpreted as sea pens, but this comparison is unconvincing (486, 487). Namapoikia rietoogensis is a slightly younger Ediacaran

75

record from the Nama Group of northern Namibia (488, 489) and, although it exhibits colonial organization and, therefore has drawn comparison to cnidarians, its affinities are speculative nevertheless. Namacalathus hermanastes is also known from the Nama Group of central and southern Namibia (490); it has also drawn comparison to cnidarians and this comparison is equally equivocal. Less equivocal records of cnidarians are to be found among Tommotian-age small shelly faunas, represented by anabaritids (491) and, later, the tentaculitids. The affinity of both has been the subject of debate but, on the basis of the available evidence, their assignment to the cnidarian total group is compelling. An almost complete life series from embryo to adult of the putative scyphozoan cnidarian Olivooides is known to co-occur with elements of the small shelly fauna (492, 493), although its affinities are equivocal. The earliest phylogenetically secure cnidarians are corals, based on slightly older Early Cambrian records from North America, Australia, and Siberia (494, 495). The oldest coral is probably Cysticyathus tunicatus from the Tommotian of Siberia (496) and, thus, a numerical constraint on the oldest secure record of a cnidarian can be derived from the age of the top of the Tommotian, the Tommotian–Atdabanian boundary. The best available date for this is 522.5 Ma, provided by Shergold and Cooper (383), though it is an estimate based on younger and older geochronological dates, errors on which are reported in the order of 2 myr. This date is significantly younger than the oldest bilaterian record which, at 531.5 Ma, we adopt as the minimum constraint on the divergence of Eumetazoa. A soft maximum bound could be codified on the basis of the more equivocal cnidarian records outlined earlier. All of these, including the Ediacaran records, are younger than the soft maximum bound established for Bilateria (581 Ma), which also encompasses the Doushantuo record of S. guizhouensis (412). Thus, we adopt the maximum date on the embryo-bearing horizons in the Doushantuo Formation as our soft maximum constraint on the divergence of crown Eumetazoa, at 581 Ma. Cnidaria: Hydra–sea anemone (minimum = 520.5 Ma; soft maximum = 581 Ma) This divergence represents the divergence of Anthozoa (sea anemones—including Nematostella, and corals) and Medusozoa (Scyphozoa, Cubozoa, Hydrozoa—including Hydra and Staurozoa) and the establishment of crown Cnidaria (497). As discussed earlier in connection with Eumetazoa, the oldest records of crown cnidarians are represented by anabaritids (491) and the putative

76


scyphozoan cnidarian Olivooides (492, 493), although its affinities are equivocal. However, the earliest phylogenetically secure cnidarians are meduzoans from the Middle Cambrian of Utah (498) and anthozoans (corals) from the Early Cambrian of North America, Australia, and Siberia (494, 495). The oldest coral is probably C. tunicatus from the Tommotian of Siberia (496) and, thus, a numerical constraint on the oldest secure record of a cnidarian can be derived from the age of the top of the Tommotian, the Tommotian–Atdabanian boundary. The best available date for this is 522.5 Ma, provided by Shergold and Cooper (383), though it is an estimate based on younger and older geochronological dates, errors on which are reported in the order of 2 myr, yielding a minimum constraint of 520.5 Ma. A soft maximum bound could be codified on the basis of the more equivocal cnidarian records outlined earlier (see Eumetazoa). All of these, including the Ediacaran records, are younger than the soft maximum bound established for Bilateria (581 Ma), which also encompasses the Doushantuo record of S. guizhouensis (412). Thus, we adopt the maximum date on the embryobearing horizons in the Doushantuo Formation as our soft maximum constraint on the divergence of crown Cnidaria, at 581 Ma. Metazoa: Porifera–Eumetazoa (minimum = 634.97 Ma; soft maximum = 836 Ma) This divergence event coincides with the origin of crown Metazoa. Dating the divergence of sponges from the lineage leading to cnidarians and bilaterians is complicated by molecular phylogenies which, in contrast to morphology-based analyses (499, 500), have resolved Porifera as paraphyletic, composed of as many as three distinct clades of phylum status, with the homoscleromorphs, calcisponges, and demosponges as successive sister taxa to Eumetazoa (91, 501–505). In what follows, we specifically aim to constrain the date of divergence of demosponges from the lineage leading to calcisponges, homoscleromorphs, and eumetazoans (cnidarians, aceols, and triploblast bilaterians). This is because our focus is to constrain the divergence of Renieria, a demosponge which has been targeted for genome sequencing. Many of the more ancient and speculative records of cnidarians have also been attributed to the sponges, including Namacalathus (488) and Namapoikia (490), but none is entirely convincing. This includes putative sponges from the Ediacaran Doushantuo Formation (506), where the structures interpreted as evidence of

poriferan affinity are just as readily interpreted as fabric of diagenetic mineralization (414). Other problematica that have been attributed to Porifera include the archaeocyaths (507), stromatoporoids (495), and chancelloriids (508, 509), but their earliest records are younger than Paleophragmodictya (510), which is from the Ediacaran of southern Australia and, as such, it is the oldest convincing record of a sponge (495). Its precise taxonomic affi liation is of little consequence so long as it falls within the clade circumscribed by demosponges and all other metazoans which, as a hexactinellid, it does. All of these records are, however, eclipsed by a biochemical record of demosponges, the precise dating of which is unclear, but which extends between sedimentary deposits representative of the Sturtian and Marinoan glaciations (511). On this basis, the age of the Marinoan glaciation can provide a minimum constraint on the divergence of demosponges from other metazoans, dated at 635.51 Ma ± 0.54 myr (419); as dating improves for the Oman sequence from which the biomarker record occurs, this date will be revised upward by tens of millions of years. A soft maximum constraint can be provided by Neoproterozoic lagerstatten, such as the Bitter Springs Formation of central Australia (512) and the Svanbergfjellet Formation of Spitsbergen (513), that exhibit cellular-level preservation of a diversity of organisms including prokaryotes, sphearomorphic acritarchs, multicellular algae, and various problematica, but no evidence of metazoans, or anything that could even be interpreted as a stem-metazoan. The Bitter Springs and Svanbergfjellet floras have been determined to be of comparable middle Neoproterozoic age on the basis of a global carbon isotope excursion (514, 515). There is no direct dating on either formation but the Bitter Springs Formation has been correlated with the volcanic sequence in the upper Loves Creek Formation which has itself been allied with the Gairdner Dyke Swarm (516, 517), dated at 827 Ma ± 6 Ma (518). Halverson et al. (515) argue for a younger date, but this is not well substantiated. Thus, we take 836 Ma as the soft maximum constraint on the divergence of the crown Metazoa.

Conclusions We are on the verge of a new age of dating the tree of life. The decades up to now have been characterized by many improvements in methods and assumptions, but also by tension and squabbling between paleontologists and molecular clock practitioners, and within the


paleontological and molecular camps. We have suggested that a certain amount of that squabbling has been unhelpful or misguided, because people were to some extent talking past each other. We identify a number of major advances in the last years. Paleontologists are beginning to explore the quality of their data and they are learning to provide the information that is required by molecular analysts. A clearer understanding of how fossils and dates relate to phylogenetic trees, and greater clarity about stem grades and crown clades, have sharpened the debate. The vision of a molecular clock with minimal, especially paleontological, assumptions is giving way to a view that realistic dates for evolutionary events can only be obtained by integrating a greater number of less-constraining assumptions, particularly concerning the nature of fossil distribution and the nature of the rock record. Further, new insights and new algorithms are providing better tools for tree analysis that take account of the reality of the uncertainty in paleontological data. More work is needed by paleontologists and geologists to clarify specific dates, and to tighten their precision further. In addition, paleontologists must become less optimistic in their claims about “the oldest X,” even though this might mean fewer papers in Science and Nature. On the molecular side, intense study is needed to identify genes, and classes of genes, that are phylogenetically informative and phylogenetically uninformative. Of course, more sequences are needed, especially for previously unsequenced minor classes and phyla—such minor taxa, often “living fossils,” can be crucial in pinpointing the origins of major clades. Dating the tree of life is a grand enterprise, and it is a privilege to live through such times of major change and discovery.

Acknowledgments We thank S. Bengtson, N. Butterfield, G. Edgecombe, R. Jenner, J. Peel, A. Rasnitsyn, J. Repetski, I. Sansom, D. Schmidt, D. Shcherbakov, M. Siddall, and C. Wellman for valuable discussion that led to the establishment of some of the divergence constraints presented. J. Müller and two anonymous referees provided constructive reviews of the manuscript. S. Powell drafted the figures. M.J.B. and P.C.J.D. are funded by the Natural Environment Research Council. P.C.J.D. is also funded by the Biotechnology and Biological Sciences Research Council, National Endowment for Science, Technology and the Arts, and The Leverhulme Trust.

77

References 1. S. Easteal, BioEssays 21, 1052 (1999). 2. M. J. Benton, BioEssays 21, 1043 (1999). 3. M. J. Benton, P. C. J. Donoghue, Mol. Biol. Evol. 24, 26 (2007). 4. P. C. J. Donoghue, M. J. Benton, Trends Ecol. Evol. 22, 424 (2007). 5. N. Steno [Stensen], De Solido intra Solidum [Concerning solids that are naturally contained within solids]. Naturalites Contento Dissertationis Prodromus (Stellae, Florence, 1669), pp. 78. 6. M. A. Norell, M. J. Novacek, Science 255, 1690 (1992). 7. L. E. Edwards, Palaios 4, 127 (1989). 8. P. M. Sadler, Ann. Rev. Earth Planet. Sci. 32, 187 (2004). 9. A. H. Knoll, M. R. Walter, G. M. Narbonne, N. ChristieBlick, Lethaia 39, 13 (2006). 10. C. Darwin, On the Origin of Species (John Murray, London, 1859), pp. 502. 11. D. M. Raup, Science 177, 1065 (1972). 12. M. J. Benton, M. A. Wills, R. Hitchin, Nature 403, 534 (2000). 13. S. M. Holland, Paleobiology 21, 92 (1995). 14. A. B. Smith, Phil. Trans. Roy. Soc. Lond. Ser. B 356, 351 (2001). 15. A. B. Smith, J. Geol. Soc. Lond. 164, 731 (2007). 16. A. B. Smith, A. J. McGowan, Palaeontology 50, 765 (2007). 17. A. B. Smith, A. S. Gale, N. E. A. Monks, Paleobiology 27, 241 (2001). 18. A. Holmes, Principles of Physical Geology (Thomas Nelson and Sons, Edinburgh, 1944), pp. 532. 19. D. H. Erwin, Ann. Rev. Earth Planet. Sci. 34, 569 (2006). 20. G. G. Simpson, Tempo and Mode in Evolution (Columbia University Press, New York, 1944), pp. 237. 21. K. E. Omland, Evolution 51, 1381 (1997). 22. L. Bromham, M. Woolfit, M. S. Y. Lee, A. Rambaut, Evolution 56, 1921 (2002). 23. W. Hennig, Grundzuege einer Theorie der phylogenetischen Systematik (Deutscher Zentralverlag, Berlin, 1950), pp. 370. 24. W. Hennig, Phylogenetic Systematics (University of Illinois, Urbana and Chicago, 1966), pp. 263. 25. W. Hennig, Insect Phylogeny (John Wiley, New York, 1981), pp. 514. 26. R. P. S. Jefferies, in The Origin of Major Invertebrate Groups, M. R. House, Ed. (Systematics Association, 1979), Special Volume 12, pp. 443. 27. C. A. Brochu, C. D. Sumrall, J. M. Theodor, J. Paleont. 78, 1 (2004). 28. P. C. J. Donoghue, Paleobiology 31, 553 (2005). 29. C. A. Brochu, Copeia 2000, 657 (2000). 30. L. Bromham, D. Penny, M. Phillips, Trends Ecol. Evol. 14, 278 (1999). 31. J. D. Archibald, Trends Ecol. Evol. 14, 278 (1999). 32. M. J. Benton, J. Mol. Evol. 30 (1990).

78


33. J. Alroy, Syst. Biol. 48, 107 (1999). 34. M. S. Y. Lee, J. Mol. Evol. 49, 385 (1999). 35. C. J. Conroy, M. van Tuinen, J. Mammal. 84, 444 (2003). 36. D. Graur, W. Martin, Trends Genet. 20, 80 (2004). 37. R. R. Reisz, J. Müller, Trends Genet. 20, 237 (2004). 38. A. B. Smith, K. J. Peterson, Ann. Rev. Earth Planet. Sci. 30, 65 (2002). 39. P. S. Soltis, D. E. Soltis, V. Savolainen, P. R. Crane, T. G. Barraclough, Proc. Natl. Acad. Sci. U.S.A. 99, 4430 (2002). 40. E. J. P. Douzery, E. A. Snell, E. Bapteste, F. Delsuc, H. Philippe, Proc. Natl. Acad. Sci. U.S.A. 101, 15386 (2004). 41. T. J. Near, P. A. Meylan, H. B. Shaffer, Am. Nat. 165, 137 (2005). 42. T. J. Near, M. J. Sanderson, Phil. Trans. Roy. Soc. Lond. Ser. B 359, 1477 (2004). 43. S. B. Hedges, S. Kumar, M. Van Tuinen, BioEssays 28, 770 (2006). 44. A. F. Hugall, R. Foster, M. S. Y. Lee, Syst. Biol. 56, 543 (2007). 45. E. J. P. Douzery, F. Delsuc, M. J. Stanhope, D. Huchon, J. Mol. Evol. 57, S201 (2003). 46. E. Gheerbrant, J. Sudre, H. Cappetta, Nature 383 (1996). 47. E. Gheerbrant, J. Sudre, M. Iarochene, A. Moumni, J. Vert. Paleont. 21, 107 (2001). 48. R. J. Asher et al., Science 307, 1091 (2005). 49. Y. Kitazoe et al., PLoS ONE 2, e384 (2007). 50. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 51. M. S. Springer, W. J. Murphy, E. Eizirik, S. J. O’Brien, Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003). 52. O. R. P. Bininda-Emonds et al., Nature 446, 507 (2007). 53. R. R. Reisz, J. Müller, Trends Genet. 20, 596 (2004). 54. J. Müller, R. R. Reisz, BioEssays 27, 1069 (2005). 55. J. Müller, R. R. Reisz, BioEssays 28, 772 (2006). 56. S. B. Hedges, F. U. Battistuzzi, J. E. Blair, in Neoproterozoic Geobiology and Paleobiology, S. Xiao, A. J. Kaufman, Eds. (Springer, 2006), pp. 199. 57. S. B. Hedges, S. Kumar, Trends Genet. 20, 242 (2004). 58. J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 387 (2005). 59. M. J. Benton, Ed., The Fossil Record 2 (Chapman & Hall, London, 1993), pp. 845. 60. I. J. Sansom, M. P. Smith, M. M. Smith, Nature 379, 628 (1996). 61. P. D. Gingerich, M. Uhen, J. Hum. Evol. 27, 443 (1994). 62. Z. Yang, B. Rannala, Mol. Biol. Evol. 23, 212 (2006). 63. A. J. Drummond, S. Y. W. Ho, M. J. Phillips, A. Rambaut, PLoS Biol. 4, e88 (2006). 64. R. Barnett et al., Curr. Biol. 15, R589 (2005). 65. S. B. Hedges, S. Kumar, M. van Tuinen, BioEssays 28, 770 (2006). 66. M. Foote, J. P. Hunter, C. M. Janis, J. J. Sepkoski, Science 283, 1310 (1999).

67. S. Tavaré, C. R. Marshall, O. Will, C. Soligo, R. D. Martin, Nature 416, 726 (2002). 68. E. Zuckerkandl, L. Pauling, in Evolving Genes and Proteins, V. Bryson, H. J. Vogel, Eds. (Academic Press, New York, 1965), pp. 97. 69. E. Zuckerkandl, L. Pauling, in Horizons in Biochemistry, M. Kasha, B. Pullman, Eds. (Academic Press, New York, 1962), pp. 189. 70. C. H. Langley, W. M. Fitch, J. Mol. Evol. 3, 161 (1974). 71. R. J. Britten, Science 231, 1393 (1986). 72. W.-H. Li, M. Tanimura, Nature 326, 93 (1987). 73. N. Takezaki, A. Rzhetsky, M. Nei, Mol. Biol. Evol. 12, 823 (1995). 74. S. B. Hedges, P. H. Parker, C. G. Sibley, S. Kumar, Nature 381, 226 (1996). 75. F. J. Ayala, A. Rzhetsky, F. J. Ayala, Proc. Natl. Acad. Sci. U.S.A. 95, 606 (1998). 76. A. Rambaut, L. Bromham, Mol. Biol. Evol. 15, 442 (1998). 77. L. Bromham, D. Penny, Nat. Rev. Genet. 4, 216 (2003). 78. D. J. Cutler, Mol. Biol. Evol. 17, 1647 (2000). 79. L. Bromham, D. Penny, A. Rambaut, M. D. Hendy, J. Mol. Evol. 50, 296 (2000). 80. A. D. Yoder, Z. H. Yang, Mol. Biol. Evol. 17, 1081 (2000). 81. M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997). 82. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 83. J. P. Huelsenbeck, B. Rannala, J. P. Masly, Science 288, 2349 (2000). 84. S. Aris-Brosou, Z. Yang, Mol. Biol. Evol. 20, 1947 (2003). 85. M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002). 86. A. J. Drummond, A. Rambaut, BMC Evol. Biol. 7, 214 (2007). 87. A. B. Smith et al., Mol. Biol. Evol. 23, 1832 (2006). 88. K. J. Peterson et al., Proc. Natl. Acad. Sci. U.S.A. 101, 6536 (2004). 89. S. B. Hedges, S. Kumar, Trends Genet. 19, 200 (2003). 90. M. W. Gaunt, M. A. Miles, Mol. Biol. Evol. 19, 748 (2002). 91. K. J. Peterson, N. J. Butterfield, Proc. Natl. Acad. Sci. U.S.A. 102, 9547 (2005). 92. L. Bromham, Palaeont. Electr. 9 (2006). 93. O. G. Pybus, PLoS Biol. 4, e151 (2006). 94. B. Rannala, Z. H. Yang, Syst. Biol. 56, 453 (2007). 95. M. Wilkinson, J. L. Thorley, P. Upchurch, Syst. Biol. 49, 754 (2000). 96. M. Wilkinson, M. J. Benton, Phil. Trans. Roy. Soc. Lond. Ser. B 351, 1 (1996). 97. K. Cranston, B. Rannala, Heredity 94, 461 (2005). 98. J. D. Archibald, Science 272, 1150 (1996). 99. J. D. Archibald, A. O. Averianov, E. G. Ekdale, Nature 414, 62 (2001). 100. L. Van Valen, Bull. Am. Mus. Nat. Hist. 132, 1 (1966). 101. R. K. Stucky, M. C. McKenna, in The Fossil Record 2, M. J. Benton, Ed. (Chapman & Hall, London, 1993), pp. 43.


102. Z. Kielan-Jawarowska, T. M. Brown, J. A. Lillegraven, in Mesozoic Mammals, J. A. Lillegraven, Z. KielanJawarowska, W. A. Clemens, Eds. (University of California Press, Berkeley and Los Angeles, 1979), pp. 221. 103. J. Meng, Y. Hu, C. K. Li, Bull. Am. Mus. Nat. Hist. 275, 1 (2003). 104. J. W. Wible, G. W. Rougier, M. J. Novacek, R. J. Asher, Nature 447, 1003 (2007). 105. Z. X. Luo, J. R. Wible, Science 308, 103 (2005). 106. J. Meng, A. R. Wyss, J. Mammal. Evol. 8, 1 (2001). 107. H. P. Luterbacher et al., in A Geologic Time Scale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 384. 108. M. S. Springer, W. J. Murphy, Biol. Rev. 82, 375 (2007). 109. S. J. Bourlat et al., Nature 444, 85 (2006). 110. G. G. Simpson, Bull. Am. Mus. Nat. Hist. 85, 1 (1945). 111. P. J. Waddell, N. Okada, M. Hasegawa, Syst. Biol. 48, 1 (1999). 112. R. Klein, The Human Career: Human Biological and Cultural Origins (University of Chicago Press, Chicago, 1999), pp. 840. 113. M. Vanhaeren et al., Science 312, 1785 (2006). 114. F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds., A Geologic Timescale 2004 (Cambridge University Press, Cambridge, 2004), pp. 500. 115. D. Lordkipanidze et al., Nature 449, 305 (2007). 116. V. M. Sarich, A. C. I. Wilson, Science 158, 1200 (1967). 117. M. Hasegawa, H. Kishino, T. Yano, J. Mol. Evol. 22, 160 (1985). 118. S. McBrearty, N. G. Jablonski, Nature 437, 105 (2005). 119. D. C. Johanson, M. Taieb, Nature 260, 293 (1976). 120. T. D. White, G. Suwa, B. Asfaw, Nature 371, 306 (1994). 121. M. G. Leakey, C. S. Feibel, I. McDougall, A. C. Walker, Nature 376, 565 (1995). 122. Y. Haile-Selassie, Nature 412, 178 (2001). 123. B. Senut et al., C. R. Acad. Sci. 332, 137 (2001). 124. M. Brunet, F. Guy, D. Pilbeam, H. T. Mackay, A. Likius, A. Djimboumalbaye et al., Nature 418, 145 (2002). 125. M. H. Wolpoff, B. Senut, M. Pickford, J. Hawks, Nature 419, 581 (2002). 126. B. G. Richmond, W. L. Jungers, Science 319, 1662 (2008). 127. A. L. Deino, L. Tauxe, M. Monaghan, A. Hill, J. Hum. Evol. 42, 117 (2002). 128. P. Vignaud et al., Nature 418, 152 (2002). 129. Y. Sawada, M. Pickford, B. Senut et al., C. R. Palévol. 1, 293 (2002). 130. S. Kumar, A. Filipski, V. Swarna, A. Walker, S. B. Hedges, Proc. Natl. Acad. Sci. U.S.A. 102, 18842 (2005). 131. G. Suwa, R. T. Kono, S. Katoh, B. Asfaw, Y. Beyene, Nature 448, 921 (2007). 132. J. G. Fleagle, Primate Adaptation and Evolution (Academic Press, San Diego, 1999), pp. 596.

79

133. M. G. Leakey, P. S. Ungar, A. Walker, J. Hum. Evol. 28, 519 (1995). 134. J. Kappelman et al., J. Hum. Evol. 21, 61 (1991). 135. E. R. Seiffert, Proc. Natl. Acad. Sci. U.S.A. 103, 5000 (2006). 136. J. B. Rossie, L. MacLatchy, J. Hum. Evol. 50, 568 (2006). 137. T. Harrison, Y. Gu, J. Hum. Evol. 37, 225 (1999). 138. E. R. Miller, J. Hum. Evol. 36, 519 (1999). 139. L. MacLatchy, W. Downs, R. Kityo, M. Mafabi, E. Musiime, Eds., New Catarrhine Fossils from the Lower Miocene of Uganda, with Implications for the Ape–Monkey Split; online abstracts, Paleoanthropology Society Meeting, 2003. 140. D. L. Gebo et al., Science 276, 401 (1997). 141. N. M. Young, L. MacLatchy, J. Hum. Evol. 46, 163 (2004). 142. J. A. Finarelli, W. C. Clyde, Paleobiology 30, 614 (2004). 143. M. Pickford, P. Andrews, J. Hum. Evol. 10, 11 (1981). 144. P. Tassy, M. Pickford, Geobios 16, 53 (1983). 145. H. B. Boschetto, F. H. Brown, I. McDougall, J. Hum. Evol. 22, 47 (1992). 146. D. T. Rasmussen, in The Primate Fossil Record, W. C. Hartwig, Ed. (Cambridge University Press, Cambridge, 2002), pp. 203. 147. E. R. Seiffert et al., Science 310, 300 (2005). 148. R. F. Kay, D. J. Meldrum, in Vertebrate Paleontology in the Neotropics, R. F. Kay, R. H. Madden, R. L. Cifelli, J. J. Flynn, Eds. (Smithsonian Institution Press, Washington, D.C., 1997), pp. 435. 149. M. T. Silcox, J. I. Bloch, E. J. Sargis, D. M. Boyer, in The Rise of Placental Mammals: Origins and Relationships of the Major Extant Clades, K. D. Rose, J. D. Archibald, Eds. (Johns Hopkins University Press, Baltimore & London, 2005), pp. 127. 150. J. I. Bloch, M. T. Silcox, D. M. Boyer, E. J. Sargis, Proc. Natl. Acad. Sci. U.S.A. 104, 1159 (2007). 151. B. Sigé, J.-J. Jaeger, J. Sudre, M. Vianey-Liaud, Palaeontogr. Abt. A 214, 31 (1990). 152. E. R. Seiffert, Am. J. Primatol. 69, 27 (2007). 153. E. Gheerbrant et al., Palaeovertebrata 27, 155 (1998). 154. M. C. McKenna, S. K. Bell, Classification of Mammals— Above the Species Level (Columbia University Press, New York, 1997), pp. 640. 155. P. E. Kondrashov, S. G. Lucas, New Mexico Mus. Nat. Hist. Sci. Bull. 26, 209 (2004). 156. E. R. Seiffert, E. L. Simons, Y. Attia, Nature 422, 421 (2003). 157. L. Van Valen, Evolution 18, 484 (1964). 158. Z. Kielan-Jaworowska, Palaeont. Pol. 38, 3 (1978). 159. Q. Ji et al., Nature 416, 816 (2002). 160. Z. X. Luo, Q. Ji, J. R. Wible, C. X. Yuan, Science 302, 1934 (2003). 161. J. Meng, Y. Hu, C. Li, Y. Wang, Geol. J. 41, 439 (2006). 162. Z.-H. Zhou, P. M. Barrett, J. Hilton, Nature 807 (2003). 163. A. Janke, X. F. Xu, U. Arnason, Proc. Natl. Acad. Sci. U.S.A. 94, 1276 (1997). 164. J. I. Bloch, D. M. Boyer, Science 298, 1606 (2002).

80


165. C.-K. Li, Vertebr. PalAsiat. 15, 103 (1977). 166. C. Li, S. Ting, Bull. Carnegie Mus. Nat. Hist. 21, 1 (1983). 167. J. Meng, A. R. Wyss, in The Rise of Placental Mammals: Origins and Relationships of Major Extant Clades, K. D. Rose, J. D. Archibald, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 145. 168. T. Martin, J. Vert. Paleont. 24, 411 (2004). 169. K. D. Rose et al., Proc. Roy. Soc. Ser. B, Biol. Sci. doi: 10.1098/rspb.2007.1661 (2008). 170. J. Meng, Y. Hu, C. Li, Palaeont. Electr. 8–1, 1 (2005). 171. A. V. Lopatin, A. O. Averianov, Paleont. J. 40, 198 (2006). 172. J. Meng, Y. Hu, C. Li, Bull. Am. Mus. Nat. Hist. 275, 1 (2003). 173. D. Huchon et al., Mol. Biol. Evol. 19, 1053 (2002). 174. R. M. Adkins, A. H. Walton, R. L. Honeycutt, Mol. Phylog. Evol. 26, 409 (2003). 175. R. M. Adkins, E. L. Gelke, D. Rowe, R. L. Honeycutt, Mol. Biol. Evol. 18, 777 (2001). 176. N. S. Shevyreva, in [Flora and fauna of Zaysan Hollow], Gabuniya, Ed. (Metsniereba, Tbilisi, 1984), pp. 81. 177. S. G. Lucas, in Late Paleocene–Early Eocene Climatic and Biotic Events in the Marine and Terrestrial Records, M. P. Aubry, S. G. Lucas, W. A. Berggren, Eds. (Columbia University Press, New York, 1998), pp. 451. 178. J. Michaux, A. Reyes, F. Catzeflis, Mol. Biol. Evol. 18, 2017 (2001). 179. S. A. Jansa, M. Weksler, Mol. Phylog. Evol. 31, 256 (2004). 180. S. J. Steppan, R. M. Adkins, J. Anderson, Syst. Biol. 53, 533 (2004). 181. P. Chevret, F. Veyrunes, J. Britton-Davidian, Biol. J. Linn. Soc. 84, 417 (2005). 182. L. L. Jacobs, L. J. Flynn, in Interpreting the Past: Essays on Human, Primate and Mammal Evolution, D. Lieberman, R. Smith, J. Kelley, Eds. (Brill Academic Publishers, Leiden, 2005), pp. 63. 183. Y. Chaimanee, V. Suteethorn, S. Triamwichanon, J. J. Jaeger, C. R. Acad. Sci. Ser. II, Fasc. A, Sci. Terre Planètes 322, 155 (1996). 184. S. Zheng, Quaternary Rodents of Sichuan-Guizhou Area, China (Science Press, Beijng, 1993), pp. 270. 185. L. L. Jacobs, D. Pilbeam, J. Hum. Evol. 9, 551 (1980). 186. J. J. Jaeger, H. Tong, C. Denys, C. R. Acad. Sci. Se. II, Fasc. A, Sci. Terre Planètes 302, 917 (1986). 187. L. L. Jacobs, W. R. Downs, in Rodent and Lagomorph Families of Asian Origins and their Diversification, Y. Tomida, C. K. Li, T. Setoguschi, Eds. (National Science Museum Monograph, Tokyo, 1994), pp. 149. 188. N. M. Johnson, J. Stix, L. Tauxe, P. F. Cerveny, R. A. K. Tahirkheli, J. Geol. 93, 27 (1985). 189. J. C. Barry et al., Paleobiology 28, 1 (2002). 190. J. D. Archibald, Mol. Phylog. Evol. 28, 350 (2003). 191. A. Averianov, J. D. Archibald, Cret. Res. 26, 593 (2005). 192. M. J. Novacek, Am. Mus. Novit. 2833, 1 (1985). 193. P. D. Gingerich, in The Rise of Placental Mammals: Origins and Relationships of the Major Extant Clades,

194. 195.

196.

197.

198. 199. 200. 201.

202.

203. 204.

205. 206. 207.

208. 209. 210. 211. 212. 213. 214. 215. 216.

K. D. Rose, J. D. Archibald, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 234. T. J. Meehan, R. W. Wilson, J. Paleont. 76, 1091 (2002). N. B. Simmons, in Origins of the Major Clades of Placental Mammals, K. Rose, D. Archibald, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 159. D. L. Lofgren, J. A. Lillegraven, W. A. Clemens, P. D. Gingerich, T. E. Williamson, in Late Cretaceous and Cenozoic Mammals of North America, M. O. Woodburne, Ed. (Columbia University Press, New York, 2004), pp. 43. J. M. Theodor, K. D. Rose, J. Erfurt, in The Rise of Placental Mammals: Origins and Relationships of the Major Extant Clades, K. D. Rose, J. D. Archibald, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 215. P. D. Gingerich, M. ul Haq, I. S. Zalmout, I. H. Khan, M. S. Malkani, Science 293, 2239 (2001). R. J. Emry, Bull. Am. Mus. Nat. Hist. 285, 130 (2004). J. J. Flynn, J. A. Finarelli, S. Zehr, J. Hsu, M. A. Nedbal, Syst. Biol. 54, 317 (2005). J. J. Flynn, in Evolution of Tertiary Mammals of North America. Vol. 1: Terrestrial Carnivores, Ungulates, and Ungulatelike Mammals, C. M. Janis, K. M. Scott, L. L. Jacobs, Eds. (Cambridge University Press, Cambridge, 1998), pp. 110. J. J. Flynn, G. D. Wesley-Hunt, in The Rise of Placental Mammals: Origins and Relationships of the Major Extant Clades, K. D. Rose, J. D. Archibald, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 175. K. C. Beard, Bull Carnegie Mus. Nat. Hist. 34, 5 (1998). P. Robinson et al., in Late Cretaceous and Cenozoic Mammals of North America: Biostratigraphy and Geochronology, M. O. Woodburne, Ed. (Columbia University Press, New York, 2004), pp. 165. G. D. Wesley, J. J. Flynn, J. Paleont. 77, 769 (2003). J. M. Hunt, Bull. Am. Mus. Nat. Hist. 285, 139 (2004). D. R. Prothero, R. J. Emry, in Late Cretaceous and Cenozoic Mammals of North America: Biostratigraphy and Geochronology, M. O. Woodburne, Ed. (Columbia University Press, New York, 2004), pp. 156. J. H. Geisler, M. D. Uhen, J. Vert. Paleont. 23, 991 (2003). J. Gatesy, C. Matthee, R. DeSalle, C. Hayashi, Syst. Biol. 51, 652 (2002). S. Bajpai, P. D. Gingerich, Proc. Natl. Acad. Sci. U.S.A. 95, 15464 (1998). M. Marti’n-Marti’n et al., Basin Res. 13, 419 (2001). U. Arnason, A. Gullberg, S. Gretarsdottir, B. Ursing, A. Janke, J. Mol. Evol. 50, 569 (2000). A. Hassanin, E. J. P. Douzery, Mol. Phylog. Evol. 13, 227 (1999). M. H. Fernandez, E. S. Vrba, Biol. Rev. 80, 269 (2005). G. Metais, P. O. Antoine, L. Marivaux, J. L. Welcomme, S. Ducrocq, Acta Palaeont. Pol. 48, 375 (2003). J. C. Barry et al., Palaeont. Electr. 8 (2005).


217. N. Solounias, J. C. Barry, R. L. Bernor, E. H. Lindsay, S. M. Raza, J. Vert. Paleont. 15, 806 (1995). 218. E. S. Vrba, G. Schaller, Antelopes, Deer, and Relatives: Fossil Record, Behavioral Ecology, Systematics, and Conservation (Yale University Press, New Haven, 2000), pp. 356. 219. E. Gheerbrant, D. P. Domning, P. Tassy, in The Rise of Placental Mammals: Origins and Relationships of the Major Extant Clades, K. D. Rose, J. D. Archibald, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 84. 220. G. J. Scillato-Yané, An., Acad. Brasil. Ciênc. 48, 527 (1976). 221. R. Tabuce et al., Proc. Roy. Soc. Ser. B, Biol. Sci. 274, 1159 (2007). 222. E. R. Seiffert, E. L. Simons, Proc. Natl. Acad. Sci. U.S.A. 97, 2646 (2001). 223. S. P. Zack, T. A. Penkrot, J. I. Bloch, K. D. Rose, Nature 434, 497 (2005). 224. M. A. O’Leary, E. M. Roberts, M. Bouare, F. Sissoko, L. Tapanila, J. Vert. Paleont. 26, 981 (2006). 225. R. L. Cifelli, Proc. Natl. Acad. Sci. U.S.A. 90, 9413 (1993). 226. Z. X. Luo, P. Chen, G. Li, M. Chen, Nature 466, 288 (2007). 227. Z. Kielan-Jaworowska, D. Dashzeveg, Zool. Scripta 18, 347 (1989). 228. R. L. Cifelli, Nature 401, 363 (1999). 229. A. O. Averianov, P. P. Skutschas, Acta Palaeont. Pol. 46, 431 (2001). 230. M. O. Woodburne, T. H. Rich, M. S. Springer, Mol. Phylog. Evol. 28, 360 (2003). 231. M. R. Sánchez-Villagra et al., Biol. Lett. 3, 318 (2007). 232. C. Muizon, R. L. Cifelli, J. Vert. Paleont. 21, 8 (2001). 233. F. J. Goin et al., J. Vert. Paleont. 26, 505 (2006). 234. M. A. Nilsson, U. Arnason, P. B. S. Spencer, A. Janke, Gene 340, 189 (2004). 235. H. Godthelp, S. Wroe, M. Archer, J. Mammal. Evol. 6, 289 (1999). 236. F. S. Szalay, B. A. Trofimov, J. Vert. Paleont. 16, 474 (1996). 237. Z. Luo, Z. Kielan-Jaworowska, R. L. Cifelli, Acta Palaeont. Pol. 47, 1 (2002). 238. Z. Kielan-Jaworowska, R. L. Cifelli, Z.-X. Luo, Mammals from the Age of Dinosaurs (Columbia University Press, New York, 2005), pp. 630. 239. B. F. Boneham, R. J. Wyatt, Proc. Geol. Ass. 104, 123 (1993). 240. J. G. Ogg, in A Geologic Time Scale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 307. 241. A. O. Averianov, Acta Palaeont. Pol. 47, 705 (2002). 242. O. W. M. Rauhut, T. Martin, E. Ortiz-Jaureguizar, P. Puerta, Nature 416, 165 (2002). 243. J. J. Flynn, J. M. Parrish, B. Rakotosamimanana, W. F. Simpson, A. R. Wyss, Nature 401, 57 (1999). 244. M. van Tuinen, E. A. Hadly, J. Mol. Evol. 59, 267 (2004). 245. M. deBraga, O. Rieppel, Zool. J. Linn. Soc. 120, 281 (1997).

81

246. V. H. Reynoso, Phil. Trans. Roy. Soc. Lond. Ser. B 353, 477 (1998). 247. M. J. Benton, Vertebrate Palaeontology, 3rd ed. (Blackwell, Oxford, 2005), pp. 455. 248. J. H. Calder, Palaeogeog. Palaeoclim. Palaeoecol. 106, 323 (1994). 249. G. Dolby, Nova Scotia Dept. Mines Energy Open File Rep. 91-006, 39 (1991). 250. H. J. Falcon-Lang, M. J. Benton, S. J. Braddy, S. J. Davies, J. Geol. Soc. Lond. 163, 561 (2006). 251. M. Menning, D. Weyer, G. Drozdzewski, H. W. J. Amerom, I. A. Wendt, Geol. Jahrb. A156, 3 (2000). 252. V. Davydov, B. R. Wardlaw, F. M. Gradstein, in A geologic Time Scale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 222. 253. R. R. Reisz, Bull. Mus. Comp. Zool. Harvard 144, 27 (1972). 254. R. R. Reisz, S. P. Modesto, Can. J. Earth Sci. 33, 703 (1996). 255. R. R. Reisz, Handbuch der Paläoherpetologie 17A, 1 (1986). 256. J. Müller, R. R. Reisz, Syst. Biol. 55, 503 (2006). 257. J. Gauthier, A. Kluge, T. Rowe, in The Phylogeny and Classification of the Tetrapods, Volume 1: Amphibians, Reptiles and Birds, M. J. Benton, Ed. (Clarendon Press, Oxford, 1988), pp. 103. 258. M. Laurin, R. R. Reisz, Zool. J. Linn. Soc. 113, 165 (1995). 259. R. L. Carroll, Zool. J. Linn. Soc. 45, 61 (1964). 260. M. J. Benton, in The Fossil Record 2, M. J. Benton, Ed. (Chapman & Hall, London, 1993), pp. 681. 261. H. J. Falcon-Lang, M. J. Benton, M. Stimson, J. Geol. Soc. Lond. 164 (2007). 262. M. J. Benton, Zool. J. Linn. Soc. 84, 97 (1985). 263. S. E. Evans, in The Phylogeny and Classification of the Tetrapods. Volume 1: Amphibians, Reptiles and Birds, M. J. Benton, Ed. (Clarendon Press, Oxford, 1988), pp. 221. 264. M. Laurin, Zool. J. Linn. Soc. 101, 59 (1991). 265. D. W. Dilkes, Phil. Trans. Roy. Soc. Lond. Ser. B 353, 501 (1998). 266. S. E. Evans, M. E. King, Proc. Yorks. Geol. Soc. 49, 229 (1993). 267. J. G. Ogg, in A Geologic Time Scale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 271. 268. M. Roscher, J. W. Schneider, in Non-Marine Permian Biostratigraphy and Biochronology. Geological Society, London, Special Publications, 265, S. G. Lucas, G. Cassinis, J. W. Schneider, Eds. (2006), pp. 95. 269. J. M. Clark, R. Hernandez, J. Vert. Paleont. 14, 180 (1994). 270. P. J. Hancox, J. S. Rubidge, Palaeont. Afr. 33, 41 (1997). 271. M. Caldwell, Neues Jahrb. Geol. Paläont., Abh. 200, 361 (1996).

82


272. J. Müller, in Recent Advances in the Origin and Early Radiation of Vertebrates, G. Arratia, M. V. H. Wilson, R. Cloutier, Eds. (Pfeil, München, 2004), pp. 379. 273. G. W. Storrs, Bull. Peabody Mus. Nat. Hist. 44, 1 (1991). 274. J. Müller, Naturwissenschaften 90, 473 (2003). 275. J. A. Gauthier, Mem. Calif. Acad. Sci. 8, 1 (1986). 276. P. C. Sereno, Soc. Vert. Paleont. Mem. 2, 1 (1991). 277. M. J. Benton, Phil. Trans. Roy. Soc. Lond. Ser. B 354, 1423 (1999). 278. S. J. Nesbitt, Hist. Biol. 17, 19 (2005). 279. D. C. Parris, S. Hope, in Proceedings of the 5th International Meeting of the Society of Avian Paleontology and Evolution, Z. Zhou, F. Zhang, Eds. (Science Press, Beijing, 2002), pp. 113. 280. G. J. Dyke, M. Van Tuinen, Zool. J. Linn. Soc. 141, 153 (2004). 281. E. Buffetaut, J. Le Loeuff, J. Geol. Soc. Lond. 155, 1 (1998). 282. J. Cracraft et al., in Assembling the Tree of Life, J. Cracraft, M. J. Donoghue, Eds. (Oxford University Press, New York, 2004), pp. 469. 283. E. N. Kurochkin, G. J. Dyke, A. A. Karhu, Am. Mus. Novit. 3386, 1 (2002). 284. J. A. Lillegraven, M. C. McKenna, Am. Mus. Novit. 2840, 1 (1986). 285. J. A. Clarke, M. A. Norell, Am. Mus. Novit. 3447, 1 (2004). 286. J. A. Clarke, C. P. Tambussi, J. I. Noriega, G. M. Erickson, R. A. Ketcham, Nature 433, 305 (2005). 287. G. J. Dyke, Geol. J. 36, 305 (2001). 288. S. Hope, in Mesozoic Birds: Above the Heads of Dinosaurs, L. Chiappe, L. Witmer, Eds. (University of California, Berkeley, 2002), pp. 339. 289. S. Chatterjee, in Proceedings of the 5th International Meeting of the Society of Avian Paleontology and Evolution, Z. Zhou, F. Zhang, Eds. (Science Press, Beijing, 2002). 290. D. Pirrie, J. A. Crame, S. A. Lomas, J. B. Riding, Cret. Res. 18, 109 (1997). 291. J. Clack, Nature 394, 66 (1998). 292. M. Ruta, M. I. Coates, D. L. J. Quicke, Biol. Rev. 78, 251 (2003). 293. W. D. I. Rolfe, E. N. K. Clarkson, A. L. Panchen, Trans. Roy. Soc. Edinb.: Earth Sci. 84 (1993). 294. R. L. Paton, T. R. Smithson, J. A. Clack, Nature 398, 508 (1999). 295. W. Gross, Lethaia 1, 184 (1968). 296. D. Fredholm, Geol. Fören. Stockholm Förhand. 110, 237 (1988). 297. D. Fredholm, Geol. Fören. Stockholm Förhand. 110, 157 (1988). 298. T. Märss, Fossilia Baltica 1, 1 (1986). 299. T. Märss, Proc. Estonian Acad. Sci.: Geol. 50, 174 (2001). 300. M. A. Kleff ner, in Graphic Correlation, Special Publication 53, K. O. Mann, H. R. Lane, Eds. (SEPM, 1995), pp. 159. 301. B. G. Fordham, paper presented at the Sixth graptolite conference of the GWG (IPA) and the 1998 field

302. 303. 304. 305. 306. 307. 308.

309.

310.

311. 312. 313. 314. 315. 316.

317. 318. 319. 320. 321. 322. 323.

324. 325. 326. 327. 328. 329.

meeting of the International Subcommission on Silurian Stratigraphy (ICS-IUGS), (Madrid, 1998). P. Janvier, C. R. Acad. Sci. Paris, Sér. II 273, 2223 (1971). P. Janvier, Proc. Estonian Acad. Sci.: Geol. 27, 88 (1978). H. Botella, H. Blom, M. Dorka, P. E. Ahlberg, P. Janvier, Nature 448, 583 (2007). H.-P. Schultze, Bull. Br. Mus. (Nat. Hist.), Geol. 16, 343 (1968). N. Z. Wang, Z.-z. Dong, Acta Palaeont. Sin. 28, 192 (1989). M. Zhu, J.-Q. Wang, Cour. Forsch. Senck. 223, 161 (2000). H.-P. Schultze, in Fossil Fishes as Living Animals, E. Mark-Kurik, Ed. (Academy of Sciences of Estonia, Tallinn, 1992), pp. 233. H.-P. Schultze, S. L. Cumbaa, in Major Events in Early Vertebrate Evolution: Palaeontology, Phylogeny, Genetics and Development. Systematics Association Special Volume 61, P. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 315. S. V. Cherkesova, in Devonian of the World. Volume III: Paleontology, Paleoecology and Biostratigraphy, N. J. McMillan, A. Embry, F. D. J. Glass, Eds. (Canadian Society of Petroleum Geologists, Calgary, 1988), pp. 669. M. Zhu, H.-P. Schultze, Lethaia 30, 293 (1997). X. Yu, J. Vert. Paleont. 18, 261 (1998). M. Zhu, X. Yu, W. Wang, W. Zhao, L. Jia, Nature 441, 77 (2006). C.-y. Wang, Acta Palaeont. Sin. 20, 400 (1981). C. G. Miller, Palaeontology 38, 341 (1995). M. J. Melchin, R. A. Cooper, P. M. Sadler, in A Geologic Time Scale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 188. C. G. Miller, S. J. E. Sutherland, K. J. Dorning, Geol. J. 32, 69 (1997). R. D. Tucker, W. S. McKerrow, Can. J. Earth Sci. 32, 368 (1995). V. Viira, R. J. Aldridge, J. Micropal. 17, 33 (1998). Z. Fang et al., J. Stratigr. 18, 81 (1994). L. Jeppsson, Geol. Palaeont. 22, 21 (1988). M. Zhu, X. Yu, P. Janvier, Nature 397, 607 (1999). M. Zhu, H.-P. Schultze, in Major Events in Early Vertebrate Evolution: Palaeontology, Phylogeny, Genetics and Development. Systematics Association Special Volume 61, P. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 289. M. Zhu, X. Yu, P. E. Ahlberg, Nature 410, 81 (2001). S. Q. Wang, K. McKenzie, Senck. Leth. 79, 589 (2000). W. Gross, Lethaia 4, 131 (1971). W. Gross, Lethaia 2, 15 (1969). H.-P. Schultze, Paläont. Z. 51, 152 (1977). P. Janvier, Early Vertebrates, Oxford Monographs on Geology and Geophysics 33 (Oxford University Press, Oxford, 1996), pp. 393.


330. V. Karatajuté-Talimaa, N. Predtechenskyj, Bull. Mus. Natl. Hist. Nat. Paris 17, 39 (1995). 331. P. C. J. Donoghue, M. P. Smith, I. J. Sansom, in Telling the Evolutionary Time: Molecular Clocks and the Fossil Record, P. C. J. Donoghue, M. P. Smith, Eds. (CRC Press, London, 2003), pp. 190. 332. T.-D. Thanh, T. H. Phuong, A. J. Boucot, D. Goujet, J. P., C. R. Acad. Sci., Sci Terre Planètes 324, 1023 (1997). 333. Arratia, Palaeo. Ichthyol. 7, 1 (1997). 334. A. Zeiss, Stuttgarter Beitr. Naturk. 31, 1 (1977). 335. K. W. Barthels, N. H. M. Swinburne, S. Conway Morris, Solnhofen: A Study in Mesozoic Palaeontology (Cambridge University Press, Cambridge, 1990), pp. 236. 336. G. Arratia, H.-P. Schultze, in Mesozoic Fishes 2: Systematics and the Fossil Record, G. Arratia, H.-P. Schultze, Eds. (Pfeil, München, 1999), pp. 265. 337. D. Nolf, Mém. Acad. R. Belg., Classe Sci. 19, 1 (1988). 338. C. Patterson, in The Fossil Record 2, M. J. Benton, Ed. (Chapman & Hall, London, 1993), pp. 621. 339. L. Sorbini, Boll. Mus. Civ. Stor. Nat. Verona 6, 1 (1979). 340. J. C. Tyler, L. Sorbini, Smithson. Contr. Paleobiol. 82, 1 (1996). 341. U. Hückel, Neues Jahrb. Geol. Paläont., Abh. 135, 113 (1970). 342. A. Cherchi, R. Schroeder, Eclogae Geol. Helv. 97, 441 (2004). 343. J. D. Obradovich, T. Matsumoto, T. Nishida, Y. Inoue, Proc. Japan Acad., Ser. B, Phys. Biol. Sci. 78, 149 (2002). 344. J. G. Ogg, F. P. Agterberg, F. M. Gradstein, in A Geologic Time Scale 2004, F. M. Gradstein, J. G. Ogg, A. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 344. 345. L. Sorbini, Boll. Mus. Civ. Stor. Nat. Verona 8, 1 (1981). 346. N. I. Holcroft, Mol. Phylog. Evol. 34, 525 (2005). 347. F. Santini, J. C. Tyler, Zool. J. Linn. Soc. 139, 565 (2003). 348. J. C. Tyler, A. F. Bannikov, Proc. Biol. Soc. Wash. 107, 97 (1994). 349. K. Gürgey, Mar. Petrol. Geol. 20, 1119 (2003). 350. Y. G. Leonov et al., Eds., Late Eocene–Early Oligocene Geological and Biotical Events on the Territory of the Former Soviet Union. Part II. The Geological and Biotical Events (GEOS, Moscow, 1998), pp. 250. 351. S. Spezzaferri, I. Premoli Silva, Palaeogeog. Palaeoclim. Palaeoecol. 83, 217 (1991). 352. P. S. Willumsen, Bull. Geol. Soc. Denmark 52, 141 (2004). 353. J. Sansom, R. J. Aldridge, M. M. Smith, Proc. Roy. Soc. Edinb.: Earth Sci. 90, 255 (2000). 354. J. Sansom, N.-Z. Wang, M. Smith, Zool. J. Linn. Soc. 144, 379 (2005). 355. T. Märss, M. V. H. Wilson, R. Thorsteinsson, Proc. Estonian Acad. Sci.: Geol. 51, 88 (2002). 356. T. Märss, M. V. H. Wilson, R. Thorsteinsson, Spec. Pap. Palaeont. 75, 1 (2006). 357. R. F. Miller, R. Cloutier, S. Turner, Nature 425, 501 (2003).

83

358. G. Maisey, M. E. Anderson, J. Vert. Paleont. 21, 702 (2001). 359. M. E. Anderson, J. E. Almond, F. J. Evans, J. A. Long, J. Afr. Earth Sci. 29, 179 (2001). 360. A. M. C. Duméril, Zoologie analytique, ou méthode naturelle de classification des animaux (Didot, Paris, 1806), pp. 344. 361. S. Løvtrup, The Phylogeny of the Vertebrata (Wiley, New York, 1977), pp. 330. 362. M. W. Hardisty, The Biology of Cyclostomes (Chapman & Hall, London, 1979), pp. 428. 363. M. W. Hardisty, in The Biology of Lampreys (Academic Press, 1982), pp. 165. 364. J. Mallatt, C. J. Winchell, Mol. Phylog. Evol. 43, 1005 (2007). 365. C. Delarbre, V. Barriel, P. Janvier, G. Gachelin, Mol. Phylog. Evol. 22, 184 (2002). 366. F. Delsuc, H. Brinkmann, D. Chourrout, H. Philippe, Nature 439, 965 (2006). 367. J. Mallatt, J.-y. Chen, J. Morph. 258, 1 (2003). 368. N. D. Holland, J.-Y. Chen, BioEssays 23, 142 (2001). 369. D.-G. Shu et al., Nature 402, 42 (1999). 370. D.-G. Shu et al., Nature 421, 526 (2003). 371. D.-G. Shu, Chinese Sci. Bull. 48, 725 (2003). 372. X. G. Hou, R. J. Aldridge, D. J. Siveter, X. H. Feng, Proc. Roy. Soc. Ser. B, Biol. Sci. 269, 1865 (2002). 373. P. C. J. Donoghue, P. L. Forey, R. J. Aldridge, Biol. Rev. 75, 191 (2000). 374. B.-D. Erdtmann, B. Weber, H.-P. Schultze, S. Egenhoff, J. Vert. Paleont. 20, 394 (2000). 375. G. C. Young, J. Vert. Paleont. 17, 1 (1997). 376. A. Ritchie, J. Gilbert-Tomlinson, Alcheringa 1, 351 (1977). 377. I. J. Sansom, P. C. J. Donoghue, G. L. Albanesi, Biol. Lett. 2, 446 (2005). 378. G. L. Albanesi, J. L. Benedetto, P.-Y. Gagnier, Bol. Acad. Nac. Cienc. Córdoba 60, 519 (1995). 379. R. A. Cooper, P. M. Sadler, in A Geologic Time Scale 2004, F. M. Gradstein, J. Ogg, A. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 165. 380. D.-G. Shu, L. Chen, J. Han, X.-L. Zhang, Nature 411, 472 (2001). 381. J. Y. Chen et al., Proc. Natl. Acad. Sci. U.S.A. 100, 8314 (2003). 382. X. G. Hou et al., The Cambrian Fossils of Chengjiang, China: The Flowering of Animal Life (Blackwell Science Ltd, London, 2004), pp. 233. 383. J. H. Shergold, R. A. Cooper, in A Geologic Timescale 2004, F. M. Gradstein, J. G. Ogg, A. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 147. 384. S. Conway Morris, Atlas of the Burgess Shale (Palaeontological Association, London, 1982), pp. 31. 385. D.-G. Shu, S. Conway Morris, X.-L. Zhang, Nature 384, 157 (1996).

84


386. J. Chen, J. Dzik, G. D. Edgecombe, L. Ramsköld, G.-Q. Zhou, Nature 377, 720 (1995). 387. P. C. J. Donoghue, M. A. Purnell, Trends Ecol. Evol. 20, 312 (2005). 388. H. Gee, in Major Events in Early Vertebrate Evolution: Palaeontology, Phylogeny, Genetics and Development, P. E. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 1. 389. D.-G. Shu, S. Conway Morris, J. Han, Z.-F. Zhang, J.-N. Liu, Nature 430, 422 (2004). 390. D.-G. Shu et al., Science 299, 1380 (2003). 391. D.-G. Shu et al., Nature 414, 419 (2001). 392. D.-G. Shu, S. Conway Morris, Science 300, 1372d (2003). 393. G. E. Budd, S. Jensen, Biol. Rev. 74, 253 (2000). 394. O. Varol, S. D. Houghton, J. Micropal. 15, 135 (1996). 395. T. C. Lacalli, BioEssays 24, 208 (2002). 396. A. B. Smith, Nature 430, 411 (2004). 397. D. E. G. Briggs, B. S. Lieberman, S. L. Halgedahl, R. D. Jarrard, Palaeontology 48, 681 (2005). 398. J. G. Gehling, Alcheringa 11, 337 (1987). 399. R. Mooi, Can. J. Zool. 79, 1209 (2001). 400. T. P. Crimes, Geol. Mag. 124, 97 (1987). 401. G. E. Budd, S. Jensen, in Telling the Evolutionary Time: Molecular Clocks and the Fossil Record, P. C. J. Donoghue, M. P. Smith, Eds. (Taylor & Francis, London, 2003). 402. V. V. Khomentovsky, A. K. Val’kov, G. A. Karlova, in Poddniy dokembriy i ranniy paleozoy Sibiri, V. V. Khomentovsky, A. S. Gibsher, Eds. (Voprosy regional’noy Stratigrafi Institut Geologiii i Geofiziki, Sibirskoe Otdelenie, Akademiya Nauk SSSR, Novosibirsk, 1990), pp. 3. 403. M. A. Fedonkin, B. M. Waggoner, Nature 388, 868 (1997). 404. D. Shu et al., Science 312, 731 (2006). 405. X. Yuan et al., Doushantuo Fossils: Life on the Eve of Animal Radiation (University of Science and Technology of China Press, 2002), pp. 171. 406. S. Xiao, B. Shen, C. Zhou, G. Xie, X. Yuan, Proc. Natl. Acad. Sci. U.S.A. 102, 10227 (2005). 407. J. Chen et al., Proc. Natl. Acad. Sci. U.S.A. 97, 4457 (2000). 408. J. Chen et al., Dev. Biol. 248, 182 (2002). 409. J. Chen et al., Science 305, 218 (2004). 410. J.-Y. Chen, P. Oliveri, E. Davidson, D. J. Bottjer, Science 306, 1291b (2004). 411. D. J. Bottjer, Sci. Am. 2005, 42 (2005). 412. S. Xiao, X. Yuan, A. H. Knoll, Proc. Natl. Acad. Sci. U.S.A. 97, 13684 (2000). 413. S. Bengtson, in The New Panorama of Animal Evolution, A. Legakis, S. Sfenthourakis, R. Polymeni, M. ThessalouLegaki, Eds. (Pensoft, Moscow, 2003), pp. 289. 414. S. Bengtson, G. Budd, Science 306, 1291a (2004). 415. E. C. Raff, J. T. Villinski, F. R. Turner, P. C. J. Donoghue, R. A. Raff, Proc. Natl. Acad. Sci. U.S.A. 103, 5846 (2006). 416. S. Xiao, Y. Zhang, A. H. Knoll, Nature 391, 553 (1998). 417. J. W. Hagadorn et al., Science 314, 291 (2006).

418. S. Xiao, J. W. Hagadorn, C. Zhou, X. Yuan, Geology 35, 115 (2007). 419. D. Condon et al., Science (2005). 420. A. H. Knoll, Precamb. Res. 100, 3 (2000). 421. M. Brasier, J. Antcliffe, Science 305, 1115 (2004). 422. B. L. Stinchcomb, N. A. Angeli, J. Paleont. 76, 965 (2002). 423. S. R. Westrop, A. R. Palmer, A. Runkel, J. Paleont. 79, 72 (2005). 424. P. J. Wagner, Smiths. Contr. Paleobiol. 88, 1 (2002). 425. D. M. Rohr, E. A. Measures, W. D. Boyce, I. Knight, Curr. Res., Newfoundland Dept. Mines Energy 2001, 113 (2001). 426. R. J. Ross et al., USGS Prof. Pap. 1579-A, 1 (1997). 427. C. Grande, J. Templado, J. L. Cervera, R. Zardoya, Mol. Biol. Evol. 21, 303 (2004). 428. C. Grandé, J. Templado, J. L. Cervera, R. Zardoya, Mol. Phylog. Evol. 33, 378 (2004). 429. B. Knudsen, A. B. Kohn, B. Nahir, C. S. McFadden, L. L. Moroz, Mol. Phylog. Evol. 38, 459 (2006). 430. T. H. Struck et al., BMC Evol. Biol. 7, 57 (2007). 431. V. Rousset, F. Pleijel, G. W. Rouse, C. Erseus, M. E. Siddall, Cladistics 23, 41 (2007). 432. D. G. Mikulic, D. E. G. Briggs, J. Kluessendorf, Science 228, 715 (1985). 433. D. G. Mikulic, D. E. G. Briggs, J. Klussendorf, Phil. Trans. Roy. Soc. Lond. Ser. B 311, 78 (1985). 434. H. Kozur, Lethaia 3, 225 (1970). 435. J. Kusta, Böhmisch Ges. Wissens., Math.-Naturw. Cl., Sitzungsber. 1887, 561 (1888). 436. G. W. Rouse, F. Pleijel, Polychaetes (Oxford University Press, Oxford, 2001), pp. 354. 437. D. Eibye-Jacobsen, Lethaia 37, 317 (2004). 438. O. Hints, M. E. Eriksson, Palaeogeog. Palaeoclim. Palaeoecol. 245, 95 (2007). 439. I. Thompson, Palaeontogr. Abt. A 163, 23 (1979). 440. K. Fitzhugh, S. D. Sroka, S. Kruty, M. D. Henderson, A. A. Hay, in Richardson’s Guide to the Fossil Fauna Of Mazon Creek, C. W. Shabica, A. A. Hay, Eds. (Northern Illinois University Press, Chicago, 1997), pp. 308. 441. R. A. Peppers, Geol. Soc. Am. Mem. 188, 1 (1996). 442. H. W. Pfefferkorn, in Mazon Creek Fossils, M. H. Nitecki, Ed. (Academic Press, New York, 1979), pp. 129. 443. R. H. Wagner, in Biostratigraphy: Compte Rendu, Neuvième Congrès International de Stratigraphie et de Géologie du Carbonifère, P. K. Sutherland, W. L. Manger, Eds. (Washington and Champaign-Urbana, 1984), Vol. 2, pp. 109. 444. N. J. Butterfield, Nature 369, 477 (1994). 445. I. Hinz, Palaeontogr. Abt. A 198, 41 (1987). 446. D. J. Siveter, M. Williams, D. Waloszek, Science 293, 479 (2001). 447. D. J. Siveter, D. Waloszek, M. Williams, in Trilobites and their Relatives (2003), Palaeontological Association Special Paper, pp. 9. 448. D. Grimaldi, M. S. Engel, Evolution of the Insects (Cambridge University Press, New York, 2005), pp. 755.


449. A. P. Rasnitsyn, in History of Insects, A. P. Rasnitsyn, D. L. J. Quicke, Eds. (Kluwer, Dordrecht, 2002), pp. 157. 450. M. S. Engel, D. A. Grimaldi, Nature 427, 627 (2004). 451. C. H. Wellman, Trans. Roy. Soc. Edinb.: Earth Sci. 97, 167 (2007). 452. M. R. House, F. M. Gradstein, in A Geologic Timescale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 202. 453. D. E. Shcherbakov, Y. A. Popov, in History of Insects, A. P. Rasnitsyn, D. L. J. Quicke, Eds. (Kluwer, Dordrecht, 2002), pp. 143. 454. J. Kukalová-Peck, R. Willmann, Can. J. Earth Sci. 27, 459 (1990). 455. D. E. Shcherbakov, Paleont. J. 34, S251 (2000). 456. J. Zajic, Cour. Forsch. Senck. 223, 563 (2000). 457. B. R. Wardlaw, V. Davydov, F. M. Gradstein, in A Geologic Timescale 2004, F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds. (Cambridge University Press, Cambridge, 2004), pp. 249. 458. O. Béthoux, A. Nel, J. Syst. Palaeont. 2, 285 (2005). 459. Z.-q. Lei, Acta Bot. Sin. 20, 229 (1978). 460. R. Willmann, Abd. Senck. Natur. Ges. 544, 1 (1989). 461. A. P. Rasnitsyn, Paleont. Zh. 1964, 88 (1964). 462. A. P. Rasnitsyn, Trudy Paleont. Inst. Akad. Nauk SSSR 123, 1 (1969). 463. E. F. Riek, Austr. J. Zool. 3, 654 (1955). 464. T. Schlüter, Paläont. Z. 74, 75 (2000). 465. I. A. Dobruskina, Stratigraficheskoye polozhenie triasovykh floronosnykh otlozheniy Evrazii (Stratigraphic position of Triassic plant-bearing beds of Eurasia). Trudy Geol. Inst. Akad. Nauk SSSR 346 (Nauka, Moscow, 1980), pp. 164. 466. I. A. Dobruskina, Triasovye flory Evrazii (Triassic floras of Eurasia). Trudy Geol. Inst. Akad. Nauk SSSR 365 (Nauka, Moscow, 1982), pp. 196. 467. A. P. Rasnitsyn, in History of Insects, A. P. Rasnitsyn, D. L. J. Quicke, Eds. (Kluwer, Dordrecht, 2002), pp. 1. 468. C. Ping, Palaeont. Sin. 13, 1 (1928). 469. A. P. Rasnitsyn, H. Zhang, Palaeontology 47, 1507 (2004). 470. Y. Q. Liu, Y. X. Liu, S. A. Ji, Z. Q. Yang, Chinese Sci. Bull. 51, 2634 (2006). 471. Y. X. Liu, Y. Q. Liu, H. Zhang, Acta Geol. Sin. 80, 733 (2006). 472. W. Krzeminski, E. Jarzembowski, Polski Pismo Entomol. 68, 445 (1999). 473. M. W. Hounslow, P. E. Posen, G. Warrington, Palaeogeog. Palaeoclim. Palaeoecol. 213, 331 (2004). 474. J. Ansorge, Polski Pismo Entomol. 68, 431 (1994). 475. W. Krzeminski, Mitt. Schweiz. Entomol. Ges. 65, 39 (1992). 476. W. Krzeminski, E. Krzeminski, Acta Zool. Cracov. 46 (Supplement), 153 (2003).

85

477. W. Krzeminski, N. L. Evenhuis, in Manual of Palaearctic Diptera. Volume 1. General and Applied Dipterology, L. Papp, B. Darvas, Eds. (Science Herald, Budapest, 2000), pp. 535. 478. F. Papier, A. Nel, L. Grauvogel-Stamm, J.-C. Gall, Geodiversitas 27, 181 (2005). 479. F. Papier, L. Grauvogel-Stamm, Palaeontogr. Abt. A 235, 141 (1995). 480. W. C. Sweet, S. M. Bergström, Ann. Rev. Earth Planet. Sci. 14, 85 (1986). 481. J. Ansorge, W. Krzeminski, Acta Zool. Cracov. 37, 115 (1994). 482. W. Krzeminski, J. Ansorge, Polski Pismo Entomol. 69, 231 (2000). 483. W. Krzeminski, E. Krzeminski, F. Papier, Acta Zool. Cracov. 37, 95 (1994). 484. D. Grazhdankin, G. Gerdes, Lethaia 40, 201 (2007). 485. S. Jensen, J. G. Gehling, M. L. Droser, S. W. F. Grant, Lethaia 35, 291 (2002). 486. J. B. Antcliffe, M. D. Brasier, J. Geol. Soc. Lond. 164, 49 (2007). 487. J. B. Antcliffe, M. D. Brasier, Palaeontology 51, 11 (2008). 488. J. P. Grotzinger, W. A. Watters, A. H. Knoll, Paleobiology 26, 334 (2000). 489. W. A. Watters, J. P. Grotzinger, Paleobiology 27, 159 (2001). 490. R. A. Wood, J. P. Grotzinger, J. A. D. Dickson, Science 296, 2383 (2002). 491. A. Kouchinsky, S. Bengtson, Acta Palaeont. Pol. 47, 431 (2002). 492. S. Bengtson, Z. Yue, Science 277, 1645 (1997). 493. Z. Yue, S. Bengtson, Lethaia 32, 181 (1999). 494. M. Hicks, J. Paleont. 80, 609 (2006). 495. F. M. Debrenne, J. Reitner, in The Ecology of the Cambrian Radiation, A. Y. Zhuravlev, R. Riding, Eds. (Columbia University Press, New York, 2001), pp. 301. 496. F. M. DeBrenne, J. LaFuste, A. Zhuravlev, Bull. Mus. Nat. Hist. Nat. Paris 12, 17 (1990). 497. A. G. Collins et al., Syst. Biol. 55, 97 (2006). 498. P. Cartwright et al., PLoS ONE 2, e1121 (2007). 499. K. J. Peterson, D. J. Eernisse, Evol. Dev. 3, 170 (2001). 500. J. Zrzavý, S. Mihulka, P. Kepka, A. Bezdek, D. Tietz, Cladistics 14, 249 (1998). 501. E. A. Sperling, D. Pisani, K. J. Peterson, J. Geol. Soc. Lond. Special Publication, 286, 355 (2008). 502. A. Wallberg, M. Thollesson, J. S. Farris, U. Jondelius, Cladistics 20, 558 (2004). 503. T. Cavalier-Smith, E. E.-Y. Chao, J. Mol. Evol. 56, 540 (2003). 504. C. Borchiellini et al., J. Evol. Biol. 14, 171 (2001). 505. M. Medina, A. G. Collins, J. D. Silberman, M. L. Sogin, Proc. Natl. Acad. Sci. U.S.A. 98, 9707 (2001). 506. C. Li, J. Y. Chen, T. E. Hua, Science 279, 879 (1998).

86


507. S. M. Rowland, J. Paleont. 75, 1065 (2001). 508. S. Bengtson, X. Hou, Acta Palaeont. Pol. 46, 1 (2001). 509. S. Bengtson, in Evolving Form And Function: Fossils and Development, D. E. G. Briggs, Ed. (Yale Peabody Museum, New Haven, 2005), pp. 101. 510. J. G. Gehling, J. K. Rigby, J. Paleont. 70, 185 (1996). 511. G. D. Love et al., Goldschmidt Conf. Abstr. A371 (2006). 512. J. W. Schopf, J. Paleont. 42, 651 (1968). 513. N. J. Butterfield, A. H. Knoll, K. Swett, Fossils Strata 34, 1 (1994).

514. G. P. Halverson, P. F. Hoff man, D. P. Schrag, A. C. Maloof, A. H. N. Rice, Geol. Soc. Am. Bull. 117, 1181 (2005). 515. G. P. Halverson, A. C. Maloof, D. P. Schrag, F. Ö. Dudás, M. Hurtgen, Chem. Geol. 237, 5 (2007). 516. A. C. Hill, K. L. Cotter, K. Grey, Precamb. Res. 100, 281 (2000). 517. A. C. Hill, M. R. Walter, Precamb. Res. 100, 181 (2000). 518. M. T. D. Wingate, I. H. Campbell, W. Compston, G. M. Gibson, Precamb. Res. 87, 135 (1998).

TIMETREES


Life S. Blair Hedges Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA ([email protected])

Abstract Life on Earth arose from a single source, ~4400–4200 million years ago (Ma), and quickly achieved a prokaryotic level of complexity. An initial split (~4200 Ma) led to the Superkingdoms Eubacteria and Archaebacteria. Theories for eukaryote origins fall into two classes, merger and deeproot models, with the former having broadest support. Under the two-merger model, an archaebacterium joined a eubacterium (possibly ~2700 Ma) to form the nucleus and hence first eukaryote. A subsequent merger with another eubacterium (~2000 Ma), an alphaproteobacterium, formed the mitochondrion. The one-merger model postulates that both the nucleus and mitochondrion formed at about the same time (~2000 Ma).

This is the earliest and most controversial portion of the tree of life and few details can be regarded as well established. It involves the relationships and times of origin of the three superkingdoms, Eubacteria, Archaebacteria, and Eukaryota. But the key to understanding how these earliest events unraveled is to know how eukaryotes arose and their relationship to prokaryotes. Despite the availability of hundreds of completely sequenced genomes from prokaryotes and dozens from eukaryotes, the answers are not yet in hand. New models for the origin of eukaryotes appear frequently. For ease of discussion they are classified here as merger models (1–11) and deep-root models (12, 13). The former ascribe the origin of eukaryotes to a merger between two prokaryotes whereas the latter minimize the role of mergers and instead postulate an ancient origin of eukaryotes, at least as old as the earliest divergences among living prokaryotes. The goal of this brief synopsis is to review the evidence bearing on the earliest aspects of the timetree of life. Details concerning differences among the many proposed models, as well as different points of view, can be found elsewhere (2, 14, 15). Life on Earth encompasses an estimated ~1.8 million described species (16) and a much larger number is

thought to be undescribed (>10 million species). Almost all of the described species are eukaryotes, and most of those are arthropods. Of prokaryotes, there are ~9400 recognized species of eubacteria and ~300 of archaebacteria, based on the latest compilations (17, 145). Some organisms have been found as deep as ~800 m below the ocean floor in subsurface sediments (18) and others (e.g., bacterial spores) have been found as high as 41 km above sea level in the atmosphere (19). The most abundant organism is probably a ubiquitous marine eubacterium (20) (Fig. 1). Different authors use different names for the three major groups, and this requires an explanation. When it was first recognized that the methanogens and relatives formed a distinct group they were given the name archaebacteria, with the remainder of prokaryotes named eubacteria (21). These were appropriate names because they were given the stem (“bacteria”) which indicated that they were both comprised of prokaryotes. Later they were renamed archaea and bacteria, respectively (22) “to avoid any connotation that eubacteria and archaebacteria are related to one another.” This was done because

Fig. 1 Cryo-electron tomographic image of a marine eubacterium, Pelagibacter ubique (alphaproteobacteria), one of the smallest self-replicating cells (1354 genes) and most abundant of organisms (20). Dimensions of the cell are ~900 × 280 nm. Credits: D. Nicastro, Brandeis University, and J. R. McIntosh; cell from S. Giovannoni.

S. B. Hedges. Life. Pp. 89–98 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

90

THE TIMETREE OF LIFE A 2

Archaebacteria Eukaryotes

3

1

4

Eubacteria

B 2

Archaebacteria Eukaryotes

1

4

HD

Ea

4000

Pa Ma Na ARCHEAN 3000

Eubacteria

Pp Np Pz Mp PH PROTEROZOIC 2000

1000

0 Million years ago

Fig. 2 Two versions of the timetree of life based on competing merger models for the origin of eukaryotes. Times of divergence are from Table 1. (A) The two-merger model. Node 1 is the divergence of eubacteria and archaebacteria (LUCA, the last universal common ancestor). Node 2 is the divergence of two types of archaebacteria, one eventually leading to the origin of eukaryotes. Node 3 is the origin of eukaryotes and represents a merger between an archaebacterium and a eubacterium that led to the eukaryote nucleus and possibly the incorporation of eubacterial genes in the genome of eukaryotes. This

pre-mitochondrial event is not yet well established (see text). Node 4 is the symbiotic event that led to the mitochondrion of eukaryotes and the transfer of genes from the eubacterial symbiont (an alphaproteobacterium) to the nuclear genome of eukaryotes. (B) The one-merger model. This is identical to the two-merger model except that the formation of the nucleus and mitochondrion are combined into a single step. Abbreviations: Ea (Eoarchean), HD (Hadean), Ma (Mesoarchean), Mp (Mesoproterozoic), Na (Neoarchean), Np (Neoproterozoic), Pa (Paleoarchean), PH (Phanerozoic), Pp (Paleoproterozoic), and Pz (Paleozoic).

archaebacteria were shown to cluster with eukaryotes in the small subunit (SSU) ribosomal RNA (rRNA) tree (22) rooted by emerging gene duplication evidence (see later). However, in the last two decades, analyses of complete genomes have contradicted that interpretation, leading to the general—although not unanimous (23)—view that eukaryotes are cytological and genomic chimeras of prokaryotes (2, 14, 15), which will be discussed at length later. For this reason, many evolutionary biologists use the first proposed names, archaebacteria and eubacteria (e.g., 14, 24–29). There is also a tradition in taxonomy that the first proposed valid names should be the ones used, even in case of informal higher-level classification. A separate reason for avoiding the name “bacteria” is that it is identical to the widely used common name for all prokaryotes, thus creating confusion. Nonetheless, there are no rules preventing anyone from using any of these names. Related to the renaming of the superkingdoms is a parallel debate over the word “prokaryote,” with advocates of the “rRNA tree of life” arguing that it should

be abandoned for the same reasons, and replaced by “microbe” (141, 142). Others reject those criticisms, claiming that the word prokaryote has useful biological meaning (e.g., 29, 143). They also correctly note that, even if replacement were justified, “microbe” would be a poor alternate because many eukaryotes are microbes. Without examples of life from other worlds, it is not possible to say from a comparative standpoint that life on Earth arose from a single source (i.e., monophyletic). Parallel evolutionary pathways (convergence) can occur at all stages of biochemical evolution. However, the great similarity in the genomes of all organisms on Earth suggests a single origin. Was that single origin from Earth or elsewhere? The dynamics of planetary ejecta show that it is very unlikely that life on Earth was seeded from another solar system, although the vehicle for transport (planetary ejecta) was abundantly available in our own solar system (30). Venus and Mars have been discussed as possible sources, although the latter more frequently. Based on the physical conditions of a Mars (or Venus) to Earth transfer, survival of some cells would have been

Life

91

Table 1. Divergence times (Ma) among major groups of life. Timetree Node

Time

Estimates Ref. (56)

Ref. (132)(a)

Ref. (132)(b)

Time

Time

Ref. (60) CI

Ref. (85) Time

Time

Ref. (99) CI

Time

Time

>3970

–

–

>4112

–

–

–

1

4200

3784

2

3806

2409

3970

4597–3343

–

3806

4486–2900

–

–

3

2730

–

2730

3122–2338

–

–

–

–

–

4

2000

1961

1570

>2020

Note: See text for details. Only multigene studies are shown. Times with > and < symbols pertain to time estimates other than the node in question, but which help constrain the nodal time. For Node 2 (3806 Ma), the time is the midpoint of the two constraining nodes, the Crenarchaeota/Euryarchaeota divergence of 4112 Ma (4486–3314 Ma) and earliest split among crenarchaeotans, 3500 Ma (3839–2900 Ma), with the CI (4486–2900) being derived from the CIs of those two nodes.

possible (31); yet the combination of radiation (32), heat (33), and impact shock (34) would have greatly reduced the fraction of such cells that survived. Therefore, all else equal, the probability that Earth life arose on Earth is much more likely. Nonetheless, it remains possible that conditions for the origin of life—in general—were more favorable on our neighboring planets than on early Earth, and (or) that the conditions for the origin of life occurred earlier on those planets, providing a lead time. Either or both of those factors could have made it more likely that Earth life arose on those planets. However, until the early history of Mars and Venus are better understood, and the necessity of having a lead time is established, the probability that Earth life arose on Earth is more parsimonious. Rooting the tree of life is critical for evolutionary interpretations. Most illustrations of this that appear in the scientific and popular literature—and even in textbooks—show an unrooted SSU rRNA tree, claiming or inferring the existence of three groups (archaebacteria, eubacteria, and eukaryotes). However, this is incorrect and misleading because an unrooted tree has no evolutionary direction and therefore no evolutionary groups can be inferred from such a diagram. For example, if the root were in the middle of archaebacteria, then archaebacteria would not be a natural (monophyletic) group. Initially, the use of duplicated genes suggested a root between the Superkingdoms Eubacteria and Archaebacteria (35, 36). Most discussions since 1990 have assumed that root to be correct, but it is not universally accepted. For example, some have proposed that the root is between the sulfur and non-sulfur green bacteria (37, 38) while others have suggested that it lies within a phylum (Firmicutes) of eubacteria based on insertion–deletion (indel) events

(39–41). Still others have presented evidence for a root within archaebacteria (42). In defense of the green bacteria root, it has been argued that duplicate gene rooting is problematic because of sequence rate variation (37). However, gene content phylogenies (43), presumably less susceptible to sequence rate variation—because they use the presence or absence of genes as characters and not sequence data—are similar to sequence phylogenies and not to the green bacteria/ Neomura tree (37, 38); other problems with the Neomura model are discussed later. The rooting by indels also has received criticism, mainly concerning the alignments (42, 44). Recently, the rooting by gene duplications was revisited with a bioinformatics approach and a broader survey of prokaryote taxa (45). The majority of gene data sets supported a root between eubacteria and archaebacteria. This node in the tree is also referred to as the cenancestor, the most recent common ancestor (MRCA), or more commonly the last universal common ancestor (LUCA). While the question of the root remains an active area of study, the current consensus is that it lies between archaebacteria and eubacteria. That the eukaryote cell arose from prokaryote cells through mergers (symbioses or fusions) has been a working hypothesis since it was elaborated decades ago (46) from ideas put forth in the nineteenth century. Abundant evidence from genetics has supported cell biology in this regard, demonstrating that mitochondria (47) and plastids (48, 49) are descendants of eubacterial endosymbionts, the former of an alphaproteobacterium and the latter of a cyanobacterium. However, the discovery of large-scale transfers of symbiont genes to the eukaryote nucleus was not predicted by cell biology. Endosymbiotic gene transfer differs from more typical

92


horizontal gene transfer (HGT) in that it involves large numbers of genes rather than one or a few. When evidence for this began to appear in the 1980s and early 1990s (50, 51), a new view of the tree of life emerged, replacing—in the eyes of evolutionary biologists but not necessarily microbiologists—the existing concept based on analysis of the SSU rRNA gene (22). By the mid-1990s, the truly hybrid nature of the eukaryote genome became widely recognized (3, 52–58). Besides acknowledging the mixing of genes from different sources, it was soon realized that some genes— particularly those involved in information transfer (replication, transcription, and translation)—arose from an archaebacterial ancestor and that other genes—those with a metabolic function—arose from a eubacterial ancestor. From the onset, there has been good evidence that both contributions were substantial, involving at least hundreds of genes (53, 58–60). It has been estimated that ~80% of eukaryote genes that are not eukaryote-specific came from eubacteria, with the remaining 20% coming from archaebacteria (25, 61). This also argues against the notion of the rRNA tree as being the tree of life. Attention has turned in recent years to testing two general types of merger models—each with multiple versions—and both types involving a merger between an archaebacterium and a eubacterium. They are referred to here as the two-merger and one-merger models, based on their major difference. The two-merger model—that is, nucleus first and mitochondrion second—was the first to be proposed. In one version based on cell biology, the first event joined an archaebacterium with a spirochete (1, 2), the latter partner providing cell motility. This led to a nucleated cell and hence the first eukaryote, but one that still lacked a mitochondrion. Other versions of the two-merger model have been proposed, involving different combinations of prokaryote partners in the initial merger (3–7). Although not a requirement of the twomerger model, it raises the possibility that some living eukaryotes are primarily amitochondriate (“archaezoa”) (62–64). A derivative of the two-merger model that seems equally possible is a multimerger model, whereby a series of three or more mergers contributed different components of the eukaryote cell, and perhaps different sets of genes in the eukaryote nuclear genome. In contrast, the one-merger model of eukaryote origins suggests that the nucleus and mitochondrion originated at about the same time. Different versions of this model have been articulated as well (9–11, 65, 66), although all involve an archaebacterial host and eubacterial symbiont. The survival of the one-merger model in

any form requires that no living eukaryotes exist that are primitively amitochondriate species (i.e., diverged from an early eukaryote lineage before the origin of the mitochondrion) rather than secondarily amitochondriate (i.e., those whose ancestors possessed a mitochondrion). This requirement led to an intense search for evidence during the last decade of any mitochondrial ancestry among living amitochondriate eukaryotes. The efforts bore fruit in that organelles—hydrogenosomes and mitosomes— believed to be relicts of mitochondria were discovered in several amitochondriate species (15, 67). They share with mitochondria a similar protein import system and iron– sulfur cluster assembly (15, 68), and the mechanisms by which they function in such a reduced state are becoming better known (69). Besides these merger models, at least two other models have been proposed that are referred to here as “deeproot” models because they postulate an ancient origin for eukaryotes, dating to the LUCA (12, 23, 70, 71). While deep-root models acknowledge the existence of mergers (e.g., the origin of mitochondria), they consider them to be less important for the definition of eukaryotes than the genes and cell components inherited from a much earlier eukaryote ancestor. A particularly controversial aspect of these models is the claim that prokaryotes are reduced versions of an ancestral state (~LUCA) that resembled a eukaryote. The existence of many eukaryote-specific proteins, not related to either eubacteria or archaebacteria, has been considered primary evidence (12, 71). However, criticism of the deep-root models has centered on this evidence by pointing out that eukaryote-specific structures and proteins which show no relationship to prokaryotes should not be used to infer ancient relationship (72, 73). The proponents of the models replied by noting that eukaryote-specific proteins are found throughout the cell of even parasites with reduced genomes, indicating their importance (13, 74). Although it is an intriguing possibility that eukaryotes are ancient and evolved before prokaryotes, more evidence will be needed before the deep-root models are considered as serious challengers of the merger models. The diplomonad Giardia lamblia has been viewed as the most deeply branching of all eukaryotes (51) and has generally resisted stringent efforts to find a higher place for it in the tree, as was found for microsporidia which also lacks a mitochondrion (75, 76). Early analyses of several genes suggested that it once harbored a mitochondrion (77, 78). Recently, its genome was sequenced and analyzed (79). Although Giardia have mitosomes, it may be difficult to prove beyond doubt that those organelles

Life

are relic mitochondria, since the mitosome lacks a genome and it may have arisen from another symbiotic event with an aerobic eubacterium, hence confounding what could be called a mitochondrial character. The latest multigene evolutionary tree showing the position of Giardia (79) supports an early branching, before the split of plants and animals. That analysis also showed that Giardia has few if any genes linked to the mitochondrial symbiotic event. The authors concluded that “a parsimonious explanation of this pattern is that Giardia never had any components of what may be considered ‘eukaryotic machinery,’ not that it had and lost them through genome reduction as is evident for Encephalitozoon. Taking a whole-evidence approach, one sees that these data reflect early divergence, not a derived genome” (79). However, the tree of eukaryotes is far from resolved (80–84) and more evidence will be needed before conclusions can be drawn regarding the position of Giardia. One earlier multigene study supported the deep-branching of Giardia (85) while another did not (86). A recent study claimed that the root of the tree had been resolved (87) based on molecular characters, but those characters were missing from groups central to the debate over the root, such as diplomonads and kinetoplastids. The most taxon-rich multigene study (84) showed a lack of resolution for many clades thought to be monophyletic, demonstrating that that the root of the tree—and hence position of Giardia—remains an open question (88). As for whether it is a primitively or secondarily amitochondriate eukaryote, the current weight of the evidence (discussed earlier) argues for the latter. The two questions are related, but not firmly connected, and therefore Giardia could very well be a deeply branching but secondarily amitochondriate eukaryote. Although the status of Giardia as a primitively or secondarily amitochondriate eukaryote is crucial for the onemerger model, it is not crucial for the two-merger model because descendants of the pre-mitochondrial eukaryote stage may have become extinct like many major lineages in the history of life (74, 89), or remain undiscovered. The existence of non-phagotrophic intracellular symbioses involving two species of bacteria (90, 91) supports both models because each requires a merger of two prokaryotes in the initial formation of the eukaryote cell. A completely separate question is whether there is evidence in the genomes of eukaryotes that traces to a pre-mitochondrial event, regardless of whether a living eukaryote exists that is primarily amitochondriate. Molecular evidence of a possible pre-mitochondrial event came with an early bioinformatics analysis of complete

93

genomes (60). In that study, contributions of alphaproteobacterial and archaebacterial genes in eukaryotes were identified by phylogenetic analysis, leaving a substantial contribution from eubacteria other than alphaproteobacteria or cyanobacteria. Based on divergence time (see later), their origin preceded that of the alphaproteobacterial genes, but a specific close relative among eubacteria could not be identified (60). Later analyses with larger numbers of genomes continued to discern a second category of eubacterial genes in the eukaryote genomes that were not of alphaproteobacterial or cyanobacterial origin (25, 61, 92). However, phylogenetic analysis is complex and susceptible to substitutional biases such as lineage-specific rate differences leading to long-branch attraction (93) as well as differential base composition and site-specific rate differences (94). Although the complex models currently used in phylogenetic analysis can account for some of these biases, it is possible that the apparent pre-mitochondrial signal in these bioinformatics analyses is the result of such biases. A separate potential source of bias is HGT, which can have a blending effect on genomes of prokaryotes (95–97) and may make it difficult to distinguish the source of a eukaryote gene among eubacterial lineages (98). One strategy for avoiding this bias has been to use core genes, which are those that show little or no evidence of HGT (e.g., 60, 81, 99). The number of such core genes typically used in studies involving all three superkingdoms is small (20–40), mainly for practical reasons of orthology determination and avoidance of missing data, not because 99% of the genome has undergone HGT (140); a large fraction of any genome is made up of genes unique to a branch of the tree of life, leaving a smaller number shared among all genomes. HGT has yet to be rigorously quantified, globally, but the fact that phyla and classes are well defined structurally, corresponding to groups identified in molecular phylogenies (e.g., 17, 145), suggests that more than 1% of prokaryote evolution is vertical. It is also possible that the origin of the nucleus did occur first (e.g., a two-merger model) but that few if any genes from the initial eubacterial partner (or partners) were incorporated in the eukaryote nuclear genome, which would make distinguishing this model from the one-merger model more difficult. Nonetheless, genomic data and analyses should eventually help to discriminate between the two-merger and one-merger models for the origin of eukaryotes, and possibly among different variants of each model. The identity of the archaebacterial partner in the merger that created eukaryotes is another unresolved

94


question. Different closest relatives have been proposed (2, 3, 6, 11, 59), but phylogenies of archaebacterial genes in eukaryotes have yielded conflicting results (25, 60, 99). A study of 32 core proteins significantly supported a close relationship between archaebacterial-like genes in eukaryotes and one of the two major subgroups of archaebacteria, Crenarchaeota (99). However, the potential impact of substitutional biases and long-branch attraction with such highly divergence sequences cannot be ruled out, especially given known rate differences (60, 100), and therefore this question deserves continued scrutiny with additional taxa and genes. If the archaebacterial partner turns out to be even more deeply branching, and is the closest relative of all archaebacteria, what should it be called? The proponents of deep-root models (12, 13) would call it a eukaryote whereas the proponents of merger models would call it a prokaryote. In the latter case, it could be placed in archaebacteria or a new superkingdom could be erected for it. Because essentially all of the cytological and environmental arguments for its identity, based on merger models, argue that it is an archaebacterium (1–3, 11), the most appropriate classification of this organism would be within archaebacteria, whether or not it is the closest relative of archaebacteria or related to one living lineage. The fossil record and biosignatures in the geologic record offer some clues as to the timescale of life, pertinent to these early divergences. The earliest eukaryotes in the fossil record are dated to ~1850 Ma (101) but are not taxonomically assignable to living groups. Recently, paleontologists have debated the possibility that the earliest eukaryote fossils, from the Paleoproterozoic (2500–1600 Ma), are prokaryotes (102–105). As was pointed out earlier (85), the molecular clock date for the origin of plastids, ~1600–1400 Ma, also supports the interpretation of those Paleoproterozoic fossils as prokaryotes rather than photosynthetic eukaryotic algae, unless they acquired photosynthetic abilities through independent (earlier) symbiotic events. The presence of steranes in much older sedimentary rocks, ~2700 Ma, from Pilbara, Australia, has been argued to be a biosignature of photosynthetic eukaryotic algae (106). While molecular clock data are not in conflict with the presence of eukaryotes at that time (60), the much later origin of plastids, ~1600–1400 Ma (85, 107, 108) again argues against the presence of eukaryotic algae at an earlier time. It is more likely that this sterane biosignature was either produced by prokaryotes or infiltrated the rock at some later time. The hopane biosignature from the same rocks, originally proposed

for cyanobacteria (109), has turned up in other groups recently (110), thus removing it as a unique biosignature of cyanobacteria. The first evidence for eubacteria (photosynthesismediated sediment deposition) is at 3400 Ma (111–113) and the first evidence for archaebacteria (methanogenesis, based on isotopically light carbon) is at 3460 Ma (114, 115). All fossil evidence for the earliest life concerns eubacteria, and it has been scrutinized heavily in recent years (116–120). While there is not complete agreement on details, there is more-or-less agreement that some fossil evidence of life exists in rocks deposited 3500 Ma (117, 118). Taken together, this evidence constrains the LUCA to be >3500 Ma. In contrast, the proponent of the Neomura model (37, 38, 121) has argued that archaebacteria is no older than 900–850 Ma, which is 2600 million years younger than the geologic evidence just noted. That claim is tied closely to a preferred phylogeny which nests archaebacteria and eukaryotes high up in the tree of eubacteria, within the phylum Actinobacteria (38). An early origin of methanogens (e.g., 3460 Ma) would create great problems for such a phylogenetic tree because it would force eukaryotes and virtually all eubacterial phyla to be older than 3500 Ma, something contradicted by too many other lines of evidence. Arguments against this model have been made elsewhere (122, 140), but some additional comments are required. First, it is true that inorganic processes can produce isotopically light carbon under certain conditions, and the earliest evidence of methanogenesis has been debated for this reason (123). However, the more abundant isotopic evidence for methanogenesis (hence archaebacteria) at ~2700 Ma (124, 125) is widely accepted (126). Secondly, claiming that the Mesoproterozoic fossils of eukaryotes must be “large and complex prokaryotes” (121) contradicts the opinion of all of the paleontologists who have been studying them for years (e.g., 102, 105, 127, 128), and the fact at least one such fossil—at 1200 Ma—is uncontroversially assigned to a specific lineage of eukaryotes, red algae (102). Thirdly, all molecular clock analyses (see details later) that have timed the origin of archabacteria have found old (>2400 Ma)—not young— times for the group, and those analyses have used different calibrations and methods. Thus, the Neomura model, and its requisite Neoproterozoic (3970 Ma. Time estimates from the study of Sheridan et al. (131), based on DNA sequences of the SSU rRNA gene, were problematic because they did not account for lineagespecific rate variation and used uncorrected distances. Other studies provide minimum and maximum time constraints for three of the four divergences (Table 1). In one (85) an estimate of 1961 Ma was obtained for

95

the earliest divergence among mitochondriate eukaryotes, based on analyses of 99 proteins and a diversity of methods. This would constrain the mitochondrial merger event to be older than that time. A second study that focused primarily on divergences among prokaryotes used sequences of 32 core proteins and 78 species from complete genomes, a Bayesian timing method, and fossil and geologic calibrations (99). The oldest date obtained, 4112 Ma (crenarchaeotans vs. euryarchaeotans), constrained the LUCA to be at least as old. In that study, the archaebacterial partner of eukaryotes was found to be the closest relative of Crenarchaeota, phylogenetically (it was not timed). Thus, the split could be constrained between 4112 and 3500 Ma (earliest divergence among crenarchaeotans). The divergence of alphaproteobacteria from other eubacteria was 2508 Ma, thus constraining the origin of mitochondria to be younger. Similar times were obtained for these same divergences among prokaryotes in analyses involving more species and fewer proteins (17, 145). In a study focused on plastid origins (132) the time of the earliest divergence among 17 diverse mitochondriate eukaryotes, using sequences of 40 proteins, was estimated as either 1570 or 2020 Ma, depending on the root. Yet another study used 129 proteins to estimate divergences among eukaryotes, resulting in a relatively young time of 1085 Ma for the earliest split (133). However, two reanalyses of that data set found methodological concerns related to calibrations (78, 134), and a separate reanalysis (see discussion in other chapters in this book, on Eukaryotes and Animals) found calibration errors. When those errors were corrected, the earliest split among mitochondriate eukaryotes was estimated as 1857 and 2216 Ma, depending on whether the root was Dictyostelium or a kinetoplastid, respectively. A timing analysis of eukaryotes, emphasizing protists and using DNA sequences of the SSU rRNA gene (135), resulted in a relatively young time of 1126 Ma (1357–948 Ma) for the earliest split among mitochondriate eukaryotes. However, another author (136) considered those dates to be underestimates resulting from the use of incorrect fossil calibration dates. A summary timetree of life (Fig. 2) presents scenarios for the two competing merger models. The date of the LUCA (4200 Ma) is approximately equal to both the minimum constraint from molecular clock studies (Table 1) and the maximum constraint from the last ocean-boiling impact event (137), while acknowledging that maximum constraints were imposed as calibrations in the studies timing prokaryote evolution (17, 99).

96


Presumably, if life evolved before the last ocean-boiling impact, it would have been annihilated. However, the date of the last ocean-boiling event is a statistical estimate with a steeply declining probability through the interval of 4440–3800 Ma (137), and therefore the last ocean-boiling impact may have occurred as much as 100–200 million years before 4200 Ma (or even later than that date). Also, life may have survived one or more ocean-boiling impacts by inhabiting the Earth’s subsurface. The early divergence of hyperthermophiles in even the most recent genome-based analyses (17) is consistent with either a high constant surface temperature at that time (138), or selection based on ocean-boiling impact events (139, 144). Nonetheless, considering the closeness of the geologic and molecular constraints on the time estimate for the LUCA, the implication is that life arose and evolved to a level of genomic and cellular complexity, comparable to living prokaryotes, in a relatively short period of time— probably less than 200 million years—early in Earth’s history. However, based on our current limited knowledge of the origin of life, such a short interval, even if only tens of millions of years, would not require that life originated at an earlier time elsewhere (e.g., on Mars) and was transported to Earth by impact ejecta. An early divergence within archaebacteria (~3800 Ma) presumably reflects nothing more than a split between two types of archaebacteria, one of which no longer survives except as the contributor of some genes in eukaryotes. Although this time is based on a close relationship with crenarchaeotans (99), the other hypotheses for the identity of the archaebacterial partner (see earlier) would yield similar times because the major clades of archaebacteria all branch deeply in the tree (99). In both merger models, the origin of eukaryotes occurred during the midlife of Earth, the late Archean and early Proterozoic (~2700–2000 Ma). In the two-merger model (Fig. 2A), the initial merger occurred ~2700 Ma followed by the mitochondrial symbiotic event (~2000 Ma) (Table 1). The timetree date for the mitochondrial event is approximate and reflects the only direct time estimate (1840 Ma) and the more recent estimates that help to constrain that estimate, mostly between ~2300 and ~2000 Ma (Table 1). These dates are consistent with the earliest undisputed eukaryote fossils at ~1600 Ma discussed earlier. In the one-merger model, the origin of eukaryotes is synonymous with the origin of mitochondria (~2000 Ma). In summary, universal agreement has not been reached on any aspect of the tree of early life, except perhaps the ancestry of the mitochondrion being from

alphaproteobacteria. Nonetheless, the most widely accepted models all involve mergers between an archaebacterium and a eubacterium and most of the current debate concerns whether this occurred in two steps or in one step. Deep-root models require more evidence before they can be considered strong competitors with merger models, but aspects of all models can be tested with genome sequence data, a continually growing resource. Recent advances in our knowledge of Earth history and the record of biosignatures have helped to constrain the timescale of the tree of life, and this has been further enhanced by molecular clocks, but much additional work is needed to estimate a robust timetree of life.

Acknowledgments The Penn State Astrobiology Research Center provided intellectual atmosphere, M. Embley commented on an earlier draft, and support was provided by the U.S. National Science Foundation and National Aeronautics and Space Administration (NASA Astrobiology Institute).

References 1. L. Margulis, Proc. Natl. Acad. Sci. U.S.A. 93, 1071 (1996). 2. L. Margulis, M. Chapman, R. Guerrero, J. Hall, Proc. Natl. Acad. Sci. U.S.A. 103, 13080 (2006). 3. R. S. Gupta, Microbiol. Mol. Biol. Rev. 62, 1435 (1998). 4. D. Moreira, P. Lopez-Garcia, J. Mol. Evol. 47, 517 (1998). 5. P. Lopez-Garcia, D. Moreira, BioEssays 28, 525 (2006). 6. J. A. Lake, M. C. Rivera, Proc. Natl. Acad. Sci. U.S.A. 91, 2880 (1994). 7. T. Horiike, K. Hamada, T. Shinozawa, Genes Genet. Syst. 77, 369 (2002). 8. T. Cavalier-Smith, Int. J. Syst. Evol. Microbiol. 52, 297 (2002). 9. D. G. Searcy, in The Origin and Evolution of the Cell, H. H. Matsuno, K. Matsuno, Eds. (World Scientific, Singapore, 1992), pp. 47–78. 10. T. Vellai, G. Vida, Proc. Roy. Soc. Lond. B 266, 1571 (1999). 11. W. Martin, M. Müller, Nature 392, 37 (1998). 12. H. Hartman, A. Fedorov, Proc. Natl. Acad. Sci. U.S.A. 99, 1420 (2002). 13. C. G. Kurland, L. J. Collins, D. Penny, Science 316, 543 (2007). 14. C. de Duve, Nat. Rev. Genet. 8, 395 (2007). 15. T. M. Embley, W. Martin, Nature 440, 623 (2006). 16. P. M. Hammond et al., in Global Biodiversity Assessment, V. H. Heywood, Ed. (Cambridge University Press, Cambridge, UK, 1995), pp. 113–138. 17. F. U. Battistuzzi, S. B. Hedges, in The Timetree of Life (Oxford University Press, New York, 2009), pp. 106–115.

Life

18. B. Taylor, P. Huchon, A. Klaus, Proc. ODP Init. Rep. 180, 4 (1999). 19. M. Wainwright, N. C. Wickramasinghe, J. V. Narlikar, P. Rajaratnam, FEMS Microbiol. Lett. 218, 161 (2003). 20. R. M. Morris et al., Nature 420, 806 (2002). 21. C. R. Woese, G. E. Fox, Proc. Natl. Acad. Sci. U.S.A. 74, 5088 (1977). 22. C. R. Woese, O. Kandler, M. L. Wheelis, Proc. Natl. Acad. Sci. U.S.A. 87, 4576 (1990). 23. A. Poole, D. Penny, Nature 447, 913 (2007). 24. T. Cavalier-Smith, Nature 446, 257 (2007). 25. D. Pisani, J. A. Cotton, J. O. McInerney, Mol. Biol. Evol. 24, 1752 (2007). 26. S. B. Hedges, Nat. Rev. Genet. 3, 838 (2002). 27. J. A. Servin, C. W. Herbold, R. G. Skophammer, J. A. Lake, Mol. Biol. Evol. 25, 1 (2008). 28. W. Martin, M. J. Russell, Philos. Trans. Roy. Soc. Lond. B 362, 1887 (2007). 29. L. Margulis, M. F. Dolan, R. Guerrero, Proc. Natl. Acad. Sci. U.S.A. 97, 6954 (2000). 30. H. J. Melosh, Astrobiology 3, 207 (2003). 31. G. Horneck et al., Astrobiology 8, 17 (2008). 32. Z. P. Zagorski, Orig. Life Evol. Biosph. 37, 351 (2007). 33. A. K. Pavlov et al., Biofizika 52, 1136 (2007). 34. M. J. Burchell, J. R. Mann, A. W. Bunch, Mon. Not. Roy. Astron. Soc. 352, 1273 (2004). 35. J. P. Gogarten et al., Proc. Natl. Acad. Sci. U.S.A. 86, 6661 (1989). 36. N. Iwabe, K. Kuma, M. Hasegawa, S. Osawa, T. Miyata, Proc. Natl. Acad. Sci. U.S.A. 86, 9355 (1989). 37. T. Cavalier-Smith, Int. J. Syst. Evol. Microbiol. 52, 7 (2002). 38. T. Cavalier-Smith, Biol. Direct. 1, 19 (2006). 39. J. A. Lake, C. W. Herbold, M. C. Rivera, J. A. Servin, R. G. Skophammer, Mol. Biol. Evol. 24, 130 (2007). 40. R. G. Skophammer, C. W. Herbold, M. C. Rivera, J. A. Servin, J. A. Lake, Mol. Biol. Evol. 23, 1648 (2006). 41. R. G. Skophammer, J. A. Servin, C. W. Herbold, J. A. Lake, Mol. Biol. Evol. 24, 1761 (2007). 42. J. T. F. Wong, J. Chen, W. K. Mat, S. K. Ng, H. Xue, Gene 403, 39 (2007). 43. E. K. Lienau, R. DeSalle, J. A. Rosenfeld, P. J. Planet, Syst. Biol. 55, 441 (2006). 44. M. Di Giulio, Gene (Amst) 394, 105 (2007). 45. O. Zhaxybayeva, P. Lapierre, J. P. Gogarten, Protoplasma 227, 53 (2005). 46. L. Margulis, Origin of Eukaryotic Cells (Yale University Press, New Haven, Connecticut, 1970). 47. S. G. E. Andersson et al., Nature 396, 133 (1998). 48. L. Bonen, W. F. Doolittle, Proc. Natl. Acad. Sci. U.S.A. 72, 2310 (1975). 49. L. Bonen, W. F. Doolittle, Nature 261, 669 (1976). 50. W. Zillig et al., Endocyt. Cell. Res. 6, 1 (1989). 51. M. L. Sogin, Curr. Opin. Genet. Dev. 1, 457 (1991).

97

52. W. Martin, H. Brinkmann, C. Savonna, R. Cerff, Proc. Natl. Acad. Sci. U.S.A. 90, 8692 (1993). 53. G. B. Golding, R. S. Gupta, Mol. Biol. Evol. 12, 1 (1995). 54. W. F. Doolittle, Proc. Natl. Acad. Sci. U.S.A. 94, 12751 (1997). 55. P. J. Keeling, W. F. Doolittle, Proc. Natl. Acad. Sci. U.S.A. 94, 1270 (1997). 56. D.-F. Feng, G. Cho, R. F. Doolittle, Proc. Natl. Acad. Sci. U.S.A. 94, 13028 (1997). 57. L. A. Katz, Trends. Ecol. Evol. 13, 493 (1998). 58. M. C. Rivera, R. Jain, J. E. Moore, J. A. Lake, Proc. Natl. Acad. Sci. U.S.A. 95, 6239 (1998). 59. T. Horiike, K. Hamada, S. Kanaya, T. Shinozawa, Nat. Cell. Biol. 3, 210 (2001). 60. S. B. Hedges et al., BMC Evol. Biol. 1, 4 (2001). 61. C. Esser et al., Mol. Biol. Evol. 21, 1643 (2004). 62. Cavalier-Smith, in Endocytobiology. II. Intracellular Space as Oligogenetic, H. E. A. Schenk, W. S. Schwemmler, Eds. (Walter de Gruyter, Berlin, 1983), pp. 1027–1034. 63. T. Cavalier-Smith, Nature 326, 332 (1987). 64. P. J. Keeling, BioEssays 20, 87 (1998). 65. W. Martin, Curr. Opin. Microbiol. 8, 630 (2005). 66. W. Martin, E. V. Koonin, Nature 440, 41 (2006). 67. J. Tovar et al., Nature 426, 172 (2003). 68. A. V. Goldberg et al., Nature 452, 624 (2008). 69. A. D. Tsaousis et al., Nature (Lond) 453, 553 (2008). 70. W. F. Doolittle, Nature 272, 581 (1978). 71. C. G. Kurland, L. J. Collins, D. Penny, Science 312, 1011 (2006). 72. T. Dagan, W. Martin, BioEssays 29, 500 (2007). 73. W. Martin et al., Science 316, 542 (2007). 74. A. M. Poole, D. Penny, BioEssays 29, 74 (2007). 75. R. P. Hirt et al., Proc. Natl. Acad. Sci. U.S.A. 96, 580 (1999). 76. P. J. Keeling, M. A. Luker, J. D. Palmer, Mol. Biol. Evol. 17, 23 (2000). 77. A. Roger, Am. Nat. 154, S146 (1999). 78. A. J. Roger, L. A. Hug, Philos. Trans. Roy. Soc. Lond. B 361, 1039 (2006). 79. H. G. Morrison et al., Science 317, 1921 (2007). 80. P. J. Keeling et al., Trends. Ecol. Evol. 20, 670 (2005). 81. F. D. Ciccarelli et al., Science 311, 1283 (2006). 82. F. Burki et al., PLoS ONE 2, e270 (2007). 83. J. D. Hackett et al., Mol. Biol. Evol. 24, 1702 (2007). 84. H. S. Yoon et al., BMC Evol. Biol. 8, 14 (2008). 85. S. B. Hedges, J. E. Blair, M. L. Venturi, J. L. Shoe, BMC Evol. Biol. 4, 2 (2004). 86. N. Arisue, M. Hasegawa, T. Hashimoto, Mol. Biol. Evol. 22, 409 (2005). 87. T. A. Richards, T. Cavalier-Smith, Nature 436, 1113 (2005). 88. P. J. Keeling, Science 317, 1875 (2007). 89. A. M. Poole, D. Penny, BioEssays 29, 611 (2007). 90. D. E. Wujek, Trans. Am. Microsc. Soc. 98, 143 (1979).

98


91. C. D. von Dohlen, S. Kohler, S. T. Alsop, W. R. McManus, Nature 412, 433 (2001). 92. M. C. Rivera, J. A. Lake, Nature 431, 152 (2004). 93. J. Felsenstein, Syst. Zool. 27, 410 (1978). 94. M. Nei, S. Kumar, Molecular Evolution and Phylogenetics (Oxford University Press, New York, 2000). 95. W. F. Doolittle, Science 284, 2124 (1999). 96. E. V. Koonin, K. S. Makarova, L. Aravind, Annu. Rev. Microbiol. 55, 709 (2001). 97. W. F. Doolittle, E. Bapteste, Proc. Natl. Acad. Sci. U.S.A. 104, 2043 (2007). 98. C. Esser, W. Martin, T. Dagan, Biol. Lett. 3, 180 (2007). 99. F. U. Battistuzzi, A. Feijao, S. B. Hedges, BMC Evol. Biol. 4, 44 (2004). 100. J. M. Kollman, R. F. Doolittle, J. Mol. Evol. 51, 173 (2000). 101. T.-M. Han, B. Runnegar, Science 257, 232 (1992). 102. N. J. Butterfield, Paleobiology 26, 386 (2000). 103. J. Samuelsson, N. J. Butterfield, Precambrian. Res. 107, 235 (2001). 104. V. N. Sergeev, L. M. Gerasimenko, G. A. Zavarzin, Microbiology 71, 623 (2002). 105. A. H. Knoll, E. J. Javaux, D. Hewitt, P. Cohen, Philos. Trans. Roy. Soc. Lond. B 361, 1023 (2006). 106. J. J. Brocks, G. A. Logan, R. Buick, R. E. Summons, Science 285, 1033 (1999). 107. H. S. Yoon, J. D. Hackett, C. Ciniglia, G. Pinto, D. Bhattacharya, Mol. Biol. Evol. 21, 809 (2004). 108. D. Bhattacharya, H. S. Yoon, S. B. Hedges, J. D. Hackett, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 116–120. 109. R. E. Summons, L. L. Jahnke, J. M. Hope, G. A. Logan, Nature 400, 554 (1999). 110. S. E. Rashby, A. L. Sessions, R. E. Summons, D. K. Newman, Proc. Natl. Acad. Sci. U.S.A. 104, 15099 (2007). 111. J. F. Allen, W. Martin, Nature 445, 610 (2007). 112. D. E. Canfield, M. T. Rosing, C. Bjerrum, Philos. Trans. Roy. Soc. Lond. B 361, 1819 (2006). 113. M. M. Tice, D. R. Lowe, Geology 34, 37 (2006). 114. E. Bapteste, C. Brochier, Y. Boucher, Archaea 1, 353 (2005). 115. Y. Ueno, K. Yamada, N. Yoshida, S. Maruyama, Y. Isozaki, Nature 440, 516 (2006). 116. M. D. Brasier et al., Nature 416, 76 (2002). 117. M. Brasier, N. McLoughlin, O. Green, D. Wacey, Philos. Trans. Roy. Soc. Lond. B 361, 887 (2006). 118. J. W. Schopf, Philos. Trans. Roy. Soc. Lond. B 361, 869 (2006). 119. J. W. Schopf, M. R. Walter, C. Ruiji, Precambrian. Res. 158, 139 (2007). 120. J. W. Schopf, A. B. Kudryavtsev, A. D. Czaja, A. B. Tripathi, Precambrian. Res. 158, 141 (2007).

121. T. Cavalier-Smith, Philos. Trans. Roy. Soc. Lond. B 361, 969 (2006). 122. S. Gribaldo, C. Brochier-Armanet, Philos. Trans. Roy. Soc. Lond. B 361, 1007 (2006). 123. B. S. Lollar, T. M. McCollom, Nature 444, E18 (2006). 124. J. M. Hayes, in Early life on Earth, S. Bengston, Ed. (Columbia University Press, New York, 1994), pp. 220–236. 125. K. U. Hinrichs, J. M. Hayes, S. P. Sylva, P. G. Brewer, E. F. DeLong, Nature 398, 802 (1999). 126. A. H. Knoll, Life on a Young Planet: The First Three Billion Years of Evolution on Earth (Princeton University Press, Princeton, New Jersey, 2003). 127. E. J. Javaux, A. H. Knoll, M. R. Walter, Nature 412, 66 (2001). 128. S. M. Porter, A. H. Knoll, Paleobiology 26, 360 (2000). 129. R. F. Doolittle, D.-F. Feng, S. Tsang, G. Cho, E. Little, Science 271, 470 (1996). 130. X. Gu, Mol. Biol. Evol. 14, 861 (1997). 131. P. P. Sheridan, K. H. Freeman, J. E. Brenchley, Geomicrobiol. J. 20, 1 (2003). 132. J. D. Hackett, H. S. Yoon, N. J. Butterfield, M. J. Sanderson, D. Bhattacharya, in Evolution of Primary Producers in the Sea, P. G. Falkowski, A. H. Knoll, Eds. (Elsevier, Burlington, 2007), pp. 109–131. 133. E. J. P. Douzery, E. A. Snell, E. Bapteste, F. Delsuc, H. Philippe, Proc. Natl. Acad. Sci. U.S.A. 101, 15386 (2004). 134. L. A. Hug, A. J. Roger, Mol. Biol. Evol. 24, 1889 (2007). 135. C. Berney, J. Pawlowski, Proc. Roy. Soc. Lond. B 273, 1867 (2006). 136. C. de Vargas, M.-P. Aubry, I. Probert, J. Young, in Evolution of Primary Producers in the Sea, P. G. Falkowski, A. H. Knoll, Eds. (Elsevier Academic Press, Burlington, Massachusetts, 2007), pp. 251–285. 137. N. H. Sleep, K. J. Zanhnle, J. F. Kasting, H. J. Morowitz, Nature 384, 139 (1988). 138. D. W. Schwartzman, Life, Temperature, and the Earth (Columbia University Press, New York, 1999), pp. 241. 139. L. Delaye, A. Becerra, A. Lazcano, in The Genetic Code and the Origin of Life, L. R. Pouplana, Ed. (Springer, New York, 2005), pp. 34–47. 140. T. Dagan, W. Martin, Genome Biol. 7, 118 (2006). 141. N. R. Pace, Nature 441, 289 (2006). 142. N. Goldenfeld, C. Woese, Nature 445, 369 (2007). 143. W. Martin, E. V. Koonin, Nature 442, 868 (2006). 144. M. Gogartenboekels, E. Hilario, J. P. Gogarten, Orig. Life Evol. Biosph. 25, 251 (1995). 145. F. U. Battistuzzi, S. B. Hedges, Mol. Biol. Evol. 26, 335 (2009).

SUPERKINGDOMS


Archaebacteria Fabia U. Battistuzzia,b,* and S. Blair Hedgesa a

Department of Biology, 208 Mueller Laboratory, The Pennsylvania State University, University Park, PA 16802-5301, USA; bCurrent address: Center for Evolutionary Functional Genomics, The Biodesign Institute, Arizona State University, Tempe, AZ 85287-5301, USA *To whom correspondence should be addressed (Fabia.Battistuzzi@ asu.edu)

Abstract The Superkingdom Archaebacteria (~300 species) is divided into two phyla, Euryarchaeota and Crenarchaeota, with two other phyla (Korarchaeota and Nanoarchaeota) under consideration. Most large-scale phylogenetic analyses agree on a topology that clusters (i) Methanomicrobia, Halobacteria, Archaeoglobi, and Thermoplasmata and (ii) Methanobacteria, Methanococci, and Methanopyri. A molecular timetree estimated here shows divergences among classes in the Archean Eon, 3500–2500 million years ago (Ma), and family divergences in the Proterozoic Eon, 2394–829 Ma. The timetree also suggests that methanogenesis had arisen by the mid-Archean (>3500 Ma) and that adaptation to thermoacidophilic environments occurred before 1000 Ma.

Extremophiles are common among species in the Superkingdom Archaebacteria (also called “Archaea”, Fig. 1) (1). For example, the species referred to as “Strain 121” can survive temperatures up to 121°C, higher than any other organism (2), and hyperacidophiles are found in the Family Thermoplasmataceae, where two species (Picrophilus oshimae and Picrophilus torridus) are the only known organisms capable of living at a pH as low as zero (3, 4). Archaebacteria show many other phenotypes including the unique ability to produce methane (methanogenesis). They have cell wall structures formed either by pseudopeptidoglycan (i.e., a material similar to the peptidoglycan of eubacteria), polysaccharides, or glycoproteins (S-layer) (5), which resemble the single-layer structure (i.e., cell membrane plus cell wall) present in gram-positive eubacteria. Furthermore, archaebacteria have a unique cell membrane structure composed of ether-linked glycerol diethers or tetraethers that confer a

higher stability to extreme conditions (5). Chemotrophy is the most widely used metabolism, although phototrophic members of the Halobacteriaceae can use light to produce ATP (6). Six families also have the unique ability of obtaining energy by combining carbon dioxide (or other carbon compounds) and hydrogen into methane (5). The Superkingdom Archaebacteria, comprising ~300 species, is subdivided into two recognized phyla, Euryarchaeota and Crenarchaeota (7). Two other phyla have been proposed based on environmental sequences only (Korarchaeota) and environmental sequences plus one fully sequenced genome (Nanoarchaeota) (8–11) but have not been officially recognized and their phylogenetic position is uncertain. The molecular information

Fig. 1 Halobacteria (rod-shaped Halobacterium) from Mono Lake, California; mixed sample (upper left); and close-up of a cell (upper right). Methanococci (round-shaped Methanococcus); Cluster of cells (lower left) and close-up of two cells (lower right). Credits: D. J. Patterson, provided by micro*scope (http:// microscope.mbl.edu) under creative commons license (upper images); and Electron Microscopy Laboratory, University of California, Berkeley (lower images).

F. U. Battistuzzi and S. B. Hedges. Archaebacteria. Pp. 101–105 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

102


Methanospirillaceae Thermoplasmataceae

12

4

Picrophilaceae Methanopyraceae (Methanopyri)

5 3

Methanobacteriaceae (Methanobacteria) 7

Methanocaldococcaceae

11

Methanococcaceae

2

Thermococcaceae (Thermococci) 1

Sulfolobaceae (Thermoprotei) Nanoarchaeum (Nanoarchaeota)

HD

Ea

4000

Pa

Ma Na

ARCHEAN 3000

Pp

Np

Mp

PROTEROZOIC 2000

Euryarchaeota

Methanosarcinaceae 10

Crenarchaeota

6

Methanococci Thermoplasmata

Halobacteriaceae (Halobacteria) 9

Methanomicrobia

Archaeoglobaceae (Archaeoglobi) 8

Pz PH

1000

0 Million years ago

Fig. 2 A timetree of archaebacteria. Divergence times are shown in Table 1. Abbreviations: Ea (Eoarchean), HD (Hadean), Ma (Mesoarchean), Mp (Mesoproterozoic), Na (Neoarchean), Np (Neoproterozoic), Pa (Paleoarchean), PH (Phanerozoic), Pp (Paleoproterozoic), and Pz (Paleozoic).

available for Korarchaeota is based only on a few tens of environmental sequences (small subunit ribosomal RNA, SSU rRNA) as none of the Korarchaeota has been successfully cultivated. The sequences available place this group before the Euryarchaeota/Crenarchaeota divergence and show the presence of five major clusters within this putative phylum (8, 12). On the other hand, the genome of one species of Nanoarchaeota, Nanoarchaeum equitans, has been fully sequenced (13) and it is routinely used in multiple-gene phylogenies and indel analyses (9, 14–18). These studies show contrasting topologies for Nanoarchaeota, with some placing it basal to Euryarchaeota and Crenarchaeota (13), others clustering it with Crenarchaeota (17, 18), and other studies clustering it within Euryarchaeota (9, 14, 16). These different phylogenetic positions affect the classification of this species, which is either being considered a member of a new phylum (i.e., Nanoarchaeota) or a fast-evolving lineage of Euryarchaeota (14). The phylogeny of taxa within the Phyla Euryarchaeota and Crenarchaeota is mostly stable. Three hypotheses have been proposed to explain this property of archaebacteria that sets them apart from eubacteria: (i) a younger origin of archaebacteria compared to eubacteria (i.e.,

weaker biases given by a shorter evolutionary history), (ii) a lower known taxonomic diversity (e.g., nine vs. 34 classes for archaebacteria and eubacteria, respectively) (7) which could mask contrasting phylogenetic signals, and (iii) speciation events more evenly distributed throughout time that did not result in rapid adaptive radiations (i.e., no or few short internal branches that are difficult to resolve phylogenetically) (14). While the geologic record and molecular clock studies do not support the first hypothesis (i.e., archaebacteria and their metabolisms appear early in Earth’s history), it is not possible to discard either of the other two hypothesis as a possible cause for the apparently more stable phylogeny of archaebacteria. Species of Crenarchaeota are placed in six families, three of which (Thermoproteaceae, Desulfurococcaceae, and Sulfolobaceae) have been used in multiple-gene analyses of relationships. These analyses consistently have found Sulfolobaceae and Desulfurococcaceae as closest relatives to the exclusion of Thermoproteaceae, with significant support (14, 15, 17–19). Within Euryarchaeota, topological differences are more common depending on the genes and methods used to build the phylogenetic tree. Gene content studies, for example, show the

Archaebacteria

103

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among archaebacteria. Timetree Node

Estimates Ref. (19)

Time

This study

Ref. (32)

Time

CI

Time

CI

Time

4193

–

–

4193

4200–4176

–

2

4187

4112

4486–3314

4187

4199–4163

3460

3

3594

–

–

3594

3691–3503

–

4

3468

–

–

3468

3490–3460

–

5

3313

–

–

3313

3388–3232

–

6

3160

3085

3514–2469

3160

3257–3056

–

7

3093

3124

3520–2522

3093

3210–2968

–

8

2799

2625

3102–2037

2799

2936–2656

–

9

2430

–

–

2430

2596–2256

2740

10

2216

–

–

2216

2394–2034

–

11

1676

–

–

1676

1875–1475

–

12

992

–

–

992

1174–829

–

1

Note: Node times in the timetree are from this study.

Class Halobacteria as a deep-diverging lineage at the base of Euryarchaeota and Crenarchaeota (20, 21) followed by the Class Thermoplasmata, either alone or clustering with Crenarchaeota. Because both Halobacteria and Thermoplasmata are classified as euryarchaeotes (22), their phylogenetic position determines the monoor paraphyly of this phylum. Multiple-gene phylogenies show Halobacteria as closely related to the Class Methanomicrobia, a derived position that is highly supported by maximum likelihood (ML), Bayesian, and supertree phylogenies with different sets of genes (14, 17, 18, 23, 53). Thermoplasmata, instead, are alternatively supported at the base of Euryarchaeota (17, 18) or clustering with Archaeoglobi, Halobacteria, and Methanomicrobia (14, 23, 53). Only this last position is significantly supported by all analyses (Fig. 2). Recent studies with large data sets have solved the issue of the position of the Class Methanopyri. This group was initially found to be an early diverging lineage within Euryarchaeota based on SSU rRNA and indel analyses (19, 24, 25). Later studies of the complete genome of its only known representative, Methanopyrus kandleri (26) and its translation and transcription apparatus (27) have shown, instead, that it is more closely related to other methanogens, such as the Methanococci and Methanobacteria. This relation was confirmed by

genome-scale analyses with different tree-building methods (supertrees, ML, Bayesian) (17, 18, 53). All of the previous evidence points toward the presence of two main clusters within Euryarchaeota: one formed by the classes Methanomicrobia, Halobacteria, Archaeoglobi, and Thermoplasmata, and another by the classes Methanococci, Methanobacteria, and Methanopyri. These clusters are the same found in a study of sequences from 25 core proteins shared by 218 prokaryote species (including archaebacteria and eubacteria) (53). In that analysis, all available species with a complete genome were included and a complete matrix of genes and species was constructed. Single gene trees were manually screened for orthology and vertical inheritance (i.e., genes not showing paraphyly of archaebacteria and eubacteria or significantly supported deepnesting of one class within another). Site homology of the multiple sequence alignment was established with GBlocks (28) and nonconserved sites were deleted. ML (29) and Bayesian (30) methods were then applied to the final alignment (6884 sites) to estimate phylogenetic relationships. A ML phylogeny was also constructed from an alternative alignment with only slow-evolving positions (16,344 sites) and showed an identical archaebacterial topology. ML and Bayesian phylogenies were found to be identical for archaebacteria with high bootstrap

104


values (>70%) for the majority of the nodes and are also identical to previous studies, except for the position of Nanoarchaeota (14). As is the case for eubacteria, there have been very few molecular clock studies applied to archaebacteria (19, 31–33). For this reason, we estimated divergence times (Table 1) and constructed a molecular timetree (Fig. 2) for 12 families and one phylum with a Bayesian timing method (34) and using the only two calibration points available within this superkingdom: (i) a minimum of 3460 million years (Ma) for the origin of methanogenesis based on isotopically light carbon (23, 35) and (ii) a maximum of 4200 Ma for the fi rst divergence within archaebacteria based on the midpoint of the range of the last ocean-vaporizing impact (36). Besides the initial evidence for biological methane production at 3460 Ma, there is additional evidence for that metabolism later in the Archean, at ~2700 Ma (37–41). The topology of the timetree is taken from our latest phylogenetic analysis of sequences from 25 core proteins (53), although it is similar to earlier studies of complete genome sequences (e.g., 19). Divergence times from another study (32), using SSU rRNA sequences, a global clock method, a single calibration point, and uncorrected distances, are shown for comparison at relevant nodes in Table 1. Although we estimate an early divergence between Crenarchaeota and Euryarchaeota (4187, 4199–4163 Ma), all divergences among classes are later in the Archean (3500–2500 Ma), with the divergence of Halobacteria and Methanomicrobia occurring near the Archean–Proterozoic boundary (2430; 2596–2256 Ma). The Phylum Nanoarchaeota is basal to all archaebacteria in our phylogeny and its divergence from other archaebacteria is estimated at 4193 Ma (4200–4176 Ma). However, given its uncertain phylogenetic position this time estimate should be considered with caution. Only three classes (Methanomicrobia, Methanococci, and Thermoplasmata) have representatives of more than one family. These family-level divergences are within the Proterozoic, between 2216 and 992 Ma (CI, 2394–829 Ma). The deepest branches of both Crenarchaeota and Euryarchaeota are occupied by hyperthermophilic organisms. Although our knowledge of the diversity of archaebacteria is limited and it is possible that mesophilic deep-branching species will be discovered, the current phylogenetic pattern suggests that the ancestor of this superkingdom was adapted to high temperature environments.

The distribution of methanogenesis among families supports multiple losses of this metabolism during evolution (e.g., Thermoplasmata, Halobacteria). The common ancestor of all methanogens (Methanobacteriaceae, Methanocaldococcaceae, Methanococcaceae, Methanosarcinaceae, Methanospirillaceae, and Methanopyraceae) is estimated to have evolved by the mid-Archean regardless of the calibration used for the molecular clock (i.e., eukaryotic or archaebacterial) (19, 53). This early evolution of methanogenesis is not only in agreement with the geologic record (42) but also lends support to one of the hypotheses addressing the faint young sun paradox (43). These suggest that a greenhouse effect was present in the early history of Earth to compensate for the lower luminosity of the Sun. Among the greenhouse gases proposed are carbon dioxide (44) and methane (45–47) with the latter, according to our timetree, being of biologic origin. Genes involved in methanogenesis and methylotrophy (i.e., methanopterin and methanofuran-linked C1 transfer genes) are shared by methanogens and at least two groups of eubacteria (the Phylum Proteobacteria and the Class Planctomycetacia) (48). Contrary to a previous hypothesis (49), the late divergence of the eubacterial Planctomycetacia (53) suggests horizontal gene transfer (HGT) from archaebacteria to eubacteria as a possible cause of their current distribution. An alternative possibility is the presence of this pathway in the ancestor of eubacteria and archaebacteria. This cannot be discarded with the current information, especially in light of the recent discovery of these genes in yet to be classified lineages (49, 50). Another ecological innovation that evolved in archaebacteria is the adaptation to thermoacidophilic environments (i.e., pH < 3; temperature > 50°C). Strict thermoacidophiles are present only in the Class Thermoplasmata (Families Thermoplasmataceae, Ferroplasmaceae, and Picrophilaceae) and some Thermoprotei (e.g., Sulfolobaceae) (51, 52) with evidence for extensive HGTs between these two classes. Because the only representative of Thermoprotei is a member of the Family Sulfolobaceae, it is not possible to constrain the time estimate for this metabolism with an upper limit. However, the divergence of Thermoplasmataceae and Picrophilaceae at 992 Ma (1174–829 Ma) sets a minimum time for the origin of this metabolism. Compared to the Superkingdom Eubacteria, the Superkingdom Archaebacteria is not as well known in terms of its taxonomic and environmental diversity. Nonetheless, the current timetree shows an early origin

Archaebacteria

of these lineages and evolutionary innovations (e.g., methanogenesis) that are likely to have played a fundamental role in the habitability and colonization of the planet.

Acknowledgment Support was provided by the U.S. National Science Foundation and National Aeronautics and Space Administration (NASA Astrobiology Institute) to S.B.H.

References 1. P. Forterre, C. Brochier, H. Philippe, Theoret. Pop. Biol. 61, 409 (2002). 2. K. Kashefi, D. R. Lovley, Science 301, 934 (2003). 3. C. Schleper, G. Puhler, H. P. Klenk, W. Zillig, Int. J. Syst. Bact. 46, 814 (1996). 4. K. J. Edwards, P. L. Bond, T. M. Gihring, J. F. Banfield, Science 287, 1796 (2000). 5. M. T. Madigan, J. M. Martinko, J. Parker, Brock Biology of Microorganisms (Prentice-Hall, New Jersey, 10th ed., 2003). 6. D. A. Bryant, Frigaard, N.-U., Trends Microbiol. 14, 488 (2007). 7. G. M. Garrity et al. Taxonomic outline of the Bacteria and Archaea Release 7.7 (Michigan State University, East Lansing, Michigan, 2007). 8. T. A. Auchtung, C. D. Takacs-Vesbach, C. M. Cavanaugh, Appl. Environ. Microb. 72, 5077 (2006). 9. C. Brochier, S. Gribaldo, Y. Zivanovic, F. Confalonieri, P. Forterre, Genome Biol. 6, R42 (2005). 10. G. Webster et al., FEMS Microbiol. Ecol. 62, 78 (2007). 11. N. R. Pace, Science 276, 734 (1997). 12. S. M. Barns, C. F. Delwiche, J. D. Palmer, N. R. Pace, Proc. Natl. Acad. Sci. U.S.A. 93, 9188 (1996). 13. E. Waters et al., Proc. Natl. Acad. Sci. U.S.A. 100, 12984 (2003). 14. S. Gribaldo, C. Brochier-Armanet, Philos. Trans. Roy. Soc. Lond. B 361, 1007 (2006). 15. B. Gao, R. S. Gupta, BMC Genomics 8, 86 (2007). 16. K. S. Makarova, E. V. Koonin, Curr. Opin. Microbiol. 8, 586 (2005). 17. D. Pisani, J. A. Cotton, J. O. McInerney, Mol. Biol. Evol. 24, 1752 (2007). 18. F. D. Ciccarelli et al., Science 311, 1283 (2006). 19. F. U. Battistuzzi, A. Feijão, S. B. Hedges, BMC Evol. Biol. 4, 44 (2004). 20. B. Runnegar, C. H. House, S. T. Fitz-Gibbon, Geobiology 1, 15 (2003). 21. Y. I. Wolf, I. B. Rogozin, N. V. Grishin, E. V. Koonin, Trends Genet. 18, 472 (2002). 22. G. M. Garrity, Ed., Bergey’s Manual of Systematic Bacteriology (Springer, New York, 2001).

105

23. E. Bapteste, C. Brochier, Y. Boucher, Archaea 1, 353 (2005). 24. S. Burggraf, K. O. Stetter, P. Rouviere, C. R. Woese, Syst. Appl. Microbiol. 14, 346 (1991). 25. M. C. Rivera, J. A. Lake, Int. J. Syst. Bact. 46, 348 (1996). 26. A. I. Slesarev et al., Proc. Natl. Acad. Sci. U.S.A. 99, 4644 (2002). 27. C. Brochier, P. Forterre, S. Gribaldo, Genome Biol. 5, R17 (2004). 28. J. Castresana, Mol. Biol. Evol. 17, 540 (2000). 29. A. Stamatakis, Bioinformatics 22, 2688 (2006). 30. F. Ronquist, J. P. Huelsenbeck, Bioinformatics 19, 1572 (2003). 31. D.-F. Feng, G. Cho, R. F. Doolittle, Proc. Natl. Acad. Sci. U.S.A. 94, 13028 (1997). 32. P. P. Sheridan, K. H. Freeman, J. E. Brenchley, Geomicrobiol. J. 20, 1 (2003). 33. S. B. Hedges et al., BMC Evol. Biol. 1, 4 (2001). 34. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 35. Y. Ueno, K. Yamada, N. Yoshida, S. Maruyama, Y. Isozaki, Nature 440, 516 (2006). 36. N. H. Sleep, K. J. Zahnle, J. F. Kasting, H. J. Morowitz, Nature 342, 139 (1989). 37. J. M. Hayes, in Earth’s Earliest Biosphere: Its Origin and Evolution, J. W. Schopf, Ed. (Princeton University Press, Princeton, New Jersey, 1983) pp. 291–301. 38. J. M. Hayes, in Early Life on Earth, S. Bengston, Ed. (Columbia University Press, New York, 1994), pp. 220–236. 39. K. U. Hinrichs, J. M. Hayes, S. P. Sylva, P. G. Brewer, E. F. DeLong, Nature 398, 802 (1999). 40. K.-U. Hinirichs, 11th Annual V.M. Goldschmidt Conference Program and Abstracts abs. 3461 (2001). 41. K. U. Hinrichs, Geochem. Geophys. Geosyst. 3, 1042 (2002). 42. Y. Ueno, K. Yamada, N. Yoshida, S. Maruyama, Y. Isozaki, Nature 440, 516 (2006). 43. J. F. Kasting, D. Catling, Ann. Rev. Astron. Astrophys. 41, 429 (2003). 44. H. Ohmoto, Y. Watanabe, K. Kumazawa, Nature 429, 395 (2004). 45. A. A. Pavlov, J. F. Kasting, L. L. Brown, K. A. Rages, R. Freedman, J. Geophys. Res. 105, 11981 (2000). 46. A. A. Pavlov, M. T. Hurtgen, J. F. Kasting, M. A. Arthur, Geology 31, 87 (2003). 47. J. F. Kasting, A. A. Pavlov, J. L. Siefert, Orig. Life Evol. Biosph. 31, 271 (2001). 48. L. Chistoserdova, S. W. Chen, A. Lapidus, M. E. Lidstrom, J. Bacteriol. 185, 2980 (2003). 49. L. Chistoserdova et al., Mol. Biol. Evol. 21, 1234 (2004). 50. M. G. Kalyuzhnaya, L. Chistoserdova, Methods Enzymol. 397, 443 (2005). 51. C. Bertoldo, C. Dock, G. Antranikian, Eng. Life Sci. 4, 521 (2004). 52. A. Angelov, W. Liebl, J. Biotechnol. 126, 3 (2006). 53. F. U. Battistuzzi, S. B. Hedges, Mol. Biol. Evol. 26, 335 (2009).

Eubacteria Fabia U. Battistuzzia,b,* and S. Blair Hedgesa Department of Biology, 208 Mueller Laboratory, The Pennsylvania State University, University Park, PA 16802-5301, USA; bCurrent address: Center for Evolutionary Functional Genomics, The Biodesign Institute, Arizona State University, Tempe, AZ 85287-5301, USA *To whom correspondence should be addressed (Fabia.Battistuzzi@ asu.edu)

Abstract The ~9400 recognized species of prokaryotes in the Superkingdom Eubacteria are placed in 25 phyla. Their relationships have been difficult to establish, although some major groups are emerging from genome analyses. A molecular timetree, estimated here, indicates that most (85%) of the phyla and classes arose in the Archean Eon (4000−2500 million years ago, Ma) whereas most (95%) of the families arose in the Proterozoic Eon (2500−542 Ma). It also points to an early origin of phototrophy (3400–3200 Ma) followed by colonization of land (3041–2755 Ma), origin of oxygenic photosynthesis (2920–2592 Ma) and of aerobic methanotrophy (2630–2371 Ma).

The Superkingdom Eubacteria (also called “Bacteria,” Fig. 1) is subdivided into 25 recognized phyla (1), although other classifications have been proposed (2) and putative lineages await taxonomic assignment (e.g., 3, 4). In contrast to most eukaryotic groups, eubacterial phylogeny has been difficult to resolve at the highest level (i.e., phyla and class relationships) despite the availability of many complete genome sequences. The small subunit (SSU) ribosomal RNA (rRNA) gene has been widely used to study prokaryote phylogeny and first revealed the distinction of eubacteria from archaebacteria (Archaea) (5, 6). However, because of its limited length and the large genetic distances among prokaryote species, it is also subject to biases such as long-branch attraction, base composition differences, and a generally poor resolving power. These issues mostly have been addressed in more recent years by using genome-scale data sets (e.g., multiple genes, gene content, and insertion–deletion events) and increased taxonomic sampling (7–19, 88). While these studies have

shown increasing support for lower-level phylogenetic clusters (e.g., classes and below), they have also shown the susceptibility of eubacterial phylogeny to biases such as horizontal gene transfer (HGT) (20, 21). In recent years, three major approaches have been used for studying prokaryote phylogeny with data from complete genomes: (i) combining gene sequences in a single analysis of multiple genes (e.g., 7, 9, 10), (ii) combining trees from individual gene analyses into a single “supertree” (e.g., 22, 23), and (iii) using the presence or absence of genes (“gene content”) as the raw data to investigate relationships (e.g., 17, 18). While the results of these different approaches have not agreed on many details of relationships, there have been some points of agreement, such as support for the monophyly of all major classes and some phyla (e.g., Proteobacteria and Firmicutes). These findings, although criticized by some (e.g., 24, 25), suggest the presence of a detectable evolutionary signal for prokaryotes when using carefully selected genes (e.g., vertically inherited) and appropriate methodologies (e.g., genome-scale data sets).

Fig. 1 Spirillum, a betaproteobacterium (upper left); Beggiatoa, a gammaproteobacterium (upper right); an unidentified cyanobacterium (lower left); and an unidentified spirochete (lower right). Credits: D. J. Patterson, L. Amaral-Zettler, and V. Edgcomb (upper left and right); D. J. Patterson (lower left); and L. Amaral-Zettler, L. Olendzenski, and D. J. Patterson (lower right). Images provided by micro*scope (http://microscope.mbl. edu) under creative commons license.

F. U. Battistuzzi and S. B. Hedges. Eubacteria. Pp. 106–115 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, New York, 2009).

Eubacteria Staphylococcaceae 50

Bacillaceae 61

Listeriaceae

30

Lactobacillaceae 54

21

Streptococcaceae

Mycoplasmataceae-1 46

Mycoplasmataceae-2 85

15

Entomoplasmataceae

Firmicutes

Acholeplasmataceae 29

Symbiobacterium 24

Carboxydothermus 33

Peptococcaceae 44

20

8

Thermoanaerobacteriaceae-1

60

Synechococcaceae-1 Synechococcaceae-2

66

6

75

Synechococcaceae-3

76

Nostocaceae

80

Chloroflexi

Gloeobacteraceae

11

Cyanobacteria

Dehalococcoides

Merismopediaceae Deinococcaceae Propionibacteriaceae Frankiaceae

63

Corynebacteriaceae 82

Mycobacteriaceae 88

59 43

Nocardiaceae Streptomycetaceae

69

Nocardiopsaceae Bifidobacteriaceae

48

Cellulomonadaceae 71

Microbacteriaceae

(continued on next page) HD

Ea

4000

Pa

Ma Na

ARCHEAN 3000

Fig. 2 Continues

Pp

Mp

Np

PROTEROZOIC 2000

1000

Pz PH 0 Million years ago

Actinobacteria

53

13

Deinococcus-Thermus

Thermoanaerobacteriaceae-2

Terrabacteria

Clostridiaceae 36

107

108

THE TIMETREE OF LIFE (continued from previous page)

Francisellaceae Shewanellaceae 73

42 70

Vibrionaceae

77 81

Enterobacteriaceae Pasteurellaceae

Colwelliaceae Pseudoalteromonadaceae

37

Moraxellaceae 57

Pseudomonadaceae 67

Hahellaceae

32

Coxiellaceae 51

Legionellaceae Chromatiaceae

38 4 26

Hydrobacteria

47

Gammaproteobacteria

Idiomarinaceae 74 79

Methylococcaceae

34

Piscirickettsiaceae

39

Xanthomonadaceae

Rhodocyclaceae 55

68

Comamonadaceae

72

Burkholderiaceae

78

58

Alcaligenaceae Nitrosomonadaceae

64

Betaproteobacteria

Neisseriaceae

Hydrogenophilaceae

(continued on next page)

HD

Ea

4000

Pa

Ma Na

ARCHEAN 3000

Pp

Mp

Np

PROTEROZOIC 2000

1000

Pz PH 0 Million years ago

Fig. 2 Continues

Phylogenetic analyses that have focused on subgroups of prokaryotes (e.g., classes) also have supported, consistently, particular groupings. For example, cyanobacteria and low GC gram positives (Firmicutes) have been united based on maximum likelihood (ML) mapping of 21 orthologous genes (26); Fusobacteria is the most closely related group to Firmicutes based on combined analyses of SSU and large subunit (LSU) rRNA sequences and ribosomal proteins (27); Bacteroidetes, Chlorobi, and Fibrobacteres form a group based on insertion–deletion analysis (28, 29); and Planctomycetacia, Chlamydiae, and Spirochaetes form a group based on concatenated

ribosomal proteins, DNA-directed RNA polymerase subunits (30), genome trees (31), and gene content analysis (18). Other taxa, such as Aquificae and Thermotogae, have been more difficult to place, appearing at the base of the tree in some analyses (4, 7, 9, 32) and in a higher, nested position in other analyses (10, 33, 34). Generally concordant relationships were found between past studies and a recent ML and Bayesian analysis of multiple core gene sequences (88). In that study, a complete matrix of 25 vertically inherited orthologous genes shared by 197 fully sequenced eubacteria and 21 fully sequenced species of archaebacteria was built.

Eubacteria (continued from previous page)

Erythrobacteraceae

17 41

Rhodobacteraceae

45

Caulobacteraceae

65

35

Rhizobiaceae 86 84

Bartonellaceae

87

14

Phyllobacteriaceae

25

Brucellaceae Rhodospirillaceae

49

Acetobacteraceae

3

SAR11 27

28

Chlorobiaceae 23

Crenotrichaceae 31

Bacteroidaceae

7

83

Porphyromonadaceae

1

Planctomycetaceae 10

Spirochaetaceae 19

Leptospiraceae Fusobacteriaceae Aquificaceae Thermotogaceae

4000

Ma Na

ARCHEAN 3000

Pp

Mp

Np

PROTEROZOIC 2000

1000

Fig. 2 A timetree of eubacteria. Divergence times are shown in Table 1. Codes for paraphyletic and/or polyphyletic groups are as follows: Mycoplasmataceae-1 (Mycoplasma genitalium), Mycoplasmataceae-2 (Mycoplasma capricolum), Synechococcaceae-1 (Synechocococcus JA-2-3Ba), Synechococcaceae-2 (Thermosynechococcus elongatus), Synechococcaceae-3 (Synechococcus elongatus),

Pz

Epsilonproteobacteria

Chlamydiaceae 12

Bacteroidetes

Helicobacteraceae

Planctomycetes

Campylobacteraceae 62

2

Aquificae

Geobacteraceae

5

Thermotogae Spirochaetes

Pelobacteraceae 52

Chlorobi

Bdellovibrionaceae

22

Hydrobacteria

Myxococcaceae

18

Deltaproteobacteria

Desulfovibrionaceae

Chlamydiae

Solibacteraceae 16

Fusobacteria

Rickettsiaceae

9

Pa

Acidobacteria

Anaplasmataceae 40

Ea

Alphaproteobacteria

Bradyrhizobiaceae

56

HD

109

PH 0 Million years ago

Thermoanaerobacteriaceae-1 (Moorella thermoacetica), Thermoanaerobacteriaceae-2 (Thermoanaerobacter tengcongensis). Abbreviations: Ea (Eoarchean), HD (Hadean), Ma (Mesoarchean), Mp (Mesoproterozoic), Na (Neoarchean), Np (Neoproterozoic), Pa (Paleoarchean), PH (Phanerozoic), Pp (Paleoproterozoic), and Pz (Paleozoic).

110


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among eubacteria. Timetree Node

Estimates Ref. (7 )

Time Time

Timetree This study

CI

Time

CI

Node

Estimates Ref. (7 )

Time Time

This study CI

Time

CI

1

4189

–

–

4189

4200–4159

45

1498

–

–

1498

1644–1351

2

4179

–

–

4179

4197–4141

46

1482

–

–

1482

1656–1313

3

3306

3186

3634–2801

3306

3447–3165

47

1481

1387

1763–1060

1481

1586–1372

4

3134

–

–

3134

3265–2987

48

1436

–

–

1436

1600–1270

5

2979

3096

3539–2723

2979

3116–2835

49

1432

–

–

1432

1580–1283

6

2908

–

–

2908

3041–2755

50

1429

1561

2012–1166

1429

1597–1269

7

2897

–

–

2897

3040–2747

51

1420

–

–

1420

1539–1297

8

2874

–

–

2874

3012–2716

52

1413

–

–

1413

1573–1257

9

2849

2800

3223–2452

2849

2983–2706

53

1402

–

–

1402

1557–1241

10

2762

–

–

2762

2926–2595

54

1392

–

–

1392

1573–1223

11

2761

–

–

2761

2920–2592

55

1386

1470

1873–1114

1386

1511–1262

12

2739

–

–

2739

2894–2575

56

1326

1537

1942–1151

1326

1471–1183

13

2739

–

–

2739

2897–2570

57

1306

–

–

1306

1420–1190

14

2687

–

–

2687

2815–2549

58

1224

–

–

1224

1349–1101

15

2607

2688

3108–2360

2607

2753–2448

59

1189

1380

1798–1038

1189

1332–1040

16

2579

–

–

2579

2715–2428

60

1180

–

–

1180

1334–1035

17

2504

2508

2928–2154

2504

2630–2371

61

1121

1282

1690–916

1121

1276–977

18

2421

–

–

2421

2563–2267

62

1104

1244

1694–856

1104

1271–943

19

2339

–

–

2339

2517–2154

63

1069

–

–

1069

1209–928

20

2281

–

–

2281

2439–2115

64

1055

–

–

1055

1184–932

21

2233

2305

2729–1944

2233

2401–2067

65

1042

–

–

1042

1180–912

22

2173

–

–

2173

2321–2018

66

1037

–

–

1037

1186–902

23

2099

–

–

2099

2261–1932

67

1030

–

–

1030

1145–917

24

2047

–

–

2047

2215–1873

68

1028

–

–

1028

1150–911

25

2042

2030

2449–1656

2042

2187–1887

69

1027

–

–

1027

1167–887

26

1993

1945

2368–1573

1993

2099–1894

70

950

–

–

950

1051–849

27

1919

–

–

1919

2083–1749

71

937

–

–

937

1084–794

28

1899

–

–

1899

2078–1719

72

872

–

–

872

993–759

29

1860

–

–

1860

2040–1680

73

871

–

–

871

972–772

30

1837

1816

2261–1415

1837

2011–1673

74

812

–

–

812

910–711

31

1834

–

–

1834

2005–1665

75

793

1039

1408–702

793

923–678

32

1806

1751

2163–1390

1806

1889–1731

76

751

–

–

751

880–639

33

1775

–

–

1775

1948–1602

77

751

–

–

751

854–653

34

1753

–

–

1753

1821–1690

78

744

–

–

744

859–640

35

1753

–

–

1753

1894–1605

79

668

–

–

668

768–573

36

1747

2132

2552–1760

1747

1931–1572

80

662

756

1070–484

662

783–553

37

1673

–

–

1673

1765–1581

81

634

–

–

634

739–538

38

1653

–

–

1653

1688–1640

82

621

928

1274–644

621

734–516

39

1621

–

–

1621

1722–1520

83

616

–

–

616

734–506

40

1620

–

–

1620

1792–1449

84

594

–

–

594

699–499

Eubacteria

111

Table 1. Continued Timetree Node

Estimates Ref. (7 )

Time

Timetree This study

Time

CI

Time

Node

Estimates Ref. (7 )

Time

CI

This study

Time

CI

Time

CI

41

1613

–

–

1613

1761–1465

85

523

–

–

523

638–425

42

1612

–

–

1612

1712–1506

86

509

–

–

509

609–421

43

1579

–

–

1579

1740–1411

87

432

–

–

432

520–351

44

1554

–

–

1554

1729–1378

88

380

–

–

380

469–304

Note: Nodes times in the timetree are from this study.

Single-gene phylogenies were screened for orthology and vertical inheritance under the assumption that phylogenies showing mixing of the two superkingdoms or significantly supported deep nesting of one class within another were affected by HGT and deleted from the data set. Site homology of the multiple sequence alignment was established with GBlocks (35) and nonconserved sites were deleted. ML (36) and Bayesian (37) methods were then applied to the final alignment (6884 sites) to estimate phylogenetic relationships. A ML phylogeny was also constructed from an alternative alignment with only slow-evolving positions (16,344 sites) and showed a similar phylogeny (except for the position of Solibacteres). A topological feature common to all of these analyses is a major dichotomy in eubacteria. This is formed by the Terrabacteria (Actinobacteria, Chloroflexi, Cyanobacteria, Deinococcus-Thermus, and Firmicutes) and the Hydrobacteria (Acidobacteria, Bacteroidetes, Chlamydiae, Chlorobi, Planctomycetacia, Proteobacteria, and Spirochaetes) to the exclusion of the hyperthermophiles and, possibly, Fusobacteria (88). This definition of Terrabacteria is more inclusive than the initial definition provided in 2004, which did not include Chloroflexi and Firmicutes (7). In addition, three possible high-level clusters within Hydrobacteria are highly supported by the ML phylogeny. Bacteroidetes, Chlorobi, Chlamydiae, Planctomycetes, and Spirochaetes are associated in a single cluster, with high bootstrap support (82%), for which we propose the name Spirochlamydiae. Within this infrakingdom, two other superphyla are strongly supported: (i) Planctomycetes and Spirochaetes (93%) and (ii) Bacteroidetes and Chlorobi (100%). For these we propose the names Spiroplancti and Bacterobi, respectively. These last two clusters are present and significantly

supported in the Bayesian phylogeny as well. Multiple evidence from other analyses (2, 11, 23, 28) supports Bacterobi, while Spiroplancti is present in some but not all previous studies. However, currently no clear physiological or metabolic adaptations can be considered as unifying characteristics of these clusters. Molecular clock methods have been used infrequently with eubacteria, probably because of difficulties in establishing a robust phylogeny and in selecting reliable calibration points. Calibrations have included the fossil record of prokaryotes and eukaryotes and biomarkers in the geologic record. A problem with most eukaryote fossils, otherwise useful for calibration, is that they are from the Phanerozoic eon (543–0 Ma), much younger than many key nodes of interest to time. This exacerbates the problems with different rates of evolution among the three superkingdoms (38), an issue diminished by using calibrations within each superkingdom. Although the prokaryote fossil record extends reliably back to ~2000 Ma (39), the taxonomic resolution is not sufficient in most cases to be broadly useful for calibration. Better resolution for a few groups is obtained from the geologic record with biomarkers produced specifically and exclusively by one lineage (40). Prokaryotes that are parasitic on eukaryotes provide other opportunities for calibration (41, 42), but they are almost entirely in the Phanerozoic, again much younger than many deep nodes of interest. One additional calibration that has been used is the maximum time for life on Earth (7), which can be set either to the origin of Earth at 4600 Ma as an absolute maximum or perhaps more realistically to the last ocean-vaporizing event, ~4200 Ga (43). Data sets using different types of molecular information (i.e., rRNA, enzymes, core genes) have been used to estimate divergence times. A divergence time analysis of

112


prokaryotes in the mid-1990s used amino acid sequences of 64 proteins, calibrated within eukaryotes (44). Divergences among Cyanobacteria, Gram positives (Firmicutes and Actinobacteria), and Gram negatives (e.g., Proteobacteria) were estimated to be in the early Proterozoic, 2500−2100 Ma. This time interval includes what would be later identified as the major rise in atmospheric oxygen (45) caused by the evolution of oxygenic photosynthesis in Cyanobacteria. However, the times in that study are problematic because lineage-specific rate differences (14) were not taken into account. In 2001, another prokaryote timescale was constructed from the limited genome data available at that time and again calibrated within eukaryotes (14). Although the focus in that study was on the origin of eukaryotes and eukaryotic genes from among different groups of prokaryotes, a divergence time for the split of Cyanobacteria and its closest relative among eubacteria was estimated as 2560 (2810–2300) Ma. Thus two studies using different data and methods obtained a similar, “young” date for the origin of Cyanobacteria, contrasting with the ~3500 Ma fossils (46) widely believed to be of cyanobacteria at that time. Subsequently, those fossils were reexamined and shown to be either of some other organism (not cyanobacteria) or an artifact of preservation (47–50). Divergence times among 98 prokaryote species were also estimated with sequences of the SSU rRNA gene, a single calibration, and a global clock method (51). However, the results were problematic because uncorrected distances were used which greatly bias time estimates. A genome-based molecular clock study appeared in 2004 which used sequences of 32 “core” proteins shared by 69 prokaryote species and three eukaryotes. A Bayesian clock method was used with multiple calibrations. The origin of Firmicutes was estimated at 2688 Ma (credibility interval, CI: 3108−2360 Ma) and the origins of alpha-, beta-, and gammaproteobacteria were estimated at 2508 Ma (CI: 2928–2154 Ma) (Table 1). Here, a detailed timetree of 197 eubacterial species (81 families) was estimated from the ML phylogeny of our latest data set (88) using archaebacteria as an outgroup (Fig. 2). This rooting follows analyses of paralogous genes that found the root of the tree of life between archaebacteria and eubacteria (52, 53), a root position that has been generally accepted for the last two decades and supported in a recent, expanded analysis (54). Although analyses of insertions and deletions in paralogous genes have suggested that the root might be in some other location (55, 56), others have questioned the alignments used in those

insertion–deletion analyses (57, 58). Additional evidence is needed before any root position can be considered well established. Divergence times were obtained with a Bayesian method (59) using all available calibration points within eubacteria. These included: (i) a minimum of 1640 Ma for the origin of Chromatiaceae (gammaproteobacteria) based on biomarker evidence (i.e., okenane) (60); (ii) a minimum of 1640 Ma for the divergence of Chlorobi and Bacteroidetes based on biomarker evidence (i.e., chlorobactane) (60); (iii) a minimum at 2300 Ma for the divergence of Cyanobacteria and Chloroflexi corresponding to the first significant rise in oxygen concentration in the atmosphere (45); (iv) a maximum of 4000 Ma for the earliest land-dwelling taxa corresponding to the presence of continents (61); and (v) a maximum constraint at 4200 Ma for the first divergence within eubacteria from the midpoint of the time range estimated for the last oceanvaporizing event (43). In the final time estimates, calibrations (iii) and (iv) were omitted because their absence did not significantly alter the estimates of the other nodes while it allowed inferences on these evolutionary adaptations. We used one species representative for monophyletic families and multiple species for paraphyletic families (Mycoplasmataceae, Thermoanaerobacteriaceae; Synechococcaceae); species lacking a family classification are also shown (Candidatus Pelagibacter ubique, Symbiobacterium thermophilum, Carboxydothermus hydrogenoformans). Time estimates (Table 1, Fig. 2) for divergences among higher-level taxa – phyla and classes – were mostly in the Archean Eon (4500−2500 Ma), with only the phyla Chlorobi/Bacteroidetes, and the classes Beta-/Gammaproteobacteria, and Bacilli/Mollicutes diverging in the Proterozoic (2500–543 Ma). Family divergences, instead, were evenly distributed throughout the Proterozoic with only a few of them occurring more recently in the Phanerozoic (Table 1). Apart from the deep branches of the two hyperthermophilic lineages (Aquificae and Thermotogae), most divergences of phyla and classes were closely spaced in time, 3500–2500 Ma and especially 3000–2600, suggesting the colonization and adaptation of this superkingdom to new environments by the end of the Archean. This corresponds to the radiation of major eubacterial clades found in earlier time analyses (e.g., 7, 14). The short internal branches in portions of the timetree may explain why there has been difficulty in obtaining a robust phylogeny in past studies (14, 62). The position of hyperthermophiles at the base of the eubacterial tree

Eubacteria

has been criticized on the grounds of potential artifacts caused by different nucleotide base compositions and outgroup rootings (33). However, in our most recent analysis (88) this topology was supported by multiple types of evidence. Regardless of the position of the hyperthermophiles, some recent support has been obtained for the adaptation of eubacterial proteins to thermophilic temperatures (i.e., above 50°C) (63). This suggests that the ancestor of eubacteria may have lived in a high-temperature environment. The origin of major adaptations in eubacteria, such as phototrophy, agrees well with evidence from the geologic record. The patchy distribution of phototrophic and photosynthetic genes among eubacteria suggests a combination of vertical and horizontal gene transfer (64–68). Accordingly, the ancestor of all eubacteria, except the two hyperthermophiles and Fusobacteria, most likely was a phototroph, placing the origin of this lifestyle by at least the mid-Archean (3265–2987 Ma) (Table 1). We propose the name Selabacteria (from the Greek selas, light, and bacteria, rod, in allusion to the innovation of phototrophy) for this group. The exclusion of Fusobacteria from Selabacteria should be treated with caution because of its uncertain phylogenetic position. The first evidence for photosynthesis-mediated sediment deposition and anoxygenic photosynthetic ecosystems is present at 3400 Ma (69–71). Considering the timetree and geologic evidence, we infer an early evolution of phototrophy, ~3400–3200 Ma. It is likely that the origin of phototrophy allowed eubacteria to undergo further biochemical adaptation and evolutionary radiation, as seen in the relatively rapid branching of many phyla in the timetree 3000–2600 Ma (Fig. 2). Shortly after the evolution of phototrophy, eubacteria diverged into two main groups of phyla, a gram-negative group, Hydrobacteria, and a primarily gram-positive group, Terrabacteria (Fig. 2). A previous analysis of aerobic methanotrophy showed that its origin within Hydrobacteria probably occurred by the end of the Archean, 2510 Ma (7). Th is metabolism was recently discovered in a species of Verrucomicrobium (a member of the Verrucomicrobia/Chlamydiae/Planctomycetes supergroup) (72, 73), suggesting its presence in the ancestor of Proteobacteria and the Verrucomicrobia. Although Verrucomicrobia is absent from our data set, evidence points to a relationship between it, Chlamydiae, and Planctomycetes (74), associating its ancestor with Proteobacteria, and thus the evolution of aerobic methanotrophy, in the late Archean (3116– 2835 Ma) (Table 1). Th is time estimate postdates, as

113

expected, the evolution of methanogenesis (minimum of 3460 Ma), which would have contributed to produce the substrate for methanotrophy. However, the requirement of oxygen for methanotrophy makes its evolution unlikely before the origin of the oxygen-producing cyanobacteria (2920–2592 Ma) (Table 1) and the rise in atmospheric oxygen recorded in the geologic record (45). Given this evidence we consider it more likely that the evolution of aerobic methanotrophy occurred in the ancestor of alpha- and gammaproteobacteria (2630–2371 Ma; Table 1) and then subsequently spread to other lineages by HGT. Terrabacteria show multiple adaptations to terrestrial habitats such as the formation of resistant stages (e.g., endospores in Firmicutes, akinetes in Cyanobacteria), the production of photoprotective pigments (e.g., carotenoids), and resistance to desiccation (75). Although these traits are not uniquely distributed within Terrabacteria, the coexistence of more than one within a class and the presumed terrestrial niche of the ancestor (88) suggest that terrestrial environments were colonized by the late Archean (3041–2755 Ma) (Table 1). Evidence from the geologic record supports this result because the continents are thought to have formed by the early Archean, 4000–3800 Ma (61, 76), and the first terrestrial ecosystems were present by 2600 Ma (77). The early colonization of land opens the possibility that later evolutionary adaptations (e.g., oxygenic photosynthesis) were favored by conditions in this environment. Moreover, Terrabacteria includes approximately twothirds of all prokaryote species (~9740) further suggesting a major influence of this environment on prokaryote speciation and adaptation (88). The sparse geologic record of prokaryotes and their biomarkers of carbon assimilation and lipid biosynthesis offer few opportunities to compare or test the molecular timescale (Fig. 2), especially since the best of such records were used as calibration points. We did not use the hopane biomarker initially proposed for cyanobacteria (78, 79) because it has turned up in other groups recently (80), thus removing it as a unique biomarker for that group. Our time estimate for the origin of cyanobacteria (2760 Ma), although young by classical interpretations based on early evaluations of stromatolites and fossils, is nonetheless older than—and thus consistent with—the rise of oxygen in the geologic record at ~2300 Ma (45). This time also overlaps with the possible sterane evidence for eukaryote algae and hence aerobic life at 2700 Ma (79, 81). However, the sterane biomarker evidence is contradicted by molecular clock evidence

114


that plastids—necessary for algae—did not arise in eukaryotes until ~1500 Ma (82, 83), and by the fossil record of eukaryotes that only extends reliably back to about ~1500 Ma (84–87). In summary, the sequence of biologic events inferred in the timetree is logical and consistent with the geologic record: the colonization of land after the appearance of continents and before the evidence of terrestrial ecosystems in the geologic record, methanogenesis followed by methanotrophy, and oxygenic photosynthesis followed by the major rise in atmospheric oxygen. With new genome sequences continually appearing, methods of time estimation constantly being refined, and new biomarkers discovered, the future appears bright for a well-defined timetree of eubacteria.

Acknowledgments Support was provided by the U.S. National Science Foundation and National Aeronautics and Space Administration (NASA Astrobiology Institute) to S.B.H.

References 1. G. M. Garrity et al., Taxonomic outline of the Bacteria and Archaea Release 7.7 (http://www.taxonomicoutline. org/) (Michigan State University, East Lansing, Michigan, 2007). 2. T. Cavalier-Smith, Int. J. Syst. Evol. Microbiol. 52, 7 (2002). 3. G. Webster et al., FEMS Microbiol. Ecol. 62, 78 (2007). 4. N. R. Pace, Science 276, 734 (1997). 5. G. E. Fox et al., Science 209, 457 (1980). 6. C. R. Woese, G. E. Fox, Proc. Natl. Acad. Sci. U.S.A. 74, 5088 (1977). 7. F. U. Battistuzzi, A. Feijão, S. B. Hedges, BMC Evol. Biol. 4, 44 (2004). 8. C. Brochier, S. Gribaldo, Y. Zivanovic, F. Confalonieri, P. Forterre, Genome Biol. 6, R42 (2005). 9. J. R. Brown, C. J. Douady, M. J. Italia, W. E. Marshall, M. J. Stanhope, Nat. Genet. 28, 281 (2001). 10. F. D. Ciccarelli et al., Science 311, 1283 (2006). 11. E. J. Deeds, H. Hennessey, E. I. Shakhnovich, Genome Res. 15, 393 (2005). 12. V. Daubin, M. Gouy, G. Perriere, Genome Res. 12, 1080 (2002). 13. U. Gophna, W. F. Doolittle, R. L. Charlebois, J. Bacteriol. 187, 1305 (2005). 14. S. B. Hedges et al., BMC Evol. Biol. 1, 4 (2001). 15. S. R. Henz, D. H. Huson, A. F. Auch, K. Nieselt-Struwe, S. C. Schuster, Bioinformatics 21, 2329 (2005). 16. C. H. House, B. Runnegar, S. T. Fitz-Gibbon, Geobiology 1, 15 (2003).

17. E. K. Lienau, R. DeSalle, J. A. Rosenfeld, P. J. Planet, Syst. Biol. 55, 441 (2006). 18. Y. I. Wolf, I. B. Rogozin, N. V. Grishin, E. V. Koonin, Trends Genet. 18, 472 (2002). 19. R. S. Gupta, E. Griffiths, Theor. Pop. Biol. 61, 423 (2002). 20. F. Delsuc, H. Brinkmann, H. Philippe, Nat. Rev. Genet. 6, 361 (2005). 21. W. F. Doolittle, Science 284, 2124 (1999). 22. V. Daubin, M. Gouy, G. Perrière, Genome Informat. 12, 155 (2002). 23. D. Pisani, J. A. Cotton, J. O. McInerney, Mol. Biol. Evol. 24, 1752 (2007). 24. T. Dagan, W. Martin, Genome Biol. 7, 10 (2006). 25. E. Bapteste et al., BMC Evol. Biol. 5, 33 (2005). 26. O. Zhaxybayeva, J. P. Gogarten, BMC Genomics 3, 4 (2002). 27. A. Mira, R. Pushker, B. A. Legault, D. Moreira, F. Rodriguez-Valera, BMC Evol. Biol. 4, 50 (2004). 28. R. S. Gupta, Crit. Rev. Microbiol. 30, 123 (2004). 29. R. S. Gupta, E. Lorenzini, BMC Evol. Biol. 7, 71 (2007). 30. H. Teeling, T. Lombardot, M. Bauer, W. Ludwig, F. O. Glockner, Int. J. Syst. Evol. Microbiol. 54, 791 (2004). 31. G. D. P. Clarke, R. G. Beiko, M. A. Ragan, R. L. Charlebois, J. Bacteriol. 184, 2072 (2002). 32. T. Coenye, P. Vandamme, Syst. Appl. Microbiol. 27, 313 (2004). 33. C. Brochier, H. Philippe, Nature 417, 244 (2002). 34. E. Griffiths, R. S. Gupta, Int. Microbiol. 7, 41 (2004). 35. J. Castresana, Mol. Biol. Evol. 17, 540 (2000). 36. A. Stamatakis, Bioinformatics 22, 2688 (2006). 37. F. Ronquist, J. P. Huelsenbeck, Bioinformatics 19, 1572 (2003). 38. J. M. Kollman, R. F. Doolittle, J. Mol. Evol. 51, 173 (2000). 39. A. H. Knoll, Life on a Young Planet: The First Three Billion Years of Evolution on Earth (Princeton University Press, Princeton, NJ, 2003). 40. J. J. Brocks, A. Pearson, Mol. Geomicrobiol. Rev. Mineral. Geochem. 59, 233 (2005). 41. H. Ochman, A. C. Wilson, J. Mol. Evol. 26, 74 (1987). 42. N. A. Moran, M. A. Munson, P. Baumann, H. Ishikawa, Proc. Roy. Soc. Lond. B 253, 167 (1993). 43. N. H. Sleep, K. J. Zahnle, J. F. Kasting, H. J. Morowitz, Nature 342, 139 (1989). 44. D.-F. Feng, G. Cho, R. F. Doolittle, Proc. Natl. Acad. Sci. U.S.A. 94, 13028 (1997). 45. H. D. Holland, Geochim. Cosmochim. Acta 21, 3811 (2002). 46. J. W. Schopf, Science 260, 640 (1993). 47. M. Brasier, N. McLoughlin, O. Green, D. Wacey, Philos. Trans. Roy. Soc. Lond. B 361, 887 (2006). 48. M. D. Brasier et al., Nature 416, 76 (2002). 49. J. W. Schopf, Philos. Trans. Roy. Soc. Lond. B 361, 869 (2006). 50. J. W. Schopf, A. B. Kurdryavtsev, D. G. Agresti, T. J. Wdowiak, A. D. Czaja, Nature 416, 73 (2002).

Eubacteria

51. P. P. Sheridan, K. H. Freeman, J. E. Brenchley, Geomicrobiol. J. 20, 1 (2003). 52. J. P. Gogarten et al., Proc. Natl. Acad. Sci. U.S.A. 86, 6661 (1989). 53. N. Iwabe, K. Kuma, M. Hasegawa, S. Osawa, T. Miyata, Proc. Natl. Acad. Sci. U.S.A. 86, 9355 (1989). 54. O. Zhaxybayeva, P. Lapierre, J. P. Gogarten, Protoplasma 227, 53 (2005). 55. R. G. Skophammer, J. A. Servin, C. W. Herbold, J. A. Lake, Mol. Biol. Evol. 24, 1761 (2007). 56. J. A. Servin, C. W. Herbold, R. G. Skophammer, J. A. Lake, Mol. Biol. Evol. 25, 1 (2008). 57. J. T. F. Wong, J. Chen, W. K. Mat, S. K. Ng, H. Xue, Gene 403, 39 (2007). 58. M. di Giulio, Gene 394, 105 (2007). 59. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 60. J. J. Brocks, G. D. Love, R. E. Summons, A. H. Knoll, G. A. Logan, S. A. Bowden, Nature 437, 866 (2005). 61. M. T. Rosing, D. K. Bird, N. H. Sleep, W. Glassley, F. Albarede, Palaeogeogr. Palaeoclimatol. Palaeoecol. 232, 99 (2006). 62. S. Gribaldo, C. Brochier-Armanet, Philos. Trans. Roy. Soc. Lond. B 361, 1007 (2006). 63. E. A. Gaucher, S. Govindarajan, O. K. Ganesh, Nature 451, 704 (2008). 64. L. J. Mix, D. Haig, C. M. Cavanaugh, J. Mol. Evol. 60, 153 (2005). 65. S. Sadekar, J. Raymond, R. E. Blankenship, Mol. Biol. Evol. 23, 2001 (2006). 66. D. A. Bryant, N.-U. Frigaard, Trends Microbiol. 14, 488 (2007). 67. J. Raymond, O. Zhaxybayeva, J. P. Gogarten, S. Y. Gerdes, R. E. Blankenship, Science 298, 1616 (2002).

115

68. M. X. Ruiz-Gonzalez, I. Marin, J. Mol. Evol. 58, 348 (2004). 69. J. F. Allen, W. Martin, Nature 445, 610 (2007). 70. D. E. Canfield, M. T. Rosing, C. Bjerrum, Philos. Trans. Roy. Soc. Lond. B 361, 1819 (2006). 71. M. M. Tice, D. R. Lowe, Geology 34, 37 (2006). 72. P. F. Dunfield et al., Nature 450, 879 (2007). 73. A. Pol et al., Nature 450, 874 (2007). 74. M. Wagner, M. Horn, Curr. Opin. Biotechnol. 17, 241 (2006). 75. P. V. Rokitko et al., Microbiology 72, 756 (2001). 76. C. J. Hawkesworth, A. I. S. Kemp, Nature 443, 811 (2006). 77. Y. Watanabe, J. E. Martini, H. Ohmoto, Nature 408, 574 (2000). 78. R. E. Summons, L. L. Jahnke, J. M. Hope, G. A. Logan, Nature 400, 554 (1999). 79. J. J. Brocks, G. A. Logan, R. Buick, R. E. Summons, Science 285, 1033 (1999). 80. S. E. Rashby, A. L. Sessions, R. E. Summons, D. K. Newman, Proc. Natl. Acad. Sci. U.S.A. 104, 15099 (2007). 81. J. J. Brocks, R. Buick, R. E. Summons, G. A. Logan, Geochim. Cosmochim. Acta 67, 4321 (2003). 82. H. S. Yoon, J. D. Hackett, C. Ciniglia, G. Pinto, D. Bhattacharya, Mol. Biol. Evol. 21, 809 (2004). 83. S. B. Hedges, J. E. Blair, M. L. Venturi, J. L. Shoe, BMC Evol. Biol. 4, 2 (2004). 84. N. J. Butterfield, Paleobiology 26, 386 (2000). 85. J. Samuelsson, N. J. Butterfield, Precambrian Res. 107, 235 (2001). 86. V. N. Sergeev, L. M. Gerasimenko, G. A. Zavarzin, Microbiology 71, 623 (2002). 87. A. H. Knoll, E. J. Javaux, D. Hewitt, P. Cohen, Philos. Trans. Roy. Soc. Lond. B 361, 1023 (2006). 88. F. U. Battistuzzi, S. B. Hedges, Mol. Biol. Evol. 26, 335 (2009).

Eukaryotes (Eukaryota) Debashish Bhattacharyaa,*, Hwan Su Yoonb, S. Blair Hedgesc, and Jeremiah D. Hackettd a

Department of Biological Sciences and Roy J. Carver Center for Comparative Genomics, 446 Biology Building, University of Iowa, Iowa City, IA 52242, USA; bBigelow Laboratory for Ocean Sciences, West Boothbay Harbor, ME 04575, USA; cDepartment of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA; dUniversity of Arizona, Ecology and Evolutionary Biology Department, Biosciences West 336, Tucson, AZ 85721, USA *To whom correspondence should be addressed ([email protected]).

Abstract Complex multicellular eukaryotes such as plants, animals, and fungi evolved independently from unicellular ancestors (protists). The relationships of these and other major groups of eukaryotes have proven difficult to determine because of a sparse fossil record and lack of a consensus among molecular phylogenies. Nonetheless, time estimates based on multiple nuclear genes indicate that eukaryotic photosynthetic organelles (plastids) arose from a symbiotic event with a cyanobacterium ~1600–1400 million years ago (Ma). The timetree suggests that most major groups of living eukaryotes also arose during the Mesoproterozoic (1600– 1000 Ma) in an evolutionary radiation probably associated with new niches created by eukaryotic algae.

Resolving the tree of life for eukaryotes is an important challenge for biologists. This challenge has largely been taken on by molecular evolutionists because of the increasing availability and ability of genome data to resolve ancient relationships, combined with the lack of an extensive fossil record from this period in eukaryotic history. The earliest eukaryote fossil that can be assigned to a living lineage is the sexual red algal fossil Bangiomorpha, which is dated at 1200 Ma (1, 2), and only a handful of other taxonomically resolved eukaryotes are known between then and the Ediacaran Period (~635 Ma) (3–5). All multicellular clades trace their roots to protist—eukaryotes that are not plants, animals, or fungi (6)—ancestors, therefore solving the basal splits in the tree with regard to protists (Fig. 1) as well providing a timeline for these events are of paramount importance.

The first attempts at generating a pan-eukaryotic tree of life relied on comparisons of rDNA genes and although these were highly useful, it turned out that a single-gene framework could not resolve all ancient protist relationships. In addition, like most single-gene markers, rDNA trees produced some controversial results, for example a tree with deeply branching lineages below and with a cluster of recently radiated lineages above (7, 8). This tree shape was partially explained by artifactual “longbranch” attraction of some (but not all) highly diverged protist lineages. The next step was to focus on multigene datasets of conserved well-studied proteins (e.g., 9–12) to increase the phylogenetic power. However, taxon representation in many of these analyses was sparse and marker choice was limited to a handful of proteins (e.g., actin, tubulins). In addition, attaining data from divergent phyla using degenerate PCR approaches proved to be very costly and time-consuming. This led to the current trend to generate eukaryotic trees using genome-wide (i.e., complete genome or expressed sequence tag, EST) data sets and large (e.g., >100) multiprotein alignments.

Fig. 1 Epistylis, a ciliate (upper left), Rhodosorus, a rhodophyte (upper right), Leptosiropsis torulosa, a chlorophyte (lower left), and Jakoba ibera, a jacobid excavate (lower right). Credits: D. J. Patterson and Aimlee Ladermann (upper left); B. Anderson and D. J. Patterson (upper right and lower left); and J. Cole and D. J. Patterson (lower right). Images provided by micro*scope (http://microscope.mbl.edu) under a creative commons license.

D. Bhattacharya, H. S. Yoon, S. B. Hedges, and J. D. Hackett. Eukaryotes (Eukaryota). Pp. 116–120 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

6

Chlorophyta 14

Embryophyta

3

Viridiplantae

Mycetozoa Glaucophyta

Bangiales 15

12

Bacillariophyta Dinophyceae

5

13

Aconoidasida 16

Coccidia

9

Hymenostomatida 17

Peniculida Euglenida

10

Trypanosomatidae Parabasalidea

7

Mesoproterozoic

Diplomonadida Neoproterozoic

PROTEROZOIC 1500

Alveolates

Oomycetes

2

Discicristates

Haptophyta 1

Stramenopiles

Cyanidiales

Amoebozoa

Acanthamoebidae 8

Plantae

Fungi

Rhodophyta

4

Chromalveolates

Metazoa

117

Excavates

Choanoflagellida 11

Opisthokonts

Eukaryota

1000

Pz

Mz

PHANEROZOIC 500

0 Million years ago

Fig. 2 A timetree of eukaryotes. Divergence times are shown in Table 1. Abbreviations: Mz (Mesozoic) and Pz (Paleozoic).

This phylogenomic approach has resulted in some notable successes with respect to protists (13–19), although these studies have been hampered by the quality of data (i.e., partial, single-pass EST reads), significant missing data, and sparse taxon sampling. In spite of these issues, the tree has begun to take shape and formed the basis for a classification scheme that defines six putative “supergroups” of eukaryotes. These are the Opisthokonta (e.g., animals, fungi, choanoflagellates), Amoebozoa (e.g., lobose amoebae, slime molds), Archaeplastida or hereafter Plantae (red algae, green algae and land plants, and glaucophyte algae), Chromalveolata (e.g., apicomplexans, ciliates, giant kelps), Rhizaria (e.g., cercomonads, foraminifera), and Excavata (e.g., diplomonads, parabasalids). Although the validity of some supergroups

(e.g., Chromalveolata, Excavata) is clearly in question (e.g., 20, 21), the supergroup concept is increasingly used (e.g., 22, 23) in the scientific literature and has permeated the field. Here we use primarily a recently published maximum likelihood (PhyML) phylogenetic hypothesis based on a 17-protein alignment generated by Hackett et al. (24) and a molecular clock method (Bayesian inference) that relaxes the requirement for a strict molecular clock (25) to estimate the dates of key nodes in the eukaryotic tree. Using this approach, the ML tree topology was first used to calculate the branch lengths with the program estbranches, using the JTT protein evolution model (26), before Bayesian estimation of divergence times using the program multidivtime (27). The ML tree was rooted

118


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among eukaryotes. Timetree Node

Estimates Ref. (12)

Time

1

1594

Ref. (24)(a)

Ref. (24)(b)

Time

CI

Time

CI

Time

CI

1827

–

1335

–

1620

–

2

1382

–

–

–

–

–

–

3

1379

1428

1579–1277

1250

1490–1060

1460

1740–1220

4

1368

1513

1642–1384

1370

1660–1150

1220

1440–1050

5

1345

–

–

1170

1400–980

1520

1830–1260

6

1225

–

–

1140

1350–970

1310

1560–1100

7

1220

–

–

820

1020–650

1620

2030–1310

8

1215

–

–

1250

1520–1040

1180

1400–990

9

1135

–

–

990

1200–810

1280

1570–1040

10

1090

–

–

740

930–570

1440

1790–1160

11

1020

–

–

1070

1270–910

970

1120–840

12

985

–

–

860

1060–680

1110

1370–890

13

940

–

–

820

1010–640

1060

1320–830

14

936

968

1150–786

870

1010–770

970

1130–840

15

935

–

–

870

1060–710

1000

1220–810

16

705

–

–

610

780–460

800

1020–590

17

640

–

–

560

730–410

720

940–520

Note: Node times in the timetree represent the mean of time estimates from different studies and methods. For Node 2, an average of times from three studies was used (30–32) (see text). For ref. (24), the two columns represent alternative rootings: (a) opisthokont and (b) diplomonads + parabasalids. For Node 1, see text for method; these times are averages of multiple time estimates, each with confidence and credibility intervals as presented in the original references.

on either the branch leading to the opisthokonts (28) or the diplomonads + parabasalids branch due to uncertainty about the placement of the root (29). We placed eight time constraints on this analysis based on the fossil record (24). To accommodate the variation in split times due to the use of two different rooting schemes, we calculate the mean divergence time for the two alternate roots. When available, the relevant dates from Hedges et al. (12) were included in averages (Table 1). The phylogenetic position of haptophytes has been controversial, although several recent multigene studies have agreed in their placement as the closest relative of Chromalveolates + Rhizaria rather than nested within that group (15, 16, 21). Although haptophytes were not included in the two molecular clock studies used here (Table 1), their divergence from other major groups of eukaryotes has been timed in three other studies (30–32), resulting in a wide range of estimates (1900–1047 Ma) and an average of 1382 Ma. Therefore, we tentatively place them in the timetree (Fig. 2) in that position.

For estimating the root time (~1600 Ma) of the timetree (Fig. 2), we averaged the times of origin of the six included lineages in the polytomy from the two studies, Hackett et al. (24) (1335 and 1620 Ma, depending on the root) and Hedges et al. (12) (1827 Ma). Clearly, better knowledge of the phylogenetic position of the root will improve the time estimate of this node; the current estimate should be considered tentative. In this timetree, the split of fungi and animals was at ~1368 Ma, the split between stramenopiles and alveolates was at ~1345 Ma, and the origin of the photosynthetic organelle (plastid) in the Plantae ancestor was between ~1600 and ~1400 Ma (i.e., on the branch leading from the root to this supergroup). This time range for the primary origin of plastids (33) agrees with the conclusions of earlier molecular clock studies focused on the origin of plastids (30) as well as the separate studies used in constructing the current timetree (12, 24). Moreover, it corresponds to earliest undisputed eukaryotes (algae) in the fossil record; records from

Eukaryota

before ~1600 Ma are debated as possibly being colonial prokaryotes (1, 34–36). The striking pattern evident in the timetree (Fig. 2) is that nearly all of the divergences occurred in the Mesoproterozoic and earliest Neoproterozoic (~1600– 900 Ma), in a relatively rapid evolutionary radiation. The likely explanation is the origin of plastids, thus creating eukaryotic algae, an increase in productivity, and an increase in ecological niches allowing diversification (40). Are there other pan-eukaryotic molecular clock analyses that conflict with the results described earlier? The most comprehensive work in this regard is the multiprotein analysis by Douzery et al. (37), who suggested that the initial split among living eukaryotes was only 1085 Ma (1259–950 Ma), the split between animals and fungi 984 Ma (1127–872 Ma), and the important split between red algae and Viridiplantae 928 Ma (1061– 825 Ma), with other dates equally young compared with previous estimates discussed earlier. Reanalyses of the Douzery et al. (37) data set were made by Roger and Hug (38) and Hug and Roger (39), who questioned the results and found that they were sensitive to the calibrations used. The specific calibrations and calibration methods used by Douzery et al. were also questioned elsewhere (40). In the study of Douzery et al. (37), each minimum calibration constraint was fi xed as the younger boundary of the major geologic period containing the pertinent fossil rather than to the actual (older) geologic time constraints of the fossil itself, thus causing underestimates of resulting times. Douzery et al. also fi xed maximum calibration constraints, arbitrarily, to the older boundary of the major geologic period containing the fossil rather than to an evolutionary event that might bear on the constraint. For example, the maximum calibration for the split of actinopterygian fish from mammals, 417 Ma, was essentially the same time as the oldest fossil on either branch, 416 Ma (41). However, there is little fossil information from this time period (Silurian) to establish that the divergence occurred precisely when the fossils appeared; more than likely it was much earlier, which would result in older Bayesian posterior time estimates. In addition, one of the maximum calibrations, the split between chelicerates and other arthropods (543 Ma), was fi xed within the Cambrian, which is problematic because there is not an extensive fossil record before that period, showing morphological transitions, which could provide support for the use of a maximum calibration. A separate reanalysis of the Douzery et al. (37) data set (S. B. Hedges, unpublished data) was conducted using

119

the same methods and calibration taxa as the original authors but with corrections made to minimum calibrations only, based on the fossil record. This resulted in older time estimates of 1134 Ma (Dictyostelium root) and 1265 Ma (kinetoplastid root) for the initial split among living eukaryotes. However, when the problematic Cambrian arthropod maximum calibration was removed, and the actinopterygian–sarcopterygian maximum calibration was adjusted from 417 Ma (earliest fossil) to 495 Ma (more realistic), the initial eukaryote split was much older: 1857 Ma (Dictyostelium root) and 2216 Ma (kinetoplastid root). Therefore, this reanalysis of Douzery et al. (37) and those by Roger and Hug (38) and Hug and Roger (39) all agree that it was not the data set and relaxed clock method of Douzery et al. that resulted in relatively young times but rather the calibrations used. After correcting for calibration errors, significantly older times are produced, concordant now with the results of other studies. Another analysis of note is that of Berney and Pawlowski (42), who used a broadly sampled data set of 240 small subunit rRNA genes. These authors were able to incorporate many more fossil dates (four maximum and 22 minimum constraints) in their analyses than genome-wide analyses due to the larger number of taxa in their tree that have a fossil record (e.g., coccolithophorids and diatoms). The increase in taxa and fossil constraints may have come with a loss of phylogenetic power due to the use of a single-gene framework. Other studies have documented the difficulties in inferring eukaryote-wide trees using single genes that sometimes show extreme rate variation and poor topological resolution. Moreover, according to de Vargas et al. (31), the resulting time estimates of Berney and Pawlowski (42) are underestimates attributed to the use of incorrect fossil dates for their clock calibration. The use of correct fossil calibrations would result in older time estimates (31). Discovery that the relatively young time estimates for divergences among eukaryotes, found by Douzery et al. (37) and Berney and Pawlowski (42), are the result of miscalibrations goes a long way toward reconciling the differences between those studies and others bearing on the timescale of eukaryote evolution (e.g., 12, 24, 30). In addition, the conflict regarding the time estimate (928 Ma) for the divergence of red and green algae in Douzery et al. (37), 300 million years younger than the earliest fossil at ~1200 Ma (1), is apparently resolved. Other time estimates for this divergence (Table 1, ~1380) are compatible with the fossil record. It also supports an earlier rise in organismal complexity, as measured in cell types

120


(12). In the future, studies that include a broader sample of taxa will permit more and better-constrained fossil calibrations that will presumably result in more reliable time estimates for the early evolutionary history of eukaryotes.

Acknowledgments Support was provided by U.S. National Science Foundation (NSF) and National Aeronautics and Space Administration (NASA) to D.B. by NSF and the NASA Astrobiology Institute to S.B.H. and by the U.S. National Institute of Health through training grant support to J.D.H.

References 1. 2. 3. 4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

N. J. Butterfield, Paleobiology 26, 386 (2000). N. J. Butterfield, Precambrian Res. 111, 235 (2001). N. J. Butterfield, Paleobiology 30, 231 (2004). A. H. Knoll, Life on a Young Planet: The First Three Billion Years of Evolution on Earth (Princeton University Press, Princeton, NJ, 2003). S. M. Porter, Paleontol. Soc. Pap. 10, 35 (2004). D. J. Patterson, Am. Nat. 154, S96 (1999). J. C. Edman et al., Nature 334, 519 (1988). M. L. Sogin, J. D. Silberman, Int. J. Parasitol. 28, 11 (1998). S. L. Baldauf, A. J. Roger, I. Wenk-Siefert, W. F. Doolittle, Science 290, 972 (2000). E. Bapteste, H. Philippe, Mol. Biol. Evol. 19, 972 (2002). H. S. Yoon, J. D. Hackett, G. Pinto, D. Bhattacharya, Proc. Natl. Acad. Sci. U.S.A. 99, 15507 (2002). S. B. Hedges, J. E. Blair, M. L. Venturi, J. L. Shoe, BMC Evol. Biol. 4, 2 (2004). N. Rodriguez-Ezpeleta et al., Curr. Biol. 15, 1325 (2005). J. E. Blair, P. Shah, S. B. Hedges, BMC Bioinf. 6, 53 (2005). J. D. Hackett et al., Mol. Biol. Evol. 24, 1702 (2007). F. Burki et al., PLoS ONE 2, e270 (2007). N. J. Patron, Y. Inagaki, P. J. Keeling, Curr. Biol. 17, 887 (2007). N. Rodriguez-Ezpeleta et al., Syst. Biol. 56, 389 (2007). N. Rodriguez-Ezpeleta et al., Curr. Biol. 17, 1420 (2007). L. W. Parfrey et al., PLoS Genet. 2, 2062 (2006).

21. H. S. Yoon et al., BMC Evol. Biol. 8, 14 (2008). 22. S. L. Baldauf, Science 300, 1703 (2003). 23. P. J. Keeling et al., Trends. Ecol. Evol. 20, 670 (2005). 24. J. D. Hackett, H. S. Yoon, N. J. Butterfield, M. J. Sanderson, D. Bhattacharya, in Evolution of Primary Producers in the Sea, P. G. Falkowski, A. H. Knoll, Eds. (Elsevier, Burlington, 2007), pp. 109–131. 25. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 26. D. T. Jones, W. R. Taylor, J. M. Thornton, Comput. Appl. Biosci. 8, 275 (1992). 27. H. Kishino, J. L. Thorne, W. J. Bruno, Mol. Biol. Evol. 18, 352 (2001). 28. A. Stechmann, T. Cavalier-Smith, Science 297, 89 (2002). 29. N. Arisue, M. Hasegawa, T. Hashimoto, Mol. Biol. Evol. 22, 409 (2005). 30. H. S. Yoon, J. D. Hackett, C. Ciniglia, G. Pinto, D. Bhattacharya, Mol. Biol. Evol. 21, 809 (2004). 31. C. de Vargas, M.-P. Aubry, I. Probert, J. Young, in Evolution of Primary Producers in the Sea, P. G. Falkowski, A. H. Knoll, Eds. (Elsevier Academic Press, Burlington, Massachusetts, 2007), pp. 251–285. 32. L. K. Medlin, A. G. Saez, J. R. Young, Mar. Micropaleontol. 67, 69 (2008). 33. D. Bhattacharya, H. S. Yoon, J. D. Hackett, BioEssays 26, 50 (2004). 34. J. Samuelsson, N. J. Butterfield, Precambrian Res. 107, 235 (2001). 35. V. N. Sergeev, L. M. Gerasimenko, G. A. Zavarzin, Microbiology 71, 623 (2002). 36. A. H. Knoll, E. J. Javaux, D. Hewitt, P. Cohen, Philos. Trans. Roy. Soc. Lond. B 361, 1023 (2006). 37. E. J. P. Douzery, E. A. Snell, E. Bapteste, F. Delsuc, H. Philippe, Proc. Natl. Acad. Sci. U.S.A. 101, 15386 (2004). 38. A. J. Roger, L. A. Hug, Philos. Trans. Roy. Soc. Lond. B 361, 1039 (2006). 39. L. A. Hug, A. J. Roger, Mol. Biol. Evol. 24, 1889 (2007). 40. S. B. Hedges, F. U. Battistuzzi, J. E. Blair, in Neoproterozoic Geobiology and Paleobiology, S. Xiao, A. J. Kaufman, Eds. (Springer, New York, 2006), pp. 199–229. 41. M. J. Benton, P. C. J. Donoghue, Mol. Biol. Evol. 24, 26 (2007). 42. C. Berney, J. Pawlowski, Proc. Roy. Soc. Lond. B 273, 1867 (2006).

PROTISTS


Haptophyte algae (Haptophyta) Linda K. Medlin Marine Biological Association of the UK, The Citadel, Plymouth PL1 2PB, UK ([email protected])

Abstract Haptophytes are members of the marine phytoplankton involved in many important biochemical cycles. They possess two smooth flagella and another organelle, called a haptonema inserted between the flagella. The cells are covered by organic scales, which are calcified in one order, the Coccolithales, permitting molecular clock calibration. Time estimates place the divergence of the two classes in the Neoproterozoic, ~800 million years ago (Ma), with order-level diversification occurring in the Phanerozoic, ~340–120 Ma. Selective survival of different orders across major extinction events may be related to the ability of the cells to switch their mode of nutrition from autotrophy to mixotrophy.

Haptophytes (Fig. 1) occur in all seas and are often major components of the nanoplankton (1, 2). They are important primary producers, and some species in the genera Emiliania, Gephyrocapsa, Phaeocystis, Chrysochromulina, and Prymnesium may form extensive blooms with major biogeochemical, ecological, or economic impact. Most species are marine, but a few thrive in freshwater. Most are unicellular, planktonic biflagellates, but palmelloid, coccoid, amoeboid, colonial, and benthic forms also occur (3). Nearly all are photosynthetic, but phagotrophy and mixotrophy appears to be common in some genera (e.g., Chrysochromulina) (4). In most species, at least one stage in their haplo-diplont life cycle possesses two flagella that are similar in form and have no tubular hairs. Between the flagella is a unique organelle, called a haptonema, which differs structurally from the flagellum. Its length varies and it has been secondarily lost in some species. It can coil or bend, but not beat, and can attach to a substratum and may be involved in food handling. Cells are typically covered by one to several layers of organic scales and in the coccolithophorids these are calcified. These are preservable and constitute the feature that leaves a fossil record for calibration of a molecular

clock. Species identification within Haptophyta is largely based on scale morphology and often requires electron microscopy. Two molecular clocks have been made for the haptophytes by Medlin and her coworkers: a strict molecular clock using the Lintree program that averages the rate of evolution across all lineages (5, 6) and a relaxed molecular clock (r8s) where the rate of evolution is allowed to vary across the lineages (7, 8). Both clocks were calibrated using at least three calibration points from the coccolith fossil record: the character-based constraint of 195 Ma for the emergence of all coccolithophores, and the divergence-based constraints of 64 Ma for the divergence of Coccolithus from Cruciplacolithus and

Fig. 1 Chrysochromulina (Prymnesiales) with arrow indicating long haptonema (upper left), Phaeocystis (Phaeocystales) with arrow indicating short haptonema (upper right), colony of Phaeocystis antarctica (lower left), and coccolithophore Emiliania huxleyi (lower right). Credits: W. Eikrem (upper left), L. K. Medlin (upper right and lower left), and J. Green (lower right).

L. K. Medlin. Haptophyte algae (Haptophyta). Pp. 123–126 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

124


Isochrysidales 5

7

Prymnesiales-3

6 2

Prymnesiales-2 8

Prymnesiales-1 Phaeocystales-2

1

9

Phaeocystales-1 Pavlovales-2

3

Neoproterozoic PROTEROZOIC 800

600

Paleozoic

Pavlovales-1 Mesozoic

Cz

Pavlovophyceae

Prymnesiales-4

4

Prymnesiophyceae

Coccolithales

PHANEROZOIC 400

200

0 Million years ago

Fig. 2 A timetree of Haptophyta. Divergence times are shown in Table 1. Pavlovales-1 = Exanthemachrysis with pigment type A (15); Pavlovales-2 = Pavlova and Diacronema with pigment type B and C (15); Phaeocystales-1 = unicellular Phaeocystis sp. (6); Phaeocystales-2 = colonial Phaeocystis sp. (6); Prymnesiales-1 = Clade B2 saddle-shaped Chrysochromulina sp. including

C. parva; Prymnesiales-2 = Clade B2 other saddle-shaped Chrysochromulina sp.; Prymnesiales-3 = Clade B1 Imantonia sp. and OLI clones; and Prymnesiales-4 = Clade B1 round, not saddle-shaped Chrysochromulina sp., Prymnesium sp., and Platychrysis sp. Details of species in each clade are presented elsewhere (1). Abbreviation: Cz (Cenozoic).

50 Ma for the divergence of Helicosphaeraceae from Pontosphaeraceae. We have constructed a molecular clock from the small subunit ribosomal RNA (SSU rRNA) gene and extrapolated dates for some of the undated nodes where there is no fossil evidence. The SSU rRNA tree appears to be evolving in a clocklike manner as judged by the relative rate tests performed in addition to the use of the Lintree program (9). Another molecular clock study of haptophytes has been done, using the SSU and large subunit (LSU) rRNA genes (10). Dates for the divergences in that study are slightly older than those found by Medlin and coworkers using a relaxed molecular clock (8). Although not treated in detail here, Haptophyta is a group that diverged from other eukaryotes deep in the Proterozoic, >1200 Ma (9–11). The long time period between the origin of haptophytes and the initial divergence (~800 Ma) of the two classes, Pavlovophyceae and Prymnesiophyceae (Table 1, Fig. 2), indicates that many of the early evolutionary branches in this group are extinct, or that they have not yet been sampled (2). The Order Phaeocystales diverged from all other Prymnesiophyceae at ~480 Ma and then the Prymnesiales diverged from the Coccolithales plus Isochrysidales at ~280 Ma, and

thus this divergence appears to be a late Paleozoic–early Mesozoic event and may be associated with the Permian– Triassic boundary (251 Ma). Modern diversifications in these lineages occurred some time after the lineage origin so many taxa were presumably lost during this time. Within the Order Phaeocystales, the divergence of the cold water clades from the warm water clades occurs at 30 Ma, when the Drake Passage opened to isolate the Antarctic Continental waters, and dispersal to the Artic occurred across the equator during a cooling trend at 15 Ma, which were separated by a warming trend that then isolated the two polar clades (6). Molecular diversification occurred earlier within the Prymnesiales than within the Coccolithales plus Isochrysidales where most of these latter divergences occurred fairly late in the haptophyte timetree (Fig. 2). The diversification within the Coccolithales plus Isochrysidales occurred predominantly after the Mesozoic–Cenozoic boundary (66 Ma), as predicted by the fossil record. Mesozoic coccolithophores have been intensively studied and at the Mesozoic–Cenozoic boundary an abrupt extinction is documented in the fossil record with ~90% of end-Cretaceous species disappearing (e.g., 12, 13). Subsequently, there was a major

Eukaryota; Haptophyta

125

Table 1. Divergence times among haptophytes. Timetree

Estimates

Node

Time

Ref. (8) Time

Ref. (9) Time

Ref. (10)(a) Time

Ref. (10)(a) Time

1

800

800

500

870

1000

2

480

480

200

400

290

3

280

280

–

300

330

4

280

280

–

300

210

5

200

200

–

250

150

6

200

200

–

180

150

7

175

175

–

–

–

8

150

150

–

–

–

9

120

120

–

–

–

Note: The node times in the timetree are based on ref. (8). Estimates from ref. (10) are based on (a) SSU rRNA and (b) LSU rRNA analyses.

radiation in the early Cenozoic with new clades rapidly diversifying and forming the origins of the modern coccolithophore biota (e.g., 14). One novel inference from our molecular tree is that the Mesozoic–Cenozoic boundary extinction does not seem to have affected the Prymnesiales, Phaeocystales, or Pavlovales to the same degree as the Coccolithales. These orders do not have a fossil record so we can only make this statement by comparing the depth of clade diversification. In each of these noncalcifying groups, there are numerous clades/lineages that cross the Mesozoic–Cenozoic boundary (8). There is no evidence of major diversification of these clades in the Cenozoic. On this basis, one would expect that the noncalcified haptophytes would have the same rate of extinction as the calcified ones, because no group produces resting stages, although some species have benthic littoral to sublittoral stages as part of their dimorphic life cycle. No haptophytes are known to produce specialized resting cells or zygotes analogous to dinoflagellate cysts or diatom resting spores. There is no evidence of bottlenecking in the noncalcified taxa at this time, as illustrated by the many clades with deeper divergences (Fig. 2). One possible explanation for this difference in their survival may lie in the mode of nutrition in the haptophyte lineages. The noncalcifying haptophytes are known for their ability to switch between autotrophic and heterotrophic nutrition (3). Thus, when nutrients are plentiful, they photosynthesize. However, when the

reverse is true, they engulf prey and survive heterotrophically. At the Mesozoic–Cenozoic boundary, it is likely that light quality was reduced and photosynthetic ability was impaired. Therefore, those taxa with either the ability to form resting stages, such as the diatoms and the dinoflagellates, or the ability to switch their mode of nutrition could have an adaptive advantage over those that did not have either of these traits. Coccolithophores are not known to form resting stages, in the strictness sense, and it appears that they are predominantly obligate autotrophs. Thus, at the Mesozoic–Cenozoic boundary, the stress induced by reduced light quantity and quality could have shut down photosynthesis. Cells that could switch nutrition or form resting stages would have had a better chance of survival. In summary, the haptophytes are a major eukaryotic group of microalgae whose closest relative is unclear. The initial class level divergence occurred in the Neoproterozoic and divergence of the orders appears to be associated with the Permian–Triassic boundary. Because this is a host lineage with a red algal plastid, it is likely that the group radiated at the Permian–Triassic boundary when the ocean chemistry changed to give the red algal plastid an adaptive advantage over host cells with a green algal plastid, which were common in the plankton before the end Permian. There appears to be a selective extinction of the Order Coccolithales at the Mesozoic–Cenozoic boundary where calcified organisms were affected by ocean chemistry, and the uncalcified

126


lineages likely switched to mixotrophy to take advantage of the poor light conditions at this extinction event.

References 1. 2.

3.

4.

5. 6.

B. Edvardsen et al., Phycologia 39, 19 (2000). B. Edvardsen. L. K. Medlin, in Unravelling the Algae—The Past, Present and Future of Algal Molecular Systematics, J. Brodie, J. Lewis, Eds. (Systematics Association, London, 2007), pp. 183–196. D. J. Hibberd, in Developments in Marine Biology, E. R. Cox, Ed. (Elsevier North Holland, New York, 1980), pp. 273–317. H. L. J. Jones, B. S. C. Leadbeater, J. C. Green, in The Haptophyte Algae, J. C. Green, B. S. C. Leadbeater, Eds. (Systematics Association and Clarendon Press, Oxford, 1994), pp. 247–263. N. Takezaki, A. Rzhetsky, M. Nei, Mol. Biol. Evol. 12, 823 (1995). L. K. Medlin, A. Zingone, Biogeochemistry 83, 3 (2007).

7.

8. 9.

10.

11. 12. 13. 14.

15.

M. Sanderson, r8s version 1.71. Estimating Rates of Molecular Evolution, http://ginger.ucdavis.edu/r8s (University of California, Davis, 2006). L. K. Medlin, A. G. Saez, J. R. Young, Mar. Micropaleontol. 67, 69 (2007). L. K. Medlin, W. C. H. F. Kooistra, D. Potter, G. W. Saunders, R. A. Andersen, Plant Syst. Evol. (Suppl) 11, 187 (1997). C. de Vargas, M. P. Aubry, I. Probert, J. Young, in Evolution of Primary Producers in the Sea, P. G. Falkowski, A. H. Knoll, Eds. (Elsevier, Amsterdam, 2007), pp. 251–285. H. S. Yoon, J. D. Hackett, C. Ciniglia, G. Pinto, D. Bhattacharya, Mol. Biol. Evol. 21, 809 (2004). N. MacLeod et al., J. Geol. Soc. Lond. 1254, 265 (1997). P. R. Bown, Geology 33, 653 (2005). P. R. Bown, J. A. Lees, Y. R. Young, in Coccolithophores— From Molecular Processes to Global Impact, H. R. Thierstein, J. R. Young, Eds. (Springer-Verlag, BerlinHeidelberg, 2004), pp. 481–508. K. Van Lenning et al., J. Phycol. 39, 379 (2003).

Diatoms (Bacillariophyta) Linda K. Medlin Marine Biological Association of the UK, The Citadel, Plymouth PL1 2PB, UK ([email protected])

Abstract The diatoms are one of the best characterized microalgal groups because of the unique features of their silicified cell wall, which is preservable. Modern diatoms constitute three classes, which have an extensive fossil record since the late Cretaceous (100–66 million years ago, Ma). Diatoms from the early Cretaceous (145–100 Ma) are rare and differ in the morphology of their silica wall from modern diatoms. Molecular clocks indicate that the divergences of classes and orders took place in the Triassic (251–200 Ma) and Jurassic (200–145 Ma), following the Permian–Triassic extinctions (251 Ma).

The diatoms are one of the most easily recognizable groups of major eukaryotic algae, because of their unique silicified cell walls (frustules), which consist of two overlapping thecae, each in turn consisting of a valve and a number of hooplike or segmental girdle bands (1). Well-preserved frustules are found in the earliest known deposits of fossil diatoms, from the early Albian (~125 Ma) of what is now the Weddell Sea, Antarctica, but these diatoms bear no resemblance to modern diatoms in their morphology (2). Molecular sequence data show that diatoms are heterokont algae. Heterokonts are chlorophyll a+c containing algae whose motile cells have two heterodynamic flagella, one covered with tripartite hairs and the other smooth (3). In diatoms, the flagellar apparatus is reduced or absent; indeed, only the spermatozoids of the oogamous “centric” diatoms are flagellated and these are uniflagellate (4), lacking all trace of a smooth posterior flagellum. Today, the diatoms are found in almost all aquatic and most wet terrestrial habitats. Existing hypotheses of diatom origins tend to agree that the pre-diatom or “Ur-diatom” developed from a scaly ancestor, not in pelagic habitats, but in shallow marine environments and were tychoplanktonic (bottom dwelling). Phylogenetic analyses have documented several invasions and

radiations into freshwater (1) and have documented that their closest relative is a flagellated picoplankter, the Bolidophyceae. Most diatomists have long assumed that the diatoms contain two groups: the centrics and the pennates, which can be distinguished by their pattern centers or symmetry, mode of sexual reproduction, and plastid number and structure (5). The oogamous centric diatoms, with radially symmetrical ornamentation on their valves and with numerous discoid plastids, are distinct from the isogamous pennate diatoms with bilaterally symmetrical pattern centers and generally fewer, platelike plastids. These groups are known to most aquatic and cell biologists under these terms, and each term conveys a distinct image of a particular type of diatom valve.

Fig. 1 Diatoms. Typical representatives of a radial centric, Class Coscinodiscophyceae (upper left), bipolar multipolar centric, Class Mediophyceae (upper right), an araphid pennate, Class Bacillariophyceae (lower left) and a raphid pennate, Class Bacillariophyceae (lower right) diatom. Credits: L. K. Medlin.

L. K. Medlin. Diatoms (Bacillariophyta). Pp. 127–130 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Fragillariophycidae-3 8

Fragillariophycidae-2

6 5

Bacillariophycidae

7

Eunotiophycidae Fragillariophycidae-1

3

Attheyaceae Radial Centrics Group

2

Tr

Ellerbeckiaceae Jurassic

Cretaceous

MESOZOIC 200

150

Paleogene

Ng

Coscinodiscophyceae

Multipolar Centrics Group 4

1

Mediophyceae


Bacillariophyceae

128

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of diatoms. Divergence times are from Table 1. Taxa shown are at the subclass and order level. Fragillariophycidae-1 are the basal araphids (17), Fragillariophycidae-2 are the core araphids with labiate processes, and Fragillariophycidae-3 are the core araphids wiithout labiate processes. Abbreviations: Ng (Neogene) and Tr (Triassic).

Historically, centric and pennate diatoms have been separated as two classes or orders. Round et al. (5), however, recognized three classes—Coscinodiscophyceae (centric diatoms), Fragilariophyceae (araphid pennate diatoms), and Bacillariophyceae (raphid pennate diatoms)—giving equal ranking to the raphid pennate diatoms (those with a slit opening, raphe, in the cell wall for movement) and the araphid pennate diatoms (those without this slit). Williams (6) has traced the historical classification of the diatoms. Recently, Medlin and colleagues (literature summarized by Medlin and Kaczmarska, 7) have divided the diatoms into two groups on the basis of molecular sequence data. What was initially called “Clade 1” in early molecular work contains those centric diatoms with essentially radial symmetry of valve shape and structure. “Clade 2” consists of two groups, the first of which contains the bi- or multipolar centrics and the radial Thalassiosirales (“Clade 2a”), and the second, the pennates (“Clade 2b”) (Fig. 1). Morphological and cytological support for these clades was reviewed in Medlin et al. (8) and Medlin and Kaczmarska (7). Clades 1 and 2 are now recognized at the subdivision level (Fig. 2), as the Coscinodiscophytina and Bacillariophytina, respectively, and Clades 1, 2a, and 2b are now recognized at the class level, as the Coscinodiscophyceae, Mediophyceae, and Bacillariophyceae (7) (Fig. 2). The first two classes

are not recovered as monophyletic if alignments are not performed using the secondary structure of the rRNA genes as a guide. There are correlations between the principal molecular clades and certain cytological features. For example, on the whole the Bacillariophytina have a perinuclear arrangement of the Golgi apparatus, whereas in the Coscinodiscophytina, the Golgi stacks are usually in Golgi–endoplasmic reticulum–mitochondrion (G-ER-M) units (7, 8, 9). However, there are some exceptions (8). The best independent, nonmolecular support for three classes comes from auxospore structure, the specialized zygote of the diatoms that swells to restore the cells to their original cell size, which has diminished with each vegetative division (7, 10). Isodiametric auxospores that can swell in all directions and have only scales are characteristic of Clade 1, anisodiametric auxospores with scales and hoops or bands (a properizonium) to restrict the swelling to bipolar or multipolar directions are found in Clade 2a, and anisodiametric auxospores that form a complex tubular perizonium, usually consisting of transverse hoops and longitudinal bands, are found in Clade 2b. The rate of evolution in Clade 1 diatoms has been calculated using two different means of calibrating molecular trees: by fossil dates within the entire clade (11), and by biomarker compounds for the clade containing

Eukaryota; Bacillariophyta

129

Table 1. Divergence times (Ma) among diatoms. Timetree

Estimates

Node

Time

Ref. (11)(a) Time

Ref. (11)(b) Time

Ref. (14)(a) Time

Ref. (14)(b) Time

Ref. (15) Time

Ref. (16)(a) Time

Ref. (16)(b) Time

1

207.5

120

200

380

330

–

180

235

2

198.5

–

–

–

–

–

172

225

3

195.0

86

159

–

–

–

170

220

4

190.0

–

–

–

–

–

165

215

5

161.0

–

–

–

–

–

140

182

6

146.5

84

167

–

–

96.5

127

166

7

145.0

–

–

–

–

–

125

165

8

136.5

–

–

–

–

–

118

155

Note: The node times in the timetree are the average of the minimum (a) and maximum (b) times reported in ref. (16). Average (a) and upper limit (b) are reported from ref. (11), and results from the analysis of SSU rRNA (a) and large subunit (LSU) rRNA (b) are reported from ref. (14).

Rhizosolenia (12). Both methods suggest that these diatoms are evolving very quickly (1% per 21.5 and 14 Ma for the rRNA gene, respectively) and this could explain why the morphology of the diatoms changes so rapidly across the Cretaceous, between Lower (146–100 Ma) and Upper (100–66 Ma) Cretaceous floras. Several people have constructed molecular clocks for the diatoms and most are concerned with dating the origin of the group and the diversification of its major clades, which would be recognized at the class level, when these are recovered as monophyletic groups. Kooistra and Medlin (11) made the first molecular clock using a linearized tree where the rate of evolution was averaged across the tree using other pigmented heterokonts as the outgroup. Using the Hillis and Morris model for their clock (11), they were able to calculate an average age and an earliest possible age given a 95% confidence interval around any undated node. In this clock, they inferred the origin of the diatoms to be 266–164 Ma (earliest to average). The major clades that constitute the two subdivisions of the diatoms were estimated to have diverged between 200 and 120 Ma (earliest to average). Clade 2 comprising the bipolar centrics and the pennates diverged between 159 and 86 Ma. The clock prepared by Phillippe et al. (13) used the ciliates as the closest relative and the origin of the diatoms calculated in that paper corresponded to the origin of the heterokonts, which is the division to which the diatoms belong, probably because too distant an outgroup was used to root the ingroup. This places in doubt the 300million-year gap in the diatom fossil record concluded

in that study. In a second clock paper by Sörhannus from the same group (14), the chrysophytes were used as the closest relative to the diatoms, which are likely still too far away from the ingroup, but this gave a more reliable date of 400–330 Ma for the origin of the diatoms based on two genes. The most recent study from Sörhannus (15) was based on a single gene, the small subunit (SSU) rRNA gene and used the closest known relative of the diatoms, the Bolidophyceae, as the outgroup. A relaxed molecular clock (PATHd8) was used, which was calibrated from single dating points sequentially. Resulting dates for the origin of the diatoms ranged from 250 to 183 Ma, which is similar to the time estimates Kooistra and Medlin (11). Sato et al. (16), using four genes and the Bayesian program Multidivtime, with designated maximum and minimum divergence times of 250 and 190 Ma, respectively, found much older divergence times for all of the classes, especially the pennates. Their clock using an ML tree as input suggests that the radial centrics, Class Coscinodiscophyceae emerged 235–180 Ma, the bipolar centrics, Class Mediophyceae split from the Class Bacillariophyceae 220–170 Ma (maximum to minimum). Those results also indicate that the early divergence of the pennates into three major clades, basal araphid, core araphid, and raphid diatoms, took place in a very short period. All major clades (orders or families) of araphid diatoms appeared by the end of the Cretaceous in all analyses. All of the divergences of these lineages took place long before there is a fossil record of diatoms. However, modern diversifications of the genera in these lineages

130


usually coincide with the first appearances of the modern genera. The reconciliation of molecular diversification with first appearances of selected genera of diatoms is discussed in Sörhannus (15), Kooistra and Medlin (11), and Sims et al. (1). The modern classification of the araphid diatoms will need to be extensively revised because the group is paraphyletic. Among the raphid diatoms, most orders and the families therein have been found to be monophyletic groups, but there are some genera that are not monophyletic and revisions will be needed there as well. It seems that features of the living cell—the zygote morphology and development—define and better support the deeper branches of the tree, whereas the details of the silica cell wall, upon which the classification of the diatoms is based, can best be used to define the tips and younger branches of the tree. Many well-documented genera, such as Diatoma, Fragilariopsis, and Pleurosigma, arise from within other genera in the trees, making the parent genus paraphyletic. This is problematic for taxonomists and therefore extensive generic redefinitions will be needed unless those paraphyletic definitions are accepted. In summary, the molecular diversification of the diatoms appears to be much earlier than the first appearance of these taxa in the fossil record at 180 Ma. None of the clocks rooted with appropriate taxa have placed the origin of diatoms earlier than 250 Ma, which likely corresponds to the Permian–Triassic (PT) extinction, 251 Ma. The heterokont algae, of which the diatoms are a member, radiated when the ocean trace metal chemistry changed at the PT boundary and gave the host plants with a red algal plastid an adaptive advantage. Red and green algal plastids differ greatly in their need for certain trace metals. The abundance of Fe after the PT boundary favors the growth of the red algal plastid, which has the Fe-containing cytochrome c6 in their photosynthetic electron carrier complex instead of the Cu-containing plastocyanin found in the photosystems of other algae (17).

References 1. 2.

3.

4. 5.

6.

7. 8.

9. 10.

11. 12. 13. 14. 15. 16. 17.

P. A. Sims, D. G. Mann, L. K. Medlin, Phycologia 45, 361 (2006). V. L. Nikolaev, D. M. Harwood, N. I. Samsonov, Early Cretaceous Diatoms (Russian Academy of Sciences, Komorov Botanical Institute, 2001). C. van den Hoek, D. G. Mann, H. M. Jahns, Algae. An Introduction to Phycology (Cambridge University Press, Cambridge, 1995). H. A. Von Stosch, Nature 165, (1950). F. E. Round, R. M. Crawford, D. G. Mann, The Diatoms. Biology and Morphology of the Genera (Cambridge University Press, Cambridge, 1990). D. M. Williams, in Unravelling the Algae, J. Brodie, J. Lewis Eds. (CRC Press, Boca Raton, FL, USA, 2007), pp. 57–91. L. K. Medlin, I. Kaczmarska, Phycologia 43, 245 (2004). L. K. Medlin, W. H. C. F. Kooistra, A. M.-M. Schmid, in The Origin and Early Evolution of the Diatoms: Fossil, Molecular and Biogeographical Approaches, A. Witkowski, A. Siemimska, J. Siemimska, Eds. (Szafer Institute of Botany, Polish Academy of Science, Cracow, Poland, 2000), pp. 13–35. A. M.-M. Schmid, Plant Syst. Evol. 158, 211 (1988). I. Kaczmarska, J. M. Ehrman, S. S. Bates, in Proceedings of the 16th International Diatom Symposium, A. Economou-Amilli, Ed. (University of Athens, Greece, 2001), pp. 153–168. W. H. C. F. Kooistra, L. K. Medlin, Mol. Phylogenet. Evol. 6, 391 (1996). J. S. Sinninghe-Damsté et al., Science 304, 584 (2004). H. Philippe, U. Sorhannus, A. Baroin, R. Perasso, F. Gasse, A. Adoutte. J. Evol. Biol. 7, 247 (1994). U. Sörhannus, Micropaleontology 43, 215 (1997). U. Sörhannus, Mar. Micropaleont. 65, 1 (2007). S. Sato, W. H. C. F. Kooistra, D. G. Mann, S. Mayama, L. K. Medlin, Mol. Phylogenet. Evol. Accepted (2008). P. G. Falkowski, O. Schonfeld, M. E. Katz, B. van de Schootbrugge, A. H. Knoll, in Coccolithophores—From Molecular Processes to Global Impact, H. Thierstein, J. R. Young Eds. (Elsevier, Amsterdam, 2004), pp. 429–453.

PL ANTS


Land plants (Embryophyta) Susana Magallóna,* and Khidir W. Hilub a

Departamento de Botánica, Instituto de Biología, Universidad Nacional Autónoma de México, 3er Circuito de Ciudad Universitaria, Del. Coyoacán, México D.F. 04510, Mexico; bDepartment of Biological Sciences, Virginia Tech, Blacksburg, VA, 24061, USA *To whom correspondence should be addressed (s.magallon@ ibiologia.unam.mx)

Abstract The four major lineages of embryophyte plants are liverworts, mosses, hornworts, and tracheophytes, with the latter comprising lycophytes, ferns, and spermatophytes. Their relationships have yet to be determined. Different studies have yielded widely contrasting views about the time of embryophyte origin and diversification. Some propose an origin of embryophytes, tracheophytes, and euphyllophytes (ferns + spermatophytes) in the Precambrian, ~700– 600 million years ago (Ma), whereas others have estimated younger dates, ~440–350 Ma. In spite of large differences in absolute timing, there is agreement that the major lineages of embryophytes and their key vegetative, physiological, and reproductive innovations evolved shortly after embryophyte origin.

Land plants (embryophytes) constitute a monophyletic group that is well supported by morphological and molecular characters. Numerous vegetative and reproductive traits directly associated with life on land characterize the group (1) (Fig. 1). Embryophytes are mainly diagnosed by the presence of multicellular sporophytes, cuticule, archegonia, antheridia, and sporangia, as well as by details of spermatozoid ultrastructure and cell division, and the presence of sporopollenin in spore walls (2). All living land plants are placed in four major taxa: Marchantiophyta (liverworts; 5000–8000 species), Bryophyta (mosses; ca. 13,000 species), Anthocerophyta (hornworts; 100 species), and Tracheophyta (vascular plants; 285,000 species) (3). The living tracheophytes, in turn, are distributed in four groups. The earliest diverging lineage constitutes the Lycopsida (lycophytes, or club mosses; 1230 species), which is the closest relative to a clade that includes ferns and spermatophytes (seed plants). Ferns (sometimes called monilophytes; ca. 10,000

species) include whisk ferns, horsetails, and eusporangiate and leptosporangiate ferns. Spermatophytes include cycads (105 species), ginkgos (one species), conifers (540 species), gnetophytes (96 species), which are the gymnosperms, and angiosperms (Magnoliophyta, or flowering plants, 270,000 species). Angiosperms represent the vast majority of the living diversity of embryophytes. Here, we review the relationships and divergence times of the major lineages of embryophytes. We follow a classification of embryophytes based on phylogenetic relationships among monophyletic groups (2, 3). Whereas much of the basis of the classification is robust, emerging results suggest some refinements of higher-level relationships among the four major groups. This includes the inversion of the position of Bryophyta and Anthocerophyta, different internal group relationships within ferns, and different relationships among spermatophytes. The phylogenetic classification provides characters that are useful for establishing taxonomic definitions based on shared-derived characters. The living liverworts are characterized by the presence of oil bodies, a distinctive spermatozoid ultrastructure, and possibly lunularic acid.

Fig. 1 A moss (Braunia squarrulosa) from Mexico displaying the gametophyte (green) and sporophyte (red) phases. Credit: C. Delgadillo Moya.

S. Magallón and K. W. Hilu. Land plants (Embryophyta). Pp. 133–137 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Angiosperms 4

Gymnosperms

3

Ferns

2

Lycopsida 1

Tracheophyta

THE TIMETREE OF LIFE Spermatophyta

134

Anthocerophyta Bryophyta Marchantiophyta Np

Cm

PR 600

O S

D

PALEOZOIC 500

C

P

Tr

J

K

MESOZOIC


100

Pg CZ 0

Fig. 2 A timetree of Embryophyta. Divergence times are from Table 1. Phylogenetic relationships among the major land plant lineages are not resolved; therefore, the deepest node in the embryophyte tree is depicted as a polytomy. Abbreviations: C (Carboniferous), Cm (Cambrian), CZ (Cenozoic), D

(Devonian), J (Jurassic), K (Cretaceous), Np (Neoproterozoic), O (Ordovician), P (Permian), Pg (Paleogene), PR (Proterozoic), S (Silurian), and Tr (Triassic). The divergence times for Nodes 1 and 2 are similar but their branching order is shown as resolved.

Mosses are distinguished by multicellular gametophytic rhizoids, gametophytic leaves, and a particular spermatozoid ultrastructure. Hornworts posses many unique features, including a distinctively shaped apical cell, a pyrenoid in chloroplasts, mucilage cells and cavities in the talus, and an intercalary meristem at the base of the sporangium. The tracheophytes are characterized by the presence of branching sporophytes with multiple sporangia and independent alternation of generations. Living tracheophytes are further characterized by annular or helical thickenings in tracheids and possibly by lignin deposition on the inner surface of the tracheid wall. Among liverworts, mosses, and hornworts, the gametophyte (haploid phase) is dominant through the life cycle, and the sporophyte (diploid phase) depends on it for nutrients and support. The early members of the lineage leading to tracheophytes, all now extinct, displayed an alternation of independent and apparently equally dominant gametophytes and sporophytes. Among tracheophytes, the sporophyte is dominant; the gametophyte can either constitute a small independent plant (among ferns) or be embedded in sporophytic tissues (among spermatophytes). Phylogenetic analyses of morphological and molecular data have generally supported the monophyly of most of the traditionally recognized major taxonomic groups of embryophytes, but there have been unexpected associations of taxa, and some relationships still remain unresolved. With a few exceptions, the four major groups of embryophytes have been found to be monophyletic (2–7).

Although a few morphological analyses have found a clade formed by liverworts, mosses, and hornworts to the exclusion of tracheophytes (5, 8), virtually all molecular analyses show those three lineages as a paraphyletic grade of early diverging land plants. Nevertheless, their branching sequence and their relationship with tracheophytes remain unclear. Either liverworts (2, 4, 7, 9–11) or hornworts (5, 11) have been found to be the earliest diverging lineage of land plants. The closest relative of tracheophytes has been identified as being the hornworts (7, 9, 11), the mosses (2, 4), a clade including mosses and hornworts (12), or a clade including mosses and liverworts (5, 11, 12). Tracheophytes have been ubiquitously recognized as a monophyletic group in which the deepest split segregates the lycophytes and the euphyllophytes (ferns plus spermatophytes) (2, 7, 11, 13). A relatively novel result is the recognition that all euphyllophytes lacking seeds, that is, the eusporangiate ferns, leptosporangiate ferns, whisk ferns, and horsetails, are more closely related to each other than to spermatophytes (2, 13). A major departure from traditional ideas about relationships is the recognition that the eusporangiate Ophioglossidae and Marattidae ferns, and the leptosporangiate Polypodiidae ferns do not constitute a monophyletic assemblage, but rather, that Ophioglossidae (moonworts) and Psilotidae (whisk ferns) are closest relatives, and that horsetails are more closely related to Marattidae and/or Polypodiidae (5, 13). Whereas spermatophyte monophyly is uncontested, deciphering the relationships of the gnetophytes (a group

Eukaryota; Viridiplantae; Streptophyta; Embryophyta

135

Table 1. Divergence times (Ma) and confidence/credibility intervals (CI) among embryophytes. Timetree Node

Estimates Ref. (21, 29)

Time

Ref. (22)

Ref. (23)

Ref. (26, 27)

Ref. (30, 31)

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

1

593

703 (21)

791–615

707.0

805–609

631.8

798–465

486.5 (26)

493–480

438.8

–

2

603

–

–

–

–

603

813–393

–

–

–

–

3

466

411.8 (29)

409–415

–

–

572.2

790–354

–

–

415.5

414–418

4

355

346.1 (29)

339–353

–

–

404.6

524–285

319.7 (27)

339–301

350.3

345–355

Note: Node times in the timetree represent the mean of time estimates from different studies. In ref. (21), estimates are derived from pairwise distances among 50 nuclear protein sequences. For ref. (23), average age and confidence intervals (obtained from the averaged dates) are derived from maximum parsimony and maximum likelihood branch length optimization for the combined sequences of four plastid and nuclear genes for a sample of land plant lineages using nonparametric rate smoothing. The age for Node 1 corresponds to the crown group of embryophytes (23). For ref. (26), the average age and confidence interval (obtained from the averaged dates) of the divergence between liverworts and seed plants are derived from applying one or two calibration points in penalized likelihood dating for a data set of 27 protein-coding genes. For ref. (22), the average date and confidence interval for the divergence of mosses and angiosperms were derived from six different rate-constant and rate-variable methods

using protein sequence data from 51 genes. For ref. (27), the average date and confidence interval (obtained from the averaged dates) for the divergence of gymnosperms and angiosperms are derived from penalized likelihood analyses focused on ferns and on angiosperms (constraining angiosperm age) (27). In ref. (29), averaged dates and confidence intervals (obtained from the averaged dates) are obtained with penalized likelihood for four chloroplast genes, using 1st + 2nd, and 3rd codon positions separately, for a sample across vascular plants. For refs. (30, 31), averaged dates and confidence intervals (obtained from the averaged dates) are derived from a combined data set of five chloroplast protein-coding genes for a sample across land plants, with penalized likelihood implementing branch-pruning and fossil-based rate smoothing. The age for Node 1 corresponds to the average of the divergence of mosses and the divergence of hornworts.

of gymnosperms) to the angiosperms and other gymnosperms has proven difficult. Spermatophytes are represented in the present day by only four or five evolutionary lineages, which are survivors of a much larger historical diversity. Phylogenetic analyses of morphological data have yielded a variety of hypotheses regarding relationships among living and extinct spermatophytes, but they agree in indicating a close relationship between gnetophytes and angiosperms (14), which together with a few particular extinct lineages formed a clade named the anthophytes. Molecular analyses have only rarely supported a phylogenetic closeness between gnetophytes and angiosperms to the exclusion of all other living seed plants (15), and, when they did, the result has been shown to be highly improbable with molecular data (16). Nevertheless, other than rejecting an anthophyte association, molecular data have so far failed to provide a single universally supported hypothesis of relationship among living spermatophytes. More recently, two hypotheses, which differ in the position of gnetophytes, have emerged as the main competitors (15, 17–19). In one, gnetophytes are the closest relatives of a group containing all other living spermatophytes (“gnetophyte-sister” hypothesis). In the second one, gnetophytes are closely related to conifers within a clade that includes all living gymnosperms. This hypothesis is obtained in two variants: with gnetophytes closest to the

conifer Family Pinaceae, rendering the conifers paraphyletic (“gnepine” hypothesis); or with gnetophytes closest to the monophyletic conifers (“gnetifer” hypothesis). Recent results have shown that the “gnetophyte-sister” signal is provided by sites with high substitution rates, and that this result is not obtained if rapidly changing sites are excluded from analysis, or if data are analyzed with optimization methods that are less prone to the long-branch attraction problem (16, 17, 19). The “gnepine” and “gnetifer” relationships are prevalent results from analyses using sites with intermediate to low rates, or using all types of data analyzed with parametric optimization criteria (17–19). Whereas the “gnepine” result is more frequent than the “gnetifer” one, the two topologies are very similar and statistically indistinguishable from the point of view of the phylogenetic signal in the data (19). Although the position of gnetophytes is particularly unstable, currently prevalent molecular results point at the phylogenetic closeness between gnetophytes and conifers, and suggest that all living gymnosperms (including gnetophytes) are more closely related to each other than to angiosperms. Relatively few studies have dated embryophytes or the major divergences within them (20). Heckman et al. (21) used the sequences of 50 nuclear proteins for a taxonomic sample, including one moss and seven angiosperms, to estimate the time of divergence of mosses and tracheophytes. By estimating pairwise distances calibrated with

136


multiple external secondary calibrations, a Precambrian (703 ± 45 Ma) divergence time was estimated. Subsequently those data were analyzed using Bayesian and likelihood rate smoothing methods (22), and a similar date was obtained (707 ± 98 Ma). Soltis et al. (23) evaluated the impact of different genes, calibration points, and branch lengths on ages of embryophytes as a whole, and of tracheophyte lineages. Their study was based on three plastid protein-coding genes and the nuclear 18S ribosomal DNA. Using the topology obtained by Pryer et al. (13), with the three bryophyte lineages forming a clade, and branch lengths estimated with maximum parsimony (MP) and with maximum likelihood (ML), divergence times across the tree were estimated with a nonparametric rate smoothing method. Estimated divergence times differed substantially depending on the calibrations used and on estimation conditions. By calibrating the tree at the divergence between the two sampled angiosperms at 125 MY, embryophytes were estimated to have originated in the Precambrian (546.8 or 716.8 Ma with MP and ML branch lengths, respectively). Tracheophytes were estimated to be of Precambrian (710.1 Ma) or Upper Cambrian (495.9) age, and euphyllophytes of Precambrian (683.4 Ma) or Middle–Upper Ordovician (460.9 Ma) age, with MP and ML branch lengths, respectively. Spermatophytes were estimated as considerably younger, of Middle Ordovician (465.4 Ma) or Mississippian (343.7 Ma) age, with MP and ML branch lengths, respectively. All the preceding dates are considerably older than spore-containing plant fragments from the Late Ordovician (24) and of microscopic dispersed spores from the Middle Ordovician (25), which are the oldest generally accepted reports of land plants (24). Sanderson (26) used 27 plastid protein-coding genes for 10 land plants and an algal outgroup to estimate the age of embryophytes. By using the semiparametric penalized likelihood method implementing two alternative internal calibrations points, the divergence between liverworts and seed plants was found to be of Lower Ordovician age (483 or 490 Ma, depending if one or two calibration points were applied). Schneider et al. (27) investigated the timing of diversification of polypod ferns and angiosperms using two independent data sets, one with rbcL and rps4 sequences for 45 ferns and outgroups, and another with atpB, rbcL, and nuclear 18S rDNA sequences for 95 angiosperms and outgroups. Ages were estimated with penalized likelihood by fi xing the age of euphyllophytes at 380 Ma, and constraining several nodes with fossil-derived minimum ages. The divergence between gymnosperms and angiosperms was

dated as Pennsylvanian (310 Ma) in the fern-based analysis and as Mississippian (329 or 333 Ma, depending on whether the angiosperm age was fi xed or not at 132 Ma) in the angiosperm-based analysis. These dates are relatively close to the Mississippian (Namurian) age of Cordaitales (28), presumably the oldest fossil member of the group containing the living lineages of spermatophytes. Magallón and Sanderson (29) conducted a study including all tracheophyte lineages, one liverwort and one charophycean outgroup, using the plastid protein-coding genes atpB, psaA, psbB, and rbcL. Ages were estimated with penalized likelihood, implementing a 419 Ma calibration to tracheophytes, derived from the age of the oldest fossil pertaining to that divergence (2). The impact of different gene and codon position partitions, and that of including or excluding fossil-derived constraints on the ages of 20 nodes, was evaluated. Estimates were found to vary substantially depending on the data and the constraints used. In fossil-constrained estimations, the average age of euphyllophytes was Lower Devonian (411.8 ± 3 Ma), and Mississippian (346.1 ± 7 Ma) for spermatophytes. Hilu et al. (30, 31) expanded the data set of Magallón and Sanderson (29) to include all major lineages of embryophytes, and the sequences of matK, a plastid protein-coding gene with a relatively fast substitution rate, representing so far the single dating study that encompasses the embryophytes as a whole and its major lineages. Dating was based on a Bayesian tree in which liverworts are the earliest diverging land plants, and hornworts are closest to tracheophytes. Dating was conducted with penalized likelihood implementing optimal rate smoothing derived from branch-pruning and fossil-based cross validations. Calibration and age constraints were very similar to those in Magallón and Sanderson (29), including calibration at the tracheophyte node and 22 minimum age constraints, but also imposing a maximal 464 Ma age to embryophytes, derived from Late Ordovician remains of presumed crown group embryophytes (24). The divergence of mosses from all other embryophytes was estimated as Upper Ordovician (446 ± 1 Ma), and the split between hornworts and tracheophytes as Silurian (Llandovery; 432 ± 1 Ma). The age of the tracheophyte node was fi xed for calibration at 421 Ma (27). Euphyllophyte age was estimated as Lower Devonian (415.5 ± 2 Ma), which is close to the age of Pertica, the oldest euphyllophyte fossil (2). Spermatophytes were estimated to be Mississippian (350.3 ± 5 Ma), older but relatively close to the age of Cordaitales. Studies about the time of origin and early diversification of embryophytes suggest two widely contrasting

Eukaryota; Viridiplantae; Streptophyta; Embryophyta

views. One set of studies (21–23) jointly suggest embryophyte origin and differentiation of liverworts, bryophytes, hornworts, tracheophytes, and euphyllophytes during a short time in the Precambrian. Spermatophyte differentiation is estimated to have occurred substantially later, in the Ordovician. Another set of studies (26, 27, 29–31) suggest embryophyte origin and diversification in the Paleozoic, in a period of ~70 Ma spanning from the Lower Ordovician origin of embryophytes, to the Lower Devonian origin of euphyllophytes. Spermatophyte origin is estimated as being somewhat younger, from the Mississippian. In spite of the considerable differences in absolute timing, available evidence congruently suggests the differentiation of embryophyte lineages occurred shortly after embryophyte origin. Whereas major innovations necessary for survival in the terrestrial environment most likely evolved before the differentiation of embryophyte lineages, a substantial amount of morphological and physiological innovation took place during the initial phase of embryophyte evolution. The shift from a dominant gametophytic phase to a dominant sporophytic phase, including the evolution of branched sporophytes with multiple sporangia, and the evolution of lignitized cells, particularly of the type that characterize living tracheophytes, occurred with the differentiation of tracheophytes. With the differentiation of lycophytes and euphyllophytes, two evolutionary types of leaves evolved: simple leaves vascularized by veins that do not form a gap in the main vascular strand of the stem, in the lycophyte lineage; and “true” leaves, derived from modification of branching axes, in the euphyllophyte lineage. The origin of spermatophytes apparently lagged behind the initial embryophyte diversification. Whereas heterospory most likely evolved several times among tracheophytes, the sequence of innovations necessary to give rise to the seed habit from a heterosporous reproductive system occurred only once (32). With the origin of spermatophytes, the major vegetative and reproductive innovations of embryophytes, including dominant sporophytes, vascularized systems, and seeds, had evolved. The substantially different estimates of the timing of embryophyte origin and early diversification suggest that the investigation of this question would greatly benefit from a comprehensive integration of fossils and molecular clocks.

References 1.

2.

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

29. 30. 31.

Acknowledgment We thank J. A. Barba for his help in preparing the initial version of Fig. 2.

137

32.

L. E. Graham, in Microspores: Evolution and Ontogeny, S. Blackmore, R. B. Knox, Eds. (Academic Press, London, 1990). P. Kenrick, P. R. Crane, The Origin and Early Diversification of Land Plants, A Cladistic Study. (Smithsonian Institution Press, Washington and London, 1997), pp. 441. P. R. Crane, P. Kenrick, Aliso 15, 87 (1997). B. D. Mishler, S. P. Churchill, Brittonia 36, 406 (1984). K. S. Renzaglia et al., Phil. Trans. Roy. Soc. Lond. B 355, 769 (2000). J. Shaw, K. S. Renzaglia, Am. J. Bot. 91, 1557 (2004). Y.-L. Qiu et al., Proc. Natl. Acad. Sci. U.S.A. 103, 15511 (2006). D. J. Garbary, K. S. Renzaglia, J. G. Duckett, Plant Syst. Evol. 188, 237 (1993). L. A. Lewis, B. D. Mishler, R. Vilgalys, Mol. Phylogenet. Evol. 7, 377 (1997). Y.-L. Qiu et al., Nature 394, 671 (1998). D. L. Nickrent et al., Mol. Biol. Evol. 17, 1885 (2000). R. J. Duff, D. L. Nickrent, Am. J. Bot. 86, 372 (1999). K. M. Pryer et al., Nature 409, 618 (2001). P. R. Crane, Ann. MO Bot. Gard. 72, 716 (1985). C. Rydin, M. Källersjö, E. M. Friis, Int. J. Plant. Sci. 163, 197 (2002). M. J. Sanderson et al., Mol. Biol. Evol. 17, 782 (2000). S. Magallón, M. J. Sanderson, Am. J. Bot. 89, 1991 (2002). D. E. Soltis, P. S. Soltis, M. J. Zanis, Am. J. Bot. 89, 1670 (2002). J. G. Burleigh, S. Mathews, Am. J. Bot. 91, 1599 (2004). M. J. Sanderson et al., Am. J. Bot. 91, 1656 (2004). D. S. Heckman et al., Science 293, 1129 (2001). S. B. Hedges, J. E. Blair, M. L. Venturi, J. L. Shoe, BMC Evol. Biol. 4, 2 (2004). P. S. Soltis et al., Proc. Natl. Acad. Sci. U.S.A. 99, 4430 (2002). C. H. Wellman, P. L. Osterloff, U. Mohiuddin, Nature 425, 282 (2003). P. K. Strother, S. Al-Hajri, A. Traverse, Geology 24, 55 (1996). M. J. Sanderson, Am. J. Bot. 90, 954 (2003). H. Schneider et al., Nature 428, 553 (2004). T. N. Taylor, E. L. Taylor, The Biology and Evolution of Fossil Plants (Prentice Hall, Englewood Cliffs, New Jersey, 1993). S. Magallón and M. J. Sanderson, Evolution 59, 1653 (2005). K. W. Hilu, S. Magallón, D. Quandt, XVII International Botanical Congress Abstracts, p. 200 (2005). K. W. Hilu, S. Magallón, D. Quandt, Annual Meeting of the Botanical Society of America Abstracts, p. 46 (2006). K. M. Pryer, H. Schneider, S. Magallón, in Assembling the Tree of Life, J. Cracraft, M. J. Donoghue, Eds. (Oxford University Press, New York, 2004), pp. 138–153.

Mosses (Bryophyta) Angela E. Newtona,*, Niklas Wikströmb, and A. Jonathan Shawc a

Department of Botany, Natural History Museum, London SW7 5BD, UK; bDepartment of Systematic Botany, Evolutionary Biology Centre, Uppsala University, Norbyvägen 180 75236, Uppsala, Sweden; c Department of Biology, Duke University, Durham, NC, 27708 USA *To whom correspondence should be addressed (a.newton@nhm. ac.uk)

Abstract Living mosses (ca. 13,000 sp.) constitute the Phylum Bryophyta, with eight classes divided into acrocarpous mosses (23 orders), not an evolutionary group, and pleurocarpous mosses (4–7 orders, 42% of extant species). Two subclasses, Dicranidae (acrocarpous haplolepidae) and Bryidae (diplolepidous-alternate mosses, with both acrocarpous and pleurocarpous members), account for 90% of extant species. The moss timetree shows lineage origin at ~380 million years (Ma) ago, the split between Haplolepidae and Diplolepidae at 220–195 Ma, and appearance of the first pleurocarp lineages at ~173 Ma. Major diversification occurred in the Cretaceous, 140–90 Ma.

Mosses are photosynthetic plants that exhibit a wide range of morphologies, based on the life cycle of alternation of haploid and diploid generations. The gametophyte generation starts with haploid spores that develop into threadlike protonema, from which grow erect or creeping axes usually up to a few centimeters tall. The plants are often branched, with leaves that are mostly one cell thick and usually radially arranged (Fig. 1). They form small cushions, velvety swards, open turfs, tufts, or deep mats. Long-lived clonal growth and specialized vegetative reproduction are common and widespread, with sexual reproduction rare or unknown in some taxa. Plants are male or female, or bisexual, with various different arrangements of gametangia. Male and female gametangia are distinct, with a basic morphology common to all mosses (and to most of the early-diverging land-plant lineages) but with variation in some groups. Motile spermatozoa travel though surface water to fertilize sedentary eggs. The resulting diploid embryo grows into a sporophyte that is largely dependent on the gametophyte, with

a single spore capsule and often with a complicated dehiscence (dispersal) mechanism (peristome). Spores are formed by meiosis and usually dispersed as monads. The taxonomic diversity of living mosses reflects the morphological diversity, and is arranged in a pectinate grade of several small or very small but highly distinct clades, a few rather larger but also distinct clades (e.g., Sphagnales, Polytrichales) and two very large and poorly resolved clades, the Dicranidae and Bryidae, that together account for 90% of extant moss species diversity. The basal grade, the Dicranidae and part of the Bryidae are acrocarpous, with the principal vegetative axis terminated by gametangia, and consequently with terminal formation of sporophytes. A monophyletic group within the Bryidae is pleurocarpous, with gametangia

Fig. 1 Wall screw-moss (Tortula muralis) from a brick wall in Richmond, England. Credit: A. E. Newton.

A. E. Newton, N. Wikstrom, and A. J. Shaw. Mosses (Bryophyta). Pp. 138–145 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Viridiplantae; Bryophyta

139

Plagiotheciaceae Meteoriaceae Fontinalaceae Cryphaeaceae Pterobryaceae

Entodontaceae Hypnaceae Brachytheciaceae

16

28

Hylocomiaceae Leucodontaceae Neckeraceae

Hookeriales/Hypnales

Thuidiaceae 26

Hypnaceae Stereophyllaceae Bryidae

Fabroniaceae 15

Lepyrodontaceae Catagoniaceae Leucomiaceae 29

Pilotrichaceae

27

13

Hookeriaceae 23

Trachylomataceae Rutenbergiaceae Hypopterygiaceae Ptychomniaceae (Ptychomniales)

Pterobryellaceae 25

Braithwaiteaceae

21

Racopilaceae Hypnodendraceae

Hypnodendrales

Aulacomniaceae (Aulacomniales) 11

Orthodontiaceae (Orthodontiales) 8

Rhizogoniaceae (Rhizogoniales)


D

C

P

PALEOZOIC 300

Tr

J

K

MESOZOIC 200 Million years ago

Pg

Ng

CZ 100

0

Fig. 2 Continues

terminating specialized lateral branches, so that sporophytes develop on branches. This innovation has been suggested to be a key in the evolution of the branching structure, contributing to the enormous species diversity in this group (1).

A similar morphological transition is seen in the structure of the dehiscence mechanism of the sporophyte. The earliest diverging lineages mostly have simple linear dehiscence while later diverging lineages have circumscissile dehiscence and (usually) a peristome. The peristome

140


(continued from previous page)

Hedwigiaceae (Hedwigiales)

Bryaceae

Splachnaceae 19

Meesiaceae

6

Timmiaceae (Timmiales) 17

Splachnales

Orthotrichaceae (Orthotrichales)

10

Bryidae

Mniaceae

Bryales

Bartramiaceae (Bartramiales)

Scouleriaceae (Scouleriales) Pottiaceae (Pottiales) 20

18

Fissidentaceae

5

Dicranales

14

Dicranaceae

Schistostegaceae Seligeriaceae (Seligeriales) Ptychomitriaceae 24

Grimmiaceae Funariaceae (Funariales)

4

9

Grimmiales

12

22

Dicranidae

Ditrichaceae

7

Encalyptaceae (Encalyptales) Diphysciaceae (Diphysciales) Buxbaumiaceae (Buxbaumiales)

3

Tetraphidaceae (Tetraphidales) Polytrichaceae (Polytrichales)

2

Oedipodiaceae (Oedipodiales) Andreaeaceae (Andreaeales)

1

Andreaeobryaceae (Andreaeobryales) Takakiaceae (Takakiales) Sphagnaceae (Sphagnales)

D

C

P

PALEOZOIC 300

Tr

J

K

200 Million years ago

Pg

Ng

CZ

MESOZOIC 100

0

Fig. 2 A timetree of Bryophyta. Classification follows Bell et al. (35) for the pleurocarpids and Goffinet and Buck (4) for the remaining taxa. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

itself shows a transition from massive teeth composed of multiple cells (nematodontous) to more delicate and flexible teeth composed of cell walls (arthrodontous). In the arthrodontous mosses the peristome is composed of one or two rings of structures, the outer exostome teeth and the inner endostome segments, derived from the cell

walls between three concentric rings of cells. The patterns of the cell wall remnants on the peristome surfaces have allowed the correspondence of the structures to be identified (2). Mosses with both peristome rings are termed diplolepidous, and have an exostome that is robust and flexes with changes in humidity, while the endostome is


Table 1. Divergence times and their confidence/ credibility intervals (CI) among Bryophyta, based on ref. (1). Timetree Node

Time

CI

1

379.0

400–362

2

329.0

342–304

3

317.0

334–295

4

292.0

316–280

5

246.0

272–230

6

219.0

243–205

7

203.0

220–176

8

195.0

216–181

9

187.0

209–162

10

184.0

204–165

11

173.0

194–161

12

156.0

188–144

13

151.0

173–141

14

149.0

175–130

15

143.0

165–131

16

141.0

157–123

17

136.0

159–111

18

121.0

145–101

19

116.0

138–95

20

111.0

141–96

21

111.0

124–88

22

109.0

134–88

23

107.0

136–93

24

105.0

131–82

25

88.0

102–67

26

71.0

96–56

27

71.0

92–61

28

67.0

118–54

29

47.0

61–39

more delicate but less flexible. Changes in positions and symmetry of division patterns of the cell lineages during development (3) have allowed the recognition of opposite or alternate forms of the basic diplolepidous pattern. In the haplolepidous mosses only the endostome is well developed, consisting usually of 16 or more simple or bifid teeth, usually flexible and often highly ornamented with papillae. The exostome may be present as a very reduced ring referred to as a prostome. Other variations are also known, and in particular reduction or loss of parts of

141

the peristome is known in many different taxa. Both the haplolepidous- and diplolepidous-alternate forms appear to be derived from the diplolepidous-opposite group, but this area of the topology is not yet strongly resolved or supported. However, these forms comprise very large monophyletic groups, the haplolepidous Dicranidae and the diplolepidous-alternate Bryidae (sensu 4), with 30% and 60%, respectively, of extant species diversity (1). Classification of the mosses has fluctuated wildly over the centuries since the starting-point publication by Hedwig in 1801 (5), depending on whether the sporophyte or gametophyte generation was given precedence (6). Certain groups in which the peristome is highly reduced or absent have been particularly problematic. However, some degree of stability was introduced with the work of Brotherus (7) and Fleischer (8) in the early twentieth century, although numerous small and not-so-small changes have continued to be made in taxon circumscription and relationships at all levels, and opinion has differed on particular taxa. For example, the Polytrichales, which have relatively well-developed vascular tissue, have been placed near the beginning (9) or end (8) of classifications, with implications of a primitive or derived status, respectively. The relatively small size and simple structures of mosses appears to have led to extensive parallelism and convergence, making the use of morphological characters for classification particularly difficult. Cladistic methodology was adopted early by bryologists, one of the earliest applications of Hennigian principles was a generic revision (10) of the moss Family Mniaceae in 1968, and in 1984 a morphological cladistic analysis (11) of the bryophytes established the very basic elements of the pectinate grade (Sphagnales (Andreaeales (Tetraphidales (Polytrichales (Buxbaumiales (Bryales)))))) most of which is still accepted. A pioneering series of cladistic analyses (using morphological data) explored the relationships of the pleurocarpous mosses (12). However, it was not until the advent of DNA sequencing that sufficient data were available to explore relationships in detail. Since the late 1990s studies using single plastid or multigene phylogenies (often including morphological data) have established the topology now widely recognized. These have included both larger studies that established the “backbone” topology (13–17) and others that provided resolution of smaller taxonomic groups (18–20). However, although the general pattern emerged quite quickly, certain problem areas have resisted resolution. A long-standing problem has been resolution of the relationships of the green algae and basal land plants, which also impacts the question of whether the mosses

142


are monophyletic. Many combinations of the four terrestrial groups (liverworts, mosses, hornworts, and vascular plants), with subdivisions, have been retrieved with strong support, using different data and taxon sampling (21–23). The genus Takakia, historically placed in the liverworts, was recognized as a moss when sporophytes were finally discovered (24), supported by molecular data that often showed a weak close relationship (15, 25) with Sphagnum as close to all other mosses. Different arrangements have been found, including Andreaea and the nematodontous mosses (Polytrichales, Tetraphis, etc.) close to Sphagnum, Takakia and all other mosses (25) but these do not contradict the evidence for monophyly. The sparse and relict monotypic genus Oedipodium, found in cool temperate or montane habitats, lacks a peristome and had been placed with the diplolepidousopposite Funariales until molecular evidence proposed a relationship either close to all peristomate mosses (15, 26), or basal in the Polytrichales clade close to all remaining peristomate mosses (26). These positions are both plausible morphologically. Within the basal arthrodontous mosses relationships are still not conclusively resolved. The diplolepidousalternate (Bryidae) and haplolepidous mosses (Dicranidae) both clearly form monophyletic clades, although the circumscription of the latter is not yet strongly supported (27). However, the mosses with arthrodontous peristomes that are neither clearly haplolepidous nor diplolepidousalternate (Funariaceae, Disceliaceae, Encalyptaceae, Gigaspermaceae, Catascopiaceae, and Timmiaceae, see 28) have been variously placed. Some have traditionally been regarded as diplolepidous-opposite while others are anomalous or lack peristomes. Two hypotheses were suggested by Goffinet and Buck (4, p. 210): (1) Funariales and Encalyptaceae form a clade close to the haplolepidous mosses (1, 16, 27) and these share an “opposite” peristome arrangement and (2) the haplolepidous and diplolepidous-alternate mosses (Dicranidae plus Bryidae) form a clade (28–30). The diplolepidous-opposite and anomalous mosses have also been shown to be paraphyletic relative to the combined clade of Dicranidae and Bryidae (15–17). However, recent studies are beginning to converge on support for a topology with the clade of diplolepidousalternate mosses close to a clade composed of paraphyletic diplolepidous-opposite and anomalous mosses with a monophyletic clade of haplolepidous mosses (1, 27). Within the Dicranidae a number of clades have been established (27) although support for several of these is weak, and the backbone topology is mostly unsupported. The Bryidae is strongly supported as monophyletic (14–16, 31) and includes several groups that are

morphologically distinct and in which some relationships have been established. However, with the exception of certain critical distal nodes, the relationships between most of the groups are obscure, either unresolved or unsupported, and with little congruence between analyses. In particular, the positions of the Orthotrichales and Hedwigiales, and the identity of the closest relative of the pleurocarpous mosses, are not clear. Several taxa have been proposed as candidates for closest relative of the pleurocarp clade, including Orthotrichaceae (28), Mniaceae (1, 32), Bartramiaceae (31), Bartramiaceae with Hedwigiaceae (17), or Bartramiaceae, Hedwigiaceae, and Mniaceae (33), Orthodontium (4), or Aulacomnium (14, 34). Recent work (35) has shown the Orthodontiaceae and Aulacomniaceae to be included within the monophyletic clade of pleurocarpous mosses, with the Orthotrichales closest to the pleurocarps and the Bartramiaceae placed in a more distant position. This relationship for the Orthotrichales has previously been proposed by other authors (32), but there is also evidence for a position of the Orthotrichales in a more basal position in the Bryinae (32). The monophyly of the clade including the lineages of pleurocarpous mosses (pleurocarpids = Aulacomniales, Orthodontiales, Rhizogoniales, Hypnodendrales, and Hypnidae) and the relationships of the basal clades of core pleurocarps (Hypnodendrales and Hypnidae) has recently been established (32, 33, 35, 36). However, problems remain unresolved in the homocostate pleurocarps, the vast majority of taxa traditionally classified in the Hypnales, Hookeriales, and Leucodontales, and characterized by an undifferentiated costa. The revelation that the Leucodontales and Hypnales s.s. are both polyphyletic (37) was tempered by the continued recognition of the Hookeriales and establishment of a new order, the Ptychomniales (38), but there is now increasing evidence (1, 39, 40) that the Hypnales s. l. are also paraphyletic relative to the Hookeriales. In addition to the Ptychomniales, two controversial lineages may be recognized—the Hypopterygiaceae as an entity separate from the Hookeriales (1, 34, 40), and a clade containing the Hookeriales and various taxa of Hypnales s. l. but of uncertain affinity, such as Catagonium, Pseudocryphaea, Rutenbergia, and Trachyloma. The circumscription of this clade has yet to be finalized, but where sampling was adequate and the topology not constrained, several published and unpublished studies have supported its existence (1, 28, 34). Finally, within the Hypnales s. l., there is very little resolution of relationships. Where published molecular studies exist, individual families may be shown to be monophyletic (e.g., Brachytheciaceae and Meteoriaceae, 41, 42) or polyphyletic (Amblystegiaceae,


43–45), frequently with changes in circumscription (46). A very few well-supported clades are consistently found by different studies, for example that consisting of exemplars of the Rigodiaceae, Lembophyllaceae, Echinodiaceae, Thamnobryaceae, Leptodontaceae, and Neckeraceae (28, 34, 37, 45, 47) although again the circumscription of this clade has yet to be finalized since the circumscription of the individual families is also being reassessed. As yet there is only one published study proposing molecular divergence times for mosses (1), although a study on liverworts (48) includes a single relevant estimate, for the divergence of mosses from liverworts. A study of the pattern of diversification in the pleurocarpous mosses (18) did not include divergence time estimates. Until recently the lack of a resolved topology for the mosses hindered such work, but the shortage of moss fossils suitable for use for calibration or as constraints (1, 49) is also a problem, necessitating the use of fossils outside the group. Both studies providing divergence time estimates used a penalized likelihood approach using r8s and cross validation, for topologies derived by Bayesian inference with only nodes ≥95% PP recognized. However, the topologies for the relationships of the basal land plants and the calibration ages differed. The first study used a fi xed calibration age of 450 Ma for the origin of the land plants, based on a conservative date for the appearance of spore tetrads (50), and mosses were estimated to have diverged from other land plants at 403 Ma. The second study used a calibration point based on the earliest split in the vascular plants at 430 Ma (51), with a fi xed maximum age constraint of 475 Ma for the oldest fossils (controversially) accepted as land plants (52). Here mosses were estimated to have diverged from other land plants (liverworts) at 454 Ma. However, the methodological differences between the studies prevent meaningful comparison. The moss timetree (Fig. 2, Table 1) shows the first split in the lineage (Node 1—between Sphagnum/Takakia and the other mosses) at around 380 Ma, and the shift from predominantly linear to circumscissile sporophyte dehiscence (Node 4) at 292 Ma, with the appearance of the peristome presumably shortly thereafter. In the nematodontous mosses the Family Polytrichales shows diversification from at least the Triassic (226 Ma), with the earliest diverging extant taxon (1) a relict species (Alophosia azorica, not shown) found only on Macaronesia. This species lacks the photosynthetic lamellae and peristome structures characteristic of the order, while its closest relatives have the lamellae poorly developed and are also relictual in their distribution (26). One of the few

143

moss fossils sufficiently well characterized to be placed in phylogenetic topologies is Eopolytrichum (53), from the Cretaceous (Campanian) at 80 Ma. This species shows derived features that allow it to be placed in a quite distal position in the family (26, 54), but there is not yet sufficient resolution to allow its generic relationships to be finalized, and it therefore cannot yet be used as a constraint in molecular divergence analyses. The first taxa with early forms of the arthrodontous peristome had appeared (Node 5) by 246 Ma, but this area of the topology is very sparse despite almost complete generic sampling. This relates to the period of the Permian through to the late Triassic, and it is assumed that very few lineages survived the Permian–Triassic extinction, although there is fossil evidence (55, 56) for a diverse Permian bryoflora markedly different from extant taxa. Only 10 lineages passed through this bottleneck to the present day (1), and of these the majority are either extremely depauperate or show evidence of recent divergence (e.g., Sphagnales, 57), while just one lineage, the arthrodontous mosses, contains 95% of extant species diversity. The arthrodontous mosses diverged (Node 6) around 219 Ma into various forms, including haplolepidous, diplolepidous-opposite, and diplolepidous-alternate. The haplolepidous mosses were found (1) to be derived from within a group of mosses with diplolepidous-opposite or anomalous peristomes, but the relationship was not supported. Considerable variation in peristome morphology is apparent in taxa in all the earlier diverging lineages, but during the Jurassic two major forms appear to have stabilized, the diplolepidous-alternate (Bryidae) from 195 Ma (Node 8) and the haplolepidae (Dicranidae) from 156 Ma (Node 12). Apart from the Orthotrichales, which have a distinctive morphology and probably diverged from the Bryidae from around 184 Ma (Node 10), further variation seems to have mostly been limited to differences in sculpturing and relative development of parts, with derived reduction and loss of features occurring in parallel in many lineages. The circumscription of the Dicranidae and relationships within the clade have recently been established (27). The clades resolved by this analysis are mostly congruent with those recognized in the timetree, differing mainly in the placement of Timmia. The haplolepidae diverged (Node 7) around 203 Ma from the proto-haplolepidae (27), represented in the timetree by Scouleria and Timmiaceae, and further divergence of the major clades (Nodes 12, 14, 18, 20, 22, 24) occurred between 156 and 105 Ma. The split between the Grimmiaceae plus Ptychomitriaceae (27, Clade B) and the remaining Dicranidae (Node 12) occurred in the

144


Jurassic at 156 Ma, while further splits between genera representing major groups occurred in the late Jurassic through to early Cretaceous, at 105 Ma (Node 24). The Bryidae diverged from the other arthrodontous mosses in the late Triassic, at 219 Ma, followed by the split (Node 8) at 195 Ma between a clade composed of the Splachnales, Orthotrichales, and the Bryaceae, and the remaining diplolepidous-alternate mosses, including the other members of the (paraphyletic) Bryales. Due to the lack of resolution of the backbone topology and low sampling, it was not possible to draw many conclusions about divergence times between most of these groups, but the few exemplars included show a pattern of divergence between families from the late Jurassic to the late Cretaceous. Although dates for the earlydiverging Bryalean lineages could not be proposed, the pleurocarpid node (Node 11) is well supported by both molecular and morphological data. This node, at 173 Ma, represents the divergence of the lineages containing pleurocarpous taxa (with gametangia on lateral specialized branches). Not every lineage is exclusively pleurocarpous, and there are several variations in morphology, indicating that this feature was unstable between about 173 and 151 Ma. Again, most of the early-diverging nodes within the pleurocarpids were not sufficiently well supported for divergence times to be proposed, but the appearance of the homocostate pleurocarps (Node 13) at 151 Ma was followed by two divergences in a very short period. Within the homocostate pleurocarps there is a grade of four lineages diverging between 151 and 141 Ma, starting with the Ptychomniales and followed by the Hypopterygiaceae, which diverged from the remaining pleurocarps (Node 15) at 143 Ma. The most substantial split (Node 16) occurred at 141 Ma, between the clade containing the Hookeriales with some associated Hypnalean taxa, and the remaining members of the Hypnales. Within the expanded hookerialian clade (Nodes 23, 27, and 29) divergence occurred at about the family level from 107 to 47 Ma, that is through the Cretaceous and into the Cenozoic. Major diversification in the Hypnales clade also seems to have occurred in the Cretaceous but sampling within this very large clade was not sufficiently dense to allow more than two divergence estimates (Nodes 26 and 28, late Cretaceous). While it is not yet possible to date the divergence of most clades at the family level, several nodes in the original analysis (1) provide support for the divergence of genera within clades, indicating that divergences at this level have occurred, in different groups, over a wide range of time, so that some “species” may be older than “genera” in other groups.

After the split between the Ptychomniales and the remaining homocostate pleurocarps there were few changes in gametangial position, but greatly increased variation in branching architecture, leaf morphology, and peristome reduction and ornamentation. Much of this variation seem to have been associated with changes in habitat utilization, such as colonization of semiaquatic habitats, and the appearance of epiphytism as the increasing diversity of the angiosperm forest provided novel habitats (41). Studies of diversification patterns (18) in the pleurocarpous mosses indicate the possibility that rates of diversification were elevated early in the history of this group, but decreased in later periods. However, evaluation of these alternative hypotheses will depend on adequate resolution of the topology and estimates of divergence times for critical nodes of this very large clade.

Acknowledgments Support was provided by U.K. National Environment Research Council to A.E.N. by the Swedish Research Council to N.W. and by the U.S. National Science Foundation to A.J.S.

References 1. A. E. Newton, N. Wikstrom, N. Bell, L. L. Forrest, M. S. Ignatov, in Pleurocarpous Mosses: Systematics and Evolution, A. E. Newton, R. S. Tangney, Eds. (CRC Press, Boca Raton, 2007), pp. 337–366. 2. B. Goffinet, J. Shaw, L. E. Anderson, B. D. Mishler, Bryologist 102, 581 (1999). 3. A. J. Shaw, L. E. Anderson, B. D. Mishler, Syst. Bot. 14, 24 (1989). 4. B. Goffinet, W. R. Buck, in Molecular Systematics of Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Gardens Press, St. Louis, 2004), pp. 205–239. 5. J. Hedwig, Species Muscorum Frondosorum Descriptae (Johannis Ambrosii Barthii, Lipsiae, 1801). 6. W. R. Buck, in Pleurocarpous Mosses: Systematics and Evolution, A. E. Newton, R. S. Tangney, Eds. (CRC Press, Boca Raton, 2007), pp. 1–18. 7. V. F. Brotherus, in Die natürlichen Pflanzenfamilien, Vol. 1 (Part 3), A. Engler, K. Prantl, Eds. (Verlag von W. Engelmann, Leipzig, 1901–1909), pp. 277–1246. 8. M. Fleischer, Die Musci der Flora von Buitenzorg (Brill, Leiden, 1904). 9. H. N. Dixon, in Manual of Bryology, F. Verdoorn, Ed. (Martinus Nijhoff, The Hague, 1932), pp. 397–412. 10. T. Koponen, Ann. Bot. Fenn. 5, 117 (1968). 11. B. D. Mishler, S. P. Churchill, Brittonia 36, 406 (1984). 12. L. Hedenäs, Bryologist 97, 225 (1994).


13. S. Beckert, S. Steinhauser, H. Muhle, V. Knoop, Plant Syst. Evol. 218, 179 (1999). 14. S. Beckert, H. Muhle, D. Pruchner, V. Knoop, Mol. Phylogenet. Evol. 18, 117 (2001). 15. A. E. Newton, C. J. Cox, J. G. Duckett, J. A. Wheeler, B. Goffinet, T. A. J. Hedderson, B. D. Mishler. Bryologist 103, 187 (2000). 16. B. Goffinet, C. J. Cox, Bryologist 103, 212 (2000). 17. C. Cox, B. Goffinet, A. E. Newton, A. J. Shaw, T. A. J. Hedderson, Bryologist 103, 224 (2000). 18. A. J. Shaw, C. J. Cox, B. Goffinet, W. R. Buck, S. B. Boles, Evolution 57, 2226 (2003). 19. J. Hyvonen, S. Koskinen, G. L. Smith Merrill, T. A. J. Hedderson, S. Steinroos, Mol. Phylogenet. Evol. 31, 915 (2004). 20. C. La Farge-England, B. D. Mishler, J. A. Wheeler, D. P. Wall, K. Johannes, A. J. Shaw, Bryologist 103, 257 (2000). 21. R. J. Duff, D. L. Nickrent, Am. J. Bot. 86, 372 (1999). 22. M. Groth-Malonek, D. Pruchner, F. Grewe, V. Knoop, Mol. Biol. Evol. 22, 117 (2005). 23. T. Nishiyama, P. G. Wolf, M. Kugita, R. B. Sinclair, M. Sugita, C. Sugiura, T. Wakasugi, K. Yamada, K. Yoshinaga, K. Yamaguchi, K. Ueda, M. Hasebe, Mol. Biol. Evol. 21, 1813 (2004). 24. D. K. Smith, P. G. Davison, J. Hattori Bot. Lab. 73, 263 (1993). 25. T. A. J. Hedderson, R. Chapman, C. Cox, in Bryology for the Twenty-First Century, J. W. Bates, N. W. Ashton, J. G. Duckett, Eds. (Maney and the British Bryological Society, Leeds, 1998), pp. 65–77. 26. J. Hyvonen, T. A. J. Hedderson, G. L. Smith Merrill, J. G. Gibbings, S. Koskinen, Bryologist 101, 489 (1998). 27. T. A. J. Hedderson, D. J. Murray, C. J. Cox, T. L. Nowell, Syst. Bot. 29, 29 (2004). 28. B. Goffinet, C. J. Cox, A. J. Shaw, T. A. J. Hedderson, Ann. Bot. 87, 191 (2001). 29. T. A. J. Hedderson, R. Chapman, W. L. Rootes, Plant Syst. Evol. 200, 213 (1996). 30. C. J. Cox, B. Goffinet, A. J. Shaw, S. B. Boles, Syst. Bot. 29, 234 (2004). 31. C. J. Cox, T. A. J. Hedderson, Plant Syst. Evol. 215, 119 (1999). 32. T. J. O’Brien, in Pleurocarpous Mosses: Systematics and Evolution, A. E. Newton, R. S. Tangney, Eds. (CRC Press, Boca Raton, 2007), pp. 19–40. 33. N. E. Bell, A. E. Newton, in Molecular Systematics of Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Gardens Press, St. Louis, 2004), pp. 41–64. 34. W. R. Buck, B. Goffinet, A. J. Shaw, Mol. Phylogenet. Evol. 16, 180 (2000). 35. N. E. Bell, D. Quandt, T. J. O’Brien, A. E. Newton, Bryologist 110, 533 (2007).

36. 37.

38. 39.

40. 41.

42.

43. 44. 45.

46. 47. 48. 49.

50. 51.

52. 53. 54.

55. 56. 57.

145

N. E. Bell, A. E. Newton, Syst. Bot. 30, 34 (2005). E. De Luna, W. R. Buck, H. Akiyama, T. Arikawa, H. Tsubota, D. Gonzalez, A. E. Newton, A. J. Shaw, Bryologist 103, 242 (2000). W. R. Buck, C. J. Cox, A. J. Shaw, B. Goffinet, Syst. Biodiv. 2,121 (2004). D. Quandt, Pleurocarps Evolutionary Research, http:// www.pleurocarps.eu/research.htm (accessed April 21, 2008). W. R. Buck, B. Goffinet, A. J. Shaw, Bryologist 103, 774 (2000). S. Huttunen, M. S. Ignatov, K. Muller, D. Quandt, in Molecular Systematics of Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Gardens Press, St. Louis, 2004), pp. 328–364. S. Huttunen, D. Quandt, in Pleurocarpous Mosses: Systematics and Evolution, A. E. Newton, R. S. Tangney, Eds. (CRC Press, Boca Raton, 2007), pp. 145–162. A. Vanderporten, L. Hedenäs, C. J. Cox, A. J. Shaw, Taxon, 51, 115 (2002). A. Vanderporten, L. Hedenäs, C. J. Cox, A. J. Shaw, Mol. Phylogenet. Evol. 23, 1 (2002). M. S. Ignatov, A. A. Gardiner, V. K. Bobrova, I. A. Milyutina, S. Huttunen, A. V. Troitsky, Pleurocarpous Mosses: Systematics and Evolution, A. E. Newton, R. S. Tangney, Eds. (CRC Press, Boca Raton, 2007), pp. 177–213. D. Quandt, S. Huttunen, H. Streimann, J.-P. Frahm, W. Frey, Mol. Phylogenet. Evol. 32, 435 (2004). D. Quandt, S. Huttunen, R. Tangney, M. Stech, Systematic Botany, in press. J. Heinrichs, J. Hentschel, R. Wilson, K. Feldberg, H. Schneider, Taxon 56, 31 (2007). M. S. Ignatov, D. E. Scherbakov, in Pleurocarpous Mosses: Systematics and Evolution, A. E. Newton, R. S. Tangney, Eds. (CRC Press, Boca Raton, 2007), pp. 321–336. J. Gray, Palaeogeogr. Palaeoclimatol. Palaeoecol. 104, 153 (1993). P. Kenrick, P. R. Crane, The Origin and Early Diversification of Land Plants (Smithsonian Institution Press, Washington, D.C., 1997). C. H. Wellman, P. L. Osterloff, U. Mohiuddin, Nature, 425, 282 (2003). A. S. Konopka, P. S. Herenden, G. L. Smith Merrill, P. R. Crane, Int. J. Plant Sci. 158, 489 (1997). S. Koskinen, J. Hyvonen, Molecular Systematics of Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Garden Press, St. Louis, 2004). pp. 255–269. M. F. Neuberg, Trudui Geologic. Inst., S. S. S. R. 19, 1 (1960). M. S. Ignatov, Palaeontographica Abt. B. 217, 147 (1990). A. J. Shaw, C. J. Cox, I. Melosik, Molecular Systematics of Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Gardens Press, St. Louis, 2004).

Liverworts (Marchantiophyta) Niklas Wikströma,*, Xiaolan He-Nygrénb, and A. Jonathan Shawc a

Department of Systematic Botany, Evolutionary Biology Centre, Norbyvägen 18D, Uppsala University, Norbyvägen 18D 75236, Uppsala, Sweden; bBotanical Museum, Finnish Museum of Natural History, University of Helsinki, P.O. Box 7, 00014 Helsinki, Finland; c Department of Biology, Duke University, Durham, NC 27708, USA *To whom correspondence should be addressed (niklas.wikstrom@ ebc.uu.se)

Abstract Liverworts (Phylum Marchantiophyta) include 5000–8000 species. Phylogenetic analyses divide liverworts into Haplomitriopsida, Marchantiopsida, and Jungermanniopsida. Complex thalloids are grouped with Blasiales in Marchantiopsida, and leafy liverworts are grouped with Metzgeriidae and Pelliidae in Jungermanniopsida. The timetree shows an early Devonian (408 million years ago, Ma) origin for extant liverworts. The complex thalloid habit originated sometime in the Triassic (246–203 Ma). Both leafy and epiphytic habits are indicated as old features, but analyses also indicate possible extinctions during the Permian and Triassic (299–200 Ma) and rapid family and genus-level divergences during the Cretaceous and early Cenozoic (145–50 Ma).

The evolution of land plants marks one of the most important events in earth history. Because of their lengthy and well-documented fossil record, the major patterns in early land plant evolution have mainly been interpreted using macrofossil evidence from the vascular plant lineage (1). Liverworts (Fig. 1), in contrast, have a limited fossil record, they are easily neglected due to their small size, and their role in early land plant evolution is rarely emphasized. Nevertheless, growing evidence (1–7) indicates an early split in land plant evolution between the liverworts and all other land plants. This implies that liverworts occupy a critical position, and that they may help us understand the morphological and reproductive changes that favored the successful radiation of land plants and their adaptations to life in a terrestrial environment. Here we review recent progress in

our understanding of phylogenetic relationships among major lineages and the origin and divergence times of those lineages. Altogether, liverworts (Phylum Marchantiophyta) comprise an estimated 5000–8000 living species (8, 9). Early and alternative classifications for these taxa have been numerous [reviewed by Schuster (10)], but the arrangement of terminal taxa (species, genera) into larger groups (e.g., families and orders) based on morphological criteria alone began in the 1960s and 1970s with the work of Schuster (8, 10, 11) and Schljakov (12, 13), and culminated by the turn of the millenium with the work of Crandall-Stotler and Stotler (14). Three morphological types of plant bodies (gametophytes) have generally been recognized and used in liverwort classifications: “complex thalloids” including ~6% of extant species diversity and with a thalloid gametophyte that is organized into distinct layers; “leafy liverworts”, by far the most speciose group, including ~86% of extant species diversity and with a gametophyte that is differentiated into stem and leaves; and “simple thalloids” including ~8% of extant species diversity and with a more or less anatomically undifferentiated thalloid gametophyte.

Fig. 1 Leafy liverwort Schistochila aligera (Nees & Blume) J.B. Jack & Stephani, Schistochilaceae, Jungermanniales. Credit: X. He-Nygrén.

N. Wikström, X. He-Nygrén, and A. J. Shaw. Liverworts (Marchantiophyta). Pp. 146–152 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Viridiplantae; Streptophyta; Marchantiophyta

147

Calypogeiaceae 28

Arnelliaceae Gymnomitriaceae

27

23

Jungermanniaceae-1 Jungermanniaceae-2

19

Jungermanniaceae-3 Balantiopsidaceae

Cephaloziellaceae

20 12

Cephaloziaceae

16

Adelanthaceae

Jungermanniidae

Scapaniaceae 25

Lophocoleaceae 26

Plagiochilaceae Herbertaceae

9

18

Lepicoleaceae Lepidoziaceae

15

Trichocoleaceae Pseudolepicoleaceae Schistochilaceae

Jungermanniopsida

Myliaceae

Jungermanniales

Acrobolbaceae

Lejeuneaceae 22

6

Jubulaceae

17

Radulaceae Porellaceae

Porellales

Frullaniaceae

13

Lepidolaenaceae

5

24

Goebeliellaceae Ptilidiaceae


D

P

C

Tr

PALEOZOIC 400

350

300

J

K

250

200

150

Pg Ng CZ

MESOZOIC 100

50

0

Million years ago

Fig. 2 Continues

Our understanding of liverwort evolution has improved tremendously over the last 5–10 years, partly through the rapid accumulation of molecular sequence data. A series of large-scale and more fine-scaled phylogenetic analyses (9, 15–24) have contributed to a robust and well-supported hypothesis of liverwort relationships

(Fig. 2). Together these analyses indicate the presence of three major groups (classes): Haplomitriopsida, Marchantiopsida, and Jungermanniopsida (15). As expected, they do not strictly correspond to the three types of gametophyte, and the presence of a more or less undifferentiated simple thalloid gametophyte

Aneuraceae

14

Metzgeriaceae

7

4

Pleuroziaceae

Hymenophytaceae

2

Fossombroniaceae Pelliaceae Marchantiales 11

1

Sphaerocarpales

8

Blasiales Treubiales 3 D

Haplomitriales C

P

PALEOZOIC 400

350

300

Tr

J

K

MESOZOIC 250 200 150 Million years ago

Pg Ng

Pelliidae

Pallaviciniaceae 10

Haplomitriopsida

Makinoaceae

21

Marchantiopsida


Jungermanniopsida


Metzgeridae

148

CZ 100

50

0

Fig. 2 A timetree of liverworts. Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic). Jungermanniaceae-1 includes Nardia scalaris and Jungermannia obovata; Jungermanniaceae-2 includes Jungermannia ovato-trigona and Leiocolea collaris; and Jungermanniaceae-3 includes Saccogyna viticulosa, Harpanthus flotovianus, and Geocalyx graveolens in the original analyses by Heinrichs et al. (23).

may be an ancestral feature for liverworts as a whole, although formal reconstructions of ancestral morphological states have not been undertaken. The ancestral body plan is ambiguous because the Haplomtriopsida, which is well resolved as closest to Marchantiopsida and Jungermanniopsida, includes both leafy and more or less thallose types. Haplomitriopsida, the least diverse group in terms of living taxa, includes only the three genera Haplomitrium, Treubia, and Apotreubia. The Marchantiopsida include a monophyletic complex thalloid clade that is closest to the Blasiales, which has a simple thalloid morphology. The Jungermanniopsida, by far the most diverse group and possibly comprising 86% of extant species diversity, includes three distinct groups (subclasses): Pelliidae (including simple thalloid taxa and the more or less leafy Fossombroniales), Metzgeriidae (including simple thalloid taxa but also the leafy Pleuroziales), and Jungermanniidae (including all the leafy liverworts except Pleuroziales). Clearly, transitions between thallose and leafy body types have happened multiple times. With the exception of the analyses by Wheeler (24) and Boisselier-Dubayle et al. (17), there are no recent studies targeting the resolution of relationships among complex thalloid taxa in Marchantiopsida. Standard molecular

markers, commonly used in the more inclusive analyses, display too little variation to establish well-supported relationships in this group (9). Some analyses place Sphaerocarpales as the closest relative of all remaining taxa (15, 16, 22, 23), and this is also reflected in the classification by He-Nygrén et al. (16), who recognized two orders in the complex thalloid clade, Sphaerocarpales and Marchantiales. Relationships within each of the three Jungermanniopsida groups are better explored. A series of analyses have included a comprehensive sample of the simple thalloid groups Pelliidae and Metzgeriidae (9, 20, 22). Family-level relationships are comparatively well understood in both of these groups, although some of this knowledge has yet to be incorporated in a classificatory framework. Also the leafy group (Jungermanniidae) has been the focus of several analyses (15, 16, 18, 19, 23). These consistently identify two major groups (orders), Porellales and Jungermanniales, plus a smaller order, the Pitilidiales (containing Ptilidiaceae and Neotrichocoleaceae), whose relationship to the two larger clades is currently unresolved. Some analyses place the Ptilidiales in Jungermanniales (9, 19), but others indicate a closer relationship with the Porellales (16, 18, 23). The Ptilidiales may also be closest to a clade that


149

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among liverworts. Timetree Node

Estimates Ref. (23)(a)

Time

Ref. (23)(b)

Ref. (23)(c)

Ref. (27)

Time

CI

Time

CI

Time

CI

Time

CI

407.7

407.7

411–405

–

–

–

–

–

–

2

370.9

372.6

383–362

–

–

–

–

369.2

403–338

3

360.0

–

–

–

–

–

–

360.0

396–316

4

307.9

328.5

335–322

–

–

–

–

287.3

331–262

5

290.6

308.7

317–301

–

–

–

–

272.5

310–243

1

6

265.1

–

–

288.3

290–286

274.8

277–273

232.1

275–207

7

263.3

–

–

269.9

272–268

256.6

259–255

–

–

8

245.7

–

–

–

–

–

–

245.7

268–231

9

235.5

–

–

240.5

251–230

230.5

240–221

–

–

10

229.0

–

–

235.3

269–202

222.7

252–194

–

–

11

203.0

–

–

–

–

–

–

203.0

203–203

12

200.5

–

–

204.3

213–196

196.7

205–188

–

–

13

187.7

–

–

189.4

203–176

185.9

198–174

–

–

14

180.5

–

–

174.2

175–173

166.2

167–165

201.0

245–186

15

171.8

–

–

174.8

185–165

168.8

178–160

–

–

16

168.1

–

–

171.1

179–163

165.1

172–158

–

–

17

156.4

–

–

155.9

166–146

150.6

160–141

162.7

197–147

18

144.6

–

–

147.0

155–139

142.2

149–135

–

–

19

139.7

–

–

141.7

150–134

137.6

146–130

–

–

20

135.2

–

–

137.4

144–131

132.9

138–128

–

–

21

128.4

–

–

131.3

132–130

125.5

126–125

–

–

22

122.0

–

–

131.8

140–124

127.7

135–120

106.4

134–94

23

107.4

–

–

109.2

118–101

105.5

111–100

–

–

24

102.9

–

–

108.5

119–98

97.3

110–84

–

–

25

102.1

–

–

103.5

108–99

100.7

104–97

–

–

26

82.9

–

–

84.6

94–75

81.1

90–73

–

–

27

50.4

–

–

50.8

56–45

49.9

55–45

–

–

28

50.2

–

–

50.3

52–49

50.1

51–50

–

–

Note: Node times in the timetree represent the mean of time estimates from different studies. For Node 5, maximum (317 Ma) and minimum (301 Ma) ages from Heinrichs et al. (23) (a) were used as alternative calibration points in their second series of analyses yielding maximum (b) and minimum (c) age estimates for each node. Minimum age constraints forced were 112 Ma (Node 9), 203 Ma (Node 11), 90 Ma (Node 17, ref. 23), and 50 Ma (Node 28).

includes both the Jungermanniales and the Porellales. This uncertainty bears directly on our interpretation of ventral lobe (lobule) and water sac evolution in the leafy liverworts (9). A jungermannialian relationship for the Ptilidiaceae–Neotrichocoleaceae clade indicates that the elaboration of ventral lobes into water sacs may be a feature that is shared by a larger group than Porellales, where it is most commonly seen. Schuster (8) interpreted

these lobules as functioning in water retention and as an adaptation for an epiphytic habit (hence the term water sacs). An alternative interpretation is that they serve a nutritional purpose (25). The loss of a mycorrhiza-like association with Glomeromycota, shared by all Jungermanniopsida and possibly associated with a change in habit from terrestrial to epiphytic, support such an interpretation (26).

150


He-Nygrén et al. (16) recognized three suborders in the Porellales: Ptilidiineae (Ptilidiaceae and Neotrichocoleaceae), Porellineae (Porellaceae, Goebeliellaceae, Radulaceae, Frullaniaceae, Jubulaceae, and Lejeuneaceae), and Lepidolaenineae (Lepidolaenaceae). However, other analyses do not support monophyly of Porellineae and indicate alternative placements for both Porellaceae and Goebeliellaceae (9, 15, 19, 23) (Fig. 2). Recent analyses, based on multigene data sets, support monophyly of four larger groups within the Jungermanniales (9, 16, 19, 23), and these groups largely correspond to the four suborders recognized by He-Nygrén et al. (16): Perssoniellineae, Cephaloziineae, Jungermanniineae, and Lophocoleineae. We should expect changes with respect to the circumscription and comprehensive inclusiveness of each of these groups in the future. They are all diverse groups and future analyses, with an even more comprehensive taxon sample, will likely improve on our current understanding. The Myliaceae (Mylia taylorii), for example, was only recently included in the analyses, but their relationships are still poorly supported (23). Despite the rapid accumulation of molecular sequence data for liverworts, there are only two papers published that provide molecular estimates of liverwort divergence times at the hierarchical level covered in the present review. Both used a penalized likelihood approach and accounted for uncertainties in branch lengths and topology by analyzing 100 trees and parameter estimates drawn from the Bayesian posterior distribution of their phylogenetic analyses. The first study included 34 liverwort taxa but mainly focused on the diversification of mosses (Bryophyta) (27). Their analysis used a plastid two-gene data set (rbcL and rps4 genes) and a fi xed calibration point at 450 Ma for extant land plants (27). Spore tetrads are considered diagnostic of land plants, and the calibration point was based on a conservative date for their appearance in the fossil record (1). Seven minimum-age constraints were enforced in their analyses, six among vascular plants and one among liverworts. The liverwort constraint concerned the split between Marchantiales and Sphaerocarpales, forcing all estimates to be at least 203 Ma. The second paper focused mainly on the Jungermanniopsida clade of liverworts (23). This analysis included a twofold strategy where an internal calibration point for liverworts was obtained in a first series of analyses using a broader sample of land plants. This first series of analyses used a two-gene data set (rbcL and rps4 genes) and included 56 taxa (49 liverworts, two mosses,

one hornwort, three tracheophytes, and one algal outgroup). The first split among tracheophytes was used as calibration point and fi xed at 430 Ma and eight minimum-age constraints were enforced during the analyses, all of which concerned liverworts. They also constrained the embryophyte crown group at a maximum age of 475 Ma based on the occurrence of spore monads and diads from the Ordovician (488–444 Ma) that have been considered to represent liverworts (28). The analyses established upper and lower bound age estimates for the split between the Metzgeriidae and Jungermanniidae and these estimates were used as calibration points in their second series of analyses. The second set of analyses focused on the Jungermanniidae and used a three-gene data set (rbcL, rps4, and psbA genes) and 86 taxa (75 Jungermanniidae, five Metzgeriidae, five Pelliidae, and one Marchantiopsida outgroup). No less than 10 minimum-age constraints were enforced during the analyses, and the split between the Metzgeriidae and Jungermanniidae was fi xed at 301 and 317 Ma in two consecutive analyses based on results from their first series of analyses. Age estimates are also affected in these analyses by the constraints and there are nodes that in all analyses are forced toward their constrained ages (Table 1). Furthermore, using ages obtained from one analysis as constraints and/or calibration points in subsequent analyses is risky. Nevertheless, these analyses provide working hypotheses for the time course of liverwort diversification. The liverwort timetree (Fig. 2) shows an initial split between the Haplomitriopsida and remaining taxa in the early Devonian (408 Ma), and this estimate is comparable to that indicated by the macrofossil record. Pallaviciniites devonicus has been associated with liverworts, and this fossil taxon has been documented from the late Devonian (29, 30). Krassilov and Schuster (29) considered this fossil taxon as possibly related to the Pallaviciniaceae or Hymenophytaceae, but their derived positions and considerably younger ages indicate this to be incorrect. The living lineages of Marchantiopsida diversified in the early Triassic (246 Ma), but neither analysis (23, 27) provides a molecular-based estimate for the origin of the complex thalloid clade as a whole. Newton et al. (27) constrained the node at a minimum age of 203 Ma based on macrofossil evidence, but this forces the node to be 203 Ma in all analyses. In Heinrichs et al. (23) the node was also constrained at a minimum age of 225 in their first series of analyses, but they only report results for a few nodes from this analysis (Table 1). Unconstrained


analyses would likely indicate a younger age for this node. Branch lengths are considerably shorter among the complex thalloid taxa, indicating a decrease in evolutionary rate (9), and although the analyses incorporate such a decrease, we may still underestimate the magnitude of this deceleration. This would lead to molecular age estimates that are too young and if we accept the fossil-based information (31), this is what we are seeing. Leafy liverworts are the most successful group of liverworts in terms of species diversity and possibly account for as much as 86% of all extant liverworts. Although uncertainty remains about the origin of their leafy habit, even a conservative estimate places the origin in the Paleozoic. The earliest divergence among living Porellales and Jungermanniales appears to be of Permian age (265 Ma) and the leafy habit clearly had evolved by this time. The Pleuroziaceae, in the Metzgeriidae, are also leafy and leafyness may have originated already in the Early Permian (309 Ma). Although the origin of the leafy habit appears to have been a Paleozoic event, both the leafy groups Porellales and Jungermanniales show a Permian to Triassic pattern with extensive periods of time with little or no cladogenesis. This pattern corresponds with that indicated for mosses (27), and likely relates to an increased rate of extinctions during this time (32, 33). Also epiphytism appears to have originated early during liverwort evolution. A constant presence of epiphytism in all families of Porellales indicates an origin already by the late Permian. Much of the family- and genus-level divergences in the leafy Porellales and Jungermanniales appear in the Mesozoic (Fig. 2) and early Cenozoic [see (23) for genuslevel divergences]. Heinrichs et al. (23) linked this pattern, at least in part, to the development of tropical angiosperm-dominated forests, and they speculated that it corresponds to the pattern reported for other lineages of land plants such as lycopods (34) and leptosporangiate ferns (35). Crown-group Lejeuneaceae, for example, is dated as late Cretaceous (23), and is a highly diverse group, with as many as 1000 extant species, and they are predominantly epiphytic or epiphyllous. That the development of tropical angiosperm-dominated rain forests in the late Cretaceous and early Cenozoic triggered and mediated this diversification seems entirely plausible. Although the molecular estimates of liverwort divergence times that have been published so far (23, 27) cannot capture every aspect of the dynamics associated with the origin and diversification of liverworts into their present-day diversity, they have brought new insights into our understanding of liverwort evolution. No doubt,

151

this molecular approach has opened up a new avenue for tracing the origin and evolution, not only of liverworts, but also of any group of organisms with a limited fossil record.

Acknowledgments Support was provided by the Swedish Research Council to N. W. and by the U.S. National Science Foundation to A. J. S.

References 1.

2. 3. 4. 5. 6. 7.

8. 9.

10.

11. 12. 13. 14.

15. 16. 17. 18.

P. Kenrick, P. R. Crane, The Origin and Early Evolution of Land Plants, A Cladistic Study (Smithsonian Institution Press, Washington, 1997). B. D. Mishler, S. P. Churchill, Brittonia 36, 406 (1984). B. D. Mishler et al., Ann. MO Bot. Gard. 91, 451 (1994). Y.-L. Qiu, Y. Cho, C. J. Cox, J. D. Palmer, Nature 394, 671 (1998). Y.-L. Qiu et al., PNAS 103, 15511 (2006). M. Groth-Malonek, D. Pruchner, F. Grewe, V. Knoop, Mol. Biol. Evol. 22, 117 (2005). D. G. Kelch, A. Driskell, B. D. Mishler, Molecular Systematics of the Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Garden Press, St. Louis, USA, 2004), pp. 3–12. R. M. Schuster, New Manual of Bryology (Hattori Botanical Laboratory, Nichinan, 1984). L. L. Forrest, E. C. Davis, D. G. Long, B. J. CrandallStotler, A. Clark, M. L. Hollingsworth, The Bryologist 109, 303 (2006). R. M. Schuster, The Hepaticae and Anthocerotae of North America, East of the Hundredth Meridian, Vol. 1 (Columbia University Press, New York, 1966). R. M. Schuster, Bryologist 61, 1 (1958). R. N. Schljakov, Botanicheskii Zhurnal 57, 496 (1972) [In Russian]. R. N. Schljakov, Liverworts: Morphology, Phylogeny, Classification (Akademia Nauk S.S.S.R, Leningrad, 1975). B. Crandall-Stotler, R. E. Stotler, in Bryophyte Biology, A.J. Shaw and B. Goffinet (Cambridge University Press, Cambridge, UK, 2000), pp. 21–70. J. Heinrichs, S. R. Gradstein, R. Wilson, H. Schneider, Cryptogamie, Bryologie 26, 131 (2005). X. He-Nygrén, A. Juslén, I. Ahonen, D. Glenny, S. Piippo, Cladistics 22, 1 (2006). M.-C. Boisselier-Dubayle, J. Lambourdière, H. Bischler, Mol. Phyl. Evol. 24, 66 (2002). X. He-Nygrén, I. Ahonen, A. Juslén, D. Glenny, S. Piippo, Molecular Systematics of the Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Garden Press, St. Louis, USA, 2004), pp. 87–188.

152


19. E. C. Davis, in Molecular Systematics of the Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Garden Press, St. Louis, USA, 2004), pp. 61–86. 20. L. L. Forrest, B. J. Crandall-Stotler, in Molecular Systematics of the Bryophytes, B. Goffinet, V. Hollowell, R. Magill, Eds. (Missouri Botanical Garden Press, St. Louis, USA, 2004), pp. 119–140. 21. B. J. Crandall-Stotler, L. L. Forrest, R. E. Stotler, Taxon 54, 299 (2005). 22. L. L. Forrest, B. J. Crandall-Stotler, J. Hattori Bot. Lab. 97, 127 (2005). 23. J. Heinrichs, J. Hentschel, R. Wilson, K. Feldberg, H. Schneider, Taxon 56, XX (2007). 24. J. A. Wheeler, Bryologist, 103, 314 (2000). 25. I. Ahonen, Evolutionary Relationships of Liverworts with a Special Focus on the Order Porellales and the Family Lejeuneaceae (University of Helsinki, Helsinki, Finland, 2005). 26. I. Kottke, M. Nebel, New Phytologist 167, 330 (2005).

27. A. E. Newton, N. Wikström, N. Bell, L. L. Forrest, M. S. Ignatov, in Pleurocarpous Mosses: Systematics and Evolution, A. E. Newton, T. R. De Luna, Eds. (CRC Press, Boca Raton, USA, 2007), pp. 337–366. 28. C. H. Wellman, P. L. Osterloff, U. Mohiuddin, Nature 425, 282 (2002). 29. V. A. Krassilov, R. M. Schuster, New Manual of Bryology, Vol. 2 (The Hattori Botanical Laboratory, Nichinan, 1974), pp. 1172. 30. C. Oostendorp, Bryophyt. Biblioth. 34, 5 (1987). 31. J. A. Townrow, J. S. African Bot. 25, 1 (1959). 32. P. M. Rees, Geology 30, 827 (2002). 33. S. L. Wing, in Extinctions in the History of Life, P. D. Taylor, Ed. (Cambridge University Press, Cambridge, 2004), pp. 61–98. 34. N. Wikström, P. Kenrick, Mol. Phylog. Evol. 19, 177 (2001). 35. H. Schneider, E. Schuettpelz, K. M. Pryer, R. Cranfi ll, S. Magallón, R. Lupia, Nature 428, 553 (2004).

Ferns Kathleen M. Pryer* and Eric Schuettpelz Department of Biology, Duke University, Durham, NC 27708, USA *To whom correspondence should be addressed ([email protected])

Abstract Ophioglossoids, whisk ferns, marattioids, horsetails, and leptosporangiates form a well-supported monophyletic group of seed-free vascular plants sometimes referred to as monilophytes (ferns). With approximately 10,000 species, ferns are the closest relatives of the seed plants. All five of the major extant fern lineages were present by the end of the Carboniferous (~299 million years ago, Ma). The Permian (299–251 Ma) and Triassic (251–200 Ma) witnessed the establishment of many leptosporangiate lineages. But despite these ancient origins, several successive radiations in the Cretaceous Period (146–66 Ma) and Cenozoic Era (66–0 Ma) generated most of modern fern diversity.

Extant ferns form a monophyletic group of vascular plants (1–8) that number about 10,000 species divided unequally among five major lineages—ophioglossoids, whisk ferns, marattioids, horsetails, and leptosporangiates—and recognized in 11 orders and 37 families (9) (Fig. 1). These lineages are all spore-bearing and “seedfree” (10). By tradition they were previously classified together with lycophytes (the only other seed-free lineage of vascular plants) under the umbrella-term “pteridophytes” or “ferns and fern allies.” However, it is now clear that the five fern lineages together (without lycophytes) are the closest relatives of the seed plants (1–8), a position supported by the presence of euphylls—leaves with marginal or apical meristems and an associated leaf gap in the vascular stele (11)—and a 30-kb inversion in the plastid genome (12). An obvious morphological sharedderived character for ferns is lacking, but the monophyly of this clade is supported by sperm ultrastructure (6), sporophyte anatomy (1), and DNA sequence data (2–5, 8). Here, we provide an overview of fern relationships and divergence time estimates for the major groups. Within ferns, the first dichotomy separates two robustly supported clades: one consisting of ophioglossoids plus whisk ferns and the other containing

horsetails, marattioids, and leptosporangiates (4, 13). Whisk ferns and ophioglossoids are both relatively small lineages (~100 species total in two families and six genera; 9) and both have a poor fossil record. Because of the extent of morphological simplification in both families, their close relationship was only recently recognized from molecular phylogenetic studies (2, 3). Horsetails, marattioids, and leptosporangiates are well supported as a clade, but relationships among these lineages remain equivocal (2–5, 8, 13). Horsetails are an ancient group of plants with fossil relatives dating back to the Late Devonian (385–359 Ma), but today consist of a mere 15 living species (all in Equisetum). The length of the branch (in terms of the number of DNA substitutions per site) leading to its few living species (with no other living taxa to sample) may be a complicating factor in determining the exact relationship of Equisetum to other fern lineages (8, 14). The marattioid ferns are also an ancient group, first appearing in the middle Carboniferous. In the Late Carboniferous and Permian, several large marattioid representatives originated, including Psaronius, which reached heights of about 8 m. Marattioids experienced a decline in diversity since the end of the Paleozoic (251 Ma), and today are represented by about 150 species in at least four genera (9). The best-known and largest lineage of ferns is the leptosporangiates (Fig. 1), a monophyletic group of

Fig. 1 A leptosporangiate fern (Matonia pectinata R. Br.) from Malaysia. Credit: K. M. Pryer.

K. M. Pryer and E. Schuettpelz. Ferns (Monilophyta). Pp. 153–156 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

154

THE TIMETREE OF LIFE Eupolypods II-2 19

17

Eupolypods I-2

20

Eupolypods I-1 Pteridaceae (Pteroids)

16

Salviniaceae 15

8

Marsileaceae Schizaeoids-2

11

6

Schizaeoids-1 Gleichenioids-2

9

4

Gleichenioids-1

7

Hymenophyllaceae (Filmy ferns) 2

Tree ferns

Tree ferns-1

Schizaeoids

Tree ferns-2 13

Gleichenioids Heterosporous ferns

Lindsaeoids+

12

Leptosporangiates

Dennstaedtiaceae (Dennstaedtioids)

14

10

Polypods

Eupolypods II-1

18

Osmundaceae (Osmundaceous ferns) Marattiaceae (Marattioids)

1

3

Equisetaceae (Horsetails) Psilotaceae (Whisk ferns) 5

D

Ophioglossaceae (Ophioglossoids) P

C

PALEOZOIC 350

300

Tr

J

Cretaceous

Pg

MESOZOIC 250 200 150 Million years ago

Ng

CZ 100

50

0

Fig. 2 A timetree of ferns. Divergence times are shown in Table 1. All numbered nodes (except 3, 7, and 17) are well supported. Gleichenioids-1 = Gleicheniaceae; Gleichenioids-2 = Dipteridaceae and Matoniaceae; Schizaeoids-1 = Lygodiaceae; Schizaeoids-2 = Anemiaceae and Schizaeaceae; Tree ferns-1 = Cibotiaceae, Cyatheaceae, Dicksoniaceae, and Metaxyaceae; Tree ferns-2 = Culcitaceae (not sampled in 4), Loxomataceae, Plagiogyriaceae, and Thyrsopteridaceae (not sampled in 4); Lindsaeoids+ = Lindsaeaceae and Saccolomataceae;

Eupolypods I-1 = Dryopteridaceae in part; Eupolypods I-2 = Davalliaceae, Dryopteridaceae (in part), Lomariopsidaceae, Oleandraceae, Polypodiaceae, and Tectariaceae (not sampled in 19); Eupolypods II-1 = Aspleniaceae; and Eupolypods II-2 = Blechnaceae, Onocleaceae, Thelypteridaceae, and Woodsiaceae. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

more than 9000 species. These ferns have sporangia that develop from a single cell and have mature walls only one cell thick. Most leptosporangiate ferns also possess a distinctive annulus that serves to eject the (usually 64) spores. The osmundaceous ferns (Osmundaceae) are well supported as the closest relative of all other leptosporangiates (4, 13, 15). This position is consistent with the fossil record, as the oldest leptosporangiate fossils assignable to an extant lineage are members of this clade (16, 17). The filmy ferns, composing a single large family (Hymenophyllaceae) and the gleichenioid ferns,

with three smaller families (Dipteridaceae, Matoniaceae, and Gleicheniaceae), are both clearly monophyletic (13, 15). However, the precise relationships of fi lmy ferns and gleichenioids to one another and to the remaining leptosporangiate ferns remain unclear. The schizaeoid ferns are well supported as closest to the so-called “core leptosporangiates” (4, 13, 15), a large clade comprising three monophyletic groups: heterosporous ferns, tree ferns, and polypod ferns. Although not always thought to form a natural group (18), the monophyly of polypod ferns receives

Eukaryota; Viridiplantae; Streptophyta; Ferns

Table 1. Divergence time estimates (Ma) and their confidence/ credibility intervals among ferns. Timetree Node

Estimates Ref. (4)

Time

Ref. (19)

Time

CI

Time

CI

1

364.4

364.4

364–356

–

–

2

359.6

359.6

–

–

–

3

354.0

354.0

–

–

–

4

323.1

323.1

330–310

–

–

5

305.6

305.6

318–267

–

–

6

286.2

286.2

297–272

–

–

7

272.9

272.9

283–259

–

–

8

266.3

266.3

281–250

–

–

9

263.3

263.3

–

–

–

10

220.0

220.0

232–206

–

–

11

211.6

211.6

237–188

–

–

12

210.8

210.8

–

–

–

13

182.9

182.9

195–169

–

–

14

175.8

–

–

175.8

200–163

15

173.3

173.3

186–156

–

–

16

151.4

–

–

151.4

174–137

17

144.5

–

–

144.5

167–130

18

104.7

–

–

104.7

124–91

19

94.5

–

–

94.5

112–79

20

93.6

–

–

93.6

112–81

Note: For each node, the confidence interval was calculated from ref. (4) or ref. (19), as the mean ± 1.96 times the standard deviation (when the standard deviation was equal to zero or not reported, no interval is given).

solid support in all recent analyses (4, 13, 15, 19). This clade, accounting for about 80% of all living fern species, is united by an unequivocal derived morphological character—sporangia each with a vertical annulus interrupted by the stalk. The smaller of the two clades arising from the first divergence within the polypod clade contains the lindsaeoid ferns and a few rather enigmatic fern genera (Cystodium, Lonchitis, and Saccoloma). The remaining polypods compose three well-supported clades: the small dennstaedtioid clade (Dennstaedtiaceae), the large pteroid clade (Pteridaceae), and the hyperdiverse eupolypod fern clade. The relationships among these three lineages are ambiguous. Within the eupolypods, two large clades of roughly equal size are resolved, which were recently dubbed “eupolypods I” and “eupolypods II” (19). This split is well supported by molecular data, but also by a frequently overlooked

155

morphological character, namely the vasculature of the petiole (15). Eupolypods I have three or more vascular bundles (with the exception of the grammitid ferns with one, and the genus Hypodematium with two), whereas eupolypods II have only two (with the exception of the well-nested blechnoid ferns with mostly three or more). Integrating fossil time constraints together with molecular data from living taxa, two studies have estimated divergence times broadly across ferns (4, 19) (Fig. 2, Table 1). These divergence times are largely in accord with previous ideas about the times of origin and diversification of major fern clades (16, 17, 20–23). However, some clades (e.g., whisk ferns and ophioglossoids) with sparse fossil records are estimated to have originated much earlier than their meager fossil data would imply. The initial divergence among fern lineages occurred 364 Ma (Late Devonian). All four eusporangiate lineages, as well as the leptosporangiate ferns, were present by the end of the Carboniferous. Whisk ferns and ophioglossoids diverged from one another in the Late Carboniferous (306 Ma), and the earliest divergences among their living lineages occurred in the Late Cretaceous (88 Ma) and Middle Jurassic (162 Ma), respectively (4). As indicated by the fossil record, horsetails and marattioids had diverged from one another by the end of the Devonian; however, the divergences among living lineages within these groups appear to be more recent phenomena. Extant horsetails were estimated to have diversified in the Cenozoic (38 Ma; 4, see also 14). Extant marattioid lineages first diverged from one another in the Middle Triassic (237 Ma; 4). Within leptosporangiate ferns, the earliest divergences are estimated to have occurred in the Carboniferous and Permian. These divergences gave rise to the osmundaceous, filmy, gleichenioid, and schizaeoid ferns, as well as to the “core leptosporangiates” (Fig. 2). The initial divergence within the osmundaceous ferns is estimated to have occurred by the end of the Triassic (206 Ma; 4) and the two major filmy fern lineages diverged from one another in the Jurassic (163 Ma; 4, see also 24). The earliest divergence within the gleichenioid ferns occurred in the Permian (263 Ma), but diversification within the extant gleichenioid families (Gleicheniaceae, Dipteridaceae, Matoniaceae) appears to be more recent (Cretaceous, see 4). The initial divergence within schizaeoid ferns is estimated to have occurred in the Triassic, 212 Ma. A Late Triassic diversification gave rise to the three major lineages of “core leptosporangiates” (Fig. 2)—heterosporous ferns, tree ferns, and polypod ferns; the earliest divergences within each of these lineages occurred in the Jurassic.

156


All of the major polypod fern clades—lindsaeoids+, dennstaedtioids, pteroids, eupolypods I, and eupolypods II—had their origins in the Jurassic or Early Cretaceous (Table 1, Fig. 2). However, diversification in nearly all of these clades did not begin until the Late Cretaceous. This result alone suggests that at least 80% of extant fern diversity arose only in the last 100 million years. Ferns attained remarkable levels of diversity from the Carboniferous to the Jurassic. But despite their ancient origins and early successes, it appears that several radiations in the Cretaceous and Cenozoic generated the bulk of modern fern diversity (Fig. 2; 4, 19). This timing is suggestive of an ecological opportunistic response to the rise of angiosperms, as flowering plants came to dominate terrestrial ecosystems (19). Angiosperm-dominated communities likely promoted speciation in many lineages across the tree of life by creating new ecospaces into which these lineages could diversify (25). Polypod ferns may have been able to exploit new shady forest ecospaces specifically, through the evolutionary acquisition of a novel, physiologically more versatile, photoreceptor (19, 26). Better estimates of divergence times will allow us to more carefully evaluate potential links among profound biological phenomena and will help to elucidate those key events that have led to the many large, species-rich radiations in the long history of fern life on Earth.

Acknowledgment Support was provided by U.S. National Science Foundation to K.M.P. and E.S.

References 1. P. Kenrick, P. R. Crane, The Origin and Early Diversification of Land Plants: A Cladistic Study. (Smithsonian Press, Washington, D.C., USA, 1997).

2. D. L. Nickrent, C. L. Parkinson, J. D. Palmer, R. J. Duff, Mol. Biol. Evol. 17, 1885 (2000). 3. K. M. Pryer et al., Nature 409, 618 (2001). 4. K. M. Pryer et al., Am. J. Bot. 91, 1582 (2004). 5. Y.-L. Qiu et al., Proc. Natl. Acad. Sci. U.S.A. 103, 15511 (2006). 6. K. S. Renzaglia, R. J. Duff, D. L. Nickrent, D. J. Garbary, Phil. Trans. Roy. Soc. Lond. B 355, 769 (2000). 7. G. W. Rothwell, K. Nixon, Int. J. Plant Sci. 167, 737 (2006). 8. N. Wikström, K. M. Pryer, Mol. Phylogenet. Evol. 36, 484 (2005). 9. A. R. Smith et al., Taxon 55, 705 (2006). 10. K. M. Pryer, H. Schneider, E. A. Zimmer, J. A. Banks, Trends Plant Sci. 7, 550 (2002). 11. H. Schneider et al., in Developmental Genetics and Plant Evolution, Q. C. B. Cronk, R. M. Bateman, J. A. Harris, Eds. (Taylor & Francis, Philadelphia, 2002), pp. 330–364. 12. L. A. Raubeson, R. K. Jansen, Science 255, 1697 (1992). 13. E. Schuettpelz, P. Korall, K. M. Pryer, Taxon 55, 897 (2006). 14. D. L. Des Marais, A. R. Smith, D. M. Britton, K. M. Pryer, Int. J. Plant Sci. 164, 737 (2003). 15. E. Schuettpelz, K. M. Pryer, Taxon 56, 1037 (2007). 16. M. E. Collinson, in Pteridology in Perspective, J. M. Camus, M. Gibby, R. J. Johns, Eds. (Royal Botanic Gardens, Kew, UK, 1996), pp. 349–394. 17. W. D. Tidwell, S. R. Ash, J. Plant Res. 107, 417 (1994). 18. A. R. Smith, Amer. Fern J. 85, 104 (1995). 19. H. Schneider et al., Nature 428, 553 (2004). 20. G. W. Rothwell, Am. J. Bot. 74, 458 (1987). 21. G. W. Rothwell, Rev. Palaeobot. Palynol. 90, 209 (1996). 22. J. E. Skog, Brittonia 53, 236 (2001). 23. P. S. Soltis et al., Proc. Natl. Acad. Sci. U.S.A. 99, 4430 (2002). 24. E. Schuettpelz, K. M. Pryer, Syst. Biol. 55, 485 (2006). 25. C. S. Moreau et al., Science 312, 101 (2006). 26. H. Kawai et al., Nature 421, 287 (2003).

Gymnosperms Susanne Renner Department of Biology, Menzingerstr. 67, University of Munich, Munich, Germany ([email protected])

Abstract Gymnosperms (~1010 sp.) are grouped into four taxa: Coniferophyta, Cycadophyta, Ginkgophyta, and Gnetophyta. Most molecular phylogenetic analyses support the monophyly of extant gymnosperms, although relationships of the groups are not resolved. Some analyses place the root of gymnosperms between cycads and the remaining groups, while others place it between a cycad-Ginkgo clade and a conifer-gnetophyte clade. A nesting of gnetophytes inside conifers, closest to Pinaceae, is supported by some molecular analyses, but contradicted by others and morphological data. Most major gymnosperm lineages are extinct, and the abundant fossil record has not yet been well-integrated with molecular time estimates.

Gymnosperms, also called Acrogymnospermae (1), are a group of seed-bearing plants (spermatophytes) with ovules on the edge or blade of an open sporophyll or ovuliferous scale (Fig. 1). Their closest extant relatives are the angiosperms, which have ovules enclosed in a carpel. Gymnospermae is a problematic name because, when fossils are included as is usually the case, the name is widely understood to apply to a paraphyletic group of seed plants from which the angiosperms also arose (1). There are just over a 1000 living species of gymnosperms in the taxa Cycadophyta, Ginkgophyta, Coniferophyta, and Gnetophyta. Here, the relationships and divergence times of families in these phyla are reviewed. The conifers, Coniferophyta, include ~630 species in seven families of which Pinaceae is by far the largest and most widespread (12 genera, 225 species) (2). Araucariaceae (three genera, about 35 species) mostly occur in the tropics and subtropics, and are absent from Africa. Cephalotaxaceae (two genera, 10–12 species) are confined to Asia; Cupressaceae (including Taxodiaceae, 31 genera, 173 species) occur in mesic habitats worldwide; Podocarpaceae (including Phyllocladaceae and Nageiaceae, 17–19 genera, 180 species) occur mostly in the Southern Hemisphere; Sciadopityaceae comprise a single

species in the mountains of Japan; and Taxaceae (4–5 genera, 20–25 species) are again widespread, although nowhere abundant. The cycads, Cycadophyta, contain c. 300 species in the Families Cycadaceae (one genus, 97 species), Stangeriaceae (two genera, three species), and Zamiaceae (eight genera, 200 species) (3). Ginkgophyta contain only Ginkgo biloba, while Gnetophyta comprise three genera with together 80 species, Ephedra (50 species), Gnetum (30–35 species), and Welwitschia (one species). Molecular phylogenetic evidence for the close relationship between gymnosperms and angiosperms is strong. A study of seven genes (from the chloroplast, mitochondrial, and nuclear genome), with a sampling of 18 gymnosperms, 19 angiosperms, and numerous other landplants (192 species total), yielded maximum likelihood (ML) bootstrap values of 87% and 100%, respectively, for the monophyly of gymnosperms and angiosperms (4). With

Fig. 1 Female strobili of a gymnosperm (Pinus sylvestris) Credit: R. B. Zimmer.

S. Renner. Gymnosperms. Pp. 157–160 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

158

THE TIMETREE OF LIFE Podocarpaceae 7

Cupressaceae 8

Cephalotaxaceae 9

Taxaceae

3

Welwitschiaceae 11

Gnetaceae

10

2

Ephedraceae

4

Pinaceae 1

Conifers (including Gnetales)

Araucariaceae

6

5

C

Zamiaceae

P

Tr

PALEOZOIC 350

300

J

Cretaceous

200

150

Ng

CZ

MESOZOIC 250

Pg

Cycads

Ginkgoaceae Cycadaceae

100

50

0 Million years ago

Fig. 2 A timetree of gymnosperms. Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), J (Jurassic), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

12 genes (from all three genomes) and 23 exemplars (10 of them gymnosperms), and with nine genes (from the three genomes) and 12 exemplars (nine of them gymnosperms), the mutual monophyly of gymnosperms and angiosperms had 100% ML bootstrap support (5, 6). A parsimony analysis of 42 genes (from the three genomes) plus morphological data for seven land plants (four of them gymnosperms) yielded a support of, respectively, 97% and 100% for gymnosperms and angiosperms (7), and a Bayesian analysis of 56 chloroplast genes from 36 exemplars (five of them gymnosperms) yielded posterior probabilities of 1.0 for mutual monophyly angiosperms and gymnosperms (8). While extant gymnosperms thus appear monophyletic, the relative positions of cycads, Ginkgo, and gnetophytes remain unresolved. Most recent analyses place the deepest split between cycads and all remaining clades (4–7), but the 56-gene study placed cycads as closest to Ginkgo (8). The placement of gnetophytes as closest to Pinaceae and thus embedded in conifers is still weakly supported (4–6, 8). Morphological cladistic analyses (e.g., of 102 informative characters for 48 taxa, 25 extinct and 23 partly extant; 10) usually yield the so-called anthophyte topology (cycads (Ginkgo (conifers (gnetophytes, angiosperms)))), in contradiction to the DNA-based studies (for reviews of the contradictory results, see 5, 10, 11). Contradictions

probably arise from the inclusion of fossil taxa in morphological studies; there are many more extinct gymnosperm lineages than living ones. With fossils included, gnetophytes often group with Bennettitales, Pentoxylon, Caytonia, and angiosperms (10–16). However, trees that are a few steps longer place gnetophytes in conifers (11) or in a clade with living and extinct conifers and Ginkgo (10, 16). The absence of a solid gymnosperm phylogeny almost certainly is the reason that no study has focused on deep divergence times in this clade, in spite of the generally good fossil record of woods and cones that might be used for calibration. Whether the true topology is (cycads (Ginkgo (conifers including gnetophytes)), ((cycads, Ginkgo)(conifers, gnetophytes)), or the anthophyte topology will affect molecular dating to a larger or smaller degree depending on the node of interest. For example, in a study focusing on divergence times in the gnetophyte genus Gnetum (17) estimates for the ingroup barely differed with either of four alternative seed plant topologies (the approach used was a Bayesian relaxed molecular clock approach with ancestor-descendent correlated rates). Table 1 summarizes molecular divergence time estimates among gymnosperms and Fig. 2 presents a timetree. The age of the root node, that is, the split between extant angiosperms and gymnosperms has not been the

Eukaryota; Viridiplantae; Streptophyta; Gymnosperms

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among gymnosperms, from ref (17). Timetree Node

Time

CI

1

366

382–344

2

346

369–319

3

298

324–270

4

288

316–258

5

283

307–271

6

273

303–243

7

257

287–228

8

227

265–189

9

187

231–144

10

159

196–132

11

138

175–112

Note: Time estimates are based on rbcL and matK genetic distances analyzed with a Bayesian autocorrelated rates approach. The matrix included angiosperms and was rooted on the fern Psilotum. The unconstrained topology resulting from ML analysis of these data shows gnetophytes as the closest relative of all remaining seed plants. A conifer family not represented here, Sciadopityaceae, branches off below Cupressaceae + Taxaceae (4).

focus of any study. The oldest seed precursors are 385 Ma (18). These early seeds, which have lobed integuments that are thought to reflect their origin from fused sterile telomes, are quite different from the seeds of “modern” seed plants. Modern seeds with completely fused integuments are not known until the late Mississippian (~325 Ma; J. Doyle, personal communication, July 2005). Modern seed plants thus arose 385–325 Ma. Divergence time estimates for the deepest splits in extant conifers require much larger gene and taxon sampling than currently available. Conifers are documented from the middle Pennsylvanian (~310 Ma), and Cordaitales, which are probably basal stem relatives of conifers that existed in the earliest Pennsylvanian (~318 Ma). The deepest split among extant cycadophytes, should be at least 270 Ma, based on Crossozamia from the early Permian, which is similar to modern cycads (19). The rbcL-based estimate of 92.5 Ma for the split between Cycadaceae and Zamiaceae (20) is almost certainly an underestimate. The oldest Pinaceae-type cones are 225 Ma (21), and early Tertiary fossils are often assigned to extant genera of Pinaceae (22). Very young divergence time estimates within Pinaceae (23, 24) are probably unreliable because of miscalibration (25).

159

The split between Araucariaceae and Podocarpaceae may be at least 160 Ma based on Middle Jurassic Araucariaceae cones from Argentina; probable stem relatives of Podocarpaceae (Rissikia) date to the Triassic (J. Doyle, personal communication, July 2005). The oldest fossils of modern Araucariaceae are Albian (112–100 Ma) pollen grains that resemble those of Wollemia (26), the closest relative of Agathis. An age of 89 Ma from slightly younger pollen was used as the root constraint in a study of within-Agathis divergence times (27). The divergence between Cupressaceae and the Taxaceae/ Cephalotaxaceae clade has been dated to 227 Ma (Table 1) and that between Taxaceae and Cephalotaxaceae to 187 (Table 1) or 230–192 Ma (28). The divergence of extant Ephedraceae from the remaining gnetophytes is estimated at 159 Ma old; the oldest Ephedra seeds are Barremian–Albian (29, 30). The split between extant Welwitschiaceae and Gnetaceae may date back to 138 Ma ago (Table 1); the earliest fossils of Welwitschia are 110 Ma old (31).

Acknowledgment For critical comments, I thank J. Doyle, G. Grimm, and J. Hilton.

References 1. P. D. Cantino et al., Taxon 56, 822 (2007). 2. A. Farjon, World Checklist and Bibliography of Conifers, 2nd ed. (Royal Botanic Gardens, Kew, 2001). 3. L. M. Whitelock, The Cycads (Timber Press, Portland, OR, 2002), pp. 374. 4. Y.-L. Qiu et al., Int. J. Plant Sci. 168, 691 (2007). 5. J. G. Burleigh, S. Mathews, Int. J. Plant. Sci. 168, 111 (2007). 6. M. Hajibabaei, J. Xia, G. Drouin, Mol. Phylogenet. Evol. 40, 208 (2006). 7. J. E. B. de la Torre et al., BMC Evol. Biol. 6, 48 (2006). 8. C.-S. Wu, Y.-N. Wang, S.-M. Liu, S.-M. Chaw, Mol. Biol. Evol. 24, 1366 (2007). 9. S.-M. Chaw, C. L. Parkinson, Y. Cheng, T. M. Vincent, J. D. Palmer, Proc. Natl. Acad. Sci. U.S.A. 97, 4086 (2000). 10. J. Hilton, R. M. Bateman, J. Torrey Bot. Soc. 133, 119 (2006). 11. J. A. Doyle, J. Torrey Bot. Soc. 133, 169 (2006). 12. P. Crane, Ann. Missouri Bot. Gard. 72, 716 (1985). 13. K. C. Nixon, W. L. Crepet, D. Stevenson, E. M. Friis, Ann. Missouri Bot. Gard. 81, 484 (1994). 14. G. W. Rothwell, R. Serbet, Syst. Bot. 19, 443 (1994).

160

15. 16. 17. 18. 19. 20. 21. 22. 23. 24.


J. A. Doyle, Int. J. Plant Sci. 157(Suppl.), S3 (1996). E. M. Friis et al., Nature 450, 549 (2007). H. Won, S. S. Renner, Syst. Biol. 55, 610 (2006). P. Gerrienne et al., Science 306, 856 (2004). E. D. Brenner, D. W. Stevenson, R. W. Twigg, Trends Plant Sci. 8, 446 (2003). J. Treutlein, M. Wink, Naturwissenschaften 89, 221 (2002). C. N. Miller, Bot. Rev. 65, 239 (1999). B. A. LePage, Acta Hort. 615, 29 (2003). X.-Q. Wang, D. Tank, T. Sang, Mol. Biol. Evol. 17, 773 (2000). A. Eckert, B. Hall, Mol. Phylogenet. Evol. 40, 166 (2006).

25. A. Willyard et al., Mol. Biol. Evol. 24, 90 (2007). 26. D. J. Cantrill, J. I. Raine, Int. J. Plant Sci. 167, 1259 (2006). 27. M. Knapp, R. Mudaliar, D. Havell, S. J. Wagstaff, P. J. Lockhart, Syst. Biol. 56, 862 (2007). 28. Y. Cheng, R. G. Nicolson, K. Tripp, S.-M. Chaw, Mol. Phylogenet. Evol. 14, 353 (2000). 29. C. Rydin, K. R. Pedersen, E. M. Friis, Proc. Natl. Acad. Sci. U.S.A. 101, 16571 (2004). 30. Y. Yang, et al., Am. J. Bot. 92, 231 (2005). 31. C. Rydin, B. Mohr, E. M. Friis, Proc. Roy. Soc. Lond. Ser. B. 270 (Suppl.), 29 (2003).

Flowering plants (Magnoliophyta) Susana Magallón Departamento de Botánica, Instituto de Biología, Universidad Nacional Autónoma de México, 3er Circuito de Ciudad Universitaria, Del. Coyoacán, México D.F. 04510, Mexico (s.magallon@ibiologia. unam.mx)

Abstract Flowering plants (Magnoliophyta; angiosperms) are the predominant plants in modern terrestrial ecosystems. They include 270,000 known species distributed in eight major lineages. Amborellales, Nymphaeales, and Austrobaileyales are the earliest branches. A clade that includes Chloranthales and magnoliids, and another joining monocots, Ceratophyllales and eudicots form the core angiosperms. The molecular timetree places the origin of angiosperms at the onset of the middle Jurassic (175 million years ago, Ma) and its initial diversification during the middle Jurassic (167–159 Ma). Core angiosperms originated in the late Jurassic (150 Ma) and differentiated into five lineages by the latest Jurassic (148–146 Ma).

Flowering plants (Magnoliophyta, angiosperms) are among the most successful organisms in the history of life. Not only do they encompass an exceptionally vast morphological and phylogenetic diversity, but also they are the major determinants of ecological function and biotic composition in modern terrestrial ecosystems (Fig. 1). Angiosperms constitute a monophyletic group very well supported by molecular data and by a large number of unique traits. These traits include, for example, apical meristems with a two-layered tunica-corpus construction, circular bordered pits lacking margo and torus, and paracytic stomata. Angiosperms share numerous unique reproductive attributes, including the aggregation of pollen- and ovule-producing organs into structurally and functionally integrated units, that is, flowers, a bithecal and tetrasporangiate anther, a carpel enclosing the ovules, two integuments surrounding each ovule, a double fertilization process that results in embryo and endosperm, and several whole genome duplications. Concerted efforts among the international botanical community have led to the recognition of

the major clades within angiosperms and to an understanding of their relationships at all phylogenetic levels. Although a few families and genera remain to be phylogenetically placed, and particular regions of the angiosperm tree have for a long-time defied resolution, a solid understanding of the phylogenetic affinity of the majority of living angiosperms and of relationships among clades has been achieved. This phylogenetic knowledge has been translated into a classification that reflects current understanding of angiosperm evolutionary relationships (1). In this classification, major clades within the angiosperms are treated as orders, and informally named supraordinal clades are also recognized (1, 2). Angiosperms include approximately 270,000 known species distributed in 457 families (2), but the real number of species may exceed 400,000 (3). Angiosperms are distributed in eight major lineages. Amborellales, Nymphaeales, and Austrobaileyales are the three earliest branches, which encompass only a minute proportion of their standing species richness (110 Ma (39, 41). By the end of the Paleocene, most families in the order were differentiated, although many did not start to diversify until later (9, 39, 41). The disjunct distribution of the large family Myrtaceae in the Southern Hemisphere has been explained by a combination of the breakup of Gondwana in the early Cretaceous and subsequent long-distance dispersals between South America and Australasia and Africa and the Mediterranean basin (39, 42). The diversification of this family in the late Cretaceous supports this proposition (39, 41). On the other hand, because Vochysiaceae is a younger family, the presence of some of its members in Africa has been hypothesized to be the result of long-distance dispersal from South America, not the result of the breakup of Gondwana (39).

References 1. A. Takhtajan, Diversity and Classification of Flowering plants (Columbia University Press, New York, 1997).

195

2. R. M. T. Dahlgren, Bot. J. Linn. Soc. 80, 91 (1980). 3. A. Cronquist, An Integrated System of Classification of Flowering Plants (Columbia University Press, New York, 1981). 4. D. E. Soltis, P. S. Soltis, P. K. Endress, M. W. Chase, Phylogeny and Evolution of Angiosperms (Sinauer Associates, Sunderland, 2005). 5. D. E. Soltis et al., Bot. J. Linn. Soc. 133, 381 (2000). 6. D. E. Soltis, M. A. Gitzendanner, P. S. Soltis, Int. J. Plant Sci. 168, 137 (2007). 7. S. Magallon, P. R. Crane, P. S. Herendeen, Ann. MO Bot. Gard. 86, 297 (1999). 8. B. Bremer et al., Bot. J. Linn. Soc. 141, 399 (2003). 9. N. WikstrÖm, V. Savolainen, M. W. Chase, Proc. Roy. Soc. Lond. B 268, 2211 (2001). 10. V. H. Heywood, R. K. Brummit, A. Culham, O. Seberg, Flowering Plant Families of the World (Royal Botanic Gardens, Kew, 2007). 11. X. Y. Zhu et al., BMC Evol. Biol. 7, 217 (2007). 12. D. E. Soltis et al., Am. J. Bot. 90, 461 (2003). 13. R. K. Jansen et al., BMC Evol. Biol. 6 (2006). 14. V. Ravi, J. P. Khurana, A. K. Tyagi, P. Khurana, Mol. Phylogenet. Evol. 44, 488 (2007). 15. M. J. Moore, C. D. Bell, P. S. Soltis, D. E. Soltis, Proc. Natl. Acad. Sci. U.S.A. 104, 19363 (2007). 16. R. K. Jansen et al., Proc. Natl. Acad. Sci. U.S.A. 104, 19369 (2007). 17. D. E. Soltis et al., Proc. Natl. Acad. Sci. U.S.A. 92, 2647 (1995). 18. V. Savolainen et al., Kew Bull. 55, 257 (2000). 19. V. Savolainen et al., Syst. Biol. 49, 306 (2000). 20. K. W. Hilu et al., Am. J. Bot. 90, 1758 (2003). 21. L. B. Zhang, M. P. Simmons, Syst. Bot. 31, 122 (2006). 22. C. L. Anderson, K. Bremer, E. M. Friis, Am. J. Bot. 92, 1737 (2005). 23. K. J. Sytsma et al., Am. J. Bot. 89, 1531 (2002). 24. S. M. Swensen, Am. J. Bot. 83, 1503 (1996). 25. J. I. Sprent, Plant Soil 161, 1 (1994). 26. G. P. Lewis, B. Schrire, B. Mackinder, M. Lock, Legumes of the World (Royal Botanic Gardens, Kew, 2005). 27. F. Forest, M. W. Chase, C. Persson, P. R. Crane, J. A. Hawkins, Evolution 61, 1675 (2007). 28. P. S. Herendeen, P. R. Crane, in Advances in Legume Systematics 4, P. S. Herendeen, D. L. Dilcher, Eds. (Royal Botanic Gardens, Kew, 1992), pp. 57–68. 29. S. Magallon, M. J. Sanderson, Evolution 55, 1762 (2001). 30. M. Lavin, P. S. Herendeen, M. F. Wojciechowski, Syst. Biol. 54, 575 (2005). 31. D. J. Mabberley, The Plant Book, 2nd edn. (Cambridge University Press, Cambridge, 1997). 32. P. S. Manos, K. P. Steele, Am. J. Bot. 84, 1407 (1997). 33. R. Q. Li et al., Int. J. Plant Sci. 165, 311 (2004).

196


34. E. M. Friis, K. R. Pedersen, J. Schonenberger, Plant Syst. Evol. 260, 107 (2006). 35. M. L. Matthews, P. K. Endress, Bot. J. Linn. Soc. 149, 129 (2005). 36. C. C. Davis, C. O. Webb, K. J. Wurdack, C. A. Jaramillo, M. J. Donoghue, Am. Nat. 165, E36 (2005). 37. T. Tokuoka, H. Tobe, J. Plant Res. 119, 599 (2006).

38. 39. 40. 41. 42.

P. G. Wilson, M. M. O’Brien, M. M. Heslewood, C. J. Quinn, Plant Syst. Evol. 251, 3 (2005). K. J. Sytsma et al., Int. J. Plant Sci. 165, S85 (2004). E. Conti et al., Syst. Bot. 22, 629 (1997). F. Rutschmann, T. Eriksson, K. Abu Salim, E. Conti, Syst. Biol. 56, 591 (2007). E. J. Lucas et al., Taxon 56, 1105 (2007).

Eurosid II Félix Forest* and Mark W. Chase Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3DS, UK *To whom correspondence should be addressed ([email protected])

Abstract Rosids are divided into two main assemblages, Eurosid I and II, and a certain number of unplaced families and orders. This chapter deals with Eurosid II (malvids) and the putatively closely related families and orders (Geraniales, Crossosomatales, Aphloiaceae, Ixerbaceae, Strasburgeriaceae). Eurosid II comprises the Orders Sapindales, Malvales, Brassicales, and Huerteales, and the small Family Gerrardinaceae of uncertain position within the group. Eurosid II was established based on DNA sequence studies whereas morphological characters uniting it remain elusive. They diverged from their closest relatives 104–97 million years ago (Ma) and representatives of each order were present from the early Paleocene (66–56 Ma).

Eurosid II (~15,000 sp.) is smaller than its counterpart in the rosids, Eurosid I (>40,000 sp.) (1). Several studies have provided robust support for Eurosid II (2–6); Dipentodontaceae and Gerrardinaceae were only recently included in molecular-based analyses. Here, relationships and divergence times within Eurosid II are reviewed as well as for the putatively closely related families (Aphloiaceae, Ixerbaceae, and Strasburgeriaceae) and orders (Geraniales and Crossosomatales). Despite the strong support recovered for Eurosid II in numerous phylogenetic analyses, relationships within this group remain unclear with most possible arrangements retrieved by at least one study. In some studies, Brassicales (Fig. 1) are found to be the closest relatives of Malvales, with Sapindales closely related to this duo (6–8), whereas others associate Malvales with Sapindales, together closely related to Brassicales (9, 10) or Brassicales plus Family Tapisciaceae (3, 4). The adjacent position of Brassicales and Tapisciaceae is also found elsewhere, but this time more closely related to Malvales alone, with Sapindales being the first lineage to diverge in the group (11). The early divergence within Eurosid II of Sapindales is also found in the only analysis comprising both

Dipentodontaceae and Gerrardinaceae, in which these two families form an unresolved group with Brassicales and Malvales (5). In the timetree, Tapisciaceae is the first diverging lineage followed by Brassicales and the pair Sapindales–Malvales (Fig. 2). Based on these estimates, Eurosid II started to diverge 95–88 Ma ago, with the diversification of the three main orders, Brassicales, Malvales, and Sapindales, initiated 79–71, 71–67, and 62–57 Ma, respectively (12). The reader should be aware that, to date, phylogenetic relationships within Eurosid II remain poorly resolved, thus molecular estimates of divergence times can only be taken as preliminary results and viewed with caution; further analyses of the group could result in considerably different results. Brassicales consist of 15 families, all characterized by the production of mustard oil glucosides (glucosinolates), a feature found only in Brassicales and one other family, Putranjivaceae (Malpighiales) (13). The most commonly known member of Brassicales is without doubt Brassicaceae, the cabbage family. Brassicaceae is by far the largest family in the order with about 4130 species of mostly annual and perennial herbs distributed in some 356 genera, representing more than 90% of the species found in the order. They are found

Fig. 1 A member of Brassicaceae (Heliophila juncea) growing near Springbok, South Africa. Credit: F. Forest.

F. Forest and M. W. Chase. Eurosid II. Pp. 197–202 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

198


Sapindaceae 28

Meliaceae

24 14

Rutaceae

Sapindales

Simaroubaceae

18

Burseraceae 22

Anacardiaceae Dipterocarpaceae 32 31

7

Sarcolaenaceae

Muntingiaceae

23

Thymelaeaceae

16

Malvales

Cistaceae

27

Bixaceae 20

12

Malvaceae

6

Neuradaceae Akaniaceae 19

Tropaeolaceae Bataceae 29

Brassicaceae 30

21

Resedaceae Limnanthaceae

17

Brassicales

3

Koeberliniaceae

25

9

Setchellanthaceae

10

Caricaceae 15

2

Moringaceae

1

Ixerbaceae

4

Crossosomataceae 26

Stachyuraceae

13

Staphyleaceae Melianthaceae 5

Geraniaceae 11

Cretaceous

Vivianiaceae

Paleogene

MESOZOIC 100

Geraniales

Aphloiaceae 8

Crossosomatales

Tapisciaceae

Neogene

CENOZOIC 50

0 Million years ago

Fig. 2 A timetree of Eurosid II. Divergence times are shown in Table 1.

almost everywhere in the world, but are concentrated in the temperate zone of the Northern Hemisphere. This family comprises important crops including cabbage, broccoli, cauliflower (all from Brassica oleracea),

rocket (Hesperis), and radish (Raphanus), to name just a few (14, 15). Brassicaceae also count among their members Arabidopsis thaliana, a model organism intensively used in molecular biology for the

Eukaryota; Viridiplantae; Streptophyta; Magnoliophyta; Eurosid II

199

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among families of Eurosid II. Timetree Node


Time

Ref. (12)(b)

Ref. (12)(c)

Time

CI

Time

Time 99

1

107

107

111–103

103

2

104

104

108–100

100

97

3

95

95

99–91

92

88

4

95

95

99–91

93

91

5

92

92

96–88

94

88

6

90

90

94–86

89

85

7

82

82

86–78

84

80

8

81

81

86–76

81

81

9

79

79

83–75

76

71

10

72

72

76–68

72

68

11

71

71

75–67

85

79

12

68

68

72–64

71

67

13

62

62

68–76

61

56

14

61

61

65–57

62

57

15

60

60

65–55

61

58

16

59

59

63–55

65

61

17

58

58

61–55

60

54

18

56

56

60–52

57

55

19

55

55

60–50

61

54

20

54

54

58–50

60

58

21

54

54

57–51

54

52

22

51

51

56–46

52

47

23

51

51

54–48

62

58

24

47

47

51–43

47

45

25

45

45

48–42

47

46

26

44

44

49–39

44

42

27

42

42

45–39

55

54

28

40

40

44–36

44

43

29

37

37

40–34

43

42

30

33

33

36–30

42

39

31

23

23

25–21

39

36

32

14

14

16–12

28

28

Note: Node times in the timetree are based on branch lengths computed using (a) ACCTRAN optimization in maximum parsimony in ref. (12). Estimates based on (b) DELTRAN optimization in maximum parsimony and (c) maximum likelihood method are also shown.

study of floral development, among many other purposes. Tropaeolaceae (nasturtium family) is the second largest family in Brassicales with some 90 species of herbs found in Central and South America followed by

Resedaceae, which encompasses 70 species of herbs and shrubs, some of which are used as dye, perfume oil and cultivated as ornamental. Caricaceae (papaya family) consists of 34 species of small trees found mainly

200


in Central and South America (one genus found in Africa) of which papaya is the best-known member. Limnanthaceae (poached egg plant) comprises 10 species of annual herbs divided into two genera found in seasonally wet areas across North America and widely used as ornamentals. The timetree comprises 10 of the 15 families assigned to Brassicales, and relationships among these families are representative of those found in more inclusive analyses (10, 16). Based on the timetree, Brassicales split from its closest relatives 90–85 Ma ago, an estimate in agreement with the fossil record of the group for which the oldest remains are from the Turonian (17). The pair Brassicaceae–Resedaceae, representing a large portion of the species in the Order Brassicales, diverged relatively recently, 42–33 Ma ago, a somewhat young age for such a species-rich group, suggesting a recent and rapid diversification. Malvaceae (also sometimes called the core Malvales) is beyond doubt the largest of the 10 families currently placed in Malvales. It comprises about 2330 species of trees, shrubs, and herbs, cosmopolitan in distribution, and with many economically important representatives such as cotton (Gossypium), cola (Cola), chocolate (Theobroma), and also several members widely used as ornamentals (e.g., Tilia, Hibiscus). Thymelaeaceae is the second largest family in Malvales with about 800 species of shrubs and trees, but also lianas and herbs, of cosmopolitan distribution with greater numbers found in Africa and Australia. Dipterocarpaceae contains about 500 species of mostly large trees, often the dominant component of lowland tropical forests, especially in Southeast Asia. Many species are used as timber and also as source of oleoresins and as an alternative to cocoa butter (14, 15). Relationships within Malvales have been extensively investigated (18–20), and one of the most important results of these analyses is the expansion of Malvaceae to include several other families (Bombacaceae, Sterculiaceae, Tiliaceae; 20). More recently, the placement of Cytinaceae, a family comprising 10 species of root parasites, which has been shown previously to be closely related to Malvales, was identified as the closest relative of Muntingiaceae (21). Based on the timetree, Malvales started to diverge at the end of the Cretaceous, an estimate consistent with the fossil record (17). The diversification of the expanded Family Malvaceae started 41–34 Ma ago following their split from their closest relatives, Bixaceae, between 60 and 54 Ma. The close relationship between Bixaceae and Malvaceae is not always recovered in other analyses; this estimate should therefore be viewed with caution.

Unlike Malvales and Brassicales, which are characterized by the presence of one large family and several smaller ones, Sapindales have several large families including Sapindaceae (2215 species), Rutaceae (930 species), Anacardiaceae (600 species), Burseraceae (500 species), and Meliaceae (550 species). The other four families, range in size from 5 to 115 species. Sapindaceae is the largest and most diverse family of the order. They comprise several well-known and economically important members such as maple (Acer), litchi (Litchi), horse chestnut (Aesculus), and Sapindus, used in the manufacture of soap. Representatives of the family are mainly trees and woody climbers found in several parts of the world, but concentrated in the tropics. Rutaceae is possibly the most economically important family of the order as they comprise the genus Citrus (e.g., lemon, orange, and grapefruit) and several species that are sources of essential oils (e.g., Agathosma). The family consists of trees, shrubs, climbers, and herbs found in the tropics, especially in the Southern Hemisphere (14, 15). Members of Anacardiaceae are shrubs and trees found mostly in the tropical areas of the world, but also in temperate regions. The family comprises many species with edible fruits such as mango (Mangifera), cashew (Anacardium), pistachio (Pistacia), and pink peppercorn (Schinus), and is also used as ornamentals (e.g., sumac) (14). The economic importance of Family Meliaceae resides in the fact that several of its representatives are valued timber trees, among which the most well known are mahoganies (Swietenia and Khaya). The family, pantropical in distribution, also comprises species of medicinal value and with edible fruits. Although most recent molecular studies strongly support Sapindales, the morphological characters defining this group are not obvious (13). Two main groups of families in Sapindales are consistently recovered, generally with strong support, one comprising Meliaceae, Rutaceae, and Simaroubaceae and the other Anacardiaceae and Burseraceae (3, 7, 9, 22). The relationships between these two groups in Sapindales vary between analyses (3, 7, 9, 22, 23). Nitrariaceae was found as the closest relative to Anacardiaceae and Burseraceae (11), Kirkiaceae as the closest relative to the remainder of the order (22), to Sapindaceae alone (11) or to Anacardiaceae + Burseraceae (23), and when included, Biebersteiniaceae was the earliest diverging lineage in Sapindales (11, 23). Aceraceae (maples) and Hippocastanaceae (horse chestnuts) were included in Sapindaceae based on both morphological and molecular investigations (22, 24, 25). Based on the timetree, Sapindales started to diverge in the early Paleocene

Eukaryota; Viridiplantae; Streptophyta; Magnoliophyta; Eurosid II

(62–57 Ma), an estimate concordant with fossil remains (the oldest are from the Maastrichtian–Paleocene boundary). Muellner et al. (23) obtained much older estimates for the first split in Sapindales, between 132.6 and 90.5 Ma, depending on the method used. All major families were present by the beginning of the Late Eocene (44–40 Ma) according to the timetree, while Muellner et al. place the origin of most families in the order in the Cretaceous and early Tertiary (23). Tapisciaceae is a small family of only two genera of shrubs or small trees. Tapiscia with a single species is restricted to China whereas the four species Huertea are found in Central and South America (14). Tapisciaceae alone is most commonly placed as the closest relative to Order Brassicales (3, 4, 11), but also as the closest relative to the remainder of Eurosid II (9). A more recent study places it adjacent to Dipentodontaceae, with strong support, and both are found in a polytomy with Gerrardinaceae, Brassicales, and Malvales (5). Dipentodontaceae comprises a single species of small trees with a distribution restricted to south-central Asia and together with Tapisciaceae form the order Huerteales. Only Tapisciaceae is represented in the timetree. It diverged from the rest of Eurosid II 95–88 Ma (12), but its position is uncertain, and Dipentodontaceae is not represented in the timetree; thus, this estimate should be viewed with care. The two species of Gerrardina are the only representatives of the newly described Family Gerrardinaceae (5). Based on various characters of the leaves, floral disk, and stamens, it was initially placed in Family Flacourtiaceae (now Achariaceae or Salicaceae) in Order Malpighiales, but subsequent molecular sequence analyses showed that it was better placed in Eurosid II, although the relationships of this small family with the other more species-rich members of this group are unclear (5). Gerrardinaceae was not included in the timetree analysis (12). The second installment of the Angiosperm Phylogeny Group (26) placed Crossosomatales in the rosids, but their exact position within this group is not established. APGII recognized three families as members of this order, but four others were recently integrated (13). Crossosomataceae is a small family of nine species of shrubs endemic to North America. The ~15 species of shrubs and trees forming family Stachyuraceae are found mainly in China and Japan with some species cultivated as ornamentals. The largest family in Crossosomatales, Staphyleaceae, comprises about 46 species of shrubs and trees mostly distributed in the temperate regions of the Northern Hemisphere, but also in

201

the northern parts of South America as well as Southeast Asia (14). The inclusion of four other families found to be part of Crossosomatales in various studies is not as well supported as the group formed by the three families mentioned earlier (13). The only species of Family Strasburgeriaceae is a tree endemic to New Caledonia whereas the single species of Geissolomataceae is confined to a few mountain slopes in the Cape region of South Africa. Aphloiaceae also comprises only one species of shrub or trees found in East Africa and Madagascar, the leaves of which are used as tea in the Mascarenes (14, 27). Likewise, Ixerbaceae consists of only one species, an evergreen tree endemic to New Zealand (14). An additional family, Guamatelaceae, was recently described to accommodate the genus Guamatela, a genus of one species found in Mexico and Central America once thought to be related to Rosaceae, but now found nested in Crossosomatales (28). In most phylogenetic analyses based on DNA sequence data, Crossosomatales is found to be either the closest relative of Geraniales (7) or the closest relative of Eurosid II as a whole (3, 9). All families now consigned to Crossosomatales had an eventful past in terms of their systematics (13, 14, 29). Staphyleaceae, Stachyuraceae, and Crossosomataceae consistently form a well-supported group in which the former is the first diverging lineage (3, 9, 11). Aphloiaceae and Ixerbaceae are together the closest relatives of the rest of Crossosomatales in several studies, although this relationship is poorly supported (3, 9, 11). The only study including all eight families divided the order into two groups. In the first one, Staphyleaceae are the earliest lineage followed by the newly described Guamatelaceae and the pair Crossosomataceae + Stachyuraceae; these relationships are all well supported (28). In the second group, Geissolomataceae is the first diverging lineage followed by Aphloiaceae and the pair Strasburgeriaceae + Ixerbaceae; only the last are well supported (28, 29). These relationships are in general supported by floral characters except that based on these, the position of Geissolomataceae and Aphloiaceae would be inverted (30). Little can be said about the divergence times in this order since the fossil record is poor or absent, and only five of the eight families are included in the timetree. Nevertheless, Crossosomatales and related families would have started to diverge relatively early, 95–91 Ma ago (12), although these estimates are much older than those proposed by a subsequent study (47–40 Ma) (31). The Order Geraniales comprises four families, of which Geraniaceae are by far the largest. Geraniaceae is

202


widely distributed around the world with a concentration in temperate regions. Many of the ~800 species of herbs and shrubs are cultivated as ornamentals, such as genera Pelargonium and Geranium (14). The other three families are much smaller and less well known. Melianthaceae (including Greyiaceae and Francoaceae) comprises 18 species, all of them found on the African continent south of the Sahara, except for one species found in Chile and formerly assigned to Francoaceae; all 18 species of Vivianiaceae, mostly shrubs or herbs, are restricted to South America; Ledocarpaceae contains 12 species of shrubs all found in South America, mostly in the Andes. Apart from Geraniaceae and Vivianaceae, which were previously thought to be closely related, Geraniales, as defined by DNA-based studies, had never been proposed before. Furthermore, no known morphological character supports this assemblage (13). Geraniales did not always receive strong support (7, 11), although it is retrieved as a monophyletic group in most studies sometimes with strong or moderate support (3, 7, 9, 11). Relationships between Geraniaceae, Melianthaceae, and Vivianaceae are unclear; Geraniaceae is the closest relative to the pair Melianthaceae–Vivianaceae (all well supported) (3) or Melianthaceae are closely related to Vivianaceae and Geraniaceae (9). When Ledocarpaceae is included, it is found to be the closest relative of Vivianaceae, together closely related to Melianthaceae (11, 13). In our timetree, Melianthaceae is closely related to the pair Geraniaceae– Vivianaceae; Ledocarpaceae was not included (12). The first diversification in the order took place in the Late Cretaceous, 94–80 Ma (12), and Geraniaceae and Vivianaceae diverged 85–71 Ma. These estimates indicate that these families became established a relatively long time ago.

References 1.

S. Magallon, P. R. Crane, P. S. Herendeen, Ann. MO Bot. Gard. 86, 297 (1999). 2. P. S. Soltis, D. E. Soltis, M. J. Zanis, S. Kim, Int. J. Plant Sci. 161, S97 (2000). 3. D. E. Soltis, M. A. Gitzendanner, P. S. Soltis, Int. J. Plant Sci. 168, 137 (2007).

4. X. Y. Zhu et al., BMC Evol. Biol. 7, 217 (2007). 5. M. H. Alford, Taxon 55, 959 (2006). 6. M. J. Moore, C. D. Bell, P. S. Soltis, D. E. Soltis, Proc. Natl. Acad. Sci. U.S.A. 104, 19363 (2007). 7. V. Savolainen et al., Syst. Biol. 49, 306 (2000). 8. K. W. Hilu et al., Am. J. Bot. 90, 1758 (2003). 9. D. E. Soltis et al., Bot. J. Linn. Soc. 133, 381 (2000). 10. L. P. R. De Craene, E. Haston, Bot. J. Linn. Soc. 151, 453 (2006). 11. V. Savolainen et al., Kew Bull. 55, 257 (2000). 12. N. Wikstrom, V. Savolainen, M. W. Chase, Proc. Roy. Soc. Lond. B 268, 2211 (2001). 13. D. E. Soltis, P. S. Soltis, P. K. Endress, M. W. Chase, Phylogeny and Evolution of Angiosperms (Sinauer Associates, Sunderland, 2005). 14. V. H. Heywood, R. K. Brummit, A. Culham, O. Seberg, Flowering Plant Families of the World (Royal Botanic Gardens, Kew, 2007). 15. W. S. Judd, C. S. Campbell, E. A. Kellogg, P. F. Stevens, M. J. Donoghue, Plant Systematics: A Phylogenetic Approach, 3rd edn. (Sinauer Associates, Sunderland, Massachusetts, 2007). 16. J. C. Hall, H. H. Iltis, K. J. Sytsma, Syst. Bot. 29, 654 (2004). 17. S. Magallon, M. J. Sanderson, Evolution 55, 1762 (2001). 18. W. S. Alverson et al., Am. J. Bot. 85, 876 (1998). 19. W. S. Alverson, B. A. Whitlock, R. Nyffeler, C. Bayer, D. A. Baum, Am. J. Bot. 86, 1474 (1999). 20. C. Bayer et al., Bot. J. Linn. Soc. 129, 267 (1999). 21. D. L. Nickrent, Taxon 56, 1129 (2007). 22. P. A. Gadek et al., Am. J. Bot. 83, 802 (1996). 23. A. N. Muellner, D. D. Vassiliades, S. S. Renner, Plant Syst. Evol. 266, 233 (2007). 24. W. S. Judd, R. W. Sanders, M. J. Donoghue, Harvard Pap. Bot. 5, 1 (1994). 25. M. G. Harrington, K. J. Edwards, S. A. Johnson, M. W. Chase, P. A. Gadek, Syst. Bot. 30, 366 (2005). 26. B. Bremer et al., Bot. J. Linn. Soc. 141, 399 (2003). 27. D. J. Mabberley, The Plant Book, 2nd edn. (Cambridge University Press, Cambridge, 1997). 28. S. H. Oh, D. Potter, Syst. Bot. 31, 730 (2006). 29. K. M. Cameron, Bot. Rev. 68, 428 (2003). 30. M. L. Matthews, P. K. Endress, Bot. J. Linn. Soc. 147, 1 (2005). 31. C. L. Anderson, K. Bremer, E. M. Friis, Am. J. Bot. 92, 1737 (2005).

Monocots Cajsa Lisa Andersona,* and Thomas Janßenb a

Department of Biodiversity and Conservation, Real Jardin Botanico, CSIC, Plaza de Murillo 2, 28014 Madrid, Spain; bResearch Institute Senckenberg, Department of Botany and Molecular Evolution, Senckenberganlage 25, 60325 Frankfurt, Germany *To whom correspondence should be addressed ([email protected])

Abstract Grasses, lilies, orchids, and many other plants from all biogeographical and climatic regions of the world constitute the monocotyledonous plants (monocots). They form a natural group of about 59,300 species in 81 families supported by morphological and molecular evidence and include many important crops, such as rice and corn, and ornamental plants. Previous analyses and new analyses presented here suggest a rapid radiation of all major monocot lineages during the Early Cretaceous (146–100 million years ago, Ma). Most extant monocot families were present at the Mesozoic–Cenozoic boundary (66 Ma).

The monocots are a strongly supported monophyletic group comprising about 25% of the angiosperm diversity. They number 59,300 species (1) and are classified in 81 families and 10 orders by the Angiosperm Phylogeny Group (APGII) (2). A number of morphological characters are shared by most monocots, although these may have been (secondarily) lost in some lineages. The single cotyledon, leaves with linear venation, a basal meristem, scattered vascular bundles in the shoots and a lack of secondary growth of xylem and phloem, and sieve cell plastids are among the most obvious shared-derived morphological characters. Monocots usually possess trimerous flowers and uniaperturate pollen, which is most commonly monosulcate. Monocot characters also appear in other angiosperm groups. For example, sieve cell plastids occur in some Aristolochiaceae, scattered vascular bundles in Nymphaeaceae and some Piperaceae, and trimerous flowers with two perianth whorls are present in Nymphaeaceae and some magnoliids. Several monocots from different orders and families do have reticulate venation. This is, however, a derived condition thought to

represent an adaptation to shaded habitats such as the forest understorey. Acorales is a small wetland order, consisting of only one genus, Acorus (sweet flag). The small flowers are densely placed on a thick axis forming a spadix. This inflorescence produces a strong odor attracting pollinators. The plants possess ethereal oils in specialized cells. The Order Alismatales contains 14 families, which all have a preference for aquatic or wetland habitats. They possess stems with small scales or glandular hairs within the sheathing leaves at the nodes, extrorse anthers, and a large embryo. Araceae, or the arum family (Fig. 1), includes the calla lily and taro. The inflorescence of these plants is a spadix, which is surrounded by a leaflike bract. The small floating duckweeds, Lemna and other genera, have previously been segregated into the Family Lemnaceae, but are recognized as part of Araceae by APGII (2). Alismataceae, or the water plantains, have comparatively large flowers and usually two- to polyporate pollen. The embryo is strongly curved, and

Fig. 1 Monocot representatives. Top row, from left to right: Cypripedium calceolus L., Orchidaceae; Convallaria majalis L., Asparagaceae, with a straw of Carex digitata L., Cyperaceae. Bottom row, from left to right: Dracunculus muscivorus Parl., Araceae; Tillandsia usneoides L., Bromeliaceae; Cynosurus cristatus L., Poaceae. Credits: Ola Lundström (Dracunculus) and C. L. Anderson (all other images).

C. L. Anderson and T. Janßen. Monocots. Pp. 203–212 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

204


Poaceae 63

Ecdeiocoleaceae

62

Joinvilleaceae

40

Restionaceae 50

Anarthriaceae

48

32

Centrolepidaceae

Xyridaceae 20

36

Eriocaulaceae

Poales

Flagellariaceae

24

Bromeliaceae Juncaceae 54

19

Cyperaceae

44

Thurniaceae

30

Commelinids

Typhaceae 59

21

Sparganiaceae Rapateaceae Zingiberaceae 71

Costaceae

15

Marantaceae 73

Cannaceae Streliziaceae

70

Lowiaceae

61

Zingiberales

66

Musaceae 60

12

Heliconiaceae

18

Commelinaceae

43

Haemodoraceae

11

35

Hanguanaceae

27

Commelinales

Pontederiaceae 53

Philydraceae Dasypogonaceae Arecaceae


Early K

Late K

Paleogene

MESOZOIC 100

Ng

CENOZOIC 50

0 Million years ago

Fig. 2 Continues

the plants possess white latex. Hydrocharitaceae, tape grasses, occur in both freshwater and marine habitats. The genera display several different pollination mechanisms: showy flowers above water pollinated by nectargathering insects, detached male flowers floating on the water surface until they meet a female flower, exploding

anthers spreading pollen on the water surface, and underwater pollination are just some examples. In Alismatales we also find Potamogetonaceae, perennial herbs with either floating or submerged leaves and often jointed stems, Zosteraceae (seagrasses) consisting of a dozen species with ribbonlike leaves and creeping rhizomes,

Eukaryota; Viridiplantae; Streptophyta; Magnoliophyta; Monocots (continued from previous page)

Asparagaceae

51

Alliaceae

49

Xanthorrhoeaceae

42

10

205

Xeronemataceae 37

Iridaceae

34

Tecophilaeaceae

22

Ixioliriaceae

14

Orchidaceae

Asparagales

Doryanthaceae 33

Hypoxidaceae

13

47

Blandfordiaceae

41

Asteliaceae 29

Boryaceae

28

Lanariaceae Colchiaceae 67

6

Alstroemeriaceae

65

Luzuriagaceae

31

Rhipogonaceae

23

68

Philesiaceae

57

Smilaceae

56

16

Liliales

Melanthiaceae

Liliaceae Campynemataceae


Early K

Late K

Paleogene

100

Ng

CENOZOIC

MESOZOIC 50

0 Million years ago

Fig. 2 Continues

and a number of smaller families, often with only one genus, or even one species. The core monocots include the orders Dioscoreales, Pandanales, Liliales, Asparagales and the commelinid orders Zingiberales, Commelinales, and Poales. The Families Petrosaviaceae (closest to the remainder of the core monocots) and Dasypogonaceae (closest to the Commelinales + Poales + Zingiberales) are not placed in any of the orders. Dioscoreales include the predominantly tropical Family Dioscoreaceae, twining vines with net-veined leaves, and with several members cultivated for their edible starchy tubers, called yams. Two other families belong to this order, Burmanniaceae and Nartheciaceae. Pandanales include the screw-pines, which are woody plants, branching trees, or shrubs,

where the increase of trunk diameter is the result of primary thickening growth. Liliales consists of 10 families, often with showy flowers, possessing tepals with basal nectaries. Liliaceae (tulips, lilies, and others) are herbaceous with bulbs or corms. They have actinomorphic, hypogynous flowers, often with various color patterns. Colchicaceae (autumn crocus and naked ladies) are a family of mainly seasonal perennials occurring in dry habitats in Africa and Eurasia, with a few exceptions confined to Australian rainforests and wet sclerophyllous forests. The presence of the highly poisonous alkaloid colchicine is a synapomorphy of the family. Within the Liliales we also find eight other families, examples being Alstroemeriaceae, with a distribution from Central America to southern South America,

206


Stemonaceae

25

Pandanaceae 55

Cyclanthaceae Dioscoreaceae

17

Burmanniaceae

9

Nartheciaceae

Dioscoreales

38

Pandanales

Velloziaceae

4

Petrosaviaceae Zosteraceae 75

Potamogetonaceae

72

2

Ruppiaceae

64

Posidoniaceae 58

69

Cymodoceaceae

46

Scheuzeriaceae

39

Aponogetonaceae

1

26

Hydrocharitaceae 52

Butomaceae

45

5

Alismatales

Juncaginaceae

Limnocharitaceae 74

Alismataceae

3

Tofieldiaceae Araceae Acoraceae

Early K

Late K

Paleogene

MESOZOIC 100

Ng

CENOZOIC 50

0 Million years ago

Fig. 2 A timetree of monocots. Divergence times are shown in Table 1. Abbreviations: K (Cretaceous) and Ng (Neogene).

Campynemataceae, comprising a few species of perennial herbs in New Caledonia and Tasmania, Corsiaceae, which is a family of nonphotosynthesizing herbs and Smilacaceae, with representatives that typically have woody roots and a climbing or vining growth form. Asparagales, according to APGII, include a large number of families. Their actual number varies according to the delimitation of individual families that is adopted. Orchidaceae, the orchids (Fig. 1), is the second largest of all plant families with about 20,000 known species, the majority being epiphytes in tropical rainforests. Floral organization is relatively constant with two whorls of tepals, and most often a column consisting of a single stamen adnate to the style and stigma. The flower forms and colors, however, display a great variety. Other examples

of families within the Asparagales are Iridaceae (e.g., iris and saffron), Xanthorrhoeaceae (grass-tress, aloe, and asphodels), Alliaceae (onion, leek, and garlic), and Asparagaceae (asparagus, Lily-of-the-Valley (Fig. 1), agaves), the last two families now circumscribed to include a number of previously recognized families (Fig. 1). The commelinids comprise about half of the species of monocots, and include the orders Arecales, Zingiberales, Commelinales, and Poales. The Arecales (palms) is a large order of woody tree-like or rarely climbing plants with pinnately or palmately veined leaves and trunks with primary thickening growth. Zingiberales comprise large, rhizomatous herbs with showy flowers. Well-known representatives are the banana in the Musaceae family, birdof-paradise flower in Strelitziaceae, the canna-lilies in

Eukaryota; Viridiplantae; Streptophyta; Magnoliophyta; Monocots

207

Table 1. Divergence times (Ma) among monocots. Timetree Node

Time

Estimates

Timetree

Ref. (24)(a)

Ref. (24)(b)

Ref. (24)(c)

Time

Time

Time

Node

Estimates

Time

Ref. (24)(a)

Ref. (24)(b)

Ref. (24)(c)

Time

Time

Time

1

134

134

134

134

39

99

99

108

98

2

131

131

124

131

40

99

99

97

-

3

128

128

123

128

41

98

98

39

104

4

126

126

107

126

42

97

97

53

100

5

125

125

123

124

43

96

96

97

98

6

124

124

104

124

44

96

96

97

98

7

124

124

104

124

45

96

96

92

-

8

124

124

102

124

46

93

93

106

92

9

123

123

101

123

47

93

93

39

100

10

122

122

102

122

48

91

91

92

97

11

120

120

100

120

49

91

91

45

93

12

118

118

98

119

50

89

89

90

96

13

118

118

70

-

51

89

89

40

91

14

117

117

70

119

52

88

88

78

88

15

116

116

98

117

53

86

86

97

89

16

116

116

98

112

54

86

86

97

88

17

115

115

75

117

55

86

86

41

98

18

114

114

101

116

56

86

86

40

91

19

112

112

97

114

57

85

85

40

90

20

111

111

98

113

58

83

83

98

82

21

110

110

97

112

59

83

83

70

89

22

110

110

66

112

60

78

78

36

88

23

110

110

71

-

61

78

78

35

87

24

109

109

98

-

62

76

76

97

90

25

109

109

90

-

63

75

75

97

89

26

108

108

114

-

64

75

75

92

-

27

107

107

97

110

65

75

75

59

79

28

107

107

40

113

66

75

75

35

-

29

107

107

40

109

67

72

72

59

76

30

106

106

97

-

68

71

71

30

76

31

106

106

71

109

69

70

70

70

67

32

105

105

98

108

70

70

70

35

78

33

105

105

60

108

71

70

70

35

79

34

104

104

59

107

72

66

66

85

65

35

101

101

97

104

73

61

61

26

68

36

101

101

97

105

74

58

58

46

57

75

47

47

77

47

37

101

101

55

103

38

100

100

90

108

Note: Node times in the timetree are obtained from a penalized likelihood (a) reanalysis of the data set in ref. (24). Estimates from (b) PATHd8 and (c) nonparameteric rate smoothing are also shown (24).

208


Cannaceae, and ginger and cardamom in Zingiberaceae. The largest family within the Commelinales is the Commelinaceae (spiderworts). The latter are more or less succulent herbs with colorful flowers often having fringed filaments. The Poales contain some families highly specialized for wind pollination, for example Poaceae (grasses, Fig. 1), Cyperaceae (sedges, Fig. 1), Juncaceae (rushes), and Restionaceae. Within Poales we find important crop plants like barley and rice (Poaceae) or pineapple in the Family Bromeliaceae, which contains genera with showy flowers pollinated by insects, birds, or bats. The Spanish moss (Fig. 1) and many other tropical epiphytes belong here. Despite the monocots being recognized as a group since the seventeenth century (3), relationships of and within the group were poorly understood before molecular phylogenetic studies starting in the mid-1990s. Since then several larger studies using extensive samplings and molecular data from plastid, nuclear, and mitochondrial regions have been conducted (e.g., 3–8). Today, there is a broad consensus about a stable backbone of the monocot phylogeny, even though some nodes are still not convincingly resolved. Mitochondrial phylogenies are in conflict with the generally accepted monocot phylogeny. However, the results from mitochondrial analyses have been suggested to suffer from error sources like paralog sampling and highly divergent evolutionary rates (4, 8). Chase et al. (9) recognized six major monophyletic groups: Alismatales, Dioscoreales, Pandanales, Liliales, Asparagales, and commelinids (including Poales, Commelinales, Zingiberales, Dasypogonaceae, and Arecales). Acorus was concluded to be the closest relative of the rest of the monocots, but internal relationships between the larger clades were less well supported. Relationships between the orders have been extensively analyzed since Chase et al. and most studies propose similar topologies (e.g., 1, 3–8). Current consenus has Acorales as closest to all other monocots. Alismatales attaches to the next higher node and is closest to the remainder of monocots, that is the so-called core monocots. Petrosaviaceae is branching off after Alismatales, and is hence the closest relative of all other core monocots. The remaining problematic area in the monocot topology concerns the relative position of Pandanales, Dioscoreales, and Liliales. The position of Liliales with respect to the other orders remains ambiguous. Recent multigene analyses (5, 7) suggest Pandanales and Dioscoreales to be closest relatives, although with moderate support. The same studies conclude that Asparagales and the commelinids are closest relatives, one study (7) obtaining high support

and the other (5) obtaining moderate support for this group. To this date, the majority of phylogenetic analyses tend to support Liliales as the closest relative to the Asparagales-commelinids-clade (2, 5–7). Following molecular phylogenetic studies, the traditional distinction between Liliales and Asparagales based mostly on characters of the seed and nectaries has been revised and major taxonomical rearrangements have been made. Most notably, two large families earlier classified within Liliales, the Iridaceae and Orchidaceae, are now transferred to Asparagales (2). Within the Liliales, the main Family Liliaceae is currently more narrowly circumscribed than in earlier taxonomical treatments (10). Within the commelinid clade Arecales are closest to the remainder, most likely followed by Dasypogonaceae. Poales and Zingiberales form a clade, which is closest to Commelinales (11). Detailed studies of phylogenetic relationships within monocot orders are available for Asparagales (12), Dioscoreales (13), Liliales (13, 14), Poales (11), and Zingiberales (15). A major surprise arising from molecular phylogenetic reconstructions is the position of the Family Hydatellaceae (formerly classified in the Poales), which is now suggested not to be a monocot, but most closely related to the Nymphaeales (16). The closest relative of the monocots has yet to be determined. A number of studies suggest eumagnoliids (with slightly different definitions of the latter) (8, 17, 18), other studies Ceratophyllum (19–21), and others Piperales (22) or Laurales (23). The largest data set employed for monocot dating so far has been compiled by Janssen and Bremer (24). There are some differences in the tree topology derived from the Janssen and Bremer data set (878 taxa, or “800+ data set”) as compared to the more recent phylogenetic studies. However, it has been shown (25) that the influence of minor changes in topology on divergence time estimation is small compared to the influence of alternative fossil calibrations, methods, and taxon sampling. The differences between the topology derived from the 800+ data set and recent phylogenetic studies are all within orders, not in the monocot backbone. Within the Alismatales, Aponogetonaceae branches off before Scheuchzeriaceae, instead of the opposite, and Ruppiaceae is the closest relative of the Potamogetonaceae–Zosteraceae clade instead of being closest to Posidoniaceae. The orders Dioscoreales and Pandanales are not closest relatives, but are collapsed into a trichotomy with the rest of the core monocots. The Orchidaceae is not the closest relative of the rest of the Asparagales in the 800+ data set. Instead, there is a basal split between the Orchidaceae


and the clade consisting of Boryaceae, Blandfordiaceae, Lanariaceae, Asteliaceae, and Hypoxidaceae. Internal nodes of the latter clade generally receive low support, and the branching order differs substantially between studies. Within the Commelinales, Commelinaceae is closest to Pontederiaceae, and this clade constitutes the closest relative of Haemodoraceae. Within the Poales, Restionaceae and Anarthriacae form a clade, with Centrolepidaceae as its closest relative, as opposed to the more recently suggested Restionaceae–Centrolepidaceae relationship, with Anarthriacae branching off before (7). Flagellariaceae are the the closest relative of a larger clade consisting of Restionaceae–Anarthriacae– Centrolepidaceae and Joinvilleaceae–Ecdeiocoleaceae– Poaceae, instead of being closest to the latter clade only. The divergence time of the living lineages of monocots has been estimated in a number of studies: Savard et al. (26) proposed 200 Ma, Goremykin et al. (27) 160 Ma, and Leebens-Mack et al. (28) 135–131 Ma. Bremer (29) estimated the split between Acorales and the rest of the monocots to 134 Ma, with a possible age span of 147–121 Ma. The age 134 Ma coincides with the earliest recorded fossil angiosperm pollen (30), and could therefore be regarded as plausible. This age was later used as a fi xed age constraint for the monocot root node by Janssen and Bremer (24). The Janssen and Bremer study is the dating study with the most extensive sampling (800+ data set) of monocot taxa. The authors use evidence from eight reference fossils to calibrate a nonparametric rate smoothing (31) analysis. The study revealed that all major lineages diverged in the Early Cretaceous (146–100 Ma), with most families being present at the Mesozoic–Cenozoic (M-C) boundary. Uncertainties associated with each node (divergence age estimate) were suggested to be in the order of ±10–20 million years. Developments following Janssen and Bremer’s study include new molecular dating methods and extended possibilities of incorporating fossil constraints in dating analyses, plus new fossil discoveries that may be used for constraining and calibrating divergence time estimations of monocots. We have compared the author’s original results from NPRS dating by reanalyzing the same molecular dataset using the penalized likelihood (PL) (32) as well as the PATHd8 (33) methods, and included additional age constraints from five new fossils. We have utilized the earliest occurrence of extinct lineages of Araceae (34) to provide a minimum age (120 Ma) for the split between Alismatales and the core monocots. A fossil assigned to the Family Triuridaceae (35) gives a minimum age of 90 Ma for the living lineages of Pandanales.

209

The placement of this fossil has, however, been disputed (36). Fossil grasses belonging to different lineages within Poaceae (37) assign a minimum age of 65 Ma to the living lineages of Poaceae. A fossil belonging to the palm subtribe Mauritiinae (38) attributes a minimum age of 65 Ma to the living lineages of Arecaceae. Finally, pollen belonging to the African Restionaceae clade (39) was used as a minimum age constraint for the living lineages of Restionaceae. Monocots are an assemblage of many lineages rather heterogeneous with respect to life forms, ecological preferences, and hence most likely also with respect to their evolutionary history. Divergence time analysis of the entire monocots will hence face the problem of rate heterogeneity in different parts of the tree. For example, phylograms reveal that palms have very short internal branches, whereas grasses have long branches, suggesting a slowdown in evolutionary rate in the former group, and a speedup in the latter (40). In such heterogeneous trees, the dating algorithm and the number of fossil constraints used may have considerable influence on age estimates (33). Furthermore, closest relatives with many vs. few representatives, with highly different branch lengths, with different habits (e.g., woody vs. herbaceous), and with long vs. short generation times frequently occur in the monocot tree and may account for further analytical problems. Age estimates of such lineages should be regarded as a rough approximation and should be interpreted with caution. Herein, we compare divergence time estimates from three different methods highlighting where major discrepancies occur and briefly discussing possible causes. Our reanalysis using PL yields highly similar age estimates to the NPRS study (see Table 1). Within Poales, the PL age estimates are slightly older, with deviations up to 15 million years and therefore within the suggested error range for the NPRS analysis (24). Age estimates obtained with PATHd8 sometimes differ substantially from those obtained with PL and NPRS. Except for the divergences of the families within the Alismatales from their closest relatives (see Fig. 2), PATHd8 generally obtains younger ages than PL or NPRS (see Table 1). Divergences of families within Pandanales and Dioscoreales differ in the magnitude of 20 Ma. Within Liliales, the estimates differ in the magnitude of 40 Ma. Divergences of families within the Asparagales deviate as much as 50–70 Ma from results obtained with PL and NPRS. PATHd8 also often suggests more rapid divergences. The divergence of Alismatales from its closest relative was dated to 131 Ma, and the divergence of all living

210


lineages of Alismatales to 128 Ma by Janssen and Bremer. In the PL analysis presented here, both divergence events receive the same ages, while the PATHd8 analysis yields somewhat younger ages, 124 and 123 Ma, respectively. The oldest fossil that can be assigned to the Alismatales (34) is from the Early Cretaceous, about 120–110 Ma old, which means that all of the three methods yield results in agreement with the fossil record. Janssen and Bremer estimated the divergence of Petrosaviaceae from its closest relative to 126 Ma, and the divergence of its living lineages to 123 Ma. The former age estimate is also obtained using PL, whilst the latter is then 121 Ma. PATHd8 estimates both divergence events at 107 and 41 Ma, respectively, which is substantially younger. The split of Dioscoreales from their closest relatives was dated to 124 Ma, and the divergence of living lineages to 123 Ma in Janssen and Bremer (24). We obtain the same results using PL, whereas PATHd8 estimates these divergences at 104 and 101 Ma, respectively. The NPRS derived age for the split of Pandanales from its closest relative is 124 Ma, that for the divergence of living lineages of Pandanales is 114 Ma. Using PL, the former age is identical and the latter age is 109 Ma. PATHd8 yields slightly younger ages 104 and 90 Ma, respectively. The age for the divergence of Liliales from its closest relative is estimated to 124 Ma by both NPRS and PL. The divergence of living lineages of Liliales is estimated to 117 Ma by NPRS and 115 Ma by PL. PATHd8 suggests ages that are substantially younger, 102 and 75 Ma, respectively. The divergence of Asparagales from their closest relative is dated to 122 Ma by both NPRS and PL, while PATHd8 estimates this age to 102 Ma. Both NPRS and PL suggest a gradual divergence within the living lineages of that order, starting at 119 and 118 Ma, respectively. PATHd8 suggests a much younger divergence time for the living lineages, 70 Ma, and rapid divergence at the internal nodes of the Asparagales at 40 Ma. The split of Arecaceae from their closest relative is dated to 120 Ma by both NPRS and PL, the divergence of living lineages of that family is estimated to 110 Ma by NPRS and 97 Ma by PL. PATHd8 gives the ages 100 and 65 Ma, respectively. Dasypogonaceae diverged at 120 Ma (NPRS and PL) or 100 Ma (PATHd8) from its closest relative, and the living lineages diverged at 100 Ma (NPRS), 88 Ma (PL), or 39 Ma (PATHd8), respectively. The most recent common ancestor to the most closely related Commelinales and Zingiberales receives the age 114 Ma according to NPRS, 112 Ma according to PL, and 97 Ma according to PATHd8. The living lineages of Commelinales diverged at 110 Ma (NPRS) or at 107 Ma

(PL), followed by a gradual divergence. PATHd8 estimates the divergence of Commelinales from Zingiberales to 97 Ma, with a rapid divergence of all families within Commelinales. Living lineages of Zingiberales are estimated to be younger than the living lineages of Commelinales by all methods: 88 Ma (NPRS), 78 Ma (PL), and 36 Ma (PATHd8). The young age obtained with PATHd8 suggests a very rapid divergence of the living lineages of Zingiberales, while these start to diverge in the mid-Upper Cretaceous, at about 78 Ma, according to the PL analyses. The divergence of Poales from their closest relative is estimated to 117 Ma (NPRS), 116 Ma (PL), or 98 Ma (PATHd8). The living lineages diverged at 113 Ma (NPRS), 111 Ma (PL), or 98 Ma (PATHd8). PATHd8 thereby suggests an almost “explosive” radiation of the numerous poalean families at the boundary between the Lower and Upper Cretaceous (~100 Ma), while the other methods suggest a slower radiation, starting about 15–20 Ma earlier, in the mid-Lower Cretaceous, with most family stem groups appearing in the Upper Cretaceous, 40 Ma later. The differences between PL and NPRS on one hand, and PATHd8 on the other, might be due to systematic errors (33, 41). The first two methods smoothen or minimize age differences between mother and daughter lineages, while PATHd8 minimizes age differences between closest relatives. Without enough calibration points, both approaches can result in a number of systematic errors, for example NPRS and PL overestimating ages for large groups with short branches, and PATHd8 underestimating the ages for the same group. In the monocot data set, some groups are likely to suffer from this phenomenon, and from further analytical problems (41), and their age estimates should therefore be used with caution, regardless of the method employed. These groups include Arecaceae, Orchidaceae compared to the rest of the Asparagales, age estimates within Zingiberales and Commelinales, and the family Poaceae. For further discussion of methodological issues, see (41). Several studies focusing on divergence time estimation within monocot orders and families have been published and will briefly be compared to results obtained in the analysis by Janssen and Bremer (24) and during our reanalysis of their data set. Two large studies, one focusing on Poales (11) and one on monocots as a whole (24), use NPRS to estimate divergence times and present deviating age estimates for families within the Poales. This seems to be a methodological artifact related to the number of taxa sampled. A larger taxon sampling is susceptible to yield older ages (24). However, age estimates are


in general differing less than 20 Ma for the divergences of living lineages of the respective families, and less than 10 Ma for divergences among most closely related families. The largest differences are found in Poaceae, Cyperaceae, and Juncaceae, and also in Restionaceae. Our reanalysis of the 800+ data set gives age estimates comparable to other published results, except for the divergence times of Arecaceae and Bromeliaceae that are probably being underestimated by the PATHd8 method. In a study of divergence times within Liliales (14), similar ages were obtained using NPRS and the mean-path length method (MPL) (42). Reanalysis of the 800+ data set using PATHd8 also yielded similar ages. Most differences among these three analyses are in the magnitude of 10 Ma. However, the NPRS and PL analyses of the 800+ data set suggest much older ages for the divergences from their respective closest relatives and for the divergences of living lineages for the order Liliales, as well as for the Families Liliaceae, Melanthiaceae, Campynemataceae, and Luzuriagaceae. In a study of divergence dates within Zingiberales (15) a data set comprising three genes, 24 ingroup taxa (21 of them identical to the 800+ data set), and two calibration points were used and analyzed with three methods: NPRS, PL, and a local clock approach. Similar estimates were obtained for all three methods in this study. The age estimates show considerable differences compared to the 800+ data set. All analyses suggest a rapid radiation of Zingiberales, but at different times: 110–100 Ma according to Kress and Specht (15), 88 Ma according to Janssen and Bremer (24), and 78 Ma (PL) or 36 Ma (PATHd8) according to this reanalysis of (24). Linder and Rudall (43) compiled a chronogram of Poales using Janssen and Bremer (24) and Bremer (11). Nodes lacking divergence time estimates in either of these studies were evenly spread between dated nodes, in an attempt to approximate ages within the graminid and xyrid clades. Datings have also been done for the Families Costaceae (44), Restionaceae (39), and Rapateaceae (45). Since these studies focus on nodes within monocot families, we will not review them here. Few monocot fossils have been discovered in early and mid-Cretaceous strata (36). This scarcity does not reflect a true lack of monocots in this time period, but rather taphonomic filtering of herbaceous plants (i.e., they are less likely to be preserved). Furthermore, the frequent lack of distinctive features in monocot pollen likely accounts for early monocot fossil pollen records remaining unrecognized. The earliest indisputable monocot found so far, Mayoa portugallica, from the late Barremian–early

211

Aptian (~125 Ma) of Portugal (34), is suggested to be part of the lineages of Arales that split off before the divergence of the living lineages. Accordingly the minimum age of monocots as evidenced by the fossil record is 120 Ma, which is relatively close to the calibration age of the root node used for the analysis of the 800+ data set (134 Ma). The fossil record of Alismatales also extends back to the early Cretaceous (46). The fossil angiosperm Pennistemon/Pennipollis might be related to Alismatales and the first Pennipollis-type pollen occurs around the Barremian–Aptian boundary (47). The earliest diverse monocot flora containing flowers, fruits, and stems from various monocot plants occurs in Maastrichtian (71–66 Ma) strata in India (36). From the middle of the Late Cretaceous (100–66 Ma), the monocot fossil record provides evidence that monocots were diverse and widespread (48). Still many monocot orders have sparse or no records in the Late Cretaceous. Several biogeographical studies of groups within the monocots have been published (see e.g., 49–53). A biogeographical analysis of the whole monocot group was conducted by Bremer and Janssen (54), who combined their earlier dating with the 800+ data set in a dispersal-vicariance analysis. They were focusing on the continental distribution of monocots, using widely circumscribed areas for the analysis. They concluded that a majority of the monocots have a South Gondwanan evolution, since the Australasian and South American optimizations dominate in the deeper nodes of the phylogeny. It is however not possible to specify an ancestral distribution for the most recent ancestor of all monocots, since the Alismatales have many widespread representatives, that obviously are easily spread due to their aquatic habit.

References 1. J. E. M. Baillie, C. Hilton-Taylor, and S. N. Stuart (eds.) 2004 IUCN Red List of Threatened Species. A Global Species Assessment (IUCN, Gland, Switzerland and Cambridge, UK, 2004.) 2. APGII, Bot. J. Linn. Soc. 141, 399 (2003). 3. T.-J. Givnish et al., Aliso 22, 28 (2006). 4. J.-I. Davis et al., Syst. Bot. 29, 467 (2004). 5. G. Petersen et al., Aliso 22, 52 (2006). 6. S.-W. Graham et al., Aliso 22, 3 (2006). 7. M.-W. Chase et al., Aliso 22, 63 (2006). 8. M.-R. Duvall, S. Mathews, N. Mohammad, T. Russell, Aliso 22, 79 (2006). 9. M. W. Chase, D. W. Stevenson, P. Wilkin, and P. J. Rudall, in Monocotyledons: Systematics and evolution, P. J. Rudall,

212

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.


P. J. Cribb, D. F. Cutler, and C. J. Humphries, Eds. (Royal Botanic Gardens, Kew, UK, 1995), pp. 685–730. M.-F. Fay et al., Aliso 22, 559 (2006). K. Bremer, Evolution 56, 1374 (2002). J. C. Pires et al., Aliso 22, 287 (2006). L. R. Caddick, P. J. Rudall, P. Wilkin, T. A. J. Hedderson, M. W. Chase, Bot. J. Linn. Soc. 138, 123 (2002). A. Vinnersten, K. Bremer, Am. J. Bot. 88, 1695 (2001). W. J. Kress, C.-D. Specht, Aliso 22, 621 (2006). J. M. Saarela et al., Nature 446, 312 (2007). P. S. Soltis, D. E. Soltis, M. J. Zanis, S. Kim, in Abstracts, XVI International Botanical Congress (St. Louis, 1999). Y.-L. Qiu et al., Int. J. Plant Sci. 166, 815 (2005). Y.-L. Qiu et al., Nature 402, 404 (1999). S. W. Graham, R. W. Olmstead, Am. J. Bot. 87, 1712 (2000). M. J. Zanis, D. E. Soltis, P. S. Soltis, S. Mathews, M. J. Donoghue, Proc. Nat. Acad. Sci. U.S.A. 99, 6848 (2002). M. R. Duvall, in Monocots: Systematics and evolution, K. L. Wilson, D. A. Morrison, Eds. (CSIRO Publishing, Collingwood, Victoria, Australia, 2000). C. L. Parkinson, K. L. Adams, J. D. Palmer, Curr. Biol. 9, 1485 (1999). T. Janssen, K. Bremer, Bot. J. Linn. Soc. 146, 385 (2004). K. Bremer, E. M. Friis, B. Bremer, Syst. Biol. 53, 496 (2004). L. Savard et al., Proc. Natl. Acad. Sci. U.S.A. 91, 5163 (1994). V. V. Goremykin, S. Hansmann, W. F. Martin, Plant Syst. Evol. 206, 337 (1997). J. Leebens-Mack et al., Mol. Biol. Evol. 22(10), 1948 (2005). K. Bremer, Proc. Natl. Acad. Sci. U.S.A. 97, 4707 (2000). N. F. Huges, A. B. McDougall, Rev. Palaeobot. Palynol. 50, 255 (1987). M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997). M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002). T. Britton, C. L. Anderson, D. Jacquet, S. Lundqvist, K. Bremer, Syst. Biol. 56 (2007).

34. E. M. Friis, K. R. Pedersen, P. R. Crane, PNAS 101, 16565 (2004). 35. M. A. Gandolfo, K. C. Nixon, W. L. Crepet, Am. J. Bot. 89, 1940 (2002). 36. E. M. Friis, K. R. Pedersen, P. R. Crane, Palaeogeogr. Palaeoclimatol. Palaeoecol. 232, 251 (2006). 37. V. Prasad, C.-A. E. Stromberg, H. Alimohammadian, A. Sahni, Science 310 (2005). 38. M.-M. Harley, Bot. J. Linn. Soc. 151, 39 (2006). 39. H. P. Linder, C. R. Hardy, F. Rutschmann, Mol. Phylogenet. Evol. 35, 569 (2005). 40. B. S. Gaut, B. R. Morton, B. C. McCaig, M. T. Clegg, Evolution 93, 10274 (1996). 41. C. L. Anderson, Comprehensive Summaries of Uppsala University Dissertations, Acta Universitatis Upsaliensis (2007). 42. T. Britton, B. Oxelman, A. Vinnersten, K. Bremer, Mol.. Phylogenet. Evol. 24, 58 (2002). 43. H. P. Linder, P. J. Rudall, Ann. Rev. Ecol. Evol. Systemat. 36, 107 (2005). 44. C.-D. Specht, Aliso 22, 633 (2006). 45. T. J. Givnish et al., Evolution 54, 1915 (2000). 46. R.-A. Stockey, Aliso 22, 91 (2006). 47. E. M. Friis, K. R. Pedersen, P. R. Crane, Grana 39, 226 (2000). 48. P. S. Herendeen, P. R. Crane, in Monocotyledons: Systematics and evolution, P. J. Rudall, P. J. Cribb, D. F. Cuttler, C. J. Humphries, Eds. (Royal Botanical Gardens, Kew, 1995), pp. 1–21. 49. H. P. Linder, Kew Bull. 42, 297 (1987). 50. O. Seberg, Bot. J. Linn. Soc. 96, 119 (1988). 51. M. G. Simpson, Ann. Mo. Bot. Gard. 77, 722 (1990). 52. J. G. Conran, J. Biogeogr. 22, 1023 (1995). 53. T. J. Givnish, T. M. Evans, J. C. Pires, K. J. Sytsma, Mol. Phylogenet. Evol. 12, 360 (1999). 54. K. Bremer, T. Janssen, Aliso 22, 22 (2006).

FUNGI


Fungi Jaime E. Blair Department of Biology, 139 McGuire Life Sciences Building, Amherst College, Amherst, MA, 01002 USA. Current address: Department of Biology, Franklin and Marshall College, Lancaster, PA, 17604 USA ([email protected])

Abstract Fungi are an essential constituent of modern terrestrial ecosystems, partnering with plants and other organisms in a range of symbiotic relationships. Recent phylogenetic analyses of multiple nuclear genes have challenged the traditional division of Fungi into four main groups, and a major reclassification of this kingdom has been proposed. The molecular timetree suggests that at least eight major fungal lineages originated in the Precambrian, before 542 million years ago (Ma), much earlier than their first appearance as fossils in the Ordovician, ~460 Ma. Fungi may have played a major role in the colonization of land by multicellular eukaryotes.

Members of Kingdom Fungi play an integral role in global nutrient cycling and are important symbionts with other eukaryotes and some prokaryotes. They are intimately associated with plants as mutualistic mycorrhizae and as the primary decomposers of lignin. Fungi are most closely related to Metazoa as part of the Opisthokonta, a eukaryotic supergroup which also includes unicellular choanoflagellates, icthyosporeans, and nucleariids (1, 2). While estimates vary (3), Kingdom Fungi may contain upward of several million species, with only a small fraction (~100,000) being formally described. Here I review the relationships and divergence times among the major fungal lineages. Fossil evidence for the origin of Fungi is limited, which has confounded molecular time estimates due to a lack of robust fungal calibrations. Putative lichen-like structures first appear in the fossil record around 600 Ma (4), and some acritarch assemblages from the early Neoproterozoic have been associated with “higher fungi” (5). The first taxonomically identifiable fossils appear in the mid-Ordovician, ~460 Ma, and are similar to modern Glomeromycota (6). Diverse fungal remains from a variety of ecological niches then appear in the Devonian,

~400 Ma, including the well-preserved fossils of a probable pyrenomycete (7, 8). This fossil species has served as an important and often controversial calibration in some molecular clock studies (9). Basidiomycetes with diagnostic hyphal clamp connections are not found until the mid-Pennsylvanian, ~300 Ma (10), although evidence for wood decay similar to modern-day basidiomycetous white rot is present in the Upper Devonian (11). Traditionally, fungi have been divided into four major groups: Ascomycota, Basidiomycota, Zygomycota, and Chytridiomycota. With the accumulation of molecular sequence data, it has become clear that both “zygomycetes” (12) and “chytrids” (13, 14) are not monophyletic assemblages. In addition, some lineages conventionally associated with Kingdom Fungi based on morphological similarities (e.g., oomycetes, slime molds) have been reclassified in separate eukaryotic supergroups (1). Conversely, the Microsporidia, obligate intracellular parasites once considered an ancient amitochondriate lineage, are now known to be closely related to, or perhaps nested within, Kingdom Fungi (15). Since 2003, a large, community-wide effort has established robust

Fig. 1 A basidiomycete (Coprinopsis sp.) from the Eastern United States. Credit: M. E. Hood (Amherst College).

J. E. Blair. Fungi. Pp. 215–219 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Saccharomycetes

5

Taphrinomycetes 10

Schizosaccharomycetes

4

Agaricomycotina 3

6

Ustilaginomycotina Pucciniomycotina

2

Glomeromycota

Chytridiomycota Blastocladiomycota Neocallimastigomycota

Mp

Neoproterozoic PROTEROZOIC 1000

800

Paleozoic

Mz

"chytrids"

Mucoromycotina 1

Taphrinomycotina

Pezizomycetes

7

"zygomycetes"

Eurotiomycetes

8

Basiodiomycota Saccharomycotina

Sordariomycetes 9

Ascomycota

THE TIMETREE OF LIFE Pezizomycotina

216

Cz

PHANEROZOIC 600

400

200

0 Million years ago

Fig. 2 A timetree of fungi. Taxonomic names are from a revised classification (18). Divergence times are shown in Table 1. Abbreviations: Cz (Cenozoic), Mp (Mesoproterozoic), Mz (Mesozoic), and Np (Neoproterozoic).

phylogenetic relationships among the main fungal lineages (16), as well as within most major groups (17). As a result of this phylogenetic reshuffling, a revised classification system for Fungi has been proposed (18), and will be followed here. The majority of fungal species are found in the Phylum Ascomycota, which is divided into three monophyletic subphyla. The Taphrinomycotina, previously known as the “Archiascomycetes,” is the most basal of the three subphyla, and contains a variety of ecologically and morphologically diverse taxa, including the model fission yeast Schizosaccharomyces pombe (19). The Saccharomycotina, or “Hemiascomycetes,” are the commonly known yeasts, including beneficial species (e.g., Saccharomyces cerevisiae), as well as important human pathogens (e.g., Candida albicans). Molecular phylogenies have shown that many genera within the Saccharomycotina are not monophyletic (20). The Pezizomycotina, largest of the Ascomycota subphyla, includes filamentous species with diverse ecologically specialties, such as decomposers, plant and animal pathogens, lichens, and mycorrhizal symbionts (21). The Pezizomycotina, or “Euascomycetes,” are currently divided into 10 classes, although the relationships among these lineages are still unresolved.

Taxa previously known as plectomycetes and pyrenomycetes are now found predominantly in the Eurotiomycetes and Sordariomycetes, respectively. In addition, environmental sampling has revealed a distinct lineage of soil-dwelling fungi that may represent a fourth subphylum within Ascomycota (22). The Phylum Basidiomycota (Fig. 1) also has been divided into three subphyla, but the relationships among these lineages remain ambiguous (23, 24). The Agaricomycotina, or “Hymenomycetes,” includes the largest diversity of mushroom-forming and wood-decaying species (25). The Ustilaginomycotina contains many dimorphic plant pathogen “smuts,” such as Ustilago maydis (corn smut) and Tilletia indica (karnal bunt of wheat), which cycle between saprobic yeast and parasitic hyphal stages (26). Similarly, the Pucciniomycotina, or “Urediniomycetes,” are predominantly plant pathogen “rusts,” such as Puccinia graminis (cereal rust). Rust fungi often utilize more than one host and can produce multiple spore types during their life cycle, complicating their taxonomic characterization (27). The Basidiomycota and Ascomycota form the Subkingdom Dikarya, reflecting their shared derived trait of dikaryotic (i.e., containing two haploid nuclei) and regularly septate hyphae (28).

Eukaryota; Fungi

217

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among fungi. Timetree Node

Time

Estimates Ref. (33)

Ref. (34)(a)

Ref. (34)(b)

Ref. (35)

Ref. (36)

Ref. (37)

Time

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

850

1423

1434–1270

893

928–857

1458

1595–1321

–

–

–

–

1

1156

2

1049

–

1287

1411–1216

856

909–812

1107

1217–997

947

1047–847

–

–

3

980

600

–

–

–

–

–

–

–

–

–

–

4

908

550

1206

1339–1165

786

848–757

1208

1429–996

968

1195–741

727

837–629

5

895

450

1148

1306–1108

724

812–691

1144

1295–993

1009

1152–866

–

–

6

791

500

1028

1167–966

669

728–630

966

1135–797

–

–

–

7

773

375

1072

1244–1051

657

756–632

1085

1244–926

982

1166–798

466

564–368

8

723

325

972

1163–955

570

686–535

–

–

–

–

–

–

9

673

–

930

1141–921

539

666–510

670

809–531

551

671–431

–

–

10

592

300

928

1189–770

549

689–448

–

–

–

–

–

–

Note: Node times in the timetree represent the mean of time estimates from different studies, except for Nodes 3 and 8 where Node times represent the midpoint between adjacent nodes. Estimates in ref. (34) are from a single gene study using a value of 1576 Ma (a) and 965 Ma (b) for the animal–fungal divergence calibration.

Aside from the well-supported Dikarya, the basal lineages of Fungi are currently in a state of phylogenetic and taxonomic flux. As mentioned earlier, the traditional “Zygomycota” phylum has been discarded in the current classification scheme, and its members have been placed in one new phylum, Glomeromycota (29), and four subphyla of uncertain placement (Mucoromycotina, Entomophthoromycotina, Zoopagomycotina, Kickxellomycotina). Molecular evidence suggests that Glomeromycota, which includes the arbuscular mycorrhizal symbionts of plants, may be the closest relative of Dikarya (16, 30). The Mucoromycotina, a group which includes the common molds Rhizopus and Mucor, has been shown to be the next closest relative of Dikarya + Glomeromycota (12, 16). Species with flagellated motile spores, the “chytrids,” are extremely diverse in their ecologies and represent the most basal lineages of Kingdom Fungi. These organisms have been divided among three phyla (Chytridiomycota, Blastocladiomycota, and Neocallimastigomycota), and the relationships among these lineages have not yet been resolved. Microsporidia has recently been classified as a fungal phylum, although it is still unclear whether it is nested within the kingdom (16) or as a close relative of Fungi (14). Molecular clock studies have produced a range of time estimates for the origin and diversification of Fungi. Early studies using a single gene, the small subunit ribosomal RNA, generally found young divergence

times consistent with the fossil records of fungi and plants, which were used as the source of calibration points (31, 32). Rate variation among lineages was noted even in these early studies. A later revision produced slightly older dates (33), suggesting the major lineages of fungi had diverged before the colonization of land by plants (Table 1). Also, the origin of the Pezizomycotina was found to be younger than 400 Ma, leading authors to suggest that the Devonian pyrenomycete fossil more likely represented an early “Archiascomycete” (33). Issues of rate variation and fossil calibration were addressed in a more comprehensive study of the small subunit ribosomal RNA, which used a rate smoothing clock method to estimate divergence times among 169 fungal taxa (34). In this study, two different calibration values for the animal–fungal divergence (1576 vs. 965 Ma) were tested, and produced dramatically different time estimates (Table 1). These authors claim that the use of the older estimate for the animal–fungal calibration may be more appropriate as it accommodates the Devonian pyrenomycete fossil (34). Studies using multiple genes and a variety of molecular clock methods have generally found ancient divergences among the fungal lineages. The first multigene analysis of Fungi (35) estimated deep Precambrian divergences for all major lineages. This study also estimated an origin for the land plant lineage ~700 Ma, suggesting that terrestrial ecosystems of multicellular eukaryotes may

218


have existed millions of years before the first fossil evidence of such in the Ordovician (35). A later reanalysis (36) using additional data and more complex molecular clock methods produced similar estimates (Table 1). An analysis of 129 proteins from 36 eukaryotes suggested younger divergences within Fungi, although only a few fungal lineages were included in this study (37). The availability of complete genome sequences from 15 eukaryotes allowed the divergence between Pezizomycotina and Saccharomycotina to be estimated at ~850 Ma (38). Analyses of a 50-gene data set also compiled from complete genome sequences illustrated how the use of external versus internal calibrations can affect divergence time estimates both within and among the plant, animal, and fungal kingdoms (9). The molecular timetree shows that, while some phylogenetic uncertainty exists, all major lineages of Fungi likely originated in the Neoproterozoic (Fig. 2). The timetree also supports the interpretation of the Devonian pyrenomycete fossil as a member of the Pezizomycotina, perhaps associated with the Sordariomycetes. There are significant dissimilarities between the molecular clock estimates and the fossil record, which is not unexpected due to the microscopic nature of most fungal species, their poor preservation potential, and the scarcity of mycologically trained paleontologists (39). For some groups (e.g., the classes of Basidiomycota, Microsporidia), no molecular time estimates currently exist, leaving only the fossil record to infer their minimum times of origin. The origin of Kingdom Fungi was almost certainly aquatic, but whether this initial environment was marine or terrestrial remains unclear (16). The early-diverging lineages of “chytrid” fungi retain flagellated spores, a trait believed to be homologous with their single-celled Opisthokont relatives (2, 40). This character has subsequently been lost multiple times during the evolution of Fungi (13, 16), and the vast majority of extant fungal species occupy terrestrial niches with various methods of spore dispersal. The diverse associations between fungal species and other organisms, as lichens, mycorrhizae, pathogens, and decomposers, suggest that Fungi played a crucial role in shaping the paleoecosystem as multicellular eukaryotes colonized land (41, 42), perhaps as early as the Neoproterozoic. The rapid accumulation of genome sequence data from a number of Fungi, along with improved molecular clock methods optimized to model rate variation, will allow for further refinement of fungal relationships and divergence times.

Acknowledgments This work was supported by the Howard Hughes Medical Institute Genomics Postdoctoral grant to Amherst College. I thank D. S. Hibbett and M. E. Hood for comments on this manuscript.

References 1. S. M. Adl et al., J. Eukaryot. Microbiol. 52, 399 (2005). 2. E. T. Steenkamp, J. Wright, S. L. Baldauf, Mol. Biol. Evol. 23, 93 (2006). 3. D. L. Hawksworth, Mycol. Res. 105, 1422 (2001). 4. X. Yuan, S. Xiao, T. N. Taylor, Science 308, 1017 (2005). 5. N. J. Butterfield, Paleobiology 31, 165 (2005). 6. D. Redecker, R. Kodner, L. E. Graham, Science 289, 1920 (2000). 7. T. N. Taylor, H. Hass, H. Kerp, Nature 399, 648 (1999). 8. T. N. Taylor, H. Hass, H. Kerp, M. Krings, R. T. Hanlin, Mycologia 97, 269 (2005). 9. J. W. Taylor, M. L. Berbee, Mycologia 98, 838 (2006). 10. R. L. Dennis, Mycologia 62, 578 (1970). 11. S. P. Stubblefield, T. N. Taylor, C. B. Beck, Am. J. Bot. 72, 1765 (1985). 12. M. M. White et al., Mycologia 98, 872 (2006). 13. T. Y. James et al., Mycologia 98, 860 (2006). 14. Y. Liu, M. Hodson, B. Hall, BMC Evol. Biol. 6, 74 (2006). 15. P. J. Keeling, C. H. Slamovits, Eukaryot. Cell 3, 1363 (2004). 16. T. Y. James et al., Nature 443, 818 (2006). 17. M. Blackwell, D. S. Hibbett, J. W. Taylor, J. W. Spatafora, Mycologia 98, 829 (2006). 18. D. S. Hibbett et al., Mycol. Res. 111, 509 (2007). 19. J. Sugiyama, K. Hosaka, S.-O. Suh, Mycologia 98, 996 (2006). 20. S.-O. Suh, M. Blackwell, C. P. Kurtzman, M.-A. Lachance, Mycologia 98, 1006 (2006). 21. J. W. Spatafora et al., Mycologia 98, 1018 (2006). 22. T. M. Porter et al., Mol. Phylogenet. Evol. 46, 635 (2008). 23. P. B. Matheny, J. A. Gossmann, P. Zalar, T. K. A. Kuman, D. S. Hibbett, Can. J. Bot. 84, 1794 (2006). 24. E. C. Swann, J. W. Taylor, Can. J. Bot. 73, S862 (1995). 25. D. S. Hibbett, Mycologia 98, 917 (2006). 26. D. Begerow, M. Stoll, R. Bauer, Mycologia 98, 906 (2006). 27. M. C. Aime et al., Mycologia 98, 896 (2006). 28. C. J. Alexopoulos, C. W. Mims, M. Blackwell, Introductory Mycology, 4th ed. (Wiley, New York, 1996). 29. A. Schüβler, D. Schwarzott, C. Walker, Mycol. Res. 105, 1413 (2001). 30. D. Redecker, P. Raab, Mycologia 98, 885 (2006). 31. M. L. Berbee, J. W. Taylor, Can. J. Bot. 71, 1114 (1993). 32. L. Simon, J. Bousquet, R. C. Levesque, M. Lalonde, Nature 363, 67 (1993).

Eukaryota; Fungi

33. M. L. Berbee, J. W. Taylor, in The Mycota, D. J. McLaughlin, E. G. McLaughlin, P. Lemke, Eds. (Springer-Verlag, New York, 2001), Vol. VII, Part B, pp. 229–245. 34. A. C. B. Padovan, G. F. O. Sanson, A. Brunstein, M. R. S. Briones, J. Mol. Evol. 60, 726 (2005). 35. D. S. Heckman et al., Science 293, 1129 (2001). 36. S. B. Hedges, J. E. Blair, M. L. Venturi, J. L. Shoe, BMC Evol. Biol. 4, 2 (2004). 37. E. J. P. Douzery, E. A. Snell, E. Bapteste, F. Delsuc, H. Philippe, Proc. Natl. Acad. Sci. U.S.A. 101, 15386 (2004).

219

38. J. E. Blair, P. Shah, S. B. Hedges, BMC Bioinformatics 6, 53 (2005). 39. J. W. Taylor et al., in Assembling the Tree of Life, J. Cracraft, Ed. (Oxford University Press, Cary, NC, 2004), pp. 171–194. 40. T. Cavalier-Smith, in Evolutionary Biology of the Fungi, A. D. M. Rayner, C. M. Brasier, D. Moore, Eds. (Cambridge University Press, Cambridge, 1987), pp. 339–353. 41. M. A. Selosse, F. Le Tacon, Trends Ecol. Evol. 13, 15 (1998). 42. T. N. Taylor, Trends Ecol. Evol. 5, 21 (1990).


ANIMALS


Animals (Metazoa) Jaime E. Blair Department of Biology, 139 McGuire Life Sciences Building, Amherst College, Amherst, MA, 01002 USA. Current address: Department of Biology, Franklin and Marshall College, Lancaster, PA, 17604 USA ([email protected])

Abstract The relationships and molecular divergence times among metazoan (animal) phyla have been the subject of debate for decades. The current consensus suggests that most traditional hypotheses of metazoan phylogeny based on morphology are not supported by molecular phylogenetic analyses. In addition, the steady accumulation of sequence data and the increased sophistication of molecular clock methods have led to an expanded number of studies estimating divergence times among metazoan lineages. Most molecular clock studies, or their reanalyses, have found that the earliest divergences among living metazoans occurred deep in the Precambrian, hundreds of millions of years before the first animal fossils.

The evolution of large, heterotrophic metazoans (Fig. 1) has undoubtedly had a significant impact on the history of life by increasing the complexity of trophic interactions both in marine and terrestrial ecosystems. Metazoans are most closely related to Fungi as part of Opisthokonta, a eukaryotic supergroup which also includes unicellular choanoflagellates, icthyosporeans, and nucleariids (1, 2). The closest relatives of metazoans are the choanoflagellates; morphological similarities between the collar cells of sponges and the colonial habits of choanoflagellates were noted over 150 years ago (3). In addition, important molecular characteristics traditionally thought to be unique to metazoans, such as cell signaling and adhesion protein families, have also been found in choanoflagellates (4, 5). Here I review the relationships and divergence times among the metazoan phyla. Relationships among the metazoan phyla have undergone major revisions over the past two decades (3, 6). The traditional view of simpler forms giving rise to more complex lineages has been challenged by the accumulation of developmental and molecular data. One important discovery has been the paraphyly of phylum Porifera

(the sedentary fi lter-feeding sponges) which has altered interpretations of the last common ancestor of metazoans (Fig. 2). The Calcarea, which possess calcareous skeletons, may be more closely related to the Eumetazoa (all metazoans other than sponges) than the siliceous Hexactinellida or Demospongiae (7–9). The exact relationship between the siliceous sponges has not yet been determined, although some data have suggested they may form a monophyletic group, the Silicea (3). The monophyly of Eumetazoa has been supported by molecular data as well as a number of morphological characteristics, such as the presence of body symmetry, a nervous system, and a mouth and gut. Although traditionally grouped together as the Coelenterata, molecular data have not supported a close relationship between Cnidaria and Ctenophora, and have instead placed Cnidaria as the closest relative of the Bilateria (7, 9–11). The relationships of Ctenophora and Placozoa to other metazoans have yet to be firmly established. In addition, the absence of bilateral symmetry and mesodermal tissues in the basal Eumetazoan lineages has been challenged by developmental and gene expression studies in jellyfish and complete genome analysis of the sea anemone Nematostella (12–14). These new data suggest a need to reevaluate the characteristics typically used to distinguish the diploblastic “Radiata” from the triploblastic Bilateria (15). Within the monophyletic Bilateria, substantial changes have been made to the traditional scheme of

Fig. 1 An onycophoran (Peripatus juliformis) from St. John, United States Virgin Islands. Credit: A. Sanchez.

J. E. Blair. Animals (Metazoa). Pp. 223–230 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Echinodermata 5

Arthropoda 8

Priapulida

6 2

Annelida

Ecdysozoa

Hemichordata

3

Mollusca

9

Nemertea 10

Cnidaria 1

Ctenophora Calcarea Demospongiae Hexactinellida

Mp

Neoproterozoic PROTEROZOIC

1200

1000

800

Paleozoic

Mz

"Porifera" "Radiata"

Platyhelmintha

Deuterostomia

Urochordata

4

Protostomia

Cephalochordata

Chordata

Vertebrata 7

Ambulacraria


Lophotrochozoa

224

Cz

PHANEROZOIC 600

400

200

0 Million years ago

Fig. 2 A timetree of metazoan phyla. Divergence times are shown in Table 1. Abbreviations: Cz (Cenozoic), Mp (Mesoproterozoic), and Mz (Mesozoic).

increasing grades of complexity from basal, acoelomate flatworms up to complex, segmented protostomes and deuterostomes (16). The accumulation of molecular sequence data, primarily from the small subunit ribosomal RNA, suggests instead a division of the Bilateria into three major clades: Lophotrochozoa (17), Ecydsozoa (18), and Deuterostomia. Lophotrochozoa and Ecdysozoa together form the Protostomia, whose monophyly is often assumed but has been supported only recently with large, multigene studies (19–22). Initially, these new divisions, along with the apparent lack of any basal lineages, suggested that the last common bilaterian ancestor was a large, possibly segmented organism likely to have left traces in the fossil record (23, 24). However, two lineages of bilaterians, the Acoelomorpha flatworms (25) and the enigmatic Myxozoa parasites (26, but see 27), have since been shown to be basal to the rest of Bilateria, suggesting the last common ancestor may have been less complex (28, 29). The Lophotrochozoa clade was first suggested based on the affinity of lophophorates (Brachiopoda, Bryozoa, and Phoronida) with protostomes, specifically with mollusks

and annelids (17). Further studies have supported the placement of additional groups within Lophotrochozoa, such as Rotifera, Acanthocephala, Gnathostomulida, and Gastrotricha, among others (6). The Platyhelmintha (excluding the basal Acoelomorpha) are also included in Lophotrochozoa, an unusual relationship first suggested by small subunit ribosomal RNA and Hox data (18, 30), and later supported by multigene studies (20, 31, but see 32). Two main subclades have been suggested within Lophotrochozoa; the Platyzoa, which contains acoelomate Platyhelmintha, Rotifera, Acanthocephala, Gastrotricha, and Gnathostomulida; and the Trochozoa, which includes lophophorates, annelids, mollusks, and Nemertea (33). Overall, the relationships within Lophotrochozoa have not yet been firmly established, as most studies are based on one or two genes (ribosomal RNAs) with limited taxon sampling. The second major protostome clade, Ecdysozoa, has received considerably more attention. The process of moulting, or ecdysis, has been suggested as the sharedderived character uniting this group of morphologically diverse phyla. Within Ecdysozoa, the monophyly of

Eukaryota; Metazoa

225

Table 1. Divergence times (Ma) their confidence/credibility intervals (CI) among animals. Timetree Node

Estimates Refs. (50, 60)

Time Time

Ref. (61)(a)

Ref. (61)(b)

Ref. (62)

CI

Time

CI

Time

CI

Time

CI

1

1237

1351(60)

1586–1116

766

803–731

1122

1360–932

–

–

2

1036

1298(60)

1443–1153

676

709–645

907

1070–775

902

959–845

3

910

976(60)

1166–786

643

669–615

845

993–731

627

727–527

4

842

896

1022–932

601

625–579

788

930–675

–

–

5

795

876

1074–725

548

554–534

713

847–608

–

–

6

790

–

–

619

648–594

790

921–685

–

–

7

774

843

1067–685

547

584–518

744

891–620

–

–

8

728

–

–

584

616–552

728

857–621

–

–

9

698

–

–

586

612–563

698

815–610

–

–

10

666

–

–

565

595–534

666

784–574

–

–

Node

Estimates Ref. (67) Time

Ref. (68) CI

Time

Ref. (69) CI

Time

Ref. (70) CI

Time

CI

1

–

–

–

–

–

–

–

–

2

–

–

–

–

–

–

–

–

3

813

975–651

931

1237–625

993

1084–902

581

610–557

4

–

–

–

–

–

–

536

544–524

5

–

–

–

–

–

–

–

–

6

–

–

–

–

–

–

–

–

7

–

–

–

–

–

–

–

–

8

–

–

–

–

–

–

–

–

9

–

–

–

–

–

–

–

–

10

–

–

–

–

–

–

–

–

CI

Time

Node

Estimates Ref. (73)(a) Time

Ref. (73)(b) CI

Time

Ref. (79)

CI

Time

Ref. (80) CI

1

–

–

–

–

–

–

–

–

2

–

–

–

–

–

–

–

–

3

695

761–642

1141

1389–934

955

1135–775

–

–

4

–

–

–

–

–

–

–

–

5

–

–

–

–

–

–

–

–

6

–

–

–

–

–

–

–

–

7

–

–

–

–

756

870–642

751

814–689

8

–

–

–

–

–

–

–

–

9

–

–

–

–

–

–

–

–

10

–

–

–

–

–

–

–

–

Note: Node times in the timetree represent the mean of time estimates from different studies, except for original times (a) in refs. (61, 73), where the times obtained after corrections (b) were used (see text for details).

226


the Panarthropoda subclade (Arthropoda, Tardigrada, and Onychophora) has been supported by a number of studies, while the relationships among the remaining Introverta phyla (Priapulida, Kinorhynca, Loricifera, Nematomorpha, Nematoda) have not been resolved (34). The inclusion of the pseudocoelomate Nematoda within Ecdysozoa has created the most controversy. The Ecdysozoa clade was first proposed based on an analysis of small subunit ribosomal RNA; the authors claimed that the use of a slower evolving nematode sequence overcame the typical long-branch attraction artifacts that place nematodes basal to other bilaterians (18). Additional studies of molecular sequence data (20, 35, 36), Hox gene homology (37), intron positions (38), gene expression patterns (39, 40), and other lines of evidence, have supported Ecydysozoa. On the other hand, genome-level studies utilizing the complete sequences of vertebrates and the two model organisms, Drosophila and Caenorhabditis, have supported a closer relationship between vertebrates and arthropods, which corresponds to a more traditional Coelomata hypothesis (32, 41–44). The position of Nematoda remains an active area of research and debate (e.g., 45–49), along with the affinity of Chaetognatha, or arrow worms. Phylogenetic rearrangements have also occurred within the Deuterostomia. Among the chordates, recent molecular studies have challenged the traditional position of Urochordata as the closest relative of the group containing Vertebrata and Cephalochordata, instead suggesting a closer relationship between Vertebrata and Urochordata (20, 21, 50–52). However, further analysis may be needed as this unusual relationship significantly alters traditional interpretations of chordate evolution (but see 53). Recent multigene studies have also solidified support for a close relationship between Echinodermata and Hemichordata as Ambulacraria (50, 51). This arrangement has important implications for the ancestral condition of deuterostomes, suggesting that “chordate” features such as gill slits and notochords may have been present in the last common ancestor (6, 54, 55). In addition, molecular studies have shown that the marine worm Xenoturbella is a deuterostome closely related to Ambulacraria (51, 56). Molecular divergence times among metazoan phyla have received considerable attention, with two main hypotheses emerging. The first hypothesis posits that metazoans originated only shortly before their appearance in the fossil record during the so-called “Cambrian Explosion.” A minority of molecular investigations have supported this view (e.g., 57), and have suggested that the

evolution of Hox genes and other developmental pathways in bilaterians led to a dramatic restructuring of trophic interactions in the late Precambrian oceans, culminating in the appearance of large complex fauna in the Cambrian (58). The second hypothesis, that metazoans arose hundreds of millions of years before the Cambrian, implies a long history of cryptic evolution not present in the fossil record. A majority of molecular dating studies support this second scenario (Table 1); however, this remains a very active area of investigation. Estimates for the earliest divergences within Metazoa, such as those among members of the paraphyletic Porifera and the Eumetazoans, range from 1350 to 660 Ma with an average of 1240 Ma (57, 59–61). Estimates for the divergence of Bilateria and Cnidaria (perhaps with Ctenophora and Placozoa) have ranged between 1300 and 600 Ma, with an average of 1035 Ma (57, 60–63). The divergence between protostomes and deuterostomes has dominated molecular clock studies. Most analyses have yielded divergence times that predate the Cambrian Explosion of animal phyla. Estimates range from 1200 to 580 Ma, with an average of 910 Ma (Table 1). An early study based on seven genes proposed a controversial estimate of ~1200 Ma for the divergence between protostomes and deuterostomes (64); similar studies using small numbers of genes have produced times between 900 and 600 Ma (57, 62, 65–68). The first large multigene study, incorporating rate variation among sites and rate-tested genes, estimated a divergence time of 993 Ma with 50 genes (69). Using a different approach, Otsuka and Sugaya (63) used theoretical rates of basepair changes in mitochondrial rRNA to estimate the divergence between protostomes and deuterostomes at 920 Ma. Other studies have used likelihood and Bayesian “relaxed clock” methods to estimate the divergence between protostomes and deuterostomes, and other splits in the tree of metazoan phyla. An analysis of 22 nuclear and mitochondrial genes using Bayesian methods to correct for temporal rate variation suggested a divergence time of 581 Ma (70); the results from this study, however, were affected by significant methodological biases. Specifically, a rate model was used as a Bayesian prior that was biased toward decreasing rates, causing divergences earlier than the Cambrian to be underestimated and divergences later than the Cambrian to be overestimated (71, 72). For example, divergences among living mammals were found to be in the Paleozoic and among living birds in the Jurassic, much older than has been found in other molecular clock studies.

Eukaryota; Metazoa

A large multigene study by Douzery et al. (73) using 129 proteins and a different Bayesian method calculated the divergence between protostomes and deuterostomes to be 695 Ma (741–642 Ma). Roger and Hug (74) and Hug and Roger (75) conducted reanalyses of this large data set, questioning many aspects of the study and its results. However, these reanalyses overlooked a significant issue with both the minimum and maximum constraints used in the original study, as was noted earlier (76). In the study of Douzery et al. (73), each minimum calibration constraint was fi xed as the younger boundary of the major geologic period containing the pertinent fossil rather than to the actual (older) geologic time constraints of the fossil itself, causing the resulting time estimates to be underestimates. Douzery et al. also fi xed maximum calibration constraints, arbitrarily, to the older boundary of the major geologic period containing the fossil rather than to an evolutionary event that might bear on the constraint. For example, the maximum calibration for the split of actinopterygian fish from mammals, 417 Ma, was essentially the same time as the oldest fossil on either branch, 416 Ma (77). However, there is little fossil information from this time period (Silurian) to establish that the divergence occurred precisely when the fossils appeared; more than likely it was much earlier, which would result in older Bayesian posterior time estimates. Also, one of the maximum calibrations, the split between chelicerates and other arthropods (543 Ma), was fi xed within the Cambrian, which can lead to circular reasoning when the results are then used to support a reconciling of the Cambrian Explosion in the fossil record with molecular clock times, as was proposed in that study (73). A separate reanalysis of the Douzery et al. data set (S. B. Hedges, personal communication) was conducted using the same methods as the original authors, but with corrected minimum calibrations, based on the fossil record. This led to a protostome–deuterostome divergence time of 742 Ma (817–692 Ma), 200 million years before the Cambrian boundary and 47 million years older (31% of time to Cambrian) than the date reported in the original study (73). Therefore, a simple and necessary correction of calibrations yielded a divergence time that overturns the primary conclusion of Douzery et al., the reconciliation of molecular clock times and fossil times. Furthermore, after the removal of the maximum calibration in the Cambrian (543 Ma) and keeping all corrected minimum calibrations, the time became 797 Ma (898–719 Ma), 255 million years before the Cambrian boundary and 102 million years older (67% of time to Cambrian) than the

227

date reported in the original study. Moreover, correcting the actinopterygian-mammal maximum calibration to 495 Ma (from 417 Ma), removing other unjustified maximum calibrations, and keeping all corrected minimum calibrations resulted in a time of 1141 Ma (1389– 934), 599 million years before the Cambrian boundary and 446 million years older (392% of time to Cambrian) than the date reported in the original study (73). These reanalyses provide corrected divergence time estimates of the protostome–deuterostome divergence for this large data set (Table 1), and their trends agree with those of Hug and Roger (75) in showing the sensitivity of these data to maximum calibrations. However, Hug and Roger (75) estimates are not included in Table 1 because those authors did not recommend any estimates based on their reanalysis. A seven-gene data set of mostly protostomes has been analyzed in three separate studies by Peterson et al. (78), Peterson and Butterfield (57), and Peterson et al. (61). The first two analyses resulted in young time estimates largely agreeing with a direct reading of the fossil record (Cambrian Explosion). For example, the protostome– deuterostome divergence was estimated to be 592–556 Ma (78) and 579 Ma (57). However, three reanalyses (71, 74, 75) identified methodological problems in the original studies which were responsible for the underestimation of divergence times, such as the use of uncorrected distances and fi xed calibrations (maximum = minimum). The original studies (57, 78) also used a constant rate method rather than a relaxed clock method. The most recent study by Peterson et al. (61) involved a Bayesian analysis of the seven-gene data set. The use of relaxed clock methods resulted in older divergence time estimates, although the dates were still much younger than other molecular studies. For example, the protostome–deuterostome divergence was estimated as 643 Ma (669–615 Ma); this date increased to 733 Ma when probability distributions on fossil calibrations were used (61). However, a potential problem with this study was that five maximum calibrations were used, four of which were placed in the latest Precambrian and Cambrian (the Cambrian Explosion). The resulting time estimates for animal phyla were therefore prohibited from being much older than those constraints, thus tightly linking the posteriors (time estimates) to the priors (calibrations). The conclusions drawn by Peterson et al. (61), of young time estimates consistent with the fossil record, were thus an example of circular reasoning. Although they claimed that the results were robust to the use (or not) of maximum calibrations, this was based on the

228


experimental removal of only one of the five maximum calibrations. The seven-gene data set of Peterson et al. (61) was reanalyzed (S. B. Hedges, personal communication) using the same method of analysis, all original minimum calibrations, and the mid-Phanerozoic maximum calibration (insects), but with the four remaining maximum calibrations that created the circularity removed. The resulting protostome–deuterostome time estimate (845 Ma) was 202 million years older than reported by Peterson et al. (643 Ma) (Table 1), demonstrating that their young time estimates were a direct result of the use of the suspect maximum calibrations. Other studies using both constant rate and relaxed clock methods have estimated the divergence between protostomes and deuterostomes in a narrow range: 976 Ma (60) to 955 Ma (79). The importance of establishing a robust time estimate for the divergence of protostomes and deuterostomes will undoubtedly lead to the development of additional molecular clock methods and large, genome-scale data sets for future analyses. The divergence of Chordata and Ambulacraria within Deuterostomia has been estimated between 1001 and 590 Ma (63–66). A recent multigene study estimated a time of 896 Ma for the divergence of vertebrates and echinoderms using 71 nuclear proteins (50). Divergence time estimates for the earliest divergence within Chordata range from 890 to 547 Ma (50, 59, 61, 73, 79, 80). However, this divergence may need to be reevaluated in light of the recent phylogenetic evidence for a closer relationship between Vertebrata and Urochordata. Estimates for the divergence between Echinodermata and Hemichordata range from 875 to 535 Ma (50, 57, 61). Within the Protostomia, the divergence between Lophotrochozoa and Ecdysozoa has been dated at approximately 800 Ma (57, 61, 63, 73) with the deepest divergences within Ecdysozoa and Lophotrochozoa dated at ~700 Ma (57, 61). All of these divergences predate the first fossil evidence for animal phyla in the Cambrian. Establishing a robust timetree for Metazoa has important implications for understanding evolution in the Neoproterozoic. Essentially all molecular clock studies suggest that bilaterians had already radiated 100 million years or more before the Precambrian–Cambrian boundary. Studies proposing younger divergences consistent with a literal interpretation of the Cambrian Explosion have been shown to suffer from methodological biases, such as incorrect assumptions about rate models or the use of inappropriate fossil constraints. Geological evidence has suggested that the interval between approximately

800 and 550 Ma was one of planetary unrest, with multiple rounds of global glaciation (Snowball Earth, 81), changes in sea water chemistry (82–84), and increases in atmospheric oxygen levels (85, 86). Changing environmental pressures, along with the evolution of complex genetic pathways for skeletogenesis (87) and organ formation (88), had an evident effect on the evolutionary trajectory of metazoans during this important period in Earth’s history.

Acknowledgments This work was supported by the Howard Hughes Medical Institute Genomics Postdoctoral grant to Amherst College. I thank S.B. Hedges for sharing unpublished results.

References 1. S. M. Adl et al., J. Eukaryot. Microbiol. 52, 399 (2005). 2. E. T. Steenkamp, J. Wright, S. L. Baldauf, Mol. Biol. Evol. 23, 93 (2006). 3. D. J. Eernisse, K. J. Peterson, in Assembling the Tree of Life, J. Cracraft, Ed. (Oxford University Press, Cary, NC, 2004), pp. 197–208. 4. N. King, C. T. Hittinger, S. B. Carroll, Science 301, 361 (2003). 5. N. King et al., Nature 451, 783 (2008). 6. K. M. Halanych, Annu. Rev. Ecol. Evol. Syst. 35, 229 (2004). 7. A. G. Collins, Proc. Natl. Acad. Sci. U.S.A. 95, 15458 (1998). 8. C. Borchiellini et al., J. Evol. Biol. 14, 171 (2001). 9. M. Medina, A. G. Collins, J. D. Silberman, M. L. Sogin, Proc. Natl. Acad. Sci. U.S.A. 98, 9707 (2001). 10. J. Kim, W. Kim, C. W. Cunningham, Mol. Biol. Evol. 16, 423 (1999). 11. M. Podar, S. H. D. Haddock, M. L. Sogin, G. R. Harbison, Mol. Phylogenet. Evol. 21, 218 (2001). 12. J. R. Finnerty, K. Pang, P. Burton, D. Paulson, M. Q. Martindale, Science 304, 1335 (2004). 13. J. Spring et al., Dev. Biol. 244, 372 (2002). 14. N. H. Putnam et al., Science 317, 86 (2007). 15. M. Q. Martindale, J. R. Finnerty, J. Q. Henry, Mol. Phylogenet. Evol. 24, 358 (2002). 16. L. H. Hyman, The Invertebrates (McGraw-Hill, New York, 1940). 17. K. M. Halanych et al., Science 267, 1641 (1995). 18. A. M. Aguinaldo et al., Nature 387, 489 (1997). 19. D. Baurain, H. Brinkmann, H. Philippe, Mol. Biol. Evol. 24, 6 (2007). 20. H. Philippe, N. Lartillot, H. Brinkmann, Mol. Biol. Evol. 22, 1246 (2005).

Eukaryota; Metazoa

21. C. W. Dunn et al., Nature 452, 745 (2008). 22. N. Lartillot, H. Philippe, Phil. Trans. Roy. Soc. B 363, 1463 (2008). 23. G. Balavione, A. Adoutte, Integr. Comp. Biol. 43, 137 (2003). 24. W. E. Muller, Comp. Biochem. Phys. A 129, 433 (2001). 25. I. Ruiz-Trillo, M. Riutort, H. M. Fourcade, J. Baguna, J. L. Boore, Mol. Phylogenet. Evol. 33, 321 (2004). 26. C. L. Anderson, E. U. Canning, B. Okamura, Nature 392, 346 (1998). 27. E. Jimenez-Guri, H. Philippe, B. Okamura, P. W. H. Holland, Science 317, 116 (2007). 28. A. Hejnol, M. Q. Martindale, Phil. Trans. Roy. Soc. B 363, 1493 (2008). 29. J. Baguñà, P. Martinez, J. Paps, M. Riutort, Phil. Trans. Roy. Soc. B 363, 1481 (2008). 30. G. Balavoine, C.R. Acad. Sci. III-Vie 320, 83 (1997). 31. J. Mallatt, C. J. Winchell, Mol. Biol. Evol. 19, 289 (2002). 32. J. E. Blair, K. Ikeo, T. Gojobori, S. B. Hedges, BMC Evol. Biol. 2, 7 (2002). 33. G. Giribet, Phil. Trans. Roy. Soc. B 363, 1513 (2008). 34. M. J. Telford, S. J. Bourlat, A. Economou, D. Papillon, O. Rota-Stabelli, Phil. Trans. Roy. Soc. 363, 1529 (2008). 35. M. J. Telford, M. J. Wise, V. Gowri-Shankar, Mol. Biol. Evol. 22, 1129 (2005). 36. J. Mallatt, G. Giribet, Mol. Phylogenet. Evol. 40, 772 (2006). 37. G. Balavoine, R. de Rosa, A. Adoutte, Mol. Phylogenet. Evol. 24, 366 (2002). 38. S. W. Roy, W. Gilbert, Proc. Natl. Acad. Sci. U.S.A. 102, 4403 (2005). 39. A. Haase, M. Stern, K. Wachtler, G. Bicker, Dev. Genes Evol. 211, 428 (2001). 40. J. McPartland, V. Di Marzo, L. De Petrocellis, A. Mercer, M. Glass, J. Comp. Neurol. 436, 423 (2001). 41. G. W. Stuart, M. W. Berry, BMC Bioinformatics 5, 204 (2004). 42. Y. I. Wolf, I. B. Rogozin, E. V. Koonin, Genome Res. 14, 29 (2004). 43. H. Dopazo, J. Santoyo, J. Dopazo, Bioinformatics 20, 116 (2004). 44. G. K. Philip, C. J. Creevey, J. O. McInerney, Mol. Biol. Evol. 22, 1175 (2005). 45. S. J. Longhorn, P. G. Foster, A. P. Vogler, Cladistics 23, 130 (2007). 46. I. B. Rogozin, Y. I. Wolf, L. Carmel, E. V. Koonin, Mol. Biol. Evol. 24, 1080 (2007). 47. M. Irimia, I. Maeso, D. Penny, J. Garcia-Fernandez, S. W. Roy, Mol. Biol. Evol. 24, 1604 (2007). 48. J. Zheng, I. B. Rogozin, E. V. Koonin, T. M. Przytycka, Mol. Biol. Evol. 24, 2583 (2007). 49. S. W. Roy, M. Irimia, Mol. Biol. Evol. 25, 620 (2008). 50. J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 2275 (2005). 51. S. J. Bourlat et al., Nature 444, 85 (2006).

52. 53. 54.

55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76.

77. 78. 79. 80.

229

F. Delsuc, H. Brinkmann, D. Chourrout, H. Philippe, Nature 439, 965 (2006). E. E. Ruppert, Can. J. Zool. 83, 8 (2005). A. B. Smith, K. J. Peterson, G. Wray, D. T. J. Littlewood, in Assembling the Tree of Life, J. Cracraft, Ed. (Oxford University Press, Cary, NC, 2004), pp. 365–383. B. J. Swalla, A. B. Smith, Phil. Trans. Roy. Soc. 363, 1557 (2008). S. J. Bourlat, C. Nielsen, A. E. Lockyer, D. T. Littlewood, M. J. Telford, Nature 424, 925 (2003). K. J. Peterson, N. J. Butterfield, Proc. Natl. Acad. Sci. U.S.A. 102, 9547 (2005). K. J. Peterson, M. A. McPeek, D. A. D. Evans, Paleobiology 31, 36 (2005). N. Nikoh et al., J. Mol. Evol. 45, 97 (1997). S. B. Hedges, J. E. Blair, M. L. Venturi, J. L. Shoe, BMC Evol. Biol. 4, 2 (2004). K. J. Peterson, J. A. Cotton, J. G. Gehling, D. Pisani, Phil. Trans. Roy. Soc. B 363, 1435 (2008). M. Lynch, Evolution 53, 319 (1999). J. Otsuka, N. Sugaya, J. Theor. Biol. 222, 447 (2003). G. A. Wray, J. S. Levinton, L. H. Shapiro, Science 274, 568 (1996). D. F. Feng, G. Cho, R. F. Doolittle, Proc. Natl. Acad. Sci. U.S.A. 94, 13028 (1997). F. J. Ayala, A. Rzhetsky, Proc. Natl. Acad. Sci. U.S.A. 95, 606 (1998). X. Gu, J. Mol. Evol. 47, 369 (1998). M. Nei, P. Xu, G. Glazko, Proc. Natl. Acad. Sci. U.S.A. 98, 2497 (2001). D. Y. Wang, S. Kumar, S. B. Hedges, Proc. Roy. Soc. Lond. B Biol. Sci. 266, 163 (1999). S. Aris-Brosou, Z. Yang, Mol. Biol. Evol. 20, 1947 (2003). J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 387 (2005). J. J. Welch, E. Fontanillas, L. Bromham, Syst. Biol. 54, 672 (2005). E. J. P. Douzery, E. A. Snell, E. Bapteste, F. Delsuc, H. Philippe, Proc. Natl. Acad. Sci. U.S.A. 101, 15386 (2004). A. J. Roger, L. A. Hug, Phil. Trans. Roy. Soc. B 361, 1039 (2006). L. A. Hug, A. J. Roger, Mol. Biol. Evol. 24, 1889 (2007). S. B. Hedges, F. U. Battistuzzi, J. E. Blair, in Neoproterozoic Geobiology and Paleobiology, S. Xiao, A. J. Kaufman, Eds. (Springer, Dordrecht, The Netherlands, 2006), pp. 199–229. M. J. Benton, P. C. J. Donoghue, Mol. Biol. Evol. 24, 26 (2007). K. J. Peterson et al., Proc. Natl. Acad. Sci. U.S.A. 101, 6536 (2004). J. E. Blair, P. Shah, S. B. Hedges, BMC Bioinformatics 6, 53 (2005). S. B. Hedges, in Major Events in Early Vertebrate Evolution, P. E. Ahlberg, Ed. (Taylor & Francis, New York, 2001), pp. 119–134.

230

81.


K.-H. Hoff mann, D. J. Condon, S. A. Bowring, J. L. Crowley, Geology 32, 817 (2004). 82. S. T. Brennan, T. K. Lowenstein, J. Horita, Geology 32, 473 (2004). 83. D. E. Canfield, S. W. Poulton, G. M. Narbonne, Science 315, 92 (2007). 84. C. Scott et al., Nature 452, 456 (2008).

85. D. E. Canfield, A. Teske, Nature 382, 127 (1996). 86. A. H. Knoll, Life on a Young Planet (Princeton University Press, Princeton, 2003). 87. D. J. Jackson, L. Macis, J. Reitner, B. M. Degnan, G. Worheide, Science 316, 1893 (2007). 88. L. F. Sempere, C. N. Cole, M. A. McPeek, K. J. Peterson, J. Exp. Zool. B 306B, 575 (2006).

INVERTEBR ATES


Cnidarians (Cnidaria) Alex D. Rogers Institute of Zoology, Zoological Society of London, Regent’s Park, London NW1 4RY, UK ([email protected])

Abstract Cnidarians, which show a remarkable diversity of morphology and lifestyles, are important as reef-constructors, predators, and parasites in marine ecosystems. Few data currently exist on the timing of the evolutionary events among major groups of cnidarians (~7 classes and ~25 orders) and some of these are associated with high levels of uncertainty. However, fossil evidence, and molecular estimates of divergence times among members of the subclass Hexacorallia (Class Anthozoa), indicate that past climatechange events have had a significant impact on the evolution of reef-building corals and related groups.

The Phylum Cnidaria is ancient and diverse in terms of size and body shape, and includes the sea fans, sea pens, sea anemones, corals, hydroids, and jellyfish (Fig. 1). Cnidarians possess two cell layers (diploblastic), the outer ectoderm and the inner endoderm, separated by an acellular mesoglea, or partially cellular mesenchyme (1). The animals are radially symmetrical, although this may be modified, and have two basic forms, the polyp and the medusa. The sessile polyp is sac like, with a single body cavity (coelenteron) opening through the mouth which is surrounded by one or more rows of tentacles. The pelagic medusa is umbrella- or bell-shaped with a mouth located in the center of the concave underside surrounded by tentacles positioned around the margins of the animal. In polyps and medusae, the tentacles are armed with stinging or adhesive structures called cnidae, each produced by a stinging cell, the cnidocyte (1, 2). In cnidarians, the alternation of an asexual benthic polypoid form, with the sexually reproducing medusoid phase, is the primitive life-history state in extant taxa (e.g., many hydroids). Depending on the taxon, the polypoid or medusoid phase may be reduced or completely absent. In anemones and corals, for example, the medusoid phase is eliminated with the gonads developing within the polyps. Colonies of many species of benthic cnidarians can grow or reproduce through asexual

production of new polyps or colonies (e.g., corals), the overall size of which can be large. Cnidarians are carnivores, suspension feeders, or parasites, and many species within the phylum have symbiotic intracellular algae in their tissues. They are ecologically important animals in marine environments, although some are also found in freshwater. Corals have been important frameworkbuilding species in reefs from the Paleozoic (359 million years ago, Ma) to the present day in both shallow and deep waters, although the main hermatypic groups have changed over time with a dramatic turnover from rugose and tabulate corals before the great Permian extinction (251 Ma) to the scleractinian corals from the mid-Triassic (245–228 Ma) (3). Jellyfish and siphonophores are also ancient and are important in coastal and oceanic marine ecosystems as pelagic predators. Today the phylum has a high species diversity, with the Class Anthozoa containing more than 6100 species: >3000 in Subclass Octocorallia (4); >1113 in Subclass

Fig. 1 A scyphozoan from the Irish Sea (Aurelia aurita; upper left), an actiniarian from southwest Britain (Metridium senile; upper right), and a scleractinian from Maldives (Acropora sp., lower). Credit: A. Rogers.

A. D. Rogers. Cnidarians (Cnidaria). Pp. 233–238 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

234


Scleractinia (Robust forms) 4

Scleractinia (Complex forms) 5

Cubozoa 2

Scyphozoa Hydrozoa

Ed

Cm O S

600

D

C

P

PALEOZOIC 500

400

Tr

J

K

MESOZOIC 300

200

100

Medusozoa

Corallimorpharia

1

PR

Anthozoa

Other Anthozoa 3

Pg CZ 0 Million years ago

Fig. 2 A timetree of cnidarians. Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), Cm (Cambrian), CZ (Cenozoic), D (Devonian), Ed (Ediacaran), J (Jurassic), K (Cretaceous), O (Ordovician), P (Permian), Pg (Paleogene), PR (Proterozoic), S (Silurian), and Tr (Triassic).

Hexacorallia Order Actiniaria (5); >1600 in Order Scleractinia (5); >461 in Subclass Ceriantharia, and hexacoral Orders Zoanthidea, Corallimorpharia, and Antipatharia combined (5). The remaining cnidarians of the Medusozoa comprise more than 5954 species: 2184 in Class Myxozoa (6); 1 in the Class Polypodiozoa; 51 in Class Stauromedusae (7); 32 in Class Cubozoa (8); 212 in Class Scyphozoa (9); >166 in Class Hydrozoa, Subclass Trachylinae (10); >1900 in Subclass Hydroidolina, Order Leptomedusae (10); >1200 in Order Anthomedusae (10); and 199 in Order Siphonophora (10). The groups that are holopelagic or which have a mainly pelagic life history generally are less species-rich than the groups with a benthic or parasitic lifestyle. This is probably a result of the homogeneity and open nature of the pelagic environment, offering less opportunity for niche specialization and allopatric speciation than in groups where the dominant life-history phase is parasitic or benthic. The cnidarians originated in the early stages of metazoan evolution, in the Precambrian (>542 Ma) (11, 14). Some of the oldest metazoan fossils, part of the Ediacaran biota, have been attributed to three extant cnidarian groups, including the Chondrophorina (sailors by the wind), the Pennatulacea (sea pens), and the Scyphozoa (jellyfish). Modern chondrophorinans are assigned to the Family Porpitidae, within the Anthomedusae, on the basis of morphology. However, molecular phylogenetic analyses, based on nuclear small subunit (SSU) ribosomal DNA (rDNA) sequences, have suggested that the Ediacaran fossils attributed to the Chondrophorina are from another unrelated taxon (11). This study is, however, consistent with the existence of pennatulids in

the Ediacaran, 565 Ma, and the scyphozoans in the late Ediacaran 545 Ma (12, 13). Using calibration from the fossil record, and a Quartet-based method for estimating the divergence dates between cnidarian taxa from a phylogeny based on nuclear SSU rDNA, the root of the Cnidaria has been placed in the Proterozoic, 800–1000 Ma (11). Phylogenetic analysis of all available protein sequence data and the use of well-constrained calibration points from the fossil record, to estimate secondary calibration points in the Precambrian (>542Ma), similarly estimate the divergence of the Cnidaria and Bilateria at 1298 ± 74 Ma (14). These dates confirm the origin of cnidarians at the base of the metazoan radiation and show that they have a substantial hidden Precambrian history (11). Cnidarians have been an integral part of theories relating to the origin of metazoans for more than 100 years, because they have been regarded as primitive animals (15). Whether the ancestral cnidarian was polypoid or medusoid, the nature of the relationships between the classes of the phylum has been important to resolving how the diversity of metazoan life arose. The discovery that the Cubozoa, Scyphozoa, and Hydrozoa, including the Siphonophora, possess a unique derived structural alteration in their mitochondrial DNA (linear mitochondrial DNA, mtDNA) (16) was strong evidence that the medusoid groups are monophyletic, forming the clade Medusozoa. Support for the close association of Medusozoa and Anthozoa (circular mtDNA as for all other metazoans) has been provided by nuclear SSU rRNA-based phylogenies (17). The derivation of the Medusozoa from the ancestral anthozoans is the likely

Eukaryota; Metazoa; Cnidaria

235

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among cnidarians. Timetree Node

Estimates Ref. (11)

Time

Ref. (18)

Time

CI

Ref. (19)(a)

Ref. (19)(b)

Ref. (26)

Time

CI

Time

CI

Time

CI

Time

CI

1

640.9

709

777–641

711.7

1035–389

548

519–579

595

561–626

–

–

2

537.5

–

–

537.5

943–191

–

–

–

–

–

–

3

517.5

–

–

517.5

–

–

–

–

–

–

4

264.0

–

–

–

–

–

–

–

–

264

240–288

5

121.0

–

–

–

–

–

–

–

–

121

110–132

2308–0

Note: Node times in the timetree represent the mean of time estimates from different studies. Estimates from ref. (19) are from two different nodes.

path of evolution for these taxa, because it would require a single origin of the medusa as opposed to an origin and subsequent loss in the Anthozoa if they were the derived group (17). The origin of the Medusozoa has been estimated from nuclear SSU and large subunit (LSU) rRNA as probably during the Cryogenian (850–630 Ma; Table 1, Fig. 2) in the Neoproterozoic. These dates range from the early Neoproterozoic to early Cambrian (11, 18, 19), and the dates for the origins of the Medusozoa, the Scyphozoa, and the Hydrozoa are inconsistent, so the relationships between these taxa are shown as a polytomy (Table 1, Fig. 2). The Cryogenian was a period of severe global glaciations including the so-called “Snowball Earth” events (20). Recent work has suggested that despite the extreme conditions during parts of this period it is likely that areas of the oceans remained open even during glaciations (21). Phylogenetic analyses based on partial sequences of the mitochondrial SSU rRNA (22), partial and complete sequences of the nuclear SSU rRNA (23, 24), a combination of those data (25), or the entire mitochondrial genome (26, 27), have suggested that the Subclass Octocorallia is the closest relative of the group containing all other anthozoan orders. Some studies have placed the Ceriantharia (tube-dwelling anemones) basal to all other Anthozoa (28, 23), although the majority favor this group to be the closest relative of the Hexacorallia (22, 24, 25, 27, 29). The relationships between the orders of the Hexacorallia remain unresolved. In particular, the relationships between the Zoanthidea (zoanthids—colonial anemones), Actiniaria (sea anemones), and Antipatharia (black corals) are unclear and studies based on different genes have estimated different relationships between these taxa (24, 27). Estimates from nuclear SSU and LSU

data place the origin of the hexacorallians in the Lower Cambrian (542–513 Ma) (18), a date consistent with the earliest well-preserved fossil anemones. This period was a time when early reefs, comprising archaeocyath sponges and tabulate corals developed, although these collapsed at the end of the Lower Cambrian (30). Of all the hexacorals, the scleractinians are unusual in their sudden and late appearance in the fossil record in the Middle Triassic (Anisian stage, 237 Ma, 3). This followed the great Permian extinction which destroyed 96% of all marine species (31) and wiped out the Orders Rugosa and Tabulata, the main reef-building corals of the Paleozoic (542–251 Ma) (3). The cause of the Permian extinction is still a subject of significant debate. Largescale volcanism probably led to global warming, resulting from increased atmospheric CO2 and subsequent methane hydrate release. Oceanic anoxia, together with release of toxic hydrogen sulfide from the deep ocean and decreased marine productivity, are thought to have also occurred (31). The changes in ocean biochemistry and in other environmental parameters at the end of Permian (251 Ma) led to a prolonged recovery well into the Triassic (251–200 Ma) (3, 31). The appearance of numerous higher taxa of the Scleractinia on the margins of the Tethys Sea is abrupt in the fossil record and the origination of stony corals has been a topic of much debate to which molecular studies have made a significant contribution. Several potential ancestors to scleractinians have been postulated (3), including rugosan corals, the Scleractiniomorpha, and soft-bodied Hexacorallia from the Orders Actiniaria, Corallimorpharia, or Zoanthidea (the “naked coral” hypothesis, 32). Phylogenetic analyses of the Scleractinia, based on SSU rRNA, have identified two distinct clades, the “robust”

236


and the “complex” stony corals (33–36). Sequence divergence between these clades suggested that they originated about 300 Ma (34), a date that is older than the fossil appearance of the scleractinians in the Anisian stage of the Middle Triassic (245–237 Ma). However, subsequent analyses of the mitochondrial genome sequences of Anthozoa have refined the estimate for the origination of the Scleractinia at 288–240 Ma (26). This also supports the theory that the stony corals arose from naked hexacorallian ancestors in the Permian/Triassic and expressed the ability to secrete aragonite skeletons when ocean chemistry became conducive to the accretion of aragonite from seawater (3, 32). Experiments on growing scleractinian corals in seawater with a lowered pH may support this hypothesis. These show that at least some species can respond to the absence of conditions for the accretion of a skeleton by dissociation of the colonial form and complete skeletal dissolution (37). This may provide an explanation as to how corals might have survived large-scale environmental changes in the Permian and thus appeared abruptly within the fossil record at a later date. However, more than 40% of extant Scleractinia live in deep waters where preservation potential is reduced compared to shallow environments. Thus, the early history of skeletonized corals may have been obscured in the fossil record (26). At least one hexacorallian group has been eliminated as ancestors of the scleractinians. The corallimorpharians were found to have been derived from the “complex” clade of the Scleractinia through the analyses of mitochondrial genome sequence (26) (but see 27). This was dated as to have occurred in the mid- to late-Lower Cretaceous, 132–110 Ma. This was a time when changing seawater chemistry increasingly favored the secretion of calcitic skeletons over aragonitic skeletons in marine organisms (38) and reefs became dominated by rudist bivalves (3). The corallimorpharians lost the coral skeleton presumably as an adaptation to increasingly unfavorable conditions for the uptake of calcium carbonate and secretion of aragonite. The rise and fall of reef-building cnidarians through geological time is intimately connected with changes in marine chemistry which have been driven by climate change (3, 38). Present-day climate change has impacted scleractinians through global temperature increases, causing coral bleaching. Increasing levels of ocean acidification have the potential to reduce calcification rates and increase dissolution rates of coral skeletons that form reefs (39, 40). Geological history tells us that if current trends in environmental change continue, even if corals

survive through physiological refugia as “naked” polyps, there will be profound changes in the distribution, diversity, and structure of coral reef communities. Molecular studies of the Medusozoa have led to a much greater understanding of non-anthozoan relationships, but have resulted in few data on the timing of evolutionary events within the group, because of relatively few fossil records for calibration. Phylogenetic analyses of the Class Staurozoa, including a new species from deep-sea hydrothermal vents, using SSU rRNA sequences (7), have supported morphological studies suggesting this group is the closest relative of all other Medusozoa (41). This is consistent with the hypothesis that the medusoid form evolved from benthic ancestors (15). Gene sequences of the SSU and LSU of nuclear rRNA indicate that the group containing Cubozoa and Scyphozoa (named the Acraspeda, 15) is the closest relative of Hydrozoa and that these groups may have evolved in the Cryogenian (Table 1) (18). The mean dates for the node between the Hydrozoa and Scyphozoa/Cubozoa actually predate those for the origin of the medusozoans and so this is shown as a polytomy (Fig. 2). This reflects the high levels of uncertainty in present estimates of the timing of events in the evolution of the medusozoans. The timing of divergence between the hydrozoans and scyphozoans must have come after the evolution of the first medusozoans (15, 41). These dates again point to a substantial evolutionary history of the cnidarians in the Precambrian. The cubozoans and scyphozoans show marked differences in their life histories, but their relationship is supported by morphological similarities in the medusae of both groups (15). Nuclear SSU and LSU rRNA data place the node between these classes in the Early Cambrian at 538 Ma (18). Within the scyphozoan jellyfish, the Rhizostomeae has previously been considered as derived from the Semaeostomeae, because of similarities in the radial canal systems, which has been confirmed by molecular data (15, 17). The Coronatae (crown jellyfish) is a close relative of the Semaeostomeae and the Rhizostomeae. On the basis of mitochondrial SSU rRNA and nuclear SSU and LSU rRNA sequences, the Hydrozoa have been separated into two major clades, the Trachylina (Orders Limnomedusae, Trachymedusae, and Narcomedusae) and the Hydroidolina (Orders Anthoathecata, Leptothecata, Siphonophorae) (15, 42). The Trachylina consists mainly of marine medusoid forms (a few are freshwater) with simple, reduced, or even absent polyp stages. The Trachylina generally fall into the previously

Eukaryota; Metazoa; Cnidaria

recognized orders, although there is evidence of paraphyly in the Limnomedusae and Trachymedusae (15). Although the athecate hydroids, thecate hydroids, and siphonophores form separate clades, the relationships between the orders of the Hydroidolina are largely unresolved. This may be a result of rapid evolution in the early history of this group. In these analyses, the monotypic Subclass Langiomedusae falls within the Athecata, and the classification of athecate hydroids into the suborders Capitata and Filifera is not supported (14). However, a new clade, the Aplanulata, united by a shared-derived character (development from egg to polyp via a nonciliated stereogastrula stage, rather than a ciliated planula stage, 43) is resolved by the molecular analyses (15, 42). This clade comprises the Tubulariidae, Corymorphidae, Candelabridae, and Hydridae. Other anthoathecate hydroid groups that exhibit this type of development may also fall within this group, although they are yet to be sampled for molecular studies (15). Ultrastructural studies and analyses of the SSU of the nuclear rRNA have suggested that the parasitic myxozoans, previously regarded as protists, are cnidarians (44). Phylogenetic analyses of 129 protein sequences have provided strong support for this hypothesis and furthermore suggest that the myxozoans are highly derived medusozoans (45). At present the relationship of the myxozoans to other medusozoan taxa has not been resolved, but this discovery has changed understanding of the diversity of species and lifestyles adopted by cnidarians. Myxozoans are parasites of a variety of animals including annelids, bryozoans, and fish and can be economically significant, especially in the aquaculture industry. Molecular phylogenetic approaches are now being employed to examine the evolution and systematics of cnidarians below the level of Order. Within the Anthozoa, poor correspondence between the preexisting morphological taxonomy and molecular phylogenetics trees has been discovered, especially within the Octocorallia (46–48) and Scleractinia (33–36). This suggests that previous interpretation of the homology of the characters of the skeletons of anthozoans, and other aspects of morphology, are unreliable as a result of convergent or parallel evolution (47). At the subordinal taxonomic levels, biogeographic and historical factors also become an important influence in the evolution and systematics of coral (49) and other cnidarian taxa (e.g., hydroids, 50). DNA sequence analyses are also demonstrating the existence of many cryptic taxa at the lower taxonomic levels such as Scleractinia (51), Octocorallia (48), Scyphozoa (52, 53), and Hydrozoa

237

(50, 54). Some of these groups have been viewed as having species with cosmopolitan or very wide geographic distributions resulting from pelagic life-history stages or hydrochory of sessile adults and a lack of barriers to dispersal across the oceans. This is leading to a reevaluation of the systematics, distribution, and overall species diversity of cnidarian taxa.

Acknowledgment The author thanks the Institute of Zoology, Zoological Society of London for funding as a Senior Research Fellow and for provision of the facilities required for the writing of this chapter.

References 1. R. C. Brusca, G. J. Brusca, Invertebrates, 2nd edn. (Sinauer Associates, Sunderland, Massachusetts, 2003). 2. R. Gibson, B. Hextall, A. Rogers, Photographic Guide to the Sea & Shore Life of Britain and North-West Europe (Oxford University Press, Oxford, UK, 2001). 3. G. D. Stanley, Earth-Sci. Rev. 60, 195 (2003). 4. G. C. Williams, S. D. Cairns, Systematic List of Valid Octocoral Genera, http://www.calacademy.org/research/ izg/OCTOCLASS.htm (accessed July 2007). 5. D. G. Fautin, Hexacorallians of the World, http:// geoportal.kgs.ku.edu/hexacoral/anemone2/index.cfm (accessed July 2007). 6. J. Lom, I. Dykova, Folia Parasitol. 53, 1 (2006). 7. A. G. Collins, M. Daly, Biol. Bull. 208, 221 (2005). 8. M. N. Dawson, Class Cubozoa Werner 1975, http:// thescyphozoan.ucmerced.edu/Syst/Cub/Cubomedusae. html (accessed August 2007). 9. M. N. Dawson, The Scyphozoan, http://www2.eve. ucdavis.edu/mndawson/tS/tsFrontPage.html (accessed August 2007). 10. P. Schuchert, The Hydrozoan Directory, http://www. ville-ge.ch/mhng/hydrozoa/hydrozoa-directory.htm (accessed August 2007). 11. B. Waggoner, A. G. Collins, J. Paleontol. 78, 51 (2004). 12. G. Hahn et al., Geologica Palaeontol. 16, 1 (1982). 13. J. W. Hagadorn, B. Waggoner, J. Palaeontol. 74, 349 (2000). 14. S. B. Hedges, J. E. Blair, M. L. Venturi, J. L. Shoe, BMC Evol. Biol. (2004). 15. A. G. Collins et al., Syst. Biol. 55, 97 (2006). 16. D. Bridge et al., Proc. Natl. Acad. Sci. U.S.A. 89, 8750 (1992). 17. A. G. Collins, J. Evol. Biol. 15, 418 (2002). 18. P. Cartwright, A. Collins, Integr. Comp. Biol. 47, 744 (2007). 19. K. J. Peterson et al., Proc. Natl. Acad. Sci. U.S.A. 101, 6356 (2004).

238


20. P. F. Hoff man, D. P. Schrag, Terra Nova 14, 129 (2002). 21. R. Rieu, P. A. Allen, M. Ploetze, T. Pettke, Geology 35, 299 (2007). 22. S. C. France et al., Mol. Mar. Biol. Biotech. 5, 15 (1996). 23. J.-I. Song, J. H. Won, Korean J. Biol. Sci. 1, 43(1997). 24. E. A. Berntson, S. C. France, L. S. Mullineaux, Mol. Phylogenet. Evol. 13, 417 (1999). 25. D. Bridge, C. W. Cunningham, R. DeSalle, L. W. Buss, Mol. Biol. Evol. 12, 679 (1995). 26. M. Medina et al., Proc. Natl. Acad. Sci. U.S.A. 103, 9096 (2006). 27. M. R. Brugler, S. C. France, Mol. Phylogenet. Evol. 42, 776 (2007). 28. C. A. Chen et al., Mol. Phylogenet. Evol. 4, 175 (1995). 29. M. Daly, D. L. Lipscomb, M. W. Allard, Evolution 56, 502 (2002). 30. G. D. Stanley, The History and Sedimentology of Ancient Reef Systems (Springer, New York, 2001). 31. R. V. White, Phil. Trans. Roy. Soc. Lond. A 360, 2963 (2002). 32. G. D. Stanley, D. G. Fautin, Science 291, 1913 (2001). 33. S. L. Romano, S. R. Palumbi, Science 271, 640 (1996). 34. S. L. Romano, S. R. Palumbi, J. Mol. Evol. 45, 397 (1997). 35. M. C. Le Goff-Vitry, A. D. Rogers, D. Baglow, Mol. Phylogenet. Evol. 30, 167 (2004). 36. A. M. Kerr Biol. Revs. 80, 543 (2005). 37. M. Fine, D. Tchernov, Science 315, 1811 (2007).

38. S. M. Stanley, Palaeogeogr. Palaeoclimat. Palaeoecol. 232, 214 (2006). 39. J. M. Guinotte et al., Front. Ecol. Environ. 1, 141 (2006). 40. C. Wilkinson, in Status of Coral Reefs of the World: 2004, C. Wilkinson, Ed. (Australian Institute of Marine Science, Townsville, Queensland, Australia, 2004). 41. A. G. Collins, M. Daly, Biol. Bull. 208, 221 (2005). 42. A. G. Collins, S. Winkelmann, H. Hadrys, B. Schierwater, Zool. Scripta 34, 91 (2005). 43. K. W. Petersen, Zool. J. Linn. Soc. 100, 101 (1990). 44. M. E. Siddall, D. S. Martin, D. Bridge, S. S. Desser, D. K. Cone, J. Parasitol. 81, 961 (1995). 45. E. Jiménez-Guri, H. Philippe, B. Okamura, P. W. H. Holland, Science 317, 116 (2007). 46. J. A. Sánchez, H. R. Lasker, D. J. Taylor, Mol. Phylogenet. Evol. 29, 31 (2003). 47. C. S. McFadden, S. C. France, J. A. Sánchez, P. Alderslade, Mol. Phylogenet. Evol. 41, 513 (2006). 48. C. S. McFadden et al., Invert. Biol. 125, 288 (2006). 49. H. Fukami et al., Nature 427, 832 (2004). 50. A. F. Govindarajan, K. K. Halawych, C. W. Cunningham, Mar. Biol. 146, 213 (2005). 51. M. J. H. van Oppen, B. J. McDonald, B. Willis, D. J. Miller, Mol. Biol. Evol. 18, 1315 (2001). 52. M. N. Dawson, D. K. Jacobs, Biol. Bull. 200, 92 (2001). 53. M. N. Dawson, Invert. Syst. 19, 361 (2005). 54. P. Schuchert, Mol. Phylogenet. Evol. 36, 194 (2005).

Scaphopod mollusks (Scaphopoda) Jan M. Strugnella,* and A. Louise Allcockb a

Department of Zoology, University of Cambridge, Downing St, Cambridge, CB2 3EJ, UK; b The Martin Ryan Marine Science Institute, National University of Ireland Galway, University Road, Galway, Ireland. *To whom correspondence should be addressed (jan.strugnell@ gmail.com)

Abstract The tusk shells (~500 sp.) are grouped into 14 families and two orders within the molluscan Class Scaphopoda. Only two molecular studies have focused on phylogenetic relationships within scaphopods. Estimates of divergence times among families are estimated here. The initial divergence among scaphopods, separating Gadilida and Dentaliida, is estimated to have occurred near the Devonian–Carboniferous boundary, ~359 million years ago (Ma), with the Fustiariidae, Rhabdidae, and Dentaliidae diverging in the Carboniferous (359–299 Ma). In contrast, the families included in the study from the Order Gadilida were estimated to have diverged from one another in the Cretaceous, 139–96 Ma.

The scaphopods (Phylum Mollusca, Class Scaphopoda) are known as tusk shells because of their curved shape (resembling elephant tusks), open at both ends (Fig. 1). They are relatively small, usually 3–6 cm in length. Scaphopods burrow into sediments with the wider (anterior) end of the shell oriented downward. Both the head and foot (used for burrowing) have an anterior location, whereas the viscera are posterior. There are ~500 valid species of recent scaphopods and about 800 valid fossil species. There is some argument as to when the lineage originated. Scaphopod fossils have been described from the Ordovician, Silurian, and Devonian, but many of these specimens have been reclassified as belonging to other groups. Yochelson (1) and others have suggested that scaphopods most likely evolved in the early Carboniferous. Here we review the evolutionary relationships and divergence times of the members of the Class Scaphopoda. The Class Scaphopoda consists of 14 families and two orders. Two of the families contain only fossil genera

(Baltodentialiidae and Prodentaliidae), whereas four others contain at least one fossil genus each, along with genera containing living species (Dentaliidae, Gadilinidae, Laevidentaliidae, and Gadilidae). The orders of Scaphopoda differ in the shape of the foot. The dentaliidans have a conical foot whereas gadilidans have a worm-shaped foot with a terminal disk capable of expansion. Additional distinguishing features are provided by Steiner (2). The monophyly of the two orders has been supported by morphological data (3) and by molecular analyses based on the nuclear gene for 18S ribosomal RNA (rRNA) (4) and the mitochondrial cytochrome oxidase I gene (COI) (5). The Order Gadilida comprises four recent families. Entalinidae is placed within the Suborder Entalimorpha, distinguished by a ribbed shell and by a smooth rachis in the radula. The remaining three families—Pulsellidae, Wemersoniellidae, and Gadilidae—are placed within the Suborder Gadilimorpha, distinguished by a smooth shell and by a cuspid rachis. Support for these suborders has been provided by morphological data (3, 6–9) and by molecular analyses based on 18S rRNA (4), although molecular analyses based on COI have suggested that the Gadilimorpha is paraphyletic (5). Analyses using 18S rRNA did not support the monophyly of the Gadilidae (4).

Fig. 1 Two scaphopod shells (Pictodentalium vernedei) from Taiwan (right) and two shells of an undescribed species (Pictodentalium sp.) from Broome, Australia. Credit: B. Sahlmann.

J. M. Strugnell and A. L. Allcock. Scaphopod mollusks (Scaphopoda). Pp. 239–241 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Gadilidae-2 6

Pulsellidae

5

Gadilidae-1

4

Gadilida

THE TIMETREE OF LIFE Gadilimorpha

240

Dentaliidae 3

Rhabdidae

2

Fustiariidae

C

Tr

P

PALEOZOIC 350

300

J

K

MESOZOIC 250

200

150

Pg

Dentaliida

Entalinidae 1

Ng

CZ 100

50

0 Million years ago

Fig. 2 A timetree of Scaphopoda. Divergence times are shown in Table 1. Gadilidae-1 contains the Subfamily Siphonodentaliinae and Gadilidae-2 contains the Subfamily Gadilinae of the classical Gadilidae. Abbreviations: C (Carboniferous), CZ (Cenozoic), J (Jurassic), K (Cretaceous), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

The Order Dentaliida comprises eight recent families whose interrelationships are not well resolved. Different taxon sampling within morphological studies makes comparisons difficult, and authors have expressed preferences for different character sets which have yielded conflicting results. Taxon sampling within molecular studies is particularly poor. Representatives from three dentaliid families were sequenced for 18S rRNA (4). Fustiariidae was basal to a clade composed of Dentaliidae and Rhabdidae. The monophyly of Dentaliidae was not supported because of Rhabdus (Rhabdidae) falling within a clade containing Antalis, Fustiaria, Dentalium, and Fissidentalium (Dentaliidae). A study based on COI (5) suggested Dentaliidae to be paraphyletic. In this case, Rhabdus grouped with Fissidentalium, which was closest to a clade containing Antalis and Dentalium. There are no previous published studies estimating divergence times among scaphopod families. We have therefore taken the nuclear 18S rRNA sequences from GenBank (4) and applied a penalized likelihood method of Sanderson (10) in the program “r8s” to estimate these divergence times. Cross-validation scores were examined over a range of smoothing parameters to find the optimal smoothing parameter for the analysis. Confidence intervals were estimated using a bootstrap approach. We selected only those (minimum) fossil constraints whose validity has not been questioned to date. For example, many authors (1, 4, 11–13) reject claims that Rhytiodentalium kentuckyensis and other early fossils showing “scaphopodization” are true scaphopods (12). So, we have not used them. Yet, we have made an attempt to select the earliest fossil representatives for

each relevant taxon. We included the following fossils: Prodentalium fredericae, Dentalium acutoides, Antalis torquatus, Fissidentalium pukaea, Fustiaria glabellum, Rhabdus paralelum, Entalina curvum, Cadulus groenlandicus, Polyschides arnoensis, and Pulsellum infundibulum. Where the specific age of a fossil was not given, the midpoint of the epoch/age of the fossil was used as a minimum constraint. Fossil dates used for calibration are as follows: minimum of 329 Ma for the divergence of Dentalida and Gadilida, minimum of 322 Ma for the diversification of Dentalidae (Dentalium v. Antalis), minimum 172.5 Ma for the divergence of Rhabdidae and Dentalidae, minimum of 123 Ma for the divergence of Entalinidae and Gadilidae, minimum of 28.3 Ma for the divergence of Fustariidae and Dentalidae, minimum of 18 Ma for the divergence of Gadilidae and Pulsellidae (Cadulus vs. Pulsellum), a minimum of 88 Ma for the diversification of Gadilinae (Gadilidae; Cadulus vs. Cadulus), and a minimum of 45 Ma for the diversification of Siphonodentaliinae (Gadilidae: Siphonodentalum vs. Polyschides). A variety of conflicting hypotheses has been proposed as to which molluscan crown groups are closest to the Scaphopoda. However, recent molecular evidence has supported a close relationship between Cephalopoda and Scaphopoda (4, 14). We have therefore rooted the tree with the four cephalopod species used in the previous study (4). Given that potential scaphopod fossil species have been described from as early as the Ordovician (although many are admittedly controversial), we have again taken a conservative approach and placed a maximum age constraint of 488 Ma on the divergence of

Eukaryota; Metazoa; Mollusca; Scaphopoda

Table 1. Divergence times (Ma) and their confidence/ credibility intervals (CI) among tusk shell mollusks based on analyses reported here. Timetree Node

Time

CI

1

363.3

367–359

2

329.9

354–306

3

324.0

345–303

4

139.5

154–124

5

106.2

121–91

6

96.0

110–82

Note: Estimates are based on a penalized likelihood analysis of the nuclear 18S rRNA sequences.

the cephalopods and scaphopods at the Cambrian– Ordovician border. The resulting timetree is shown in Fig. 2. The initial divergence among scaphopods, separating Gadilida and Dentaliida, is estimated to have occurred ~363 Ma, with the Fustiariidae, Rhabdidae, and Dentaliidae diverging in the Carboniferous (359–299 Ma). In contrast, the families included in the study from the Order Gadilida were estimated to have diverged from one another in the Cretaceous, 139–96 Ma. The most basal family, the Entalinidae, was estimated to have diverged close to the Jurassic–Cretaceous border, 145 Ma. Steiner and Dreyer’s (4) sequence data imply the polyphyly of the Gadilidae (see earlier). The Subfamily Gadilinae (Gadilidae-2, Fig. 2) was estimated to have diverged from the Family Pulsellidae in the Middle Cretaceous (110–82 Ma). Together this clade diverged from the Subfamily Siphonodentaliinae (Gadilidae-1, Fig. 2) slightly earlier in the Cretaceous. Whilst the dates proposed for the divergence of groups within the Dentaliida show close affi liation with dates from the fossil record, our timetree suggests that divergences within the Gadilida occurred a little earlier than is suggested by the fossil record (15).

241

The fossil record of scaphopods is extensive and is well suited to a molecular dating analysis by taking into account the abundance and distribution of fossil scaphopods. However, such an endeavor awaits further progress in molecular sequencing of the Scaphopoda. Additional sequencing of more scaphopod families and a greater number of genes would undoubtedly improve the resolution and information that could be gained from such an analysis.

Acknowledgments This work would not have been possible without the extensive catalogue of Steiner and Kabat, 2004 and without the online database of Bernd Sahlmann. J.S. is supported by a Natural Environment Research Council Antarctic Funding Initiative grant awarded to L.A and a Lloyd’s Tercentenary Fellowship.

References 1. E. L. Yochelson, Ann. Naturhist Mus. Wien, Ser. A 106, 13 (2004). 2. G. Steiner, J. Mollus. Stud. 58, 385 (1992). 3. P. D. Reynolds, A. Okusu, Zool. J. Linn. Soc. 126, 131 (1999). 4. G. Steiner, H. Dreyer, Zool. Scripta 32, 343 (2003). 5. G. Steiner, P. D. Reynolds, in Molecular Systematics and Phylogeography of Mollusks, C. Lydeard, D. L. Lindberg, Eds. (Smithsonian Books, Washington, 2003), pp. 123–139. 6. P. D. Reynolds, Zool. Scripta 26, 13 (1997). 7. G. Steiner, J. Mollus. Stud. 58, 385 (1992). 8. G. Steiner, Zool. Scripta 27, 73 (1998). 9. G. Steiner, J. Mollus. Stud. 65, 151 (1999). 10. M. J. Sanderson, R8s, Version 1.5 (University of California, Davis, 2002). 11. J. S. Peel, Acta Palaeontol. Polonica 49, 543 (2004). 12. J. S. Peel, Palaeontology 49, 1357 (2006). 13. T. Engeser, F. Riedel, Mitt Geol-Paläontol. Inst. Univ. Hamburg 79, 117 (1996). 14. Y. J. Passamneck, C. Schander, K. M. Halanych, Mol. Phylogenet. Evol. 32, 25 (2004). 15. W. K. Emerson, J. Paleontol. 36, 461 (1962).

Cephalopod mollusks (Cephalopoda) Jan M. Strugnella,*, Annie Lindgrenb, and A. Louise Allcockc a

Department of Zoology, University of Cambridge, Downing St, Cambridge, CB2 3EJ, UK; bMuseum of Biological Diversity, The Ohio State University, 1315 Kinnear Road, Columbus, OH 43215, USA; cThe Martin Ryan Marine Science Institute, National University of Ireland Galway, University Road, Galway, Ireland. *To whom correspondence should be addressed (jan.strugnell@ gmail.com)

Abstract Squids, cuttlefish, octopuses, and nautiluses (~700 species) are grouped into 47 families within the Class Cephalopoda of the Phylum Mollusca. The resolution of many higherlevel phylogenetic relationships within cephalopods has been hindered by homoplasy among morphological characters, although some recent progress has been made with molecular phylogenies and molecular clocks. The cephalopod timetree supports a Paleozoic (542–251 million years ago, Ma) origin of the Orders Vampyromorpha, Octopoda, and the majority of the extant higher-level decapodiform taxa. The major lineages within the Order Octopoda were estimated to have diverged during the Mesozoic era (251–66 Ma).

The class Cephalopoda is a monophyletic group which can be divided into two subclasses; Nautiloidea and Coleoidea. Nautiloidea contains the nautiluses (Nautilus and Allonautilus), whereas Coleoidea contains the octopuses (Fig. 1), squids, and cuttlefishes. Coleoid cephalopods differ from nautiloids most notably through the reduction (or complete loss) and internalization of the shell. Defining features of Coleoidea include a muscular mantle used for locomotion and respiration, the modification of the foot into appendages around the mouth, a closed circulatory system, and complex eyes with lenses, although many of these features have been lost or reduced in various taxa. The widely cited annotated classification of the recent Cephalopoda (1) listed over 700 valid species in 139 genera and 47 families. Here we review the evolutionary relationships and divergence times of the members of the class Cephalopoda. Several

alternative classifications have been proposed for relationships within Cephalopoda (2). Herein we follow the classification of Young et al. (3) that generally does not include ranks above the family level; however, we make certain assumptions about rank based on nomenclature and position. The Coleoidea are divided into two superorders: Decapodiformes and Octopodiformes (4). The Decapodiformes is a morphologically and ecologically diverse group comprising 31 families, 95 genera, and approximately 450 species. Four major lineages are recognized within Decapodiformes: Sepioidea (cuttlefish, bottletail, and bobtail squids), Spirulida (the Ram’s horn squid), Oegopsida (open-eye squids), and Myopsida (closedeye squids) (3, 5). Three families, Bathyteuthidae,

Fig. 1 A benthic octopus (Pareledone charcoti) from Antarctica. Photo credit: A. L. Allcock.

J. M. Strugnell, A. Lindgren, and A. L. Allcock. Cephalopod mollusks (Cephalopoda). Pp. 242–246 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Spirulidae

Ommastrephidae

2 10

Joubiniteuthidae

7 1

Pyroteuthidae

6

Oegopsida

Loliginidae

243

Decapodiformes

Sepiidae 8

3

Myopsida

Sepiolidae 4

Sepioidea

Eukaryota; Metazoa; Mollusca; Cephalopoda

Bathyteuthidae Idiosepiidae

K

Pg

300

Ng

CZ

200

0 Million years ago

100

Bolitaenidae 16

Vitreledonellidae

15

Octopodidae-3

14

Octopodidae-2

13

Octopodidae-1

11

Tremoctopodidae 12

9

Argonautidae Opisthoteuthidae

5

17

Stauroteuthidae

Octopoda

400

J MESOZOIC

Incirrata

Tr

Cirrata

P

Ctenoglossa

C PALEOZOIC

Argonautoidea

D

Vampyroteuthidae

P

Triassic

Jurassic

PZ

Cretaceous

250

200

150

Pg

Ng

CENOZOIC

MESOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of cephalopod mollusks: Decapodiformes and Octopodiformes. Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Ng (Neogene), P (Permian), Pg (Paleogene), PZ (Paleozoic), and Tr (Triassic). Octopodidae 1 = Octopus, Hapalochlaena; Octopodidae 2 = Benthoctopus, Enteroctopus; Octopodidae 3 = Adelieledone, Pareledone.

Chtenopterygidae, and Idiosepiidae, have features that do not clearly place them within one of these lineages. However, it appears that bathyteuthids and chtenopterygids are closely related and these have been grouped together in the Bathyteuthoida (3). The somewhat unique pygmy squids (6), Idiosepiidae, currently remain as a stand-alone family. Relationships among the lineages remain somewhat unclear due to conflicting or inconclusive analyses of both morphological (7) and molecular (8–11) data. Sepioidea, Idiosepiidae, and Spirulida have often been interpreted as closely related (12, 13), but have never been found to be a monophyletic group in

any modern phylogenetic study containing representatives of each of these taxa. Some molecular studies support a relationship between Bathyteuthidae, Spirulida, and Sepioidea but not Idiosepiidae (14), while others support a close relationship between Bathyteuthidae and Oegopsida (11). Sepioidea contains two groups, the bobtail squids, Sepiolida (families Sepiolidae and Sepiadariidae), and the cuttlefishes (Family Sepiidae), united by morphological features such as eyes with secondary lids and a funnel with a lateral canal (5). Molecular data have yet to support a monophyletic Sepioidea; recent molecular

244


Table 1. Divergence times (Ma) and their credibility intervals (CI) among cephalopod mollusks. Timetree Node


Time

Time

CI

Ref. (11)(b) Time

CI

1

390.0

390.0

596–236

–

–

2

349.0

349.0

544–206

–

–

3

321.0

321.0

506–188

–

–

4

297.0

297.0

473–172

–

–

5

285.5

252.0

369–175

319.0

476–206

6

282.0

282.0

454–161

–

–

7

230.0

230.0

–

–

–

8

206.0

206.0

345–109

–

–

9

195.5

174.0

262–118

217.0

336–133

10

175.0

–

–

–

–

11

156.0

136.0

208–92

176.0

277–106

12

111.5

105.0

166–67

118.0

196–66

13

110.0

100.0

155–66

120.0

196–70

14

88.5

77.0

121–49

100.0

166–57

15

77.5

67.0

107–43

88.0

146–50

16

75.0

65.0

–

85.0

–

17

75.0

65.0

–

85.0

–

Note: Node times in the timetree represent the mean of time estimates from different studies. Divergence times for the Octopodiformes are from a Bayesian analysis of (a) three mitochondrial and three nuclear genes partitioned by gene and codon and (b) three nuclear genes partitioned by gene and codon. Divergence times for Decapodiformes are from a Bayesian analysis of (a) three nuclear genes not partitioned (first and second codon positions only).

studies found Sepiidae to be closest to Spirulida (10, 14), with Sepiolidae closest to a Sepiidae + Spirulida clade. Alternatively, Lindgren et al. (9) found Sepiidae and Idiosepiidae to be closest relatives, while Strugnell et al. (10, 11) found Idiosepiidae and Sepiolidae to be closest relatives. Although their relationships are somewhat unstable across molecular analyses, Sepiidae and Sepiolida appear to be closely related on the basis of both morphological and molecular data. Spirulida contains a single living species, Spirula spirula, which exhibits a number of unique features, the most striking being the retention of a phragmocone with unusual ventral coiling, making it particularly difficult to place among extant decapodiforms. Furthermore, it lacks a cornea, which makes placing Spirulida as the closest relative of Sepiidae difficult, as all sepioids possess a

corneal covering over the eye. The absence of a cornea implies either a close relationship with Oegopsida or that loss is a convergent character in Sepiidae and Oegopsida. Naef (12) believed Spirulida to be closest to Sepiidae (cuttlefishes) due to similarity in embryonic phragmocone development. Some molecular studies have found a close relationship between Spirulida and Sepiidae (10, 14), although other molecular studies (8, 9) yielded different phylogenetic positions for Spirulida. Because of the unique morphological characters and inconsistent results with DNA sequence data, further examination is required before the position of Spirulida can be firmly established. Oegopsida and Myopsida were historically grouped together in the Suborder Teuthoidea (e.g., 12), united by similarities in gladius, branchial canal, tentacular clubs, and interstellate connective (15). The primary feature used to separate Oegopsida and Myopsida is the lack of a corneal covering in the oegopsid eye. A close relationship between Myopsida and Oegopsida was found in several topologies generated by Carlini and Graves (8) using the COI locus, but other analyses have rendered Oegopsida paraphyletic with respect to Myopsida, or found a close relationship between Myopsida and some sepioids (9–11). Myopsida and Sepioidea exhibit several features in common, such as the presence of a cornea, beak without an angled point, vena cava ventral to intestine, buccal crown with suckers (only some taxa), accessory nidamental glands, and tentacle pockets (15). Because of the lack of resolution at basal nodes in the tree, many of the features that unite Sepioidea and Myopsida may be ancestral. The lack of a consistent relationship between Myopsida and Oegopsida suggests that the taxon Teuthoidea may not be valid. Oegopsida is the most species-rich and morphologically diverse group within the Decapodiformes. Species can range in size from a few centimeters in the Family Pyroteuthidae to tens of meters in the giant squid, Architeuthis dux. The diversity in anatomy, behavior, and morphology of Oegopsida, combined with a somewhat uninformative fossil record (15), has made generating hypotheses of family-level relationships within this group very difficult. Morphology-based phylogenies have focused primarily on higher-level relationships, and have not yielded well-supported groups within Decapodiformes, because of difficulty in establishing polarity and homology (7). At the family level, molecular data have recently been employed, but the trees generated are highly dependent on gene choice, taxon sampling, and analytical method (8–10, 16, 17).

Eukaryota; Metazoa; Mollusca; Cephalopoda

The Octopodiformes contains the Orders Vampyromorpha (vampire “squid”) and Octopoda (pelagic and benthic octopuses). A close relationship between these two orders has been supported by morphological studies (7, 16, 18, 19) and some molecular studies (10, 20, 21), although some molecular phylogenies (9, 22) provide contradictory or inconclusive support for a close relationship between Vampyromorpha and Decapodiformes. The Octopoda contains the Suborders Cirrata (deep-sea finned octopuses) and the Incirrata (benthic and pelagic octopuses). A close relationship between these suborders is widely accepted (7, 23, 24). Four families are currently recognized within the Cirrata (Opisthoteuthidae, Grimpoteuthidae, Cirroctopodidae, and Cirroteuthidae) (25, 26). A molecular study using 16S rRNA suggests the first three of these families, which all possess a single web, form a clade, to the exclusion of the last one, which possesses an intermediate or secondary web (25). The phylogenetic relationships between the eight families of Incirrata have not been well investigated. Four of the pelagic families (Alloposidae, Tremoctopodidae, Argonautidae, and Ocythoidae) comprise a well-defined monophyletic clade (Superfamily Argonautoidea) linked by a detachable hectocotylus (the modified arm used in copulation) in males within this group. Although no molecular study has been published which encompasses all four families, molecular evidence to date has been supportive of this grouping (9, 10, 21, 27). Naef (12) proposed that the remaining pelagic octopods (Vitreledonellidae, Amphitretidae, and Bolitaenidae) be placed in a grouping Ctenoglossa based on the structure of the radula. The monophyly of this grouping has been confirmed by molecular studies (27) and it has been suggested that ctenoglossans have neotenous origins (21). Molecular work shows the ctenoglossans are unlikely to be closely related to the Argonautoidea (21, 27). One might therefore suppose that the remaining family, the Octopodidae, provides an evolutionary link between these two groups. However, the situation is far more complex than this. Molecular studies (10, 27) suggest that the Octopodidae is not monophyletic and this is hindering our understanding of other relationships within the Incirrata. Two molecular-based studies have estimated divergence times among the major lineages of cephalopods (11, 21). Both studies used three nuclear (rhodopsin, pax-6, octopine dehydrogenase) and three mitochondrial (16S rRNA, 12S rRNA, cytochrome oxidase I [COI]) genes. The first of these employed a penalized likelihood

245

(PL) method with a primary interest of estimating the divergence time of the grouping Ctenoglossa from the remaining Octopoda (21). The second study (11) used the same six genes and employed a Bayesian approach (28) to estimate divergence times both within octopodiforms (eight families) and decapodiforms (nine families). Both studies used constraints taken from the coleoid cephalopod fossil record. The timetree of coleoids using the Bayesian approach was based on the results obtained from a prior phylogenetic study investigating the effect of data partitioning on resolving phylogenies in a Bayesian framework (10). This study showed that the strongest phylogenetic resolution for the Octopodiformes was obtained from analyses using all six genes partitioned by gene and codon. Two topologies were presented for the Decapodiformes: one resulting from analysis of nuclear genes (all three codon positions) partitioned by codon, and the second resulting from only the first and second codon positions (not partitioned). These phylogenies and partitions were used within the Bayesian approach to estimate divergence times within coleoids. Here we present the topology obtained using first and second positions only as it is likely that the third positions are saturated and are therefore not informative in estimating deep divergences. Furthermore, to determine whether there was any difference between mitochondrial and nuclear genes on dating estimates, the octopodiform topology was also analyzed using only nuclear genes. The mean divergence times of almost all the major lineages leading to the extant decapodiform taxa were estimated to have occurred in the Paleozoic in the decapodiform topology presented (Fig. 2; Table 1). The divergence of the Spirulidae and Sepiidae is the one exception, estimated to have occurred in the Mesozoic for the decapodiform topology using first and second codon positions only. Some diversification within the Oegopsida already appears to have occurred around the Paleozoic–Mesozoic boundary (251 Ma). Such an ancient diversification of the major decapodiform lineages may have contributed to obscuring phylogenetic relationships within this group with both morphological and molecular characters becoming saturated over the last 300 million years. In contrast to the Decapodiformes, divergences among most taxa within the Octopodiformes were estimated to have occurred much more recently. The divergence of Vampyromorpha and the Octopoda was estimated to have occurred in the upper Paleozoic while the origins of the rest of the major lineages of Octopoda were estimated to have occurred in the Mesozoic. The

246


divergences of a number of lineages, including the ctenglossans, were estimated to have occurred close to the Mesozoic–Cenozoic boundary (66 Ma), while the PL analysis estimated the origin of the ctenoglossan lineage to be slightly younger (48.5 ± 7.5 Ma) (21). It is possible that the divergence of these lineages may correspond to the extinction of the ammonoids and/or the end of the Cretaceous “oceanic anoxic event” (~93 Ma), with mesopelagic depths only habitable after this time (15). The estimated divergence times within the Octopodiformes were found to be slightly older (but not significantly different) when nuclear genes only were used in the Bayesian analyses.

Acknowledgment J.S. is supported by a Natural Environment Research Council Antarctic Funding Initiative grant awarded to L.A and a Lloyd’s Tercentenary Fellowship.

References 1. M. J. Sweeney, C. F. E. Roper, Smithson. Contrib. Zool. 586, 561 (1998). 2. K. M. Mangold, R. E. Young, Smithson. Contrib. Zool. 586, 21 (1998). 3. R. E. Young, M. Vecchione, K. M. Mangold. Cephalopoda Cuvier 1797. Octopods, Squids, Nautiluses, etc. Version 01, http://tolweb.org/Cephalopoda (accessed on 2007). 4. T. Berthold, T. Engeser, Ver Naturwissenschaftliche Vereins Hamburg 29, 187 (1987). 5. R. E. Young, M. Vecchione. Amer. Malac. Bull. 12, 91 (1996). 6. A. Appellöf, Abhandlungen hrsg. Von der Senckenbergischen Naturorschenden Gesellschaft 24, 570 (1898).

7. R. E. Young, M, Vecchione, Bull. Am. Malacol. Union 12, 91 (1996). 8. D. B. Carlini, J. E. Graves, Bull. Mar. Sci. 64, 57 (1999). 9. A. R. Lindgren, G. Giribet, M. K. Nishiguchi, Cladistics 20, 454 (2004). 10. J. Strugnell, M. Norman, J. Jackson, A. J. Drummond, A. Cooper, Mol. Phylogenet. Evol. 37, 426 (2005). 11. J. Strugnell, J. Jackson, A. J. Drummond, A. Cooper, Cladistics 22, 89 (2006). 12. A. Naef, Fauna and Flora of the Bay of Naples (Keter Press, Jerusalem, 1921–1923). 13. C. F. E. Roper, R. E. Young, G. L. Voss, Smithson. Contrib. Zool. 13, 1 (1969). 14. A. R. Lindgren, M. Daly, Cladistics 23, 464 (2007). 15. R. E. Young, M. Vecchione, D. T. Donovan, S. Afr. J. Mar. Sci. (1998). 16. F. E. Anderson, Zool. J. Linn. Soc. 130, 603 (2000). 17. A. R. Lindgren, E. Amezquita, O. Katugin, M. K. Nishiguchi, Mol. Phylogenet. Evol. 36, 101 (2005). 18. S. V. Boletzky, Rev. Zool. 99, 755 (1992). 19. G. E. Pickford, Bull. Inst. Oceanographiue, Monaco 777, 1(1939). 20. D. B. Carlini, K. S. Reece, J. E. Graves, Mol. Biol. Evol. 17, 1353 (2000). 21. J. M. Strugnell, M. Norman, A. J. Drummond, A. Cooper, Curr. Biol. R300 (2004). 22. L. Bonnaud, R. Boucher-Rodoni, M. Monnerott, Mol. Phylogenet. Evol. 7, 44 (1997). 23. G. Grimpe, Zool. Anz. 52, 297 (1921). 24. J. R. Voight, J. Moll. Stud. 63, 311 (1997). 25. S. B. Piertney, C. Hudelot, F. G. Hochberg, Zool. J. Linn. Soc. 119, 348 (2003). 26. M. A. Collins, R. Villanueva, Oceanogr. Mar. Biol. 44, 277 (2006). 27. D. B. Carlini, R. E. Young, M. Vecchione. Mol. Phylogenet. Evol. 21, 388 (2001). 28. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002).

Nematodes (Nematoda)

The Nematoda is a phylum of noncoelomate invertebrates with ~23,000 described species. Traditional taxonomies recognize ~250 families, but these schemes are undergoing substantial revision based on molecular phylogenetic analyses. Nematode fossils are rare, and mostly uninformative as to deeper relationships. While the phylum has an origin in the Precambrian (>543 million years ago, Ma), and the three extant recognized classes likely diverged very early, few dates have been estimated for nematode divergences. The caenorhabditids are estimated to have diverged from the vertebrate-parasitic Strongyloidea ~380 Ma, and their divergence from the parasitic Ascaridomorpha is dated at ~540 Ma.

Nematoda in several different places within Metazoa, but recent molecular analyses are converging upon a placement within the Superphylum Ecdysozoa (3). This conflicts with opinions based on morphological characters such as the possession of a true coelom, and of metameric segmentation, but is supported by features of moulting and ciliated epithelia. The internal phylogeny of the Nematoda has been much debated also, and molecular data are being brought to bear (4–6). While no complete system has yet been devised, there are some general areas of agreement between different studies that also have morphological support (7, 8). The Nematoda has three classes: Enoplia, Dorylaimia, and Chromadoria. The relative branching order of these has not been determined. All three classes contain parasitic groups, and whole-phylum analyses suggest that animal and plant parasitisms have arisen independently at least six and three times, respectively (9). Within the Chromadoria, the free-living model species Caenorhabditis elegans is part of a radiation of terrestrial

The phylum Nematoda encompasses small and abundant organisms (Fig. 1) that inhabit most of the available habitats on Earth. A famous image derived from a pioneering, evangelical nematologist, Nathan Cobb, suggested that if all the other matter of the Earth, apart from nematodes, were taken away, it would still be possible to make out the shape of the planet, and of most organisms on it because of the abundance and ubiquity of the group. Commonly known as roundworms or eelworms, they largely escape notice except when they cause damage to humans, domestic animals, and agricultural crops; many species are successful parasites. Over 23,000 species have been described from just over 3000 genera in 250 families, but estimates of true nematode diversity range from 100,000 to 10 million species (1, 2). Nematodes are defined by their generally cylindrical body shape (secondarily modified in some parasites), a collagenous cuticle that is moulted four times in postembryonic development, and a series of anterior sensory organs that have a particular hexaradially symmetrical distribution (Fig. 1). Here I review the divergence times and evolutionary relationships of selected taxa within the Nematoda. Traditional systematics have placed the

Fig. 1 Scanning electron micrograph of the head of a soildwelling cephalobomorph nematode (Acrobeles complexus). Credit: M. Mundo-Ocampo.

Mark Blaxter Institute of Evolutionary Biology, School of Biological Sciences, University of Edinburgh, Ashworth Laboratories, King’s Buildings, Edinburgh EH9 3JT, UK ([email protected])

Abstract

M. Blaxter. Nematodes (Nematoda). Pp. 247–250 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Heligmonellidae

2

Rhabditidae

1

Anisakidae 4

Cm

O

S

D

C

P

PALEOZOIC 500

400

Ascaridae

J

Tr

K

Pg

MESOZOIC 300

200

Spirurina Rhabditida

Trichostrongylidae 3

Rhabditoidea

THE TIMETREE OF LIFE Ascaridomorpha Strongylida

248

CZ

100

0 Million years ago

Fig. 2 A timetree of Nematoda. Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), Cm (Cambrian), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), O (Ordovician), P (Permian), Pg (Paleogene), S (Silurian), and Tr (Triassic).

microbivores that also includes animal-parasitic clades. Also in the Chromadoria are the Tylenchina (including plant and animal parasites), the Spiruromorpha which are all parasitic (10) and a paraphyletic group of freeliving marine and terrestrial taxa, the “chromadorids” (4). Within the Dorylaimia are marine and terrestrial microbivores, plant parasites, and animal parasites. The Enoplia includes marine and terrestrial microbivores and predators, and plant parasites. While the largest nematodes (parasites of the cetacean gut or urinary systems) are meters in length, most of the members of this soft-bodied taxon are less than a millimeter, and thus fossilize poorly. Two sources of fossil nematodes have been found. One is in ambers from the Dominican Republic (20–15 Ma; Neogene), Mexico (26 Ma; late Paleogene), the Baltic (~40 Ma; Paleogene), and Myanmar (~110–100 Ma; Cretaceous). The other is in coprolites of subfossil (~1 Ma; Quaternary) and early Cretaceous (~130 Ma) age. Fossils from amber (11) have been diagnosed as being from representatives of insect parasites from the Mermithida (12–15), and Tylenchina (16–18), and from free-living plectids and diplogasterids (16, 19). Fossils from coprolites are often poorly preserved, and the only robust identification has been of eggs with morphology typical of the Ascaridomorpha in dinosaur coprolites from Belgian strata (20). Molecular analyses of nematode phylogeny using both nuclear and mitochondrial genes have been published, but few have attempted to place estimates of divergence times on internal nodes. One feature that is evident from many analyses, particularly those using nuclear ribosomal RNAs, is that of extremes of evolutionary rate heterogeneity in different lineages. Thus the lineage that includes the model C. elegans is characterized by

an elevated rate of substitution compared to its close relatives (5). Despite these caveats, four studies have placed time estimates on the nematode tree. In the publication describing the sequencing of the genome of Caenorhabditis briggsae (21), Avril Coghlan generated an estimate of 110–80 Ma for the divergence time of this species from the model C. elegans. This estimate used 338 sets of protein-coding orthologous genes that appeared not to display elevated rates in the nematodes compared with humans and the mosquito Anopheles gambiae, and they assumed a nematode–arthropod divergence date of 1000–800 Ma. While the date of nematode–arthropod divergence is obviously open to debate, these estimates compare favorably with earlier ones (22). When a nematode–arthropod divergence of 800 Ma was assumed, the 95% confidence interval for the C. briggsae–C. elegans divergence was 90–78 Ma. More recently, Asher Cutter (23) has used estimates of neutral mutational accumulation to reassess this estimate and has presented analyses that suggest a much more recent divergence (100 Ma made in 1997 (25) had no justification other than that the split must predate between species within the genus Caenorhabditis. These estimates

Eukaryota; Metazoa; Nematoda

Table 1. Divergence times (Ma) among Nematoda, from ref. (28). Timetree Node

Time

1

541

2

383

3

235

4

198

all conflict with that made by Vanfleteren et al. (28) (see later) for the divergence of C. elegans from another group in the Rhabditomorpha, and need close revision. The only other molecular dating study used cytochrome c and nematode-type globin sequences to date the divergences of a number of parasitic groups from each other and from C. elegans (28). In this study, the unit evolutionary periods (the time represented by a 1% sequence divergence at the protein level) for cytochrome c (21 Ma) and globin (5 Ma) were derived from other, mainly vertebrate taxa, and it is thus unclear how accurate these estimates are. However, using the same method to estimate the C. elegans–C. briggsae divergence yields 112 Ma (Blaxter, unpublished observations), suggesting that if the estimates for the same divergence made using multiple genes or mutation accumulation experiments (23) are more correct (21), then the globin-dated divergences may be overestimates, particularly in the rapidly evolving chromadorid clades. Using cytochrome c or globin, this method yielded estimates of 1200–950 Ma for the divergence of Nematoda from other phyla tested (Arthropoda, Echinodermata, and Chordata). Within Nematoda, globin comparisons yielded a 383 ± 8 Ma (i.e., Devonian) divergence for C. elegans and two families of strongylid parasites (Trichostrongylidae and Heligmonemellidae). While traditional systems place Strongylida at ordinal rank, “Strongylida” in molecular analyses is deeply nested within the traditional Order “Rhabditida”: both of these families are now considered members of the Superfamily Strongyloidea within Infraorder Rhabditomorpha (8). Within the Ascaridomorpha, globin sequences were used to give an estimate of 198 ± 42 Ma (i.e., Jurassic) divergence for members of the Ascarididae and Anisakidae. The estimate for the divergence of Rhabditomorpha and Ascaridomorpha was 541 ± 11 Ma (i.e., Cambrian). The paucity of informative fossils, and the related issues of identification of molecular markers with

249

verifiable clocklike behaviors, limits knowledge of the dates of divergences of a major component of the biota, the Nematoda. Other routes to dating using external mark points such as examining the congruence of host and parasite phylogenies to transfer robust dates from the hosts (e.g., vertebrates) have not been successful, because of rampant lateral capture of hosts by parasite groups (29). However, the emerging genome sequence data sets available for a wide phylogenetic range of nematodes, and the extensive EST data sets being generated for parasitic species (30), should yield additional proteincoding genes with which analyses aimed at dating nematode radiations can be performed. These will have to be analyzed with caution, as the most extensive comparisons performed to date, between members of the genus Caenorhabditis (21, 23), have yielded date estimates that vary by several fold.

Acknowledgments I would like to thank all my colleagues in the nematode Tree of Life project (NemaToL) and G. Poinar for helpful discussions.

References 1. P. J. D. Lambshead, G. Boucher, J. Biogeogr. 30, 475 (2003). 2. P. J. D. Lambshead et al., Mar. Ecol. Prog. Ser. 194, 159 (2000). 3. A. M. A. Aguinaldo et al., Nature 387, 489 (1997). 4. B. H. Meldal et al., Mol. Phylogenet. Evol. 42, 622 (2007). 5. M. L. Blaxter et al., Nature 392, 71 (1998). 6. M. Holterman et al., Mol. Biol. Evol. 23, 1792 (2006). 7. P. De Ley and M. Blaxter, in Nematology Monographs and Perspectives, R. Cook, D. J. Hunt, Eds. (E. J. Brill, Leiden, 2004), Vol. 2, pp. 633–653. 8. P. De Ley and M.L. Blaxter, in The Biology of Nematodes, D. Lee, Ed. (Taylor & Francis, London, 2002), pp. 1–30. 9. M. Dorris, P. De Ley, and M. L. Blaxter, Parasitol. Today 15, 188 (1999). 10. S. A. Nadler et al., Parasitology 134, 1421 (2007). 11. G. Poinar, Jr., Life in Amber (Stanford University Press, Stanford, USA, 1992). 12. G. Poinar, Jr., Nematology 3, 753 (2001). 13. G. Poinar, Jr., Parasitology 125, 457 (2002). 14. G. Poinar, Jr., J. P. Lachaud, A. Castillo et al., J. Invertebr. Pathol. 91, 19 (2006). 15. G. Poinar, Jr. and R. Buckley, J. Invertebr. Pathol. 93, 36 (2006). 16. G. O. Poinar, The Natural History of Nematodes (Prentice Hall, Englewood Cliffs, New Jersey, 1983).

250

17. 18. 19. 20. 21. 22. 23. 24.


G. Poinar, Jr., J. Parasitology 70, 306 (1984). G. Poinar, Jr., Parasitology 127, 589 (2003). G. Poinar, Jr., Nematologica 23, 232 (1977). G. Poinar, Jr. and A. J. Boucot, Parasitology 133, 245 (2006). L. D. Stein et al., PLoS Biol. 1, E45 (2003). A. Coghlan, K. H. Wolfe, Genome Res. 12, 857 (2002). A. D. Cutter, Mol. Biol. Evol. (2008). D. R. Denver et al., Science 289, 2342 (2000).

25. A. Etzinger, R. Sommer, Science 278, 452 (1997). 26. A. Pires-daSilva, R. J. Sommer, Genes Dev. 18, 1198 (2004). 27. R. L. Hong, R. J. Sommer, Bioessays 28, 651 (2006). 28. J. R. Vanfleteren et al., Mol. Phylogenet. Evol. 3, 92 (1994). 29. M. L. Blaxter, Adv. Parasitol. 54, 101 (2003). 30. J. Parkinson et al., Nat. Genet. 36, 1259 (2004).

Arthropods (Arthropoda) Davide Pisani Laboratory of Evolutionary Biology, Department of Biology, The National University of Ireland, Maynooth, Co. Kildare, Ireland (davide. [email protected])

Abstract Living arthropods comprise more than 1 million species and represent the majority of the Earth’s animal diversity. This phylum of animals includes one extinct subphylum (Trilobita) and four living subphyla: Myriapoda (e.g., centipedes and millipedes), Chelicerata (e.g., spiders, mites, and horseshoe crabs), Hexapoda (e.g., insects), and Crustacea (e.g., shrimps and crabs). The relationships of the subphyla and classes remain uncertain, although some consensus is emerging for the relationships among hexapods and crustaceans. The current evidence suggests that arthropods originated before 630 million years ago (Ma), but that the divergences leading to the currently recognized subphyla occurred in the Ediacaran and the Cambrian (630–488 Ma).

With more than 1 million described species (1), arthropods are a marvelous evolutionary success. This phylum of animals includes one extinct subphylum (Trilobita, ~4000 described species) and four living subphyla: Myriapoda (e.g., centipedes and millipedes, ~11,500 species), Chelicerata (e.g., spiders, mites, and horseshoe crabs, ~70,000 species), Hexapoda (e.g., insects, ~948,000 species), and Crustacea (e.g., shrimps and crabs, ~68,000 species). The Phylum Arthropoda is characterized by species having an articulated chitinous exoskeleton divided into thick areas, corresponding to segments, joined by thin “rings” (2). Each segment typically carries a pair of articulated legs, and a number of anterior segments are fused into a cephalon. In most groups the cephalon carries eyes, and two or more pairs of limbs (2). Arthropods have been treated as a single animal group essentially since 1753, when Linnaeus called them Insecta in the first edition of the Systema Naturae (3). However, their monophyly has long been debated because of their morphological disparity (4–6). Molecular data eventually confirmed arthropod monophyly (7) and to date arthropods are classified into one extinct subphylum (the Trilobita) and four extant subphyla: Crustacea,

Hexapoda, Myriapoda, and Chelicerata (Fig. 1). Each subphylum is further divided into classes (8). Here, I will review the relationships and divergence times among the arthropod classes and subphyla. The higher-level relationships among arthropods are still uncertain. Traditionally, Hexapoda (insects and their allies) and Myriapoda (centipedes, millipedes, and their allies) were joined in the group Atelocerata. The Crustacea (crabs, shrimps, and their allies) were joined with Atelocerata in the group Mandibulata (9). Finally, Chelicerata (spiders, horseshoe crabs and their allies) were considered the closest relative of Mandibulata. Molecular phylogenetics revolutionized this scenario, supporting a derivation of Hexapoda from within a paraphyletic Crustacea (7, 10–13). The group composed of Hexapoda and Crustacea was named Pancrustacea

Fig. 1 A crab (Metopaulias depressus; Decapoda) from Jamaica (top); and a millipede (Anadenobolus arboreus; Diplopoda) from Puerto Rico (bottom). Credit: A. Sanchez.

D. Pisani. Arthropods (Arthropoda). Pp. 251–254 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

252

THE TIMETREE OF LIFE Ostracoda 11

Cirripeda Cephalocarida

9

Remipedia

7 10

Chilopoda 6

Symphyla Diplopoda

2

Picnogonida 3

Arachnida 8

Xiphosura

Paleozoic

Np PR 700

Mesozoic

Myriochelata

Branchiopoda Myriapoda

1

Hexapoda

Chelicerata

5

Pancrustacea

Malacostraca

4

Cz

PHANEROZOIC 600

500

400

300

200

100

0 Million years ago

Fig. 2 A timetree of arthopods. The divergence times are from Table 1. This figure assumes the Myriochelata hypothesis, but see text for caveats. Abbreviations: Cz (Cenozoic), Np (Neoproterozoic), and PR (Proterozoic).

(14), and Mandibulata was redefined as the group joining Myriapoda and Pancrustacea. To date, Pancrustacea is arguably the best supported group of arthropod subphyla (7, 10–13, 15–24). Molecular phylogenetics also questioned the validity of Mandibulata as many independent molecular studies recovered Myriapoda as the closest relative of Chelicerata. The group joining Myriapoda and Chelicerata was named Myriochelata by Pisani et al. (21) and Paradoxopoda by Mallat et al. (18). Support for the Myriochelata hypothesis has been found in mitochondrial genome analyses (16, 19, 25, 26), from the analysis of a concatenation of nine nuclear and 15 mitochondrial genes (21), and from the analysis of Hox genes (17). Analyses of 18S rRNA and 28S rRNA sequences have not provided conclusive support for this group (24). However, the largest sequence analysis thus far, including 40 Mb of expressed sequence tags and 21 animal phyla, supported Myriochelata (36). Putative derived characters of Myriochelata have been proposed (25, 27), but Mandibulata is generally favored in combined analyses of molecular and morphological data (23). Furthermore, concerns that Myriochelata may be a long-branch attraction artifact remain (20), and the Myriochelata vs. Mandibulata controversy cannot be considered settled.

Relationships within the subphyla of arthropods are also uncertain. Within Pancrustacea, the monophyly of the Class Hexapoda is well supported (24) and there is evidence that the closest relative of Hexapoda is most likely the Branchiopoda, the brine shrimps and their allies (22, 24). However, the relationships among the remaining classes of Pancrustacea remain uncertain. Similarly, the relationships among the myriapod classes are not yet resolved (22). The higher-level relationships within Chelicerata are better established, with Xiphosura (the horseshoe crabs) and the Arachnida (spiders, mites and their allies) forming a monophyletic group (20), closest to the Pycnogonida (sea spiders; 22). The arthropod fossil record is rich, but the tempo of early arthropod evolution is unclear. The earliest fossil arthropods are trilobites and date back to the Atdabanian stage (523–519 Ma) of the early Cambrian (28). However, the biogeographic distribution of the earliest trilobites suggests that diversification within Arthropoda must have predated the breakup of the late Neoproterozoic supercontinent Pannotia (~600–550 Ma; 29). If this was true, the earliest arthropod history must be unrecorded in the fossil record (30). Molecular clocks could be used to test the hypothesis that the earliest evolutionary history of Arthropoda was not recorded in the fossil record. However, only two

Eukaryota; Metazoa; Arthropoda

253

Table 1. Divergence times (Ma) and their credibility/confidence intervals (CI) among arthropods. Timetree Node

Estimates Ref. (21)

Time

Ref. (22)

Time

CI

Time

CI

Ref. (31)

Ref. (32)

Time

CI

1

698.5

725.0

825–634

672.0

732–612

–

–

2

642.0

642.0

765–519

–

–

–

–

3

632.0

–

–

632.0

685–573

–

–

4

587.0

–

–

587.0

634–540

–

–

5

562.0

640.0

779–569

546.0

593–499

500.0

532–508

6

524.0

442.0

540–344

606.0

666–545

–

–

7

521.0

–

–

521.0

572–470

–

–

8

510.5

475.0

578–372

546.0

593–499

–

–

9

471.0

–

–

471.0

521–421

–

–

10

470.0

–

–

–

–

470.0

434–421

11

459.0

–

–

459.0

510–408

–

–

Note: Node times in the timetree represent the mean of time estimates from different studies.

studies investigated the earliest evolutionary history of this phylum using molecular clocks (21, 22), although some pertinent divergence time estimates are available from two other studies (31, 32) (Fig. 2, Table 1). Pisani et al. (21) used 61 genes to derive divergence times under a variety of global and relaxed molecular clock methods in a limited number of taxa (seven). In contrast, Regier et al. (22) used fewer genes (three) but included a larger number of taxa (17). Despite these differences, the two studies obtained similar results (22), which are also similar to those reported in Gaunt and Miles (32). However, estimates from Aris-Brosou and Yang (31) differ significantly from other studies. They used a variable number of mitochondrial and nuclear genes (1–14 depending on the clade) to estimate divergence times among several animal groups (including four of those in Fig. 2) in an approach where the rate variation among lineages was modeled as an Ornstein–Uhlenbeck process. Their results have been debated, primarily because they reported divergence times that were significantly younger than some corresponding fossil-based divergences (33, 34), and hence they cannot be correct. Only divergence times among arthropod groups inferred by Aris-Brosou and Yang (31) that did not postdate the corresponding fossil-based divergence time were included in Table 1. The timetree of arthropod evolution (Fig. 2) indicates that Arthropoda originated in pre-Ediacaran times (~698 Ma), but that diversification of the major lineages occurred during the Ediacaran and the Cambrian

(630–488 Ma). These results are mostly in agreement with the arthropod fossil record, and with what is known of the latest Neoproterozoic biogeography (29). Indeed, the only major incongruence between the fossil record and the molecular clock-based divergence times here reported is represented by the pre-Ediacaran (~698 Ma; see Table 1 and Fig. 2) divergence between Myriochelata and Pancrustacea. However, the results presented here reject the Cambrian explosion hypothesis, in which animal Phyla are assumed to originate in the lower Cambrian, or very close to Ediacaran–Cambrian (542 Ma) boundary (35). The great arthropod radiation was essentially completed by ~459 Ma, well before the end of the Ordovician.

Acknowledgments I thank J. Cotton, A. Blumlein, and an anonymous reviewer for comments and suggestions.

References 1. R. C. Brusca, Ann. Mo. Bot. Gard. 87, 13 (2000). 2. C. Nielsen, Animal Evolution: Interrelationships of the Living Phyla, 2nd ed. (Oxford University Press, Oxford, UK, 2001). 3. C. Nielsen, in Arthropod Relationships, R. A. Fortey, R. H. Thomas, Eds. (Chapman & Hall, London, 1998), pp. 11. 4. G. Fryer, in Arthropod Relationships, R. A. Fortey, R. H. Thomas, Eds. (Chapman & Hall, London, 1998), pp. 23.

254


5. S. M. Manton, The Arthropoda (Clarendon Press, Oxford, 1977). 6. F. R. Schram, S. Koenemann, in Assembling the Tree of Life, J. Cracraft, M. Donoghue, Eds. (Oxford University Press, New York, 2004), pp. 319–329. 7. K. G. Field et al., Science 239, 748 (1988). 8. R. C. Brusca, G. J. Brusca, The Invertebrates, 2nd ed. (Sinauer Associates, Sunderland, 2003). 9. R. E. Snodgrass, Principles of Insect Morphology (McGraw-Hill, New York, 1935). 10. J. M. Turbeville, D. M. Pfeifer, K. G. Field, R. A. Raff, Mol. Biol. Evol. 8, 669 (1991). 11. J. W. Ballard et al., Science 258, 1345 (1992). 12. M. Friedrich, D. Tautz, Nature 376, 165 (1995). 13. J. L. Boore, T. M. Collins, D. Stanton, L. L. Daehler, W. M. Brown, Nature 376, 163 (1995). 14. J. Zrzavy, P. Stys, J. Evol. Biol. 10, 353 (1997). 15. J. W. Shultz, J. C. Regier, Proc. Roy. Soc. Lond. B 267, 1011 (2000). 16. U. W. Hwang, M. Friedrich, D. Tautz, C. J. Park, W. Kim, Nature 413, 154 (2001). 17. C. E. Cook, M. L. Smith, M. J. Telford, A. Bastianello, M. Akam, Curr. Biol. 11, 759 (2001). 18. J. M. Mallatt, J. R. Garey, J. W. Shultz, Mol. Phylogenet. Evol. 31, 178 (2004). 19. F. Nardi et al., Science 299, 1887 (2003). 20. D. Pisani, Syst. Biol. 53, 978 (2004).

21. D. Pisani, L. L. Poling, M. Lyons-Weiler, S. B. Hedges, BMC Biol. 2, 1 (2004). 22. J. C. Regier, J. W. Shultz, R. E. Kambic, Proc. Roy. Soc. Lond. B 272, 395 (2005). 23. G. Giribet, S. Richter, G. D. Edgecombe, W. C. Wheeler, in Crustacean Issues 16: Crustacea and Arthropod Relationships. Festschrift for Frederick R. Schram, S. Koenemann, R. A. Jenner, Eds. (Taylor & Francis, Boca Raton, FL, 2005), pp. 307–352. 24. J. Mallatt, G. Giribet, Mol. Phylogenet. Evol. 40, 772 (2006). 25. E. Negrisolo, A. Minelli, G. Valle, Mol. Biol. Evol. 21, 770 (2004). 26. A. Hassanin, Mol. Phylogenet. Evol. 38, 100 (2006). 27. D. Kadner, A. Stollewerk, Dev. Genes. Evol. 214, 367 (2004). 28. F. Gradstein, J. Ogg, A. Smith, A Geologic Time Scale 2004 (Cambridge University Press, Cambridge, UK, 2004). 29. B. S. Lieberman, Interg. Comp. Biol. 43, 229 (2003). 30. R. A. Fortey, D. E. G. Briggs, M. A. Wills, Biol. J. Linn. Soc. 57, 13 (1996). 31. S. Aris-Brosou, Z. Yang, Mol. Biol. Evol. 20, 1947 (2003). 32. M. W. Gaunt, M. A. Miles, Mol. Biol. Evol. 19, 748 (2002). 33. J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 387 (2005). 34. J. J. Welch, E. Fontanillas, L. Bromham, Syst. Biol. 54, 672 (2005). 35. G. E. Budd, S. Jensen, Biol. Rev. 75, 253 (2000). 36. C. W. Dunn et al., Nature 452, 745 (2008).

Spiders (Araneae) Nadia A. Ayoub* and Cheryl Y. Hayashi Department of Biology, University of California, Riverside, CA 92521, USA *To whom correspondence should be addressed ([email protected])

Abstract Spiders (~40,000 sp.), Order Araneae, are members of the Class Arachnida and are defined by numerous sharedderived characters including the ability to synthesize and spin silk. The last few decades have produced a growing understanding of the relationships among spider families based primarily on phylogenetic analysis of morphological characters. Only a few higher-level molecular systematic studies have been conducted and these were limited in their taxonomic sampling. Nevertheless, molecular time estimates indicate that spider diversification is ancient and that many families radiated rapidly in the early Cretaceous (146–100 million years ago, Ma) and before.

Spiders (Araneae) constitute one of the most diverse orders of animals with greater than 39,000 described species, which are found worldwide in virtually all terrestrial habitats (1). They are members of the Class Arachnida, which also includes orders such as ticks and mites (Acari), scorpions (Scorpiones), and harvestmen or daddy long legs (Opiliones). Arachnid phylogeny is poorly understood, but the current consensus is that the closest relative of spiders is the Pedipalpi, which is a group of arachnids composed of whip-scorpions (Uropygi), tailless whip-scorpions (Amblypygi), and short-tailed whipscorpions (Schizomida) (2, 3). Monophyly of spiders is strongly supported by a number of shared-derived characters, including cheliceral venom glands, male pedipalpi modified for sperm transfer, lack of a trochanter-femur depressor muscle, and abdominal spinnerets and silk glands (2, 4). Spiders are arranged into two suborders: the Mesothelae (one family with 87 species), which retains significant traces of abdominal segmentation, and the Opisthothelae (107 families with 39,638 species), which has lost all traces of abdominal segmentation (5). Here,

we review relationships and divergence times among families of the highly diverse Opisthothelae. Most systematic studies of spiders at the family level have relied exclusively on morphological characters (reviewed in 6). These studies are often hindered by many spider taxa retaining ancestral characters and exhibiting high levels of convergence or parallelism (e.g., 5, 7, 8). Spiders are thought to have arisen in the Devonian (416–359 Ma) (9), and their antiquity contributes to these problems. Fossil representatives of many extant families have been found in the early to mid-Cretaceous, 146–100 Ma (10). Despite these issues, phylogenetic analyses over the last 30 years have dramatically improved our understanding of spider relationships. Within the Opisthothelae, spiders are divided into two major groups (5): the tarantulas and their kin (Mygalomorphae; 15 families with 2564 species), and the “true” spiders (Araneomorphae; 92 families with 37,074 species). Mygalomorphs retain numerous primitive characters and are, for the most part, large, stout-bodied, dispersal-limited spiders that occupy similar ecological niches (e.g., Fig. 1). In contrast, araneomorphs display numerous derived traits and a vast array of morphological and behavioral characters, which are reflected in the tremendous species diversity of this clade compared to mygalomorphs. Although there have been several

Fig. 1 A tarantula (Aphonopelma seemanni) from Costa Rica. Credit: M. Chappell.

N. A. Ayoub and C. Y. Hayashi. Spiders (Araneae). Pp. 255–259 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

256

THE TIMETREE OF LIFE Ctenizidae 20

Cyrtaucheniidae 22

Idiopidae Theraphosidae 25

15

Microstigmatidae 23

Nemesiidae

17

Ctenizidae 14

Hexathelidae 21

Actinopodidae

11

Mygalomorphae

Barychelidae

18

Orthopalpae

16

Migidae 9

Dipluridae Hexathelidae

Deinopidae

8

Araneidae 12

7

Theridiidae Zorocratidae 24

Pisauridae

2

Hypochilidae 5

Filistatidae

3

Diguetidae 6

D

C

P

Tr

350

300

J

Cretaceous

MESOZOIC

PALEOZOIC 400

Plectreuridae

250

200

150

RTA clade

Uloboridae 13

1

Orbiculariae

Mecicobothriidae

Araneomorphae

Atypidae

10

Entelegynae

Atypoidea

Antrodiaetidae 19

Haplogynae

4

Pg CZ

100

50

0 Million years ago

Fig. 2 A timetree of spiders (Araneae). Divergence times are from Table 1. Ctenizidae and Hexathelidae are shown in part; these families are typically found to be paraphyletic or polyphyletic (7, 25, 26). Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), P (Permian), Pg (Paleogene), and Tr (Triassic).

phylogenetic studies within and among particular araneomorph families, none has treated all families simultaneously. Instead, relationships have been reconstructed by piecing together evidence from separate analyses to create a consensus phylogeny for araneomorphs (4, 6). Thus, many putative clades are uncertain and poorly resolved. The Family Hypochilidae is the presumed closest relative of all other araneomorphs, which includes two commonly recognized clades: Haplogynae (17 families) and Entelegynae (72 families) (11, 12). Over half (39) of the

families in the Entelegynae are grouped into the “RTAclade” on the basis of possessing a knob on the male palpi termed the “retrolateral tibial apophysis” (4, 13, 14). Within Entelegynae, the Orbiculariae is a very species-rich and controversial clade (15). The Orbiculariae is composed of two superfamilies, the Deinopoidea (Deinopidae and Uloboridae, 320 species) and the Araneoidea (12 families with 11,075 species), both of which include orb-web weavers. Orb webs are wagon-wheel shaped aerial nets built from spoke-like radii

Eukaryota; Metazoa; Arthropoda; Arachnida; Araneae

257

Table 1. Divergence times (Ma) and their 95% confidence/credibility intervals (CI) among spiders (Araneae). Timetree Node


Time

Time

Ref. (26)(b)

CI

Time

Ref. (29)

CI

Time

CI

1

392.0

392

–

392

–

279

307–251

2

375.5

371

381–346

380

385–350

–

–

3

307.0

301

326–282

313

332–293

–

–

4

296.0

290

318–276

302

326–286

–

–

5

269.0

–

–

269

294–246

–

–

6

238.0

230

250–201

246

272–216

–

–

7

229.5

219

242–203

240

261–225

222

231–213

8

210.5

219

242–203

202

223–192

–

–

9

207.5

202

233–186

213

240–198

–

–

10

201.5

198

232–177

205

237–185

–

–

11

190.0

190

215–174

–

–

–

–

12

175.0

200

223–180

150

172–139

–

–

13

170.5

175

203–159

166

187–152

–

–

14

167.0

167

194–157

–

–

–

–

15

158.0

158

182–146

–

–

–

–

16

152.0

152

174–139

–

–

–

–

17

146.0

146

168–133

–

–

–

–

18

142.0

142

161–129

–

–

–

–

19

139.0

142

172–130

136

161–124

–

–

20

134.5

137

158–125

132

146–118

–

–

21

118.5

122

146–110

115

139–103

–

–

22

114.5

120

138–107

109

122–96

–

–

23

113.5

114

136–103

113

130–104

–

–

24

106.9

84.8

104–75

129

153–113

–

–

25

93.5

93.5

111–82

–

–

–

–

Note: Node times in the timetree are based on the MP topology (a) for mygalomorphs and the ML topology (b) for araneomorphs (26). Node times are averaged across estimates from the two topologies.

overlaid with sticky capture spirals. Both deinopoid and araneoid orb weavers spin architecturally similar webs but differ in the type of silk they use to construct the sticky capture spiral. This difference is part of the evidence for the traditional interpretation of the orb web as a spectacular example of convergent evolution (see 16, 17). However, phylogenetic analysis of morphological and behavioral characters associated with orb-web construction supports monophyly of Orbiculariae (e.g., 15, 18). In addition, Deinopoidea and Araneoidea share silk protein-coding genes not found in other groups of spiders, corroborating monophyly of Orbiculariae and the single origin of the orb web (19).

The Mygalomorphae has received considerably less phylogenetic attention than the Araneomorphae. However, Raven (20) performed a comprehensive taxonomic overview that laid the groundwork for mygalomorph systematics by defining 15 families and proposed a hypothesis of their relationships. A later phylogenetic analysis by Goloboff (7) of a subset of mygalomorph genera assessed monophyly of the families defined by Raven (20) and provided a revised hypothesis of mygalomorph relationships. The phylogenetic analysis indicated that a number of the families described in the taxonomic overview were paraphyletic (species sharing ancestral characteristics but not forming a single evolutionary group),

258


including Ctenizidae, Dipluridae, Cyrtaucheniidae, Nemesiidae, and Barychelidae. In addition, the two hypotheses of higher-level relationships differed considerably, although both recognized a few common clades: (1) a close relationship between Antrodiaetidae and Atypidae, (2) the Rastelloidina (Migidae, Actinopodidae, Ctenizidae, Idiopidae, Cyrtaucheniidae), and (3) the Theraphosoidina (Theraphosidae, Barychelidae, and Paratropididae). Traditional taxonomists grouped Mecicobothriidae with Antrodiaetidae and Atypidae into the Atypoidea, which was thought to be closest to the remaining mygalomorphs (21–23). Both Raven’s (20) taxonomic overview and Goloboff ’s (7) phylogenetic analysis removed mecicobothriids from this clade, but the latter study recovered the grouping of Antrodiaetidae and Atypidae as the closest relative of all other mygalomorphs. Mecicobothriidae was found to be closest to a clade of the remaining families, which was referred to as the Orthopalpae. Hexathelids and diplurids were considered to form a paraphyletic grade at the base of Orthopalpae (7). Molecular studies of family-level relationships among spiders are sparse. A ribosomal RNA gene (28S rRNA) was used to reconstruct relationships of eight araneomorph families (24). These 28S data weakly supported the monophyly of two haplogyne families but strongly supported monophyly of six entelegyne families, two RTA-clade families, and three araneoid families. In contrast to morphological results, the araneoids were closest to the RTA-clade representatives, rather than the deinopoid exemplar, but this relationship was not well supported. Another investigation used 18S and 28S rRNA genes from representatives of each mygalomorph family (80 genera sampled, 25). This study found support for monophyly of the traditional Atypoidea as well as monophyly of the remaining mygalomorphs, the Orthopalpae. Additionally, the rRNA genes positioned diplurids and hexathelids as a paraphyletic grade at the base of Orthopalpae and rejected monophyly of many of the same families that cladistic analysis of morphological characters had recovered as paraphyletic (7). The rRNA data found barychelids and theraphosids to be each other’s closest relatives, but did not support monophyly of Theraphosoidina or Rastelloidina. One molecular study sampled multiple representatives of both araneomorph and mygalomorph taxa (26). These authors used a nuclear protein-coding gene, elongation factor-1 gamma (EF-1γ), to reconstruct relationships among 14 mygalomorph families and 10 araneomorph families (Fig. 2). The EF-1γ study grouped Hypochilidae

with the three families in the Haploygnae, instead of closest to all araneomorphs, perhaps as a result of longbranch attraction at the base of the araneomorph clade (26). The six entelegyne families clustered with strong support. Orbiculariae was found to be monophyletic in one analysis of the EF-1γ data, but in another, the RTA-clade respresentatives were intermingled with the sampled araneoids (Araneidae and Theridiidae). Relationships among mygalomorph families based on EF-1γ were largely consistent with those based on rRNA (25). For instance, monophyly of both Atypoidea and Orthopalpae were well supported, and nemesiids, ctenizids, and hexathelids were paraphyletic. Unexpectedly, one of the hexathelid genera was placed with an actinopodid, a previously unproposed relationship. Sampling of other families was insufficient to assess their monophyly. Rastelloid monophyly was not supported, but could not be rejected by the EF-1γ data. Analyses of EF-1γ further found barychelids to be the closest relatives of theraphosids. While relationships within the Orthopalpae differed according to optimality criterion, the parsimony analysis placed hexathelids and diplurids at the base of the Orthopalpae in agreement with phylogenetic analysis of morphological (7) and rRNA (25) characters. The most comprehensive study to estimate divergence times among spider families used EF-1γ sequences, five fossil calibration points, and a nonparametric rate smoothing method (NPRS, 27) to account for rate heterogeneity (26). Due to discrepancies in some of the phylogenetic results, this study calculated estimates using both the maximum parsimony (MP) and the maximum likelihood (ML) trees (Fig. 2, Table 1). Divergence time estimates show that spider lineages are ancient. The split of araneomorphs from mygalomorphs was estimated at 392 Ma, the maximum constraint for this node based on the oldest known spider fossil (9). However, when the common ancestor of living mygalomorphs was fi xed to the age of the oldest mygalomorph fossil (240 Ma, 28), the divergence of mygalomorphs from araneomorphs was estimated at 340 Ma. The other age estimate for this node is based on hemocyanin protein sequences assuming a strict molecular clock and using a fossil calibration point for the divergence of Xiphosura (horseshoe crabs) from Arachnida (29). This estimate of 279 (±28) Ma, which is much younger than that determined from the EF-1γ data and spider fossil constraints, is likely biased by the sparse taxon sampling (three spider families) and the use of a single non-spider calibration point. Within Mygalomorphae, the primary split between atypoids and orthopalps was estimated to have

Eukaryota; Metazoa; Arthropoda; Arachnida; Araneae

occurred 326–276 Ma, in the late Carboniferous or early Permian. While the exact timing of divergences within Orthopalpae is unclear, most family-level diversification appears to have occurred between the late Jurassic and early Cretaceous, 170–100 Ma. Many mygalomorph families, such as Migidae and Idiopidae, have classic Gondwanan distributions (30). The EF-1γ divergence times are consistent with migids (194–157 Ma) predating the initial breakup of Gondwanaland 165–150 Ma, and the divergence of African and Australian idiopids (130–90 Ma, 26) with the opening of the southern South Atlantic Ocean approximately 135 Ma (31). Despite the antiquity of mygalomorph families, perhaps even more striking is that mygalomorph diversification is recent when compared to araneomorph diversification. The estimated appearance of Haplogynae 332–282 Ma is much older than the age suggested by the fossil record (~94 Ma, 32). Even the divergence 272–201 Ma of Diguetidae and Plectreuridae, which are thought to be each other’s closest relatives (33), dates to the Triassic. The entelegyne node (261–203 Ma) also dates to the Triassic. The estimate from the EF-1γ data for this node overlaps with the estimate from hemocyanins of 231–213 Ma (29). The orbicularian superfamilies diverged in the late Triassic or early Jurassic, which implies that the orb-web architecture minimally dates to this period and has subsequently been modified, and even lost, in some araneoids and deinopoids (15, 18). As predicted from the fossil record (10), diversification of extant spider families is ancient, dating to well before or during the Cretaceous. Ancient rapid radiations, no doubt, contribute to the difficulties in reconstructing relationships among spider families. More studies with denser taxonomic sampling that are based on a broader range of independent loci and fossil calibrations than are currently available are needed to accurately reconstruct the time line of spider evolution.

Acknowledgments We thank J. Garb, J. Gatesy, and C. Griswold for comments on this chapter. Support was provided by U.S. National Science Foundation to C.Y.H.

References 1. N. I. Platnick, The World Spider Catalog, version 7.5. American Museum of Natural History, http://research. amnh.org/entomolog y/spiders/catalog/index.html (accessed 2007).

259

2. J. W. Schultz, Cladistics 6, 1 (1990). 3. G. Giribet, G. D. Edgecomb, W. C. Wheeler, C. Babbitt, Cladistics 18, 5 (2002). 4. J. A. Coddington, H. W. Levi, Ann. Rev. Ecol. Syst. 22, 565 (1991). 5. N. I. Platnick, W. J. Gertsch, Am. Mus. Nov. 2607, 1 (1976). 6. J. A. Coddington, in Spiders of North America: An Identification Manual, D. Ubick, P. Paquin, P. E. Cushing, V. Roth, Eds. (American Arachnological Society, 2005), pp. 18–24. 7. P. A. Goloboff, Am. Mus. Nov. 3056, 1 (1993). 8. J. E. Bond, B. D. Opell, Zool. J. Linn. Soc. Lond. 136, 487 (2002). 9. P. A. Selden, W. A. Shear, P. M. Bonamo, Palaeontology 34, 241 (1991). 10. D. Penney, C. P. Wheater, P. A. Selden, Evolution 57, 2599 (2003). 11. C. E. Griswold, J. A. Coddington, N. I. Platnick, R. F. Forster, J. Arachnol. 27, 53 (1999). 12. C. E. Griswold, M. J. Ramírez, J. Coddington, N. Platnick, Proc. Calif. Acad. Sci., 4th Series, 56 (Suppl. II), 1 (2005). 13. P. Sierwald, Nemouria, Occ. Pap. Delaware Mus. Nat. Hist. 35, 1 (1990). 14. C. E. Griswold, Smithson. Contrib. Zool. 539, 1 (1993). 15. J. A. Coddington, in Spiders: Webs, Behavior, and Evolution, W.A. Shear, Ed. (Stanford University Press, Stanford, 1986), pp. 319–363. 16. E. J. Kullman, Am. Zool. 12, 395 (1972). 17. W. G. Eberhard, Annu. Rev. Ecol. Syst. 21, 341 (1990). 18. C. E. Griswold, J. A. Coddington, G. Hormiga, N. Scharff, Zool. J. Linn. Soc. Lond. 123, 1 (1998). 19. J. E. Garb, T. DiMauro, V. Vo, C. Y. Hayashi, Science 312, 1762 (2006). 20. R. J. Raven, Bull. Am. Mus. Nat. Hist. 182, 1 (1985). 21. E. Simon, Histoire naturelle des araignees (Paris) 1, 1 (1892). 22. R. V. Chamberlin, W. Ivie, Ann. Entomol. Soc. Am. 38, 549 (1945). 23. W. J. Gertsch, American Spiders (Van Nostrand, New York, 1949). 24. B. Hausdorf, J. Evol. Biol. 12, 980 (1999). 25. M. Hedin, J. E. Bond, Mol. Phylogenet. Evol. 41, 454 (2006). 26. N. A. Ayoub, J. E. Garb, M. Hedin, C. Y. Hayashi, Mol. Phylogenet. Evol. 42, 394 (2007). 27. M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997). 28. P. A. Selden, J.-C. Gall, Palaeontology 35, 211 (1992). 29. A. Averdam, J. Markl, T. Burmester, Eur. J. Biochem. 270, 3432 (2003). 30. C. Griswold, J. Ledford, Occ. Pap. Calif. Acad. Sci. 151, 1 (2001). 31. I. Sanmartín, F. Ronquist, Syst. Biol. 53, 216 (2004). 32. D. Penney, Palaeontology 45, 709 (2002). 33. N. I. Platnick, J. A. Coddington, R. F. Forster, C. E. Griswold, Am. Mus. Nov. 3016, 1 (1991).

Holometabolous insects (Holometabola) Brian M. Wiegmann*, Jung-wook Kim, Michelle D. Trautwein Department of Entomology, North Carolina State University, Raleigh, NC 27695-7613, USA *To whom correspondence should be addressed ([email protected])

Abstract The Holometabola includes 11 orders that represent the vast majority of insect diversity (~850,000 species). Recent molecular and morphological treatments support holometabolan monophyly, confirm the monophyly of the major orders, and provide new evidence to place the orders in a phylogeny. Estimates of divergence time based on molecular evidence suggest an origin of the Holometabola within the Carboniferous, 359–299 million years ago, Ma, but definitive fossils first appear in the Permian, 299–280 Ma. The molecular timetree reveals striking parallel radiations of insect lineages throughout the Mesozoic (251–66 Ma).

The insect clade Holometabola (~850,000 species) includes 11 living orders that together comprise the vast majority of all insect diversity and therefore also represent a significant fraction (>60%) of all terrestrial animals (1). Holometabola includes the four largest orders of insects: Coleoptera (beetles, Fig. 1), Hymenoptera (bees, ants, and wasps), Diptera (true flies), and Lepidoptera (moths and butterflies), as well as the Neuroptera (lacewings), Megaloptera and Raphidioptera (dobsonflies and alderflies), Trichoptera (caddisflies), Mecoptera (scorpionflies), Siphonaptera (fleas), and Strepsiptera (twisted-wing insects). The name of the group reflects their defining characteristic—they undergo complete metamorphosis. Their life history is divided into discrete developmental stages, including a distinct larval (feeding) and pupal (quiescent) stage. The major developmental, morphological, and behavioral modifications that led to the holometabolous larva are thought to have arisen through extension of the prenymphal stage of hemimetabolous insects (2, 3). Metamorphosis from larval to adult morphology occurs in the pupal stage where the larval structures are broken down and adult features

(legs, wings, antennae, genitalia) then develop from specialized internal regions of subcuticular epidermal cells called imaginal discs (4). The larval cuticle is reduced or entirely lost and an adult cuticle is newly formed. The internal development of the wings is denoted in the other common name of the group, the Endopterygota. Despite its huge diversity, there are relatively few holometabolan lineages that contain exceptionally large numbers of species. Developmental specialization clearly played a major role in the expansion of holometabolan life histories, but the hyperdiversity of major clades of beetles, flies, moths, and wasps are most often attributed to independent, lineage-specific radiations enabled by unique combinations of trophic, life history, morphological adaptations, and the expansion of terrestrial plant communities (2, 5–9). For a more comprehensive perspective on insect diversity, fossil history, and evolution, see Grimaldi and Engel (2). Here, we review evidence on the phylogeny and divergence times of holometabolous insects. Phylogenetic classifications of Holometabola based on morphological features divide the group into two major subclades, the Neuropteroidea, which includes Coleoptera + the Neuropterida (Neuroptera, Megaloptera, Rhaphidioptera), and the Mecopterida (= Panorpida),

Fig. 1 A holometabolous beetle larva (Coleoptera: Chrysomelidae, Zygogramma sp.) from Arizona, USA. Photo credit: A. Wild.

B. M. Wiegmann, J. Kim, and M. D. Trautwein. Holometabolous insects (Holometabola). Pp. 260–263 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Diptera

5

Lepidoptera 10

Neuroptera 8

Rhaphidioptera 11

2

Megaloptera

4

Coleoptera

1

6

Strepsiptera

Neuropteroidea

Trichoptera

3

261

Mecopterida

Siphonaptera

7

Amphiesmenoptera

Mecoptera 9

Antliophora

Eukaryota; Metazoa; Arthropoda; Insecta; Holometabola

Hymenoptera Hemimetabolous outgroups

C

Tr

P

350

300

J

Cretaceous

Pg

MESOZOIC

PALEOZOIC 250

200

150

Ng

CZ 100

50

0 Million years ago

Fig. 2 A timetree of the holometabolous insects. Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), J (Jurassic), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

including Lepidoptera, Trichoptera, Diptera, Mecoptera, and Siphonaptera. Hymenoptera and Strepsiptera have been placed in various positions, the former often placed as closest to the Mecopterida and the latter traditionally placed either within or closest to Coleoptera (10, 11). The consensus view is that most morphological features of the Hymenoptera and the Strepsiptera are too highly modified to unequivocally resolve their phylogenetic positions. Additional widely accepted groupings are the Amphiesmenoptera (Lepidoptera + Trichoptera) (8) and the Antliophora (Diptera + Mecoptera including Siphonaptera) (12). Morphological evidence also supports a close relationship between the Mecoptera and the Siphonaptera (11). The most comprehensive review of the morphological evidence for holometabolan relationships is that of Kristensen (1999; 8) and was further evaluated in light of emerging alternative phylogenetic hypotheses by Beutel and Pohl (12). New perspectives on specific character systems such as sclerites, muscle insertions, and functional features of the wing-base (13), and mouthparts (14), as well as new paleontological findings and interpretations (2, 15) continue to add to the evidence on relationships. Molecular analyses of holometabolan phylogeny have primarily relied on 18S ribosomal DNA, and the results have been highly dependent on taxon sampling,

alignment, and methods of analysis. Previous molecular studies, most using rDNA, have recovered a monophyletic Neuropterida (16, 17), Neuropteroidea (16, 18, 19), Amphiesmenoptera (16, 17, 19–21), Mecopterida (20, 21) and, most provocatively, Halteria (Strepsiptera + Diptera) (17, 19). Two recent phylogenomic projects, with limited taxon sampling but large numbers of genes, addressed the placement of the Hymenoptera; mitochondrial genomes provided evidence that Hymenoptera is the closest relative of Mecopterida (22), while combined analysis of 185 nuclear genes strongly supports placement of the Hymenoptera as the earliest branching lineage, the closest relative of all other Holometabola (23). The most current molecular study by Wiegmann et al. (submitted) is the first to include both nuclear genes (AATS, CAD, TPI, SNF, PGD, and RNA POLII) and representative taxa from all holometabolan orders. Their findings support traditional morphological hypotheses (Neuropteroidea + Mecopterida including Amphiesmenoptera + Antliophora) and Savard et al.’s (23) early branching position for Hymenoptera. Additionally, multiple nuclear genes provide evidence for the placement of the enigmatic Order Strepsiptera as the closest relative of Coleoptera. These results add to the compounding and conflicting evidence for the placement of Strepsiptera—the most controversial issue

262


Table 1. Divergence times (Ma) and their credibility/ confidence intervals (CI) among holometabolous insects, based on Wiegmann et al. (submitted). Timetree Node

Time

CI

1

355

360–334

2

350

359–336

3

300

315–287

4

286

299–274

5

282

300–264

6

274

285–270

7

256

279–234

8

255

276–227

9

243

270–222

10

230

261–190

11

213

247–134

in holometabolan phylogenetics. Beutel (12) recently reviewed arguments surrounding the “Strepsiptera problem” (24). The original findings of Whiting and Wheeler (25) placed Strepsiptera as the closest relative of Diptera based on 18S rDNA and initiated a useful debate regarding empirical evidence for spurious grouping by “longbranch attraction” in molecular phylogenetics (26–30). Further studies supporting Halteria were based on 28S rDNA, 18S rDNA, and morphology; the most convincing morphological evidence being modifications of the wings into halteres, shared by both dipterans and strepsipterans, albeit on different thoracic segments (19, 27, 31). Several additional morphological and molecular studies reported evidence refuting the existence of Halteria (30, 32–34). The findings of Wiegmann et al. (submitted) supporting the close relationship between the Strepsiptera and the Coleoptera are robust and in agreement with traditional morphological hypotheses. The Holometabola is thought to have originated in the late Carboniferous (2, 35, 36), but definitive fossil evidence is lacking until the Permian (~280 Ma), a time when most of the extant orders had their origins (2, 37). An insect gall, presumed to be from a member of Holometabola, has been identified from the late Pennsylvanian, 302 Ma, that if accurately diagnosed provides the earliest physical evidence of their existence. A molecular analysis that relied on mitochondrial data (cox1) and maximum likelihood (ML) global and local molecular clocks to date the origin of the insects included

both dipterans and lepidopterans and found the origin of the taxon-limited Holometabola to be 351–338 Ma (38). Wiegmann et al. (submitted) estimated the divergence times of all holometabolan orders using a Bayesian phylogeny based on multiple nuclear genes, fossil calibrations, and relaxed clock Bayesian methods using the program Multidivtime (39). Congruent with the findings of Gaunt and Miles (38), multiple nuclear genes placed the origin of the Holometabola at 355 Ma, within the Carboniferous, but earlier than traditional estimates. The Hymenoptera, as the earliest branching lineage in the phylogeny, has an age of origin nearly equivalent with the age of the divergence of Holometabola from its closest relative (Fig. 2). This date is considerably older than existing fossil estimates, typical of molecular estimates (39). The split between the two major subclades Neuropteroidea and Mecopterida took place in the Permian 300 Ma, with the Amphiesmenoptera/Antliophora divergence occurring 284 Ma. The divergence of all orders (excluding the Hymenoptera) appears to have occurred in relatively rapid succession, with dates of origin falling in the range 274–213 Ma, with the earliest being the Coleoptera/ Strepsiptera divergence at 274 Ma and the most recent being the split of Rhaphidioptera and Megaloptera at 213 Ma. Though some estimates of ordinal-level divergences do not precisely correspond with traditional ages based on fossils, paleontological evidence is dramatically expanding, and thus, better fossil calibrations coupled with larger samples of genes and taxa as well as improved analytical methods should continue to refine divergence time estimates for the major holmetabolan clades. Molecular divergence time estimates and fossils agree that the Holometabola had its origins within the Paleozoic. The origination of the orders (excluding Hymenoptera) took place primarily within the Triassic with the primary split (Neuropterida + Mecopterida) occurring at the end of the Permian, and the remaining orders all appearing in the Jurassic. The huge hyperdiverse lineages of the Holometabola that contribute to the group’s reputation for evolutionary success (phytophagous and staphylinid beetles, apocritan wasps, cyclorrhaphan flies, and ditrysian Lepidoptera) may owe their species-richness to mid- and late Jurassic developments such as the radiation of angiosperms and the acquisition of specialized morphological innovations, such as a wasp-waist and the fly puparium (8, 40). Extreme diversity has made it difficult to resolve phylogenetic relationships among the major orders and conflicting lines of evidence continue to make holometabolan phylogeny one of the most challenging problems in insect phylogenetics.

Eukaryota; Metazoa; Arthropoda; Insecta; Holometabola

Acknowledgment

20.

Support was provided to the authors by the U. S. National Science Foundation.

21.

References 1.

2. 3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

P. M. Hammond, in Global Biodiversity: Status of the Earth’s Living Resources, B. Groombridge, Ed. (Chapman & Hall, London, 1992), pp. 17–39. D. Grimaldi, M. S. Engel, Evolution of the Insects (Cambridge, New York, 2005). J. W. Truman, L. W. Riddiford, Nature 401, 447 (1999). B. D. Farrell, Science 281, 553 (1998). H. F. Nijhout, Insect Hormones (Princeton University Press, Princeton, 1994). B. D. Farrell, C. Mitter, in Species Diversity in Ecological Communities. Historical and Geographical Perspectives, R. E. Ricklefs, D. Schluter, Eds. (University of Chicago Press, Chicago, 1993), pp. 253–266. C. Labandeira. Arthropod Syst. Phyl. 64, 53 (2006). N. P. Kristensen, Euro. J. Entomol. 96, 237 (1999). C. Mitter, B. D. Farrell, D. J. Futuyma, Trends Ecol. Evol. 6, 290 (1991). R. A. Crowson, Annu. Rev. Entomol. 5, 111 (1960). R. G. Beutel, S. Gorb, J. Zool. Syst. Evol. Res. 39, 177 (2001). R. G. Beutel, H. Pohl, Syst. Entomol. 31, 202 (2006). T. Hornschenmeyer, Zool. Scripta 31, 17 (2002). H. W. Krenn, Arthropod Syst. Phyl. 65, 7 (2007). D. E. Shcherbakov, Paleontol. J. 42, 15 (2008). K. M. Kjer, Syst. Biol. 53, 506 (2004). M. F. Whiting, Zool. Scripta 31, 3 (2002). D. Carmean, L. S. Kimsey, M. L. Berbee, Mol. Phylogenet. Evol. 1, 270 (1992). N. Chalwatzis, J. Hauf, Y. V. D. Peer, R. Kinzelbach, F. K. Zimmerman, Ann. Entomol. Soc. Am. 89, 788 (1996).

22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

263

D. P. Pashley, B. A. McPheron, E. A. Zimmer, Mol. Phylogenet. Evol. 2, 132 (1993). B. M. Wiegmann, J. C. Regier, C. Mitter, T. P. Friedlander, D. L. Wagner, E. S. Nielsen, Mol. Phylogenet. Evol. 15, 242 (2000). L. R. Castro, M. Dowton, Mol. Phylogenet. Evol. 34, 469 (2005). J. Savard et al., Genome Res. 16 1334 (2006). N. P. Kristensen, Annu. Rev. Entomol. 26, 135 (1981). M. F. Whiting, W. C. Wheeler, Nature 368, 696 (1994). D. Carmean, B. J. Crespi, Nature 373, 666 (1995). M. F. Whiting, J. C. Carpenter, W. C. Wheeler, Q. D. Wheeler, Syst. Biol. 46, 1 (1997). J. P. Huelsebeck, Syst. Biol. 46, 69 (1997). J. P. Huelsebeck, Syst. Biol. 47, 519 (1998). U. W. Hwang, W. Kim, D. Tautz, M. Friedrich, Mol. Phylogenet. Evol. 9, 470 (1998). W. C. Wheeler, M. Whiting, Q. D. Wheeler, J. M. Carpenter, Cladistics 17, 113 (2001). A. Rokas, J. Kathirithamby, P. W. H. Holland, Insect Mol. Biol. 8, 527 (1999). D. C. Hayward, J. W. H. Trueman, M. J. Bastiani, E. E. Ball, Dev. Genes Evol. 215, 213 (2005). F. Bonneton, F. G. Brunet, J. Kathirithamby, V. Laudet, Insect Mol. Biol. 15, 351 (2006). J. Kukalova Peck, Can. J. Earth Sci. 27, 459 (1990). C. C. Labandeira, T. L. Phillips, Proc. Natl. Acad. Sci. U.S.A. 93, 8470 (1996). W. Hennig, Insect Phylogeny (Wiley, New York, 1981). M. W. Gaunt, M. A. Miles, Mol. Biol. Evol. 19, 748 (2002). J. L. Thorne, H. Kishino, and I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). D. K.Yeates, R. Meier, B. M. Wiegmann, Entomol. Abhandl. 61, 170 (2003).

Bees, ants, and stinging wasps (Aculeata) Seán G. Bradya, Leah Larkinb, and Bryan N. Danforthc,*

Bees, ants, and stinging wasps comprise the clade Aculeata within the Order Hymenoptera. Molecular dating analyses within Aculeata have focused primarily on ants (Formicidae; ~12,000 species) and bees (Anthophila; ~20,000 species). Published molecular divergence times for ants differ considerably. The most recent study argues for a range of 135– 115 million years ago (Ma), consistent with the known fossil record for aculeates. Dating analyses of bees have focused primarily on families containing eusocial species. These studies have revealed that eusociality first evolved ~75 Ma in corbiculate bees and as recently as 20 Ma in several independent halictid taxa.

Bethylonymidae (1). This extinct family probably represents the closest relative of all modern aculeates (2). All three modern aculeate superfamilies have an extensive fossil record extending back to the early Cretaceous (140 Ma) (2). Early Cretaceous vespoid fossils exist for several families including Sierolomorphidae, Rhopalosomatidae, Vespidae, Scoliidae, and Tiphiidae, but are notably absent for Formicidae (ants). Extinct stem-group lineages to modern ants include Sphecomyrminae (~100–70 Ma) and potentially the more distantly related Armaniidae (~100–75 Ma). The fossil record of Apoidea likewise extends to the early Cretaceous (~140 Ma). Extinct stem-group lineages referred to collectively as Angarosphecidae (3, 4) are known from Barremian (140 Ma) up until the early Eocene (54–52 Ma) (5) from sites in Europe, South America, and Canada. Bennett and Engel (6) provide a recent synopsis of the non-bee apoid fossil record. There are no published estimates of molecular divergence times among aculeate families. In order to provide a rough molecular timescale for this group, we present analyses of 18S and 28S data obtained from GenBank that include representatives of all aculeate families for which these data are currently available (Table 1; Fig. 2).

Aculeate wasps are characterized by the modification of the ovipositor into a sting. All other Hymenoptera deposit their eggs through their ovipositor, while aculeates instead lay their eggs from the base of their sting, and this structure now serves to inject venom into prey and enemies. All members of Aculeata form a monophyletic lineage comprising three superfamilies: Chrysidoidea (seven families), Apoidea (11 families; includes bees and digger wasps; Fig. 1), and Vespoidea (10 families; includes spider wasps, hornets, and ants). Aculeata contains most major groups of eusocial insects, including social wasps, bees, and ants. We review the relationships and divergence times of Aculeata with particular reference to bees and ants, the two taxa that have been the focus of previously published molecular dating studies in this group. The oldest aculeate fossils are from late Jurassic (146 Ma) compression fossils from Central Asia placed in the

Fig. 1 A halictid bee (Agapostemon virescens) from New York, USA. Photo credit: A. Wild.

a

Department of Entomology and Laboratories of Analytical Biology, National Museum of Natural History, Smithsonian Institution, Washington, DC, 20560, USA; bDepartment of Biological Sciences, University of the Pacific, Stockton, CA 95211, USA; cDepartment of Entomology, Comstock Hall, Cornell University, Ithaca, NY 14853, USA *To whom correspondence should be addressed ([email protected])

Abstract

S. G. Brady, L. Larkin and B. N. Danforth. Bees, ants, and stinging wasps (Aculeata). Pp. 264–269 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Arthropoda, Insecta; Aculeata

265

Colletidae 21

Stenotritidae

19

Halictidae

17

Apidae 14

18

Megachilidae

Apoidea

Andrenidae

16

Melittidae s.s.

12 11

Dasypodaidae

7

Crabronidae Sphecidae Sapygidae 20

2

Mutillidae

15

Tiphiidae Vespidae 10

3

Scoliidae

8 1

Vespoidea

Pompilidae

5

Formicidae

6

Bradynobaenidae

4

Bethylidae 13

Chrysididae

9

Dryinidae

Jurassic

Early K

Late K

Paleogene

Ng

CENOZOIC

MESOZOIC 150

Chrysidoidea

Rhopalosomatidae

100

50

0 Million years ago

Fig. 2 A timetree of the bees, ants, and stinging wasps (Aculeata). Divergence times are shown in Table 1. Abbreviations: Ng (Neogene) and K (Cretaceous).

Our results suggest that the three aculeate superfamilies began to diversify in the middle Jurassic (~170 Ma). Many of the modern vespoid families diverged in the late Jurassic through the early Cretaceous, although a few families including Pompilidae, Mutillidae, and Sapygidae may have originated considerably more recently. Within Apoidea, the clade containing extant bee families arose at least 120–112 Ma, and by the end of the Cretaceous all modern bee families had evolved. Detailed phylogenetic and molecular divergence dating studies have been published on two groups of aculeates—ants and bees—and we discuss each of these taxa in turn. Ants form the monophyletic Family Formicidae within Vespoidea. All ants are eusocial and collectively occupy keystone positions in many environments

(7, 8). Most work on higher-level molecular systematics in ants has focused on one or several closely related subfamilies (9–16). In 2006, however, two larger phylogenies were published encompassing a much greater portion of ant diversity: Moreau et al. (17), which included ~4.5 kilobases (kb) from six genes and 19 subfamilies; and Brady et al. (18), which included ~6 kb from seven genes and 20 subfamilies. All available molecular evidence agrees on a robust group termed the formicoid clade (19) that unites 14 of the 20 subfamilies of ants, including several subfamilies containing species displaying presumptive primitive behaviors and morphologies. Bayesian analyses of these large data sets also suggest that Leptanillinae, a rarely encountered and poorly known ant subfamily (20), is the closest relative of all other ants in the arrangement

266


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among bees, ants, and stinging wasps (Aculeata). Timetree Node

Estimates

Time

This study (a)

This study (b)

Time

CI

Time

CI

166.1

181–152

173.9

190–159

1

170.0

2

163.5

162.1

176–149

164.9

178–152

3

159.9

158.6

173–146

161.2

174–149

4

154.1

152.1

168–136

156.1

171–142

5

153.2

153.7

168–140

152.8

165–141

6

146.7

142.6

159–126

150.8

167–136

7

140.6

139.5

158–121

141.7

159–125

8

139.9

136.5

153–121

143.2

159–128

9

135.7

136.7

176–84

134.8

180–74

10

129.2

120.5

147–88

137.9

157–117

11

127.8

130.0

148–112

125.6

142–111

12

116.0

120.3

138–103

111.7

126–98

13

111.7

111.7

162-59

–

–

14

109.3

112.1

131–94

106.5

121–92

15

104.2

91.8

138–44

116.6

155–69

16

100.4

103.0

121–85

97.8

113–83

17

92.3

96.4

116–77

88.2

103–74

18

91.4

94.2

114–76

88.5

107–69

19

83.0

88.4

109–68

77.7

92–63

20

72.7

50.1

97–12

95.2

143-45

21

69.6

75.3

102–49

64.0

82-46

Note: Node times in the timetree represent the mean of time estimates from different analyses. Dates were estimated with two independent data sets, 18S rRNA (a) and 28S rRNA (b), obtained from the public databases (GenBank) and aligned by hand. The 18S data set included 177 taxa and 770 aligned base pairs (bp) and spanned Symphyta, Apocrita, and Aculeata in the Hymenoptera. The 28S data set included 157 taxa and 1314 aligned base pair, and spanned Apocrita and Aculeata with a single symphytan species. The 18S data set was date-calibrated with 19 priors and 28S with 14 priors, and the data sets were analyzed using an uncorrelated lognormal relaxed clock and a GTR+G+I model of evolution in BEAST 1.4.6 (59). Phylogenetic relationships were constrained using information from the most recently published studies available (32, 60, 61), with the important caveat that some vespoid relationships are not robust (62) and may change with new data.

(Leptanillinae + (formicoids + poneroids)), but other methods of analysis as well as removal of long-branched outgroups demonstrate that alternative relationships at the base of the ant tree are also supported by the current data (18). Sampling of additional phylogenetic characters and taxa, as well as more sophisticated methods of dealing with potential phylogenetic artifacts, will be necessary to resolve the earliest relationships among ants with any degree of certainty. The first ants belonging to modern subfamilies appear in the fossil record ~100 Ma. Although Cretaceous ant

fossils are relatively scarce, representatives of several modern subfamilies including Formicinae, Dolichoderinae, and Myrmicinae (all within the formicoid clade) are identifiable from this time period (2, 21, 22). By the Eocene ants appeared to have diversified substantially, as indicated by the many modern genera found in Baltic amber (23, 24) and other deposits (25, 26). Crozier et al. (27) represents the first attempt to infer the age of ants using molecular data. Because this study preceded the development of phylogenetic relaxed clock methods, the authors instead used linear regression of


mitochondrial DNA pairwise distances, calibrated using the fossil Cariridris [which has since been transferred from Formicidae to Ampulicidae (1)]. Their analysis suggested a Jurassic origin for ants dated at 185 ± 36 Ma. Although this estimate may be too old, this paper was notable in challenging the notion that the clade of modern ants appeared concomitantly with their first appearance in the fossil record. Several subsequent dating studies also hinted that modern ants were somewhat older than indicated by their fossil record. Bayesian divergence dating (28) studies of army ants and relatives (dorylomorphs) (10) and bulldog ants (Myrmeciinae) (12) in which the ant node was given an a priori value consistently returned older a posteriori dates for that node. However, because these studies were designed to date nodes within specific groups of ants and possessed very limited taxon sampling outside these groups, the interpretation of this analytical behavior on the origin of ants is not clear. The two large-scale molecular ant phylogenies discussed previously (17, 18) have also provided the most comprehensive dating estimates currently available for ants. Although both studies used a similar, overlapping set of over 40 ant fossils as minimum-age calibrations and both relied primarily on the same analytical method of penalized likelihood (29), these two studies resulted in substantially different age estimates. Moreau et al. (17) inferred a range of dates for the antiquity of modern ants at 168–140 Ma, with this variation reported as being caused solely by minor (~5 Ma) alterations in the minimum-age calibrations from three recent fossil strata. Brady et al. (18) inferred a younger range of ~135–115 Ma for the origin of modern ants, and argued that these dates were more accurate for several reasons. Their analyses were not influenced by these minor differences in minimum calibration ages, which altered their estimates by only 0–2 million years. Instead, the range established by Brady et al. (18) was based on calibrating their basal outgroup node (all sampled Aculeata except Chrysidoidea) with lower and upper bounds using additional information from the entire aculeate fossil record. The range of ~135–115 Ma also accords with arguments based on the overall completeness of the fossil record that modern ants originated ~120 Ma (2). These estimates correspond roughly with our own estimates based on a far smaller sample of ant taxa and just two genes. Our data puts the common ancestor of the modern ants plus their older, now extinct relatives (including Sphecomyrminae) at 137–143 Ma (Table 1). Bees comprise a monophyletic group of ~20,000 species of Aculeata specialized on floral resources such as

267

pollen, nectar, and floral oils. Bees are currently the most important pollinators of angiosperm plants and may have played an important role in angiosperm diversification in the early to mid-Cretaceous. Based on the most recent study (30), bees form the closest relative of the spheciform wasp Family Crabronidae. Monophyly of bees is supported by 14 morphological, developmental, and behavioral characters (31) and molecular evidence (32). Bees are currently divided into seven families: Andrenidae, Colletidae, Halictidae, Melittidae and Stenotritidae (which together comprise the “shorttongued” bees), and Megachilidae and Apidae (which comprise the “long-tongued” bees). Family-level bee phylogeny has been analyzed based on morphology (33, 34) as well as combined morphological and molecular data (32, 35). While Colletidae has traditionally been viewed as the basal branch of bees, a new view is emerging in which the root node of bees falls near or within the Family Melittidae (32, 35–37). The antiquity of bees remains unclear. There are no published relaxed clock-dating analyses at the level of the bee families, but the fossil record of bees would appear to greatly underestimate their true antiquity. The oldest fossil of a member of the clade of living bees (Cretotrigona prisca from New Jersey amber) is closely related to extant stingless bee groups and, in fact, was placed in an extant genus (Trigona) when first described (38, 39). The antiquity of C. prisca is somewhat controversial (40). While initially presumed to be 80 Ma (38, 39), it has since been estimated to be 70 Ma (41) and 65 Ma (42) in age. Other important bee fossil deposits include the French (Oise) Eocene amber (~53 Ma), the Baltic amber (~42 Ma), and the Dominican amber (~23 Ma). The Baltic amber includes representatives of the extant families Apidae, Megachilidae, Melittidae, and Halictidae, and one extinct family closely related to Melittidae (Paleomelittidae) (43). Dominican amber deposits include representatives of five of the seven extant bee families (Stenotritidae and Melittidae are not represented). The French (Oise) amber includes the oldest fossil melittid (Paleomacropis eocinicus) (37). This fossil shows oil-collecting hairs typical of extant members of the genus Macropis, indicating that oil collecting is an ancient trait in bees (37). Bees were certainly present as far back as 100 Ma because there are well-preserved pemphredonine (crabronid) wasp fossils from the Burmese amber (44). The presence of the presumed closest relative of bees in the Burmese amber implies that stem-group bees were also present at that time. A recent report of an apparently pollen-collecting apoid from 100 Ma old Burmese

268


amber (45) is consistent with this hypothesis. One could furthermore conjecture that bees could not have arisen before the origin of the angiosperms (~140 Ma) which they pollinate. Therefore, a conservative window for bee origins would be between 100 and 140 Ma, a range of dates consistent with our molecular dating results (Table 1; node 12: 120–112 Ma). Several studies have used relaxed clock-dating methods to estimate the antiquity of particular bee groups. Such studies have sometimes been hindered by the limited numbers of fossil calibration points available and thus there are large standard errors on some date estimates. Fossil-calibrated phylogenies exist for Xylocopini (46), Allodapini (47), Bombini (48), and Halictidae (49, 50). There is enormous variation in the antiquity of eusocial lineages of bees. Within Halictidae, eusociality appears to have arisen within a narrow, and relatively recent, time period 22–20 Ma (50). Several dates exist for the eusocial allodapine bees, a monophyletic group of Old World social and socially parasitic species (47), ranging from a mean of 59–35 Ma. Corbiculate bees arose far earlier. The oldest fossil corbiculate is C. prisca (65 Ma) but our unpublished molecular estimates place the group close to 75 Ma in age. The origins of eusociality in bees therefore spans a time period from the late Cretaceous to the Miocene. The origin of ants occurred substantially before any of these bee groups, and their antiquity is matched among social insect taxa only by the termites (Termitidae) (51) which originated >130 Ma based on fossil evidence (52, 53). There are no published estimates of the antiquity of eusocial Vespidae, another major group of aculeate social insects. A recent molecular phylogeny shows that eusociality evolved twice in this family (54), and the vespine fossil nest Celliforma favosites indicates that eusociality evolved in one of these groups >63 Ma (55). There appears to be a fairly clear correlation between antiquity and social complexity in the groups of social aculeates for which speciation times have been estimated. Halictid bees, the youngest taxa among all eusocial aculeates, show the smallest colonies, a high frequency of reversion to solitary nesting, and only limited queen/ worker dimorphism (56). The oldest taxa, corbiculate bees and ants, possess the most elaborate societies with substantial queen/worker dimorphism, advanced forms of communication, colony defense and nest founding, and no known cases of secondarily solitary species (species that have reverted back to solitary nesting from a eusocial ancestor) (57). As suggested by Wilson and Hölldobler (57, 58), the advanced eusocial groups (such

as ants and corbiculate bees) appear to have reached a “point of no return” where eusociality cannot revert to solitary nesting.

Acknowledgments Valuable comments were provided by P. Ward. Support was provided by U.S. National Science Foundation to the authors.

References 1. A. P. Rasnitsyn, in History of Insects, A. P. Rasnitsyn, D. L. J. Quicke, Eds. (Kluwer, Dordrecht, 2002), pp. 242–254. 2. D. Grimaldi, M. S. Engel, Evolution of the Insects (Cambridge University Press, Cambridge, 2005), pp. xv + 755. 3. A. P. Rasnitsyn, Trudy Paleontologicheskogo Instituta, Akademii Nauk SSSR 147, 1 (1975). 4. D. C. Darling, M. J. Sharkey, Bull. Am. Mus. Nat. Hist. 195, 123 (1990). 5. W. J. Pulawski, A. P. Rasnitsyn, D. J. Brothers, S. B. Archibald, J. Hym. Res. 9, 34 (2000). 6. D. J. Bennett, M. S. Engel, Am. Mus. Novit. 3529, 1 (2006). 7. B. Hölldobler, E. O. Wilson, The Ants (Harvard University Press, Cambridge, Massachusetts, 1990). 8. P. S. Ward, Curr. Biol. 16, R152 (2006). 9. M. Chiotis, L. S. Jermiin, R. H. Crozier, Mol. Phylogenet. Evol. 17, 108 (2000). 10. S. G. Brady, Proc. Natl. Acad. Sci. U.S.A. 100, 6575 (2003). 11. R. N. Johnson, P.-M. Agapow, R. H. Crozier, Mol. Phylogenet. Evol. 29, 317 (2003). 12. P. S. Ward, S. G. Brady, Invertebr. Syst. 17, 361 (2003). 13. C. Saux, B. L. Fisher, G. S. Spicer, Mol. Phylogenet. Evol. 33, 457 (2004). 14. P. S. Ward, D. A. Downie, Syst. Entomol. 30, 310 (2005). 15. G. D. Ouellette, B. L. Fisher, D. J. Girman, Mol. Phylogenet. Evol. 40, 359 (2006). 16. D. Kronauer, C. Schöning, L. Vilhelmsen, J. Boomsma, BMC Evol. Biol. 7, 56 (2007). 17. C. S. Moreau, C. D. Bell, R. Vila, S. B. Archibald, N. E. Pierce, Science 312, 101 (2006). 18. S. G. Brady, T. R. Schultz, B. L. Fisher, P. S. Ward, Proc. Natl. Acad. Sci. U.S.A. 103, 18172 (2006). 19. P. S. Ward, S. G. Brady, B. L. Fisher, T. R. Schultz, Myrmecol. Nachr. 7, 87 (2005). 20. B. Bolton, Syst. Entomol. 15, 267 (1990). 21. G. M. Dlussky, D. J. Brothers, A. P. Rasnitsyn, Insect Syst. Evol. 35, 1 (2004). 22. M. S. Engel, D. A. Grimaldi, Am. Mus. Novit. 3485, 1 (2005).


23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

W. M. Wheeler, Schr. Phys.-Okon. Ges. Königsb. 55, 1 (1915). G. M. Dlussky, Paleontol. J. 31, 616 (1997). F. Carpenter, Bull. Mus. Comp. Zool. 70, 1 (1930). G. M. Dlussky, A. P. Rasnitsyn, Russ. Entomol. J. 11, 411 (2003). R. Crozier, L. S. Jermiin, M. Chiotis, Naturwissenschaften 84, 22 (1997). J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002). G. A. R. Melo, Scientific Papers, Univ. Kansas Nat. Hist. Mus. 14, 1 (1999). C. D. Michener, The Bees of the World, 2nd ed. (Johns Hopkins University Press, Baltimore, MD, 2007). B. N. Danforth, S. D. Sipes, J. Fang, S. G. Brady, Proc. Natl. Acad. Sci. U.S.A. 103, 15118 (2006). A. Roig-Alsina, C. D. Michener, Univ. Kansas Sci. Bull. 55, 123 (1993). B. A. Alexander, C. D. Michener, Univ. Kansas Sci. Bull. 55, 377 (1995). B. N. Danforth, J. Fang, S. D. Sipes, Mol. Phylogenet. Evol. 39, 358 (2006). B. N. Danforth, Curr. Biol. 17, R156 (2007). D. Michez, A. Nel, J.-J. Menier, P. Rasmont, Zool. J. Linn. Soc. 150, 701 (2007). C. D. Michener, D. A. Grimaldi, Proc. Natl. Acad. Sci. U.S.A. 85, 6424 (1988). C. D. Michener, D. A. Grimaldi, Am. Mus. Novit. 2917, 1 (1988). A. P. Rasnitsyn, C. D. Michener, Ann. Entomol. Soc. Am. 84, 583 (1991). D. A. Grimaldi, Ann. Mo. Bot. Gard. 86, 373 (1999). M. S. Engel, Am. Mus. Novit. 3296, 1 (2000).

269

43. M. S. Engel, Bull. Am. Mus. Nat. Hist. 259, 1 (2001). 44. A. V. Antropov, Bull. Nat. Hist. Mus., Lond. (Geol.) 56, 59 (2000). 45. G. Poinar, Jr., B. N. Danforth, Science 314, 614 (2006). 46. R. Leys, S. J. B. Cooper, M. P. Schwarz, Biol. J. Linn. Soc. 77, 249 (2002). 47. M. P. Schwarz, S. Fuller, S. M. Tierney, S. J. B. Cooper, Syst. Biol. 55, 31 (2006). 48. H. M. Hines, Syst. Biol. 57, 58 (2008). 49. B. N. Danforth, S. G. Brady, S. D. Sipes, A. Pearson, Syst. Biol. 53, 309 (2004). 50. S. G. Brady, S. Sipes, A. Pearson, B. N. Danforth, Proc. Roy. Soc. Lond. B 273, 1643 (2006). 51. D. Inward, G. Beccaloni, P. Eggleton, Biol. Lett. 3, 331 (2007). 52. A. E. Emerson, Psyche (Cambridge) 74, 276 (1968). 53. B. L. Thorne, D. A. Grimaldi, K. Krishna, in Termites: Evolution, Sociality, Symbioses, Ecology, T. Abe, D. E. Bignell, M. Higashi, Eds. (Kluwer Academic Publishing, Dordrecht, 2000), pp. 77–93. 54. H. M. Hines, J. H. Hunt, T. K. O’Connor, J. J. Gillespie, S. A. Cameron, Proc. Natl. Acad. Sci. U.S.A. 104, 3295 (2007). 55. J. W. Wenzel, Psyche (Cambridge) 97, 21 (1990). 56. M. P. Schwarz, M. H. Richards, B. N. Danforth, Ann. Rev. Entomol. 52, 127 (2007). 57. E. O. Wilson, The Insect Societies (Harvard University Press, Cambridge, Massachusetts, 1971). 58. E. O. Wilson, B. Hölldobler, Proc. Natl. Acad. Sci. U.S.A. 102, 13367 (2005). 59. A. J. Drummond, S. Y. W. Ho, M. J. Phillips, A. Rambaut, PLoS Biol. 4, e88 (2006). 60. D. J. Brothers, Zool. Scripta 28, 233 (1999). 61. J. M. Carpenter, Zool. Scripta 28, 215 (1999). 62. F. Ronquist, Zool. Scripta 28, 3 (1999).

True flies (Diptera) Matthew A. Bertone* and Brian M. Wiegmann Department of Entomology, Campus Box 7613, North Carolina State University, Raleigh, NC 27695-7613, USA *To whom correspondence should be addressed ([email protected])

Abstract With over 150,000 described species in ~180 families, the insect Order Diptera (true flies) is one of the largest and most diverse groups of organisms. Flies exhibit an extremely wide range of morphological characters that have supported or confounded phylogenetic inferences within the group. Though molecular phylogenies exist, few have addressed macroevolutionary questions within the order and none has comprehensively addressed the order as a whole. Fossil and molecular data indicate that the earliest divergences among living dipterans occurred in the late Paleozoic, 270– 251 million years ago (Ma). Most divergences among families occurred in the Triassic and Jurassic, 251–146 Ma.

The Order Diptera (true flies) comprises an ecologically and morphologically diverse assemblage of holometabolous insects. A number of morphological characters unite this lineage (1–3), the most recognizable one being the extremely reduced, knob-like, metathoracic wings, or halteres (Fig. 1). The majority of true flies also bear specialized sponging mouthparts that differ markedly from the chewing mouthparts found in most insects. Approximately 150,000 species of Diptera have been described in ~180 families, although total species diversity undoubtedly exceeds twice that number (1, 2, 4). Myriad species of Diptera are economically important vectors of human and animal pathogens (e.g., Culicidae) and many are destructive to crops and livestock (e.g., Tephritidae and Oestridae, respectively). Flies are also important ecologically as predators, decomposers, parasitoids, and pollinators (5, 6). Here we review the relationships and divergence times of major events in dipteran evolution, including the origin of the order and its constituent suborders, infraorders, and families. Traditionally flies have been divided into two suborders: Nematocera (“thread-horn” flies) and Brachycera (“short-horn” flies). This division was based primarily

on characters of the adult antennae and larval head capsule (1, 2, 7). Although a number of shared-derived characters support the monophyly of the Brachycera, the Nematocera is now widely regarded as a paraphyletic assemblage of infraorders, or suborders (8), from/within which the Brachycera originated (7, 9, 10). For a detailed overview of the current state of Diptera systematics, see Yeates and Wiegmann (1, 2). Evolutionary relationships among the lower Diptera (= “Nematocera”) have been particularly difficult to resolve. Morphology-based hypotheses disagree with respect to the composition and interrelationships of the nematoceran infraorders (7, 9, 11–13). Interpretations of character homology, polarity, and homoplasy, as well as incongruence between adult characters vs. those of the larvae and pupae, have contributed to disagreement concerning the higher-level relationships of these flies (1, 2). Comprehensive reviews of the relationships within the nematocerous Diptera are presented elsewhere (7, 9). Determination of the closest relatives of the hyperdiverse Brachycera has been equally difficult. Hennig (12, 13) gave evidence, taken largely from adult characters,

Fig. 1 A predatory robber fly (Asilidae: Ommatius gemma) from Mississippi, USA. Credit: G. and J. Strickland.

M. A. Bertone and B. M. Wiegmann. True fl ies (Diptera). Pp. 270–277 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Arthropoda; Insecta; Diptera

271

Mycetophilidae-1 43

Sciaridae

41

Cecidomyiidae 32

Bibionidae 42

Pachyneuridae

38

25

Mycetophilidae-3

Bibionomorpha

Mycetophilidae-2

40

Scatopsidae 46

Canthyloscelidae

24 16

Anisopodidae

11

Perissommatidae 5

Psychodidae 44

Tanyderidae

15

Blephariceridae Culicidae 37

Corethrellidae

18

Dixidae

8

Ceratopogonidae 6

9

Chironomidae

Culicomorpha

Chaoboridae

27

2

Psychodomorpha

Brachycera

Simuliidae 31

Thaumaleidae Nymphomyiidae

17

Axymyiidae Ptychopteridae Tipulidae

4

Trichoceridae Deuterophlebiidae

P

Triassic

Jurassic

PZ 250

Cretaceous

MESOZOIC 200

150

Pg

Tipulomorpha

1

Ptychopteromorpha

3

Ng

CENOZOIC 100

50

0 Million years ago

Fig. 2 Continues

for a relationship between the Brachycera and his concept of the Bibionomorpha. Although not explicit in their published tree, Wood and Borkent (7) suggested a relationship between Brachycera and their Psychodomorpha (including Scatopsidae, Canthyloscelidae (as Synneuridae), Perissommatidae, and Anisopodidae, all of which were in Hennig’s Bibionomorpha). This hypothesis was based

predominately on characters of the larval mandible and was later supported (14). Oosterbroek and Courtney’s (9) analysis of characters from all life stages found a single family—Anisopodidae—to be the closest relative of the Brachycera. This group was placed at the tip of a clade they termed the “higher Nematocera + Brachycera” (the “higher Nematocera” including all of Wood and Borkent’s

Syrphidae Platypezidae

28

Dolichopodidae

45

Empididae

36

Atelestidae Therevidae

30

Scenopinidae Asilidae

19

23

Mydidae

33

Apioceridae

13

Bombyliidae 10

Acroceridae Nemestrinidae Pantophthalmidae

7

29

Xylomyidae

34 12

Stratiomyidae Tabanidae

35

Pelecorhynchidae

21

Rhagionidae

26

14

Vermileonidae Xylophagidae

Tr

Jurassic

Cretaceous

150

Ng

CENOZOIC

MESOZOIC 200

Pg

Tabanomorpha

20

Asiloidea

22

Muscomorpha

39

Stratiomyomorpha

Drosophilidae

47

Xylophagomorpha

Muscidae

48

Empidoidea Cyclorrhapha


Nemestrinoidea

272

100

50

0 Million years ago

Fig. 2 A timetree of true flies (Diptera). Divergence times are shown in Table 1. The timetree for Brachycera is continued in a separate panel. Mycetophilidae-1 (Diadocidiidae, Mycetophilidae sensu stricto), Mycetophilidae-2 (Keroplatidae

and Lygistorrhinidae), Mycetophilidae-3 (Ditomyiidae). Abbreviations: Ng (Neogene), P (Permian), Pg (Paleogene), PZ (Paleozoic), and Tr (Triassic).

Psychodomorpha and the Tipulidae). Subsequently, Michelsen (11) reunited the Brachycera with a Hennigian Bibionomorpha (as “Neodiptera”) based on adult thoracic sclerites and musculature. Divisions within the Brachycera have traditionally followed a trend of paraphyletic stem grades (e.g., Orthorrhapha and Aschiza) giving rise to monophyletic clades (e.g., Cyclorrhapha and Schizophora). The lower Brachycera (= Orthorrhapha) are generally small to very large flies, many of which are predators or parasitoids as larvae. The basal Infraorders Xylophagomorpha,

Stratiomyomorpha, and Tabanomorpha are each represented by one, three, and five families, respectively (1, 2, 15). The Tabanomorpha is of particular interest for containing three families in which at least some female flies suck vertebrate blood (Tabanidae, Athericidae, and Rhagionidae) (16). Other notably diverse families within the lower Brachycera include the Bombyliidae (bee flies), Asilidae (robber flies), Empididae (dance flies), and Dolichopodidae (long-legged flies). Woodley (17), Sinclair et al. (18), and Yeates (15) present morphological evidence supporting relationships among the lower

Eukaryota; Metazoa; Arthropoda; Insecta; Diptera

273

Table 1. Divergence times (Ma) and their credibility/confidence intervals (CI) among true flies (Diptera). Timetree Node

Time

Estimates

Timetree

Ref. (10, 68) Time

CI

Refs. (39, 40) Time

Node

Estimates Ref. (10)

Time

CI

Time

Refs. (39, 40) CI

Time

CI

1

267

267

269–260

–

–

25

155

155

195–114

–

–

2

265

265

269–256

–

–

26

147

147

190–104

–

–

3

241

241

260–224

–

–

27

146

146

188–97

–

–

4

235

235

261–221

–

–

28

143

143

173–122

–

–

5

234

234

259–209

–

–

29

139

139

186–94

–

–

6

226

226

243–215

–

–

30

133

133

170–96

–

–

7

222

222

239–195

–

–

31

130

130

181–74

–

–

8

220

220

234–212

–

–

32

126

126

168–84

–

–

9

213

213

223–210

–

–

33

123

123

159–87

–

–

10

213

213

235–188

233(40)

239–217

34

122

122

171–78

129(39)

194–83

11

210

210

243–179

–

–

35

121

121

168–80

–

–

12

210

210

234–180

–

–

36

120

120

153–92

–

–

13

202

202

226–179

–

–

37

118

118

163–71

–

–

14

198

198

227–163

–

–

38

116

116

158–75

–

–

15

197

197

225–181

–

–

39

115

115

145–89

–

–

16

196

196

230–160

–

–

40

114

114

157–74

–

–

17

195

195

236–138

–

–

41

103

103

144–64

–

–

18

190

190

216–155

–

–

42

98.6

98.6

139–60

–

–

19

184

184

209–171

–

–

43

95.6

95.6

137–57

–

–

20

178

178

203–161

–

–

44

95.2

95.2

140–57

–

–

21

172

172

211–131

–

–

45

87.7

87.7

122–59

–

–

22

166

166

193–143

–

–

46

87.1

87.1

136–45

–

–

23

165

165

194–140

–

–

47

84.5

84.5

115–71

–

–

24

160

160

200–120

–

–

48

47.9

47.9

76–29

–

–

Note: Node times in the timetree are from ref. (10, 68).

Brachycera. However, as in the lower Diptera, anatomical characters uniting major lineages are often lacking, equivocal, or convergent. The brachyceran clade Cyclorrhapha contains over half of all true flies. Extreme reduction of the larval head capsule and pupation of the third instar in the final larval skin (puparium) are the major innovations of this group (1, 2). More homogeneous in morphology than the lower Diptera, Cyclorrhapha contains the stereotypical “higher” flies and familiar members of this group include the vinegar fly (commonly called the fruit fly by geneticists, Drosophila melanogaster) and the house fly (Musca domestica).

The lower Cyclorrhapha (= Aschiza) is a paraphyletic collection of families united only by plesiomorphic characters. Included within this lineage are ~8 families of flies, some of which are highly diverse (e.g., Syrphidae and Phoridae) (19, 20). Relationships between families within the lower Cyclorrhapha remain ambiguous (21, 22). The division Schizophora contains most of the family-level diversity in the Diptera, with at least 75 described families (1, 2). All flies in this group possess a membranous head sac (ptilinum) that, when inflated, allows the adult to escape the puparium. After emergence the sac is withdrawn leaving a remnant, U-shaped, ptilinal fissure. Schizophora is further divided into two

274


main groups: Acalyptratae and Calyptratae. While the Calyptratae is a well-supported monophyletic group (1, 19), the Acalyptratae may or may not be monophyletic (1, 2, 19, 23, 24) and most dipterists now suspect the latter. Although containing only about 20% of fly species, Acalyptratae contains nearly half of the order’s familylevel diversity (~62 families). Remarkably, six common acalyptrate families (Tephritidae, Lauxaniidae, Agromyzidae, Chloropidae, Drosophilidae, and Ephydridae) make up >50% of the species diversity in the entire assemblage (2). Resolving the relationships within and among acalyptrate superfamilies has traditionally been difficult, due, in part, to the lack of convincing sharedderived characters for most major groupings. Calyptrate flies are divided into ~13 families that are important medically (e.g., Glossinidae, Muscidae and Oestridae), forensically (e.g., Calliphoridae and Sarcophagidae), or as biological control agents (e.g., Tachinidae). Calyptratae also contains several groups of specialized, vertebrate ectoparasites (Hippoboscidae, Streblidae, and Nycteribiidae). Despite the publication of 14 completed dipteran genomes (Anopheles gambiae, Aedes aegypti, and 12 Drosophila species), relatively few studies have used molecular markers to reconstruct higher-level relationships of flies. However, many new molecular studies are emerging (73) and these will certainly increase knowledge of relationships within the order. Early nucleotide-based higher-level studies compared nuclear 28S ribosomal DNA sequences among Culicomorpha (25), some Cyclorrhapha (26), and among the nematocerous Diptera (27). Major markers employed to date for phylogenetic inference include mitochondrial genes (28–30), elongation factor 1α (EF-1α) (31, 32), 28S ribosomal DNA (10, 16, 27), and the nuclear protein-coding genes CPSase (CAD) (33–36) and white (37). Friedrich and Tautz’s (27) phylogeny of the earliest dipteran lineages strongly supported several infraorders; however, the relationships among the infraorders were not well resolved (80% of families, with DNA sequences from mitochondrial 16S rRNA and cytochrome oxidase subunit I for nearly half of the taxa. They analyzed a 340taxon subset of their 1880-taxon data set using Bayesian inference, fi xed the age of the ingroup at 285 Ma in an “all compatible” version of the resulting consensus tree, and applied seven fossil age constraints to calibrate and date internal nodes using penalized likelihood. Higherlevel molecular timetrees (molecular trees calibrated with fossils) previously had been available only for the beetle superfamilies Chrysomeloidea (20, 21; long-horn beetles, leaf beetles, and allies) and Curculionoidea (20) (Table 1). Hunt et al. (18) recovered Adephaga and Polyphaga as closest relatives, themselves closest to Myxophaga plus Archostemata. This arrangement is consistent with most other molecular phylogenetic studies. They reported average ages and 95% confidence intervals (CIs) for 13 selected clades in their published timetree [Adephaga (237.2 ± 2.63 Ma), Bostrichiformia (219.4 ± 11.19 Ma)], the cerylonid series of families in Cucujoidea (flat bark beetles, flower beetles, ladybird beetles; 202.9 ± 11.44 Ma), Cucujiformia (236.2 ± 7.47 Ma), Curculionoidea; weevils (171.5 ± 27.06 Ma), Elateriformia (217 ± 10.92 Ma), Elateroidea; click beetles and allies (188.1 ± 22.23 Ma), Histeroidea; clown beetles (190.8 ± 9.42 Ma), Hydradephaga; diving beetles and whirligigs (219.8 ± 3.89 Ma), Hydrophiloidea (175.4 ± 23.36 Ma), Myxophaga + Archostemata (227.0 ±

1.68 Ma), Nitidulidae (sap-feeding beetles; 129.7 ± 12.34 Ma), and Polyphaga (270.5 ± 2.26 Ma)]. Because only these 13 age estimates and corresponding CIs were published (18), and because their published timetree lacks names for terminal taxa, timing and patterns of diversification across much of the tree, for example, for most superfamilies and all families but Nitidulidae, remain difficult to interpret. To help clarify divergence times for family-level and higher groupings, we obtained a “recreated” version of the Hunt et al. (18) timetree (their Figure 3) from the authors. Th is tree has the same topology as the published version and names for terminal taxa, but node ages estimated from the timetree differ by at least 5 Ma from the published version for seven of the aforementioned 13 nodes for which average ages were published (perhaps on account of this being a “recreated” tree) (18). These differences were greatest in Cucujiformia. Nonetheless, we reduced the timetree to family-level taxa (when possible), and obtained estimated ages for all nodes based on their relative positions in the timetree (Table 1). Age estimates reported without CIs are our own best estimates based on the timetree and corresponding timescale provided by the authors (unless otherwise noted). Based on these data, the Adephaga–Polyphaga split was estimated as ~277 Ma. The Suborder Adephaga comprised two well-supported clades, the aquatic Hydradephaga and the terrestrial Geadephaga (215.7 Ma; ground beetles and tiger beetles). Overall, relationships within Adephaga were similar to those found by other authors using molecular data (22, 23), and reconstructed divergence times are compatible with the fossil record (13). Within the Suborder Polyphaga, five series of families are traditionally recognized (4, 5); Bostrichiformia (Superfamilies Bostrichoidea and Derodontoidea), Cucujiformia (Chrysomeloidea, Cleroidea, Cucujoidea, Curculionoidea, Lymexyloidea, and Tenebrionoidea), Elateriformia (Buprestoidea, Byrrhoidea, Dascilloidea, Elateroidea, and Scirtoidea), Scarabaeiformia (Scarabaeoidea), and Staphyliniformia (Hydrophiloidea and Staphylinoidea). The Superfamilies Derodontoidea (Family Derodontidae) and Scirtoidea (Families Clambidae, Decliniidae, Eucinetidae, and Scirtidae) occupied the basal nodes in Polyphaga. This arrangement conflicts with the traditional placement of Derodontoidea in the Series Bostrichiformia and Scirtoidea in the Series Elateriformia (4), but is consistent with other recent studies (16, 17). Derodontoidea and Scirtoidea exhibit several pleisiomorphic morphological features that

Eukaryota; Metazoa; Arthropoda; Insecta; Coleoptera

support their placement at the base of Polyphaga, including paired dorsal ocelli, mesocoxal cavities partly closed by the metepisterna, a transverse metasternal suture, a trilobed aedeagus, and six free Malphigian tubules (8). The Series Staphyliniformia comprised a paraphyletic grade near the base of Polyphaga. Scarabaeiformia (~191.4 Ma) appeared within Staphyliniformia. Elateriformia minus Scirtoidea was found to be the closest relative of Bostrichiformia minus Derodontoidea. Cucujiformia was strongly supported as monophyletic. The Superfamilies Buprestoidea (~142.5 Ma; metallic wood-boring beetles), Dascilloidea (~73.1 Ma), and Elateroidea were each found to be monophyletic. Byrrhoidea was polyphyletic, with the Byrrhidae (moss beetles) resolved separately from a clade comprised of the remaining families. Lymexyloidea (ship-timber beetles) was polyphyletic and appeared near the base of Tenebrionoidea. The Families Biphyllidae and Byturidae appeared within Cleroidea (checkered beetles and allies), an arrangement previously suggested by other authors (e.g., 4) based on morphology. Cucujoidea was polyphyletic, with Sphindidae as the closest relative of Tenebrionoidea plus Lymexloidea, and Silvanidae and Phloeostichidae as the closest relatives of a clade comprised of the chrysomelid Subfamily Hispinae and a monophyletic Curculionoidea. The Superfamily Chrysomeloidea and the Family Chrysomelidae (leaf beetles and long-horn beetles) were therefore polyphyletic. Traditionally, Chrysomeloidea and Chrysomelidae are thought to be monophyletic, and Chrysomeloidea is thought to be closest to Curculionoidea (e.g., 20, 21, 24). While direct comparisons are difficult due to differences in taxon sampling and resolution, all lineages of Chrysomeloidea sampled by Hunt et al. (18), including both lineages of their polyphyletic Family Chrysomelidae (Chrysomelidae-1; ~191.3 Ma, Chrysomelidae-2; ~187.6 Ma) are estimated to have originated before 100 Ma (Table 1). This is in contrast to a recent study (21) employing 18S, 28S rRNA, and mitochondrial 16S rRNA, which places the origin of Chrysomelidae at 73–79 Ma (95% CI; 63–86 Ma), and argues that the Family Chrysomelidae radiated in the Cenozoic, long after the Jurassic to early Cretaceous origin and middle Cretaceous diversification (e.g., see 25) of their (predominantly) angiosperm host plants. Note that the estimated timing of origin of most chrysomeloid lineages in Hunt et al. (18) are at least loosely consistent with the aforementioned timing of angiosperm diversification (25). The questions remain open, therefore, as to whether and how the origin and diversification of angiosperms influenced diversification

287

in the largely herbivorous Superfamily Chrysomeloidea (>50,000 species), and other groups of beetles—e.g., Curculionoidea (>60,000 species). We recently completed an analysis of 18S rRNA sequence data with the dual goals of reconstructing higher-level relationships and divergence times in Coleoptera. While overlapping in part with the data set analyzed by Hunt et al. (18), our methods of vetting data, alignment, and fossil calibration to form a timetree differ sufficiently from Hunt et al. (18) to warrant mention. Further, the resulting timetree permits at least casual comparison with the Hunt et al. (18) timetree. We obtained all Genbank 18S rRNA sequences available for beetles as of May 2007. To this data set, we added one unpublished sequence of our own, Prolixocupes (Archostemata: Cupedidae), effectively doubling the number of Archostemata included in previous studies. All sequences with fewer than 1300 bp of sequence data were excluded from analysis. Genuslevel exemplars were randomly selected when duplicates were present, except when sequences differed by more than 100 bp in aligned length, in which case the most complete sequence available was used. Six sequences were excluded due to large numbers of Ns and unusual alignment problems indicative of low-quality data. The final data set consisted of 955 ingroup sequences representing 134 families. Genbank sequences from six neuropteroids were used to root the tree [Hemerobius (AF423790); Myrmeleon (chimera of U65137 & L10182); Oliarces (AF012527); Phaeostigma (X89494); Sialis (chimera of AY521864 & X89497); and Mantispa (chimera of AY620034 & U65189)]. DNA sequences were aligned with Clustal X (26) and manually adjusted in MacClade v.4.05 (27). An annotated secondary structural alignment for insect 18S (28) was used to further refine the alignment. Regions 4, 11A, 14A, and 14B of Kjer (28) could not be unambiguously aligned, and were excluded from analysis. The remaining aligned data consisted of 1920 nucleotide positions. A maximum likelihood (ML) search employing the GTR+I+Γ substitution model and limited to 107 generations, was implemented in GARLI v0.951 (29). Branch lengths were optimized in PAUP* v.4.03b10 (30). In the absence of clocklike molecular evolution, we used nonparametric rate smoothing (31) implemented in r8s v.1.71 (32) to generate an ultrametric tree from the ML topology (−lnL = −73081.14). Fossils used to calibrate the tree and to date internal nodes included: (i) the oldest unequivocal fossil Hydradephaga (1), applied as a minimum constraint of 225 Ma on the

288


age of Hydradephaga, (ii) the oldest unequivocal fossil Scarabaeidae (1), applied as a minimum constraint of 152 Ma on the age of Scarabaeioidea, (iii) the oldest unequivocal fossil Tenebrionidae (1), applied as a minimum constraint of 125 Ma on the age of Tenebrionoidea, (iv) the oldest unequivocal fossil Curculionoidea (1, 33), applied as a minimum constraint of 152 Ma on the age of Curculionoidea, (v) the oldest unequivocal fossil Staphylinidae (1), applied as a minimum constraint of 227.5 Ma on the age of Staphylinoidea, and (vi) the oldest unequivocal fossil Cupedidae (1), applied as a minimum constraint of 199.6 Ma on the age of Cupedidae. When the age of a given fossil was not reported in the literature, we used the upper boundary of the subdivision of the global geological record reported as having contained the fossil, as a minimum age constraint. Fossil constraints were applied conservatively, so the resulting nodal age estimates should be considered similarly conservative. Holometabolous insects are not known from before the Permian (2), so we constrained the maximum age of the ingroup to 299.0 Ma, the Carboniferous–Permian boundary. We separately applied each of two alternative maximum constraints on the age of the ingroup, the Carboniferous–Devonian boundary (359.2 Ma), and the Devonian–Silurian boundary (416 Ma), to evaluate the robustness of nodal age estimates to relaxation of this constraint. Nodal age estimates were only minimally affected, so we report our results as a range of ages (and mean) spanning the three age estimates determined for each node of interest. Based on these analyses, we determined the basal relationships of suborders to be: (Myxophaga + Archostemata, (Adephaga, Polyphaga)). This is in agreement with Hunt et al. (18) and most other analyses employing 18s rDNA (14–16). The placement of Archostemata within Myxophaga should be viewed as tentative due to the limited sampling of Archostemata in all studies to date. Sampling of additional Archostemata (e.g., Ommatidae and Micromalthidae) and Myxophaga (Lepiceridae), experimentation with outgroup taxon sampling, inclusion of data from nuclear protein coding genes, and additional analyses, may help clarify relationships between these two interesting suborders. Overall, relationships within Adephaga were very similar to those obtained by other authors (15, 17, 18, 22, 23). While relationships within Polyphaga were not well resolved, they were grossly similar to other studies, for example, obtaining Scirtoidea and Derodontoidea as the earliest branching lineages in the suborder.

The series Cucujiformia, while monophyletic, contained numerous very short internal branches, and relationships within the series were generally not well supported. It should be noted that with few exceptions, the 18S sequences of Cucujiformia and several other higher-level groups within Polyphaga exhibit relatively little overall divergence. As a consequence, 18S should be expected to be minimally informative for such relationships, especially because several of the more highly variable regions of 18S were excluded from this and most other studies. The series Cucujiformia contains nearly half of all beetle families and most beetle species, and most cucujiform beetles feed on plants. Therefore, an accurate reconstruction of relationships and timing and patterns of diversification within the series is critical to our understanding of beetle macroevolution, including the role of angiosperms in beetle diversification. Due to the lack of well-supported resolution at lower levels in Polyphaga and beyond, we did not evaluate divergence times below the subordinal level. Based on the topology we obtained, and employing the fossil age constraints described herein, we estimate that the split between the clade comprised the Suborders Myxophaga + Archostemata and the clade comprised the Suborders Adephaga and Polyphaga, occurred ~269–265 Ma (mean 266.8 Ma). Hunt et al. (18) fi xed this age at 285 Ma. We determined the Adephaga–Polyphaga split to have occurred ~269–265 Ma (mean 266.4 Ma), just slightly later than Hunt et al. (18), who estimated this split to have occurred ~277 Ma. These observations suggest that the four living suborders of beetles diverged at a time (Permian) when many other groups of terrestrial organisms, including other insects (34), underwent rapid diversification. The subordinal relationships and divergence times based on the limited numbers and kinds of genes used in the papers reviewed here appear to be robust. However, lower-level relationships and divergence times remain unsettled. Without a well-supported topology at this level, accompanied by dated nodes with confidence intervals, particularly for the most species-rich Cucujiformia, it is difficult to justify detailed evaluation of the timing, causes, and consequences of ecological diversification (e.g., the role of phytophagy, predation, or fungivory). Nonetheless, large-scale molecular phylogenetic studies such as that presented by Hunt et al. (18) promise the most comprehensive picture to date of the main branching events and their divergence times in the evolution of insects, including the famously diverse beetle order Coleoptera.

Eukaryota; Metazoa; Arthropoda; Insecta; Coleoptera

Acknowledgments The authors would like to thank A. Vogler for providing the timetree we used to estimate node ages in Hunt et al. (18). Support was provided by the U.S. National Science Foundation.

14. 15. 16. 17.

References 1. D. Grimaldi, M. S. Engel, Evolution of the Insects (Cambridge University Press, New York, 2005). 2. R. H. Arnett, Jr., M. C. Thomas, American Beetles, Vol. 1. (CRC Press, Boca Raton, 2000). 3. R. H. Arnett, Jr., M. C. Thomas, P. E. Skelley, J. H. Frank, American Beetles, Vol. 2. (CRC Press, Boca Raton, 2002). 4. J. F. Lawrence, A. F. Newton, Jr., Families and Subfamilies of Coleoptera (Muzeum i Instytut Zoologii PAN, Warszawa, 1995). 5. R. G. Beutel, R. A. B. Leschen, Coleoptera: Morphology & Systematics (W. DeGruyter, Berlin, 2005). 6. R. G. Beutel, Coleoptera: Myxophaga (DeGruyter, Berlin, 2005). 7. T. Hörnschemeyer, Coleoptera: Archostemata (DeGruyter, Berlin, 2005), pp. 29–42. 8. J. F. Lawrence, A. F. Newton, Jr., Ann. Rev. Ecol. Syst. 13, 261 (1982). 9. A. G. Ponomarenko, Trudy Paleontologicheskogo Instituta 125, 1 (1969). 10. R. Beutel, F. Haas, Cladistics 16, 103 (2000). 11. J. F. Lawrence, S. A. Slipinski, J. Pakaluk, From Latreille to Crowson: A History of the Higher-Level Classification of Beetles (Muzeum i Instytut Zoologii PAN, Warszawa, 1995), pp. 87–155. 12. R. A. Crowson, Ann. Rev. Entomol. 5, 111 (1960). 13. R. G. Beutel, in Handbook of Zoology, Vol. IV Arthropoda: Insecta. Part 38. Coleoptera, Vol. 1: Morphology and Systematics (Archostemata, Adephaga, Myxophaga, Polyphaga (partim), R. G. Beutel, R. A. B. Leschen, Eds. (DeGruyter, Berlin, 2005), pp. 1–16.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

30. 31. 32. 33. 34.

289

M. S. Caterino, V. L. Shull, P. M. Hammond, A. P. Vogler, Zool. Scripta 31, 41 (2002). V. L. Shull, A. P. Vogler, M. D. Baker, D. R. Maddison, P. M. Hammond, Syst. Biol. 50, 945 (2001). M. S. Caterino, T. Hunt, A. P. Vogler. Mol. Phylogenet. Evol. 34, 655 (2005). A. Vogler, in Handbook of Zoology, Vol. IV Arthropoda: Insecta. Part 38. Coleoptera, Vol. 1: Morphology and Systematics (Archostemata, Adephaga, Myxophaga, Polyphaga (partim), R. G. Beutel, R. A. B. Leschen, Eds. (DeGruyter, Berlin, 2005), pp. 17−22. T. Hunt, et al. Science 318, 1913 (2007). J. S. Hughes, et al. Mol. Biol. Evol. 23, 268 (2006). B. D. Farrell, Science 281, 555 (1998). J. Gómez-Zurita, T. Hunt, F. Kopliku, A. P. Vogler, PLoS ONE, 4, e360 (2007). I. Ribera, J. E. Hogan, A. P. Vogler, Mol. Phylogenet. Evol. 23, 43 (2002). D. R. Maddison, M. D. Baker, K. A. Ober, Syst. Ent. 24, 103 (1999). B. D. Farrell, A. Sequeira, Evolution 58, 1984 (2004). C. D. Bell, D. E. Soltis, P. S. Soltis, Evolution 59, 1245 (2005). J. D. Thompson, T. J. Gibson, F. Plewniak, F. Jeanmougin, D. G. Higgins. Nuc. Acids Res. 24, 4876 (1997). W. P. Maddison, D. R. Maddison, MacClade. Software v.4.05 (Sinauer, Sunderland, Massachusetts, 2002). K. M. Kjer, Syst. Biol. 53, 506 (2004). D. J. Zwickl, Genetic Algorithm Approaches for the Phylogenetic Analysis of Large Biological Sequence Datasets under the Maximum Likelihood Criterion, Ph.D. dissertation (University of Texas, Austin, 2006). D. L. Swofford, PAUP*. Software v.4.03b10 (Sinauer, Sunderland, Massachusetts, 2002). M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997). M. J. Sanderson, Bioinformatics 19, 301 (2003). V. G. Gratshev, V. V. Zherikhin, Acta Zoolog. Cracov. 46 (suppl.—Fossil Insects), 129 (2003). C. C. Labandeira, J. J. Sepkoski, Jr., Science 261, 310.

Lacewings (Neuroptera) Shaun L. Wintertona,* and Brian M. Wiegmannb a

Entomology, Queensland Department of Primary Industries & Fisheries, 80 Meiers Road, Indooroopilly, Queensland, Australia 4068; bDepartment of Entomology, North Carolina State University, Raleigh, NC 27695, USA *To whom correspondence should be addressed (wintertonshaun@ gmail.com)

Abstract Lacewings (~5700 species) are divided into 17 families distributed on all continents. This group of insects is well defined by complex larval characteristics such as modified sucking jaws, incomplete gut, and modified Malphigian tubules used for spinning a silken cocoon in which the immature pupates. Recent molecular evidence supports a Permian (299–251 million years ago, Ma) origin of the order with Coniopterygidae as the closest relative of all other neuropterans. The only well-defined group of families is Myrmeleontiformia. This group of five families diverged from the rest of the order ~184 Ma and has become the most species-rich lineage of Neuroptera.

The lacewings comprise a moderately sized group of insects whose adults are characterized by delicate wings, often large in size, with highly reticulate or lacelike venation (Fig. 1). The order is presently divided into 17, mostly well defined, families with ~5700 species worldwide. Lacewings, together with Megaloptera and Raphidioptera, constitute the Superorder Neuropterida, which is assumed to be the closest relative of Coleoptera. Three unique larval characters distinguish Neuroptera from other insects: modified jaws for sucking, incomplete gut, and modified Malphigian tubules for spinning a silken cocoon in which the immature pupates (1). Here, we review the relationships and divergence times of the families of lacewings (Fig. 2). Elucidating the evolutionary history of lacewings has proven to be highly problematic. Neuroptera classification has traditionally been founded largely on the intuitive phylogenies by Handlirsch (2) for fossils and by Withycombe (3) for larval anatomy and morphology. From these works a number of higher-level groups have

been identified, but they usually differ greatly between subsequent authors with regard to membership of families in each (1). Recent quantitative analyses of morphological characters and phylogeny by Aspöck et al. (4) led to consolidation of this classification into only three Suborders Nevrorthiformia (containing Nevrorthidae only), Myrmeleontiformia (containing five families: Nemopteridae, Psychopsidae, Nymphidae, Ascalaphidae, and Myrmeleontidae), and Hemerobiiformia (containing 11 families: Ithonidae, Polystoechotidae, Chrysopidae, Hemerobiidae, Mantispidae, Rhachiberothidae, Berothidae, Coniopteryigdae, Dilaridae, Sisyridae, and Osmylidae). The most significant aspects of this work were (a) placement of Ithonidae (moth lacewings) not as the closest relative of remaining neuropterans but rather in a more derived position closer to Myrmeleontiformia and (b) the placement of Nevrorthidae, instead, as the earliest-branching family. A monophyletic Hemerobiiformia as circumscribed by Aspöck et al. (4) has never been recovered as a natural group in any molecular analysis (5, 6). Myrmeleontiformia

Fig. 1 A split-footed lacewing (Norfolius howensis) from Australia. Credit: S. L. Winterton.

S. L. Winterton and B. M. Wiegmann. Lacewings (Neuroptera). Pp. 290–292 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Metazoa; Arthropoda; Insecta; Neuroptera

291

Hemerobiidae 11

Chrysopidae

8

Ithonidae 10

Nemopteridae

6 13

Myrmeleontidae 14

12

Ascalaphidae

9

5

Nymphidae Psychopsidae

Myrmeleontiformia

Polystoechotidae

Mantispidae 3

Berothidae

7

Rhachiberothidae Osmylidae 2

4

Nevrorthidae Sisyridae

1

Dilaridae Coniopterygidae

Triassic

Jurassic

Cretaceous

MESOZOIC 250

200

150

Pg

Ng

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of lacewings (Neuroptera). Divergences are shown in Table 1. Abbreviations: Ng (Neogene) and Pg (Paleogene).

is a strongly supported group in all published phylogenies based on morphological (3, 4, 7–9) and molecular data (Winterton and Wiegmann, submitted; 5, 6). A suite of adult and larval morphological characters define this group. Relative to morphological data, very few molecular phylogenies of Neuroptera have been published. Haring and Aspöck (5) published the first attempt at elucidating lacewing phylogeny using molecular data by sequencing representatives of most neuropteran families for amino acid sequences of cytochrome c oxidase subunit III. The results of this study recovered monophyletic Myrmeleontiformia and Nevrorthiformia, but a polyphyletic Hemerobiiformia, with considerable discord between molecular and morphological data regarding the placement of the families Osmylidae and Dilaridae. Using 18S rDNA sequences Winterton (6) also showed strong support for Myrmeleontiformia, but did not recover a monophyletic Hemerobiiformia. In this analysis Dilaridae were also placed as a basal group rather than being closely related to Mantispidae, Rhachiberothidae, and Berothidae.

The most recent analysis of lacewing phylogeny is by Winterton and Wiegmann (submitted), using both molecular and morphological data with much greater taxon sampling than in previous studies of the group. They produced the most comprehensive phylogeny of the order to date, by using DNA sequences for 16S, 18S ribosomal genes, COI and CAD, combined with morphological data for representatives of all families of Neuroptera. Results from this study show that Coniopteryigdae rather than Nevrorthidae is the earliest-branching family in the order, and that Hemerobiiformia is a paraphyletic group having a laddered topology leading to a monophyletic Myrmeleontiformia. Dilaridae is again supported as a basal group closely related to Sisyridae and Coniopteryigdae. Neuropterans have a rich fossil record with definitive neuropteridan ancestors dating back to the Permian (1, 10). Only Winterton and Wiegmann (submitted), using the aforementioned gene sequences, have estimated divergence times among lacewing families based on genetic data. Coniopterygidae diverged from the rest of the order during the late Permian (~257 Ma) followed by

292


Table 1. Divergence times (Ma) and confidence/ credibility intervals (CI) among lacewings (Neuroptera) based on Winterton and Wiegmann (submitted). Timetree Node

Time

CI

1

257

268–244

2

247

259–235

3

242

254–229

4

228

242–212

5

227

239–214

6

219

231–207

7

214

230–199

Neuroptera is a relatively ancient group of insects with a rich fossil record. This, along with tremendous morphological disparity between families and highly disjunct biogeographical patterns across related taxa, indicate that the golden age of lacewings has long since past (4). Divergence estimates support this view with all major divergences occurring during the Mesozoic (251–66 Ma) (Winterton and Wiegmann, submitted). Our understanding of the evolutionary history of this group is still developing, with morphology providing only fragmentary insights. Further studies using molecular data (e.g., more molecular markers) and intermediate fossil material are needed to fully elucidate the evolutionary history of Neuroptera.

8

213

226–200

9

208

221–195

10

197

211–184

Acknowledgment

11

197

214–177

12

187

200–176

Support was provided by the U.S. National Science Foundation to the authors.

13

172

183–167

14

153

162–150

Note: Two protein-encoding genes and two ribosomal genes were used to estimate divergence times in Winterton and Wiegmann (submitted).

all major family-level divergences occurring during the Jurassic (>145 Ma). This is reflected in the fossil record, with fossil taxa of most extant families present in Jurassic or early Cretaceous aged deposits (1, 2, 10). The estimated age of the divergence of Myrmeleontiformia from its closest relative is ~216 Ma. The most recent family-level divergence is that of Ascalaphidae from Myrmeleontidae (~153 Ma). Based on these data, interesting questions may be posed regarding the origins of certain families. One example is that of the origins of freshwater sponges and spongilla flies (Sisyridae). Freshwater sponges are classified in the Suborder Spongillina and likely colonized freshwaters from marine ancestors during the early Jurassic (210–141 Ma) (11). Considering that the origin of Sisyridae is estimated at ~247 Ma (Winterton and Wiegmann, submitted), and that this group are specialist predators on freshwater sponges and bryozoans, ancestral sisyrids possibly fed on freshwater bryozoans during the Triassic before expanding to freshwater sponges during the Jurassic.

References 1. T. R. New, Handbook of Zoology. Planipennia: Lacewings, Vol. IV. Part 30 (Walter de Gruyter, Berlin, 1989). 2. A. Handlirsch, Die fossilen Insekten Und Die Phylogenie Der Rezenten Formen: Ein Handbuch Für Paläontologen Und Zoologen (Engelmann, Leipzig, Germany, 1906–1908). 3. C. L. Withycombe, Trans. Entom. Soc. Lond. 72, 303 (1925). 4. U. Aspöck, J. D. Plant, H. L. Nemeschkal, Syst. Entom. 26, 73 (2001). 5. E. Haring, U. Aspöck, Syst. Entom. 29, 415 (2004). 6. S. L. Winterton, Entomol. Abhand. 61, 158 (2003). 7. E. G. MacLeod, A Comparative Morphological Study of the Head Capsule and Cervix of Larval Neuroptera (Insecta), Ph.D. Thesis (Harvard University, Cambridge, Massachusetts, 1964). 8. C. S. Henry, Psyche 85, 265 (1978). 9. M. W. Mansell, in Current Research in Neuropterology. Proc. Fourth Int. Symp. on Neuropterology, M. Canard, H. Aspöck, M. W. Mansell, Eds. (Privately printed, Toulouse, France, 1992), pp. 233–241. 10. D. Grimaldi, M. S. Engel, Evolution of The Insects (Cambridge University Press, Cambridge, UK, 2005). 11. R. Manconi, R. Pronzato, in Systema Porifera: A guide to the classification of sponges, J. N. A. Hooper, R. W. M. Van Soest, Eds. (Kluwer Academic/Plenum Publishers, New York, 2002), pp. 921–1021.

Crabs, shrimps, and lobsters (Decapoda) Keith A. Crandalla,b,*, Megan L. Porterc, and Marcos Pérez-Losadaa,d a

Department of Biology, Brigham Young University, Provo, UT 84602, USA; bMonte L. Bean Life Sciences Museum, Brigham Young University, Provo, UT 84602, USA; cDepartment of Biological Sciences, University of Maryland Baltimore County, Baltimore, MD 21250, USA; dCIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, Universidade do Porto, Campus Agrário de Vairão, 4485-661 Vairão, Portugal. *To whom correspondence should be addressed ([email protected])

Abstract Decapoda is the most diverse and species-rich Crustacean order. The 15,000 decapod species are organized in ~170 families, including ~20 known only from fossils. Their relationships are largely unresolved, but molecular phylogenies support the recognition of two suborders and seven infraorders of decapods. Molecular time estimates using multiple fossil and geological calibrations indicate that the first divergences among living decapods occurred in the early Silurian, ~430 million years ago (Ma). Diversification was rapid, resulting in lineages representing all infraorders by the Carboniferous, 325 Ma.

Crustaceans comprise the fourth most species-rich group of metazoans on the planet, following insects, chelicerates, and mollusks. But in terms of morphological diversity (disparity), they are unrivaled (see 1, 2). Foremost among the crustaceans in number and diversity are the decapods. With over 15,000 described species they include those crustaceans most familiar to the general public—shrimp, lobsters, crabs (Fig. 1), and crayfish— but also lesser known and unusual groups (3). The most recent classification (2) partitions ~62,000 species of extant Crustacea among 849 families (compared to 874 families for all 1.6 million species of insects). Approximately 152 of those extant families belong to the Decapoda with another 20 families known only from fossils, an enormous assemblage that has been called “the pinnacle of crustacean evolution.” More than six new families of decapods have been recognized since 2001 from both extant (e.g., 4, 5) and extinct (e.g., 6, 7) groups. Thus,

~18% of all described crustacean families belong to the decapods. Additionally, some 91 decapod families contain fossil taxa, including 27 known only from fossils (8). Decapods inhabit a broad diversity of ecological niches, including marine waters, deep-sea vents, estuaries, freshwater, caves, and terrestrial ecosystems. Accordingly, they are the subject of more published papers and controversy than all other crustacean groups combined, due in part to their species richness, economic importance, and ecological and morphological diversity (1). Indeed, the decapods have served as model organisms (including physiology, development, behavior, and morphology) for over a century (2). Decapoda is a clearly defined taxon that is generally regarded to be monophyletic within the Class Malacostraca. The decapods are usually divided into two suborders: Dendrobranchiata containing seven families and the more diverse Pleocyemata encompassing seven infraorders: Stenopodidea (two families), Caridea (37 families), Astacidea (seven families; clawed lobsters and crayfish), Thalassinidea (12 families; mud shrimp), Achelata, (five families; spiny lobsters), Anomala (16 families; hermit crabs, king crabs), and the Brachyura (93 families; true crabs). However, debate continues concerning general classification of the decapods (3–7) and specific arrangements of families within infraorders (see discussion in Martin and Davis (1)). Classification

Fig. 1 A land crab (Geocarcinus sp.) from Santa Cruz de Barahona, Dominican Republic. Credit: A. Sanchez.

K. A. Crandall, M. L. Porter, and M. Pérez-Losada. Decapod crustaceans (Decapoda). Pp. 293–297 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

294


21

Cambaridae

16

Parastacidae

9

Nephropidae

7

Palinuridae-1 18

Scyllaridae

13

Palinuridae-2

Achelata

Callianassidae 6

Thalassinidea

Cambaridae (Cambaroidinae)

Astacidea

Astacidae 22

Galatheidae 15

Chirostylidae

12

Lithodidae

10

Aeglidae 17

8

Anomala

5

Lomidae

4

Hippidae

Grapsidae

23

Cancridae

19

3

Portunidae

11

Brachyura

Potamidae 24

Palaemonidae 20

Atyidae Stenopodidae Penaeidae S

D

C

P

PALEOZOIC 400

300

Tr

J

K

CZ

MESOZOIC 200

Pg Ng

Stenopodidea

Hippolytidae

14

1

Dendrobranchiata

2

Caridea

Majidae

100

0 Million years ago

Fig. 2 A timetree of crabs, shrimps, and lobsters (Decapoda). Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Ng (Neogene), P (Permian), Pg (Paleogene), S (Silurian), and Tr (Triassic). Palinuridae-1 (Panulirus regius); Palinuridae-2 ( Jasus edwardsii).

within the Decapoda is highly unstable with further rearrangements proposed recently (4). Brusca and Brusca (8) summarize nicely the classification schemes for the Decapoda: “Rearrangement of the subtaxa within this order is a popular carcinological pastime, and the classification is far from being stable.” There are as many phylogenetic hypotheses concerning the relationships among the higher decapods as there are experts with opinions (6), with no consensus in sight. Historically, the decapod crustaceans were divided into two groups based on mode of locomotion: the Natantia

(“swimming” lineages) and the Reptantia (“crawling” lineages) (9). However, the “Natantia” was recognized early as paraphyletic and accordingly the Decapoda were reorganized into the Suborders Dendrobranchiata (penaeoid shrimp and their relatives) and Pleocyemata (all other decapods) by Burkenroad (10, 11). This taxonomic restructuring is supported by several defining morphological characters (i.e., dendrobranchiate gill structure and pleocyemate brooding of eggs on the female’s pleopods) and phylogenetic studies showing the “natant” decapods to be a paraphyletic assemblage

Eukaryota; Metazoa; Arthropoda; Crustacea; Decapoda

Table 1. Divergence times (Ma) and their credibility intervals (CI) among crabs, shrimps, and lobsters (Decapoda). Timetree Node


Time

Time

Ref. (12)(b)

CI

Time

CI

1

430.0

437

515–394

423

448–398

2

422.0

423

499–385

421

439–403

3

399.5

417

491–381

382

395–369

4

372.0

385

450–360

359

–

5

336.0

358

420–323

314

325–303

6

318.5

341

402–301

296

309–283

7

278.0

325

384–280

231

–

8

270.0

309

372–261

231

244–218

9

247.0

278

330–235

216

221–211

10

231.5

272

334–220

191

201–181

11

224.0

254

317–203

194

–

12

220.0

255

317–202

185

194–176

13

208.0

239

310–174

177

192–162

14

201.5

263

322–217

140

146–134

15

189.0

215

275–164

163

173–153

16

183.0

181

185–172

185

–

17

175.0

244

306–191

106

–

18

169.0

201

270–139

137

147–127

19

166.0

225

288–172

107

118–96

20

165.0

224

280–177

106

–

21

158.0

164

179–154

152

–

22

154.0

156

167–152

152

–

23

122.0

199

264–140

45

–

24

94.0

149

225–77

39

41–37

Note: Node times in the timetree represent the mean of time estimates from different studies. The divergence times estimated from the TK method incorporating calibrations as minimum ages (a) and the AHRS method using calibrations as fixed ages (b) are shown.

(12–14). Most of the studies investigating relationships among the major decapod lineages have been based on morphological characters, which due to the extreme diversity of form, makes it difficult to discern homologous relationships among features in a standard morphological analysis (15). Moreover, there has been surprisingly little molecular phylogenetic study of ordinal-level relationships in this group. Those molecular studies that have been completed have focused on only part of the order (i.e., the “Natantia”) and have not included adequate taxon sampling

295

within the Reptantia to evaluate the relationships of the major infraorders (12, 16). However, two recent phylogenetic studies have included all the infraorders within the Reptantia, but produced contradictory results. One based solely on molecular data, such as 16S mtDNA, 18S and 28S rDNA, and histone 3 (H3) (5) and the other combining molecular (16S mtDNA, 18S and 28S rDNA) and morphological (105 characters) data (4). Clearly, work must be done to reconcile these alternative views. Both studies lacked broad taxonomic sampling across the decapod families. A recent study introducing two new nuclear genetic markers for decapod systematics provides yet another view of relationships among the major lineages (17). In this latest study with more taxon sampling, Tsang et al. (17) found the Thalassinidea to be polyphyletic. They found the Stenopodidea and Caridea to form a clade, contrary to Porter et al. (5) and Ahyong and O’Meally (4). The remaining relationships are quite disparate amongst these three latest studies. Like most systematic controversy, this diversity of opinion stems from lack of sufficient characters coupled with inadequate taxon sampling and a disregard for the geologic history of the groups. The Porter et al. (5) study was the only study to include fossil calibration points and estimate divergence times among the major lineages. Of course, this may well be premature given the instability of relationships at the moment. Nevertheless, because this is the only study with calibration points and divergence time estimates, we take our divergence estimates from it. The decapod fossil record is continually being updated and reclassified, due to new discoveries of both fossil (18) and trace fossil evidence (19) and because many fossils are described from incomplete specimens causing uncertainty as to their phylogenetic affinities. Consequently, where possible, fossil references for this study were taken from species where descriptions were based on nearly complete specimens or where recent phylogenetic studies have placed fossil species relative to extant groups (20–23). Additionally, the fossils selected for calibration points were chosen based on the precision of the estimated date of the oldest known representative for particular clades, across several levels of divergence relative to the taxa sampling of our phylogeny. Based on these factors and the ages of fossils relative to their placement on the phylogeny, a set of seven fossils were used as calibrations in our analyses (5). Additionally, because the Bayesian method chosen for divergence time estimation (see later) requires at least one calibration to consist of an upper limit (maximum age), we set the split between

296


the crayfish Superfamilies Astacoidea and Parastacoidea as an upper limit of 185 Ma based on the splitting of Pangea (24). Decapoda divergence times were estimated using the Bayesian method of Thorne and Kishino (25) (referred to as TK) and the likelihood heuristic ratesmoothing algorithm (AHRS) of Yang (26). The decapod TK chronogram based on the single maximum likelihood topology and treating the calibration points as minimum or maximum ages places the origin of the Dendrobranchiata and Pleocyemata decapod lineages in the early Silurian (437 Ma; Fig. 2). This implies that the stem line of the decapods emerged even earlier; however, we are unable to estimate this age given our taxon sampling. Based on the molecular timescales, the radiation of the major decapod lineages occurred rapidly. The reptant lineage originated 385 Ma and all of the major reptant infraorders were present by the late Carboniferous, 60 million years later (Fig. 2, Table 1). The radiation of the extant taxa within each infraorder, however, occurred at different periods of time. The natant lineages have an early origin (423–417 Ma), however the caridean Superfamilies Alpheoidea, Atyoidea, and Palaemonoidea radiate in the early Permian (263 Ma). Among the Brachyuran superfamilies sampled, the Majoidea has the oldest lineage (254 Ma). The Achelata originate 341 Ma, with radiation of the extant lineages (Palinuridae and Scyllaridae) occurring as early as 239 Ma. The Thalassinidea appear 325 Ma, with the radiation of the Callianasoidea occurring at least 173 Ma. The anomalan lineage originated 309 Ma, with the extant superfamilies radiating between 309 and 244 Ma. The Astacidea lineage originated 325 Ma, with the divergence between the astacid lineages (Astacoidea, Parastacoidea) and the Nephropoidea occurring 278 Ma. Within the astacids, the radiation of the Parastacidae (~134 Ma) occurred earlier than the Astacidae (76 Ma) or the Cambaridae (90 Ma). The Nephropidae radiated as early as 140 Ma, with the genus Homarus appearing ~19 Ma. Although Porter et al. (5) estimated decapod divergence times without assuming a molecular clock and using multiple molecular markers and fossil calibration points, and these estimates appear to be concordant to a large degree with the decapod fossil record, these analyses come with a number of caveats. The first and most obvious concern is the instability of the phylogenetic estimate itself, upon which all the divergence time analyses are contingent. Given the recent and divergent studies on decapod phylogeny, it appears we are still far

from consensus on a stable phylogeny to base divergence estimates (3–5, 17). Alternative topologies would possibly generate different estimates for the derived nodes of the infraorders, but the two main conclusions of our analyses—that the Decapoda originated in the Silurian (437 Ma) and have experienced a fast radiation with all of the major infraorders present by the late Carboniferous (325 Ma)—would not change. Furthermore, the monophyletic Pleocyemata and the informal “Reptantia” are consistent in all hypotheses of decapod relationships, and therefore the divergence time estimates of these clades (423 and 385 Ma, respectively) can be used as common time points regardless of the particular arrangement of lineages. There are also inaccuracies associated with the fossil record that are not taken into account (27). These analyses assumed that the fossil ages are known with no error. Future advances in divergence time estimation methodologies could take advantage of the Bayesian framework to account for uncertainties in topology estimation and fossil dating and use different priors for rates and divergence times, as those included in Aris-Brosou and Yang (28). An extension of this Bayesian approach to include multiple genes and calibrations has recently been implemented (29). Rapid diversification and radiation is characteristic of the Crustacea as a whole (30), and this is a trend readily apparent in these divergence time estimates of decapod lineages (Fig. 2). Major decapod radiation events have been proposed to have occurred in the Eocene (Brachyura, 15), the Cretaceous (31), and the Triassic (macrurous forms, 15). The molecular-based divergence time estimates are older than hypotheses based solely on the fossil record, with the radiation of the “natant” infraorders occurring in the Devonian, the reptant infraorders in the Carboniferous (359–299 Ma), Anomalan diversification in the Permian–Triassic (299– 200 Ma), and the Callianassoidea and Palaemonoidea in the Cretaceous (146–66 Ma). As decapod paleontological research is a quickly expanding field of research (31), it will be most interesting to track the knowledge of decapod fossil date ranges relative to molecular-based divergence time estimations. Indeed, we hope that the above account will stimulate bringing together paleontological and evolutionary studies to shed further light on the divergence times of the decapod lineages.

Acknowledgment Support was provided by the U.S. National Science Foundation.

Eukaryota; Metazoa; Arthropoda; Crustacea; Decapoda

References 1. J. W. Martin, G. E. Davis, Nat. Hist. Mus. Los Angeles Co. Sci. Ser. 39, 1 (2001). 2. T. H. Huxley, The Crayfish: An Introduction to the Study of Zoology (D. Appleton, New York, ed. 1977, MIT Press reproduction, 1880). 3. G. Scholtz, S. Richter, Zool. J. Linn. Soc. 113, 289 (1995). 4. S. T. Ahyong, D. O’Meally, Raffles Bull. Zool. 52, 673 (2004). 5. M. L. Porter, M. Perez-Losada, K. A. Crandall, Mol. Phylogenet. Evol. 37, 355 (2005). 6. F. R. Schram, Hydrobiologia 449, 1 (2001). 7. P. K. L. Ng, D. Guinot, P. J. F. Davie, Raffles Bull. Zool. 17, 1 (2008). 8. R. C. Brusca, G. J. Brusca, Invertebrates, 2nd ed. (Sinauer, Sunderland, MA, 2002). 9. J. E. V. Boas, Danske vidensk. Selsk. Skr. (Nat.) 6, 25 (1880). 10. M. D. Burkenroad, Tulane Stud. Geol. 2, 1 (1963). 11. M. D. Burkenroad, Trans. San Diego Soc. Nat. Hist. 19, 251 (1981). 12. L. G. Abele, Mem. Queensland Mus. 31, 101 (1991). 13. L. G. Abele, B. E. Felgenhauer, J. Crust. Biol. 6, 385 (1986). 14. B. E. Felgenhauer, L. G. Abele, Crustacean Issues 1, 291 (1983).

15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

297

F. R. Schram, Crustacea (Oxford University Press, New York, 1986). W. Kim, L. G. Abele, J. Crust. Biol. 10, 1 (1990). L. M. Tsang, K. Y. Ma, S. T. Ahyong, T.-Y. Chan, K. H. Chu, Mol. Phylogenet. Evol. 48, 359 (2008). A. J. Martin et al., Gondwana Res. 14, 287 (2008). E. Bedatou, R. Melchor, E. Bellosi, J. F. Genise, Palaeogeogr. Palaeclimatol. Palaeoecol. 257, 169 (2008). F. R. Schram, C. J. Dixon, Smithsonian Contrib. Zool. 72 (2003). D. Tshudy, J. Crust. Biol. 23, 178 (2003). A. L. Rode, L. E. Babcock, J. Crust. Biol. 23, 418 (2003). L. Amati, R. M. Feldmann, J.-P. Zonneveld, J. Paleo. 78, 150 (2004). K. A. Crandall, D. J. Harris, J. W. Fetzner, Proc. Roy. Soc. Lond. B 267, 1679 (2000). J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). Z. Yang, Acta Zool. Sinica 50, 645 (2004). D. Graur, W. Martin, Trends Genet. 20, 80 (2004). S. Aris-Brosou, Z. Yang, Syst. Biol. 51, 703 (2002). A. J. Drummond, W. Y. H. Simon, J. P. Matthew, A. Rambaut, PLoS Biol. 4, 699 (2006). F. R. Schram, R. M. Feldmann, M. J. Copeland, J. Paleo. 52, 1375 (1978). R. M. Feldmann, J. Paleo. 77, 1021 (2003).

Stalked and acorn barnacles (Thoracica) Marcos Pérez-Losadaa,b, Jens T. Høegc, and Keith A. Crandallb,d,* a

CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, Universidade do Porto, Campus Agrário de Vairão, 4485-661 Vairão, Portugal; bDepartment of Biology, Brigham Young University, Provo, UT, 84602-5181, USA; cComparative Zoology and Department of Biology, Institute of Biology, University of Copenhagen, Universitetsparken 15, DK-2100, Copenhagen, Denmark; dMonte L. Bean Life Sciences Museum, Brigham Young University, Provo, UT 84602, USA. *To whom correspondence should be addressed (keith_crandall@ byu.edu)

Abstract The Superorder Thoracica is the most diverse group of barnacles, Class Maxillopoda, and contains over 1000 species in nearly 30 families. Fossil Thoracica are known from the late Carboniferous and early Permian (310–290 million years ago, Ma), while the earliest cirripeds are known from the Silurian (443–416 Ma). Molecular divergence times place the divergence of the orders of Thoracica in the early Carboniferous (340 Ma). The suborders of the polyphyletic Order Pedunculata further diversified between the early Permian (287 Ma) and the early Jurassic (198 Ma). The Order Sessilia, excluding the Brachylepadomorpha, appeared in the late Jurassic (147 Ma).

Barnacles are a member of a non-monophyletic assemblage of crustaceans including (sometimes) ostracods, copepods, barnacles, and other assorted crustaceans. Within the Class Maxillopoda, the most recent classification places the barnacles into the Subclass Thecostraca (1). Within the Thecostraca, they are further subdivided into three Infraclasses, namely the Facetotecta (with a single family), the Ascothoracida (with six families), and the Cirripedia (containing the bulk of the barnacle diversity). Within the Cirripedia (e.g., Fig. 1), there are three superorders, the boring Acrothoracica, the parasitic Rhizocephala, and the stalked and sessile Thoracica. It is the Thoracica (the stalked and acorn barnacles) that we concentrate on in this chapter with representatives from the Rhizocephala as outgroups. The Thoracica contains two orders: the stalked Pedunculata (14 families) and the

sessile Sessilia (15 families). Note that this current classification is in the midst of some flux as new phylogenetic studies come to light that bare directly on the classification of barnacles (2, 3). The Thoracica, ordinary or true barnacles, is the most diverse group of barnacles with over 1000 species found in virtually all marine and estuarine environments from intertidal pools to abyssal vents. They deviate from all other Crustacea in being permanently and irreversibly attached suspension feeders that have abandoned the normal arthropod growth pattern by being armed externally with mineralized plates that are never shed in molts but increase incrementally in size (4). Barnacles were in many respects the first model organism in evolutionary biology as reflected in Darwin’s work (5–8). Their very specialized morphologies, diverse habitats,

Fig. 1 A goose barnacle (Lepas anatifera) from Taiwan showing the cirri or feeding legs that define the infraclass Cirripedia. Credit: J. T. Høeg, B. B. Chan, and C.-H. Hsu.

M. Pérez-Losada, J. T. Høeg, and K. A. Crandall. Stalked and acorn barnacles (Thoracica). Pp. 298–301 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Neoverrucidae (Verrucomorpha-1) 14

Eolepadidae (Pedunculata-2)

6

Scalpellidae Tetraclitoidea 18

Balanoidea-1

16

Coronuloidea

17

Balanoidea-2 Chthamaloidea

11

2

Verrucidae (Verrucomorpha-2)

7

Lithotryidae

5 1

Pollicipedidae Calanticidae (Pedunculata-3) Iblidae (Iblomorpha)

C

P

Triassic Jurassic

PALEOZOIC 300

Cretaceous

Pg

MESOZOIC 250

200

150

Pedunculata-3

12

Scalpellomorpha-2

3

Scalpellomorpha-1

Heteralepadidae-2

4

Pedunculata-1

Lepadidae (Lepadomorpha-2) 9

Sessilia-1 Pedunculata-2

Oxynaspididae

8

299

Sessilia-2

Poecilasmatidae

10

Heteralepadomorpha-2

Heteralepadidae-1

13

Balanomorpha

Poecilasmatidae (Lepadomorpha-1) 15

Heteralepadomorpha-1

Eukaryota; Metazoa; Arthropoda; Crustacea; Thoracica

Ng

CZ 100

50

0 Million years ago

Fig. 2 A timetree of stalked and acorn barnacles (Thoracica). The divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), Ng (Neogene), P (Permian), and Pg (Paleogene). Heteralepadidae-1 (Heteralepadomorpha), Heteralepadidae-2 (Paralepas palinuri, Paralepas dannevigi), Balanoidea-1 (Austrominius modestus, Elminius kingie), Balanoidea-2 (Austromegabalanus psittacus, Balanus balanus, Balanus crenatus, Balanus glandula, Balanus perforates, Megabalanus californicus, Megabalanus tintinnabulum, Megabalanus spinosus, Menesiniella aquila, Semibalanus balanoides, Semibalanus cariosus).

and reproductive systems made them excellent for testing and honing his ideas on biological evolution (9, 10). Barnacles have retained the attention of biologists ever since. They are important members of many marine habitats, such as the rocky intertidal zone, and their sessile mode of life makes them the primary fouling objects on man-made structures in the sea (11). However, there is still a pervasive lack of phylogenetic information for a group that has been the focus of intense study for almost two centuries (12). A robust phylogeny is therefore pivotal in understanding how barnacles have evolved and diversified from a more conventional ancestor and also how experimental studies on single species, such as in antifouling research, can be extended to larger groups. Many hypotheses concerning barnacle evolution have been proposed (4, 13–21), but Glenner et al. (22)

were the first to apply cladistic approaches to determine their interrelationships. The morphology-based analysis of Glenner et al. (22) and the subsequent molecular studies (2, 3, 23–25) differed in their conclusions from each other and from existing taxonomies (1, 4, 14, 26). PérezLosada et al. used a more extensive taxon sampling and both molecular and morphological data sets for a thorough reassessment of evolutionary relationships (2). Timing the radiation of the main barnacle groups based on their extinct relatives has always provoked great interest. Fossils and evolutionary hypotheses have been combined previously (22, 26, 27, 28), but only two studies (2, 3) have integrated both within a statistical framework. In this study, phylogenetic procedures of time estimation and three fossil calibration points were used to date the radiation of the main thoracican clades using 18S ribosomal

300


Table 1. Divergence times (Ma) and their credibility/ confidence intervals (CI) among stalked and acorn barnacles (Thoracica), based on ref. (2). Timetree Node

Time

CI

1

340.0

415–310

2

287.0

351–241

3

273.0

333–228

4

252.0

310–207

5

239.0

295–195

6

238.0

294–194

7

198.0

242–162

8

187.0

247–139

9

177.0

236–130

10

157.0

215–109

11

147.0

153–130

12

135.0

150–114

13

135.0

190–91

14

131.0

196–73

15

116.0

170–73

16

113.0

137–89

17

103.0

128–78

18

97.0

122–73

Note: Calibrations ranged from the Carboniferous (M. Pennsylvanian, 312–307 Ma) to the Neogene (M.-L. Miocene, 23–12 Ma).

DNA sequences. However, divergence time estimates based on a single gene can be biased (29, 32). The time estimate will be biased if the phylogeny used is incorrect (33). Unbiased and accurate divergence time estimates can be obtained by integrating multiple gene loci and multiple fossil calibration points into a robust phylogeny (29, 30, 32). Therefore, Thoracica divergence times generated using the Bayesian method of Thorne and Kishino (T-K) (29) from multiple gene regions and multiple calibration points (including maximum and/or minimum age constrains) are expected to be the most reliable (2). In that study (2), 14 fossil calibrations points were used to anchor local molecular clocks (Table 1). Previous calibrations used in Pérez-Losada et al. (3) for the Heteralepadomorpha (Priscansermarinus, M. Cambrian) and the Scalpellomorpha (Pabulum, L. Carboniferous) were not used in this study, because their barnacle affinities have been questioned and because these groups were not monophyletic in our new phylogenetic analysis. Pabulum is now considered to be a

bivalve mollusk (Martin Whyte, personal communication) and there is only feeble evidence for a cirriped origin of Priscansermarinus (34, 35). Given that most fossils are dated to an age range, the midpoint of each geologic range was chosen in the divergence time estimation. A likelihood ratio test significantly (P < 0.001) rejected the null hypothesis that all genes, separately or combined, were evolving with equal rates across all lineages, requiring the use of methods that relax the molecular clock hypothesis to estimate divergence times. The thoracican T-K chronogram was estimated using the Bayesian phylogeny, three genes, and 14 fossil calibrations (Fig. 2). Multiple independent Bayesian runs produced identical mean time estimates for all the major clades, including 95% credibility intervials (CI) (Table 1). The 95% CI were large for most clades, because we used only one upper limit (Pycnolepas rigida). However, as previously shown (32), incorporating both lower and upper constraints during time estimation can reduce the standard deviation of the estimates. This analysis places the origin of the Thoracica suborders in the early Carboniferous (340 Ma). The suborders of the polyphyletic Pedunculata radiated between the early Permian (287 Ma) and the early Jurassic (198 Ma). The Sessilia (excluding the Brachylepadomorpha) appeared in the late Jurassic (147 Ma). These estimates have reasonable corroboration with the fossil record. For example, the presumptive iblomorph Illilepas damrowi (36) from the Carboniferous (359–299 Ma) (although significantly later than previous calibrations, 28) would agree reasonably well with our molecular estimates. Clearly, this timetree holds many implications for the evolution of barnacle morphology (2).

Acknowledgments This research was supported by U.S. National Science National Science Foundation (K. A. C. and M. P.-L.); the U.S.-Israel Binational, Scientific Fund (K. A. C.); and the Danish Natural Science Research Council and European Union Framework, COBICE and SYNTHESYS programs (J. T. H.).

References 1. J. W. Martin, G. E. Davis, Nat. Hist. Mus. Los Angeles Co. Sci. Ser. 39, 1 (2001). 2. M. Pérez-Losada et al., Mol. Phylogenet. Evol. 46, 328 (2008). 3. M. Pérez-Losada, J. T. Høeg, K. A. Crandall, Syst. Biol. 53, 244 (2004).

Eukaryota; Metazoa; Arthropoda; Crustacea; Thoracica

4.

5.

6.

7.

8.

9. 10. 11.

12.

13. 14.

15.

16.

D. T. Anderson, Barnacles: Structure, Function, Development and Evolution (Chapman & Hall, London, 1994). C. Darwin, A Monograph of the Fossil Lepadidae or, Pedunculated Cirripedes of Great Britain (Paleontological Society, London, 1851). C. Darwin, A Monograph of the Sub-class Cirripedia, with Figures of all the Species. The Lepadidae: Or, Pedunculated Cirripedes (Ray Society, London, 1852). C. Darwin, A Monograph of the Sub-class Cirripedia, with Figures of all the Species. The Balanidae (or Sessile Cirripedes); the Verrucidae, etc. (Ray Society, London, 1854). C. Darwin, A Monograph of the Fossil Balanidae and Verrucidae of Great Britain (Paleontological Society, London, 1855). M. T. Ghiselin, The Triumph of the Darwinian Method (University of California Press, Berkeley, 1969). J. T. Høeg, O. S. Møller, Invertebr. Reprod. Dev. 49, 125 (2006). M.-F. Thompson, R. Nagabhushanam, Barnacles: The Biofoulers (Vedams eBooks Ltd., New Delhi, India, 1999). F. R. Schram, J. T. Høeg, in Crustacean Issues: New Frontiers in Barnacle Evolution, F. R. Schram, J. T. Høeg, Eds. (A. A. Balkema, Rotterdam, The Netherlands, 1995), vol. 10, pp. 297–312. W. A. Newman, Trans. San Diego Soc. Nat. Hist. 19, 153 (1979). W. A. Newman, in Barnacle Biology, A. J. Southward, Ed. (A. A. Balkema, Rotterdam, The Netherlands, 1987), Vol. 5, pp. 3–42. W. A. Newman, A. Ross, “Revision of the Balanomorph Barnacles; including a Catalog of the Species,” Tech. Report Memoir 9 (San Diego Society of Natural History, 1976). J. S. Buckeridge, W. A. Newman, J. Paleontol. 66, 341 (1992).

17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

28. 29. 30. 31. 32. 33. 34.

35. 36.

301

J. M. Healy, D. T. Anderson, Rec. Austral. Mus. 42, 1 (1990). T. Yamaguchi, Pacific Sci. 44, 135 (1990). G. A. Kolbasov, Arthropoda Selecta 5, 3 (1996). W. A. Newman, T. Yamaguchi, Bull. Mus. Natl. Hist. Nat. 4 (ser. A 17), 211 (1995). W. A. Newman, in Traité de Zoologie, J. Forest, Ed. (Masson, Paris, 1996), Vol. 7, pp. 453–540. H. Glenner, M. J. Grygier, J. T. Høeg, P. G. Jensen, F. R. Schram, Zool. J. Linnean Soc. 114, 365 (1995). D. J. Harris, L. Maxson, L. F. Braithwaite, K. A. Crandall, J. Crust. Biol. 20, 393 (2000). R. Perl-Treves, L. Mizrahi, D. J. Katcoff, Y. Achituv, J. Crust. Biol. 20, 385 (2000). T. Spears, L. G. Abele, M. A. Applegate, J. Crust. Biol. 14, 641 (1994). W. A. Newman, in Traité de Zoologie 7, J. Forest, Ed. (Masson, Paris, 1996), Vol. 2, pp. 453–540. B. A. Foster, J. S. Buckeridge, in Barnacle Biology, A. J. Southward, Ed. (A. A. Balkema, Rotterdam, The Netherlands, 1987), Vol. 5, pp. 43–61. J. S. Buckeridge, W. A. Newman, Zootaxa 1136, 1 (2006). J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). Z. Yang, Acta Zool. Sinica 50, 645 (2004). Z. Yang, A. D. Yoder, Syst. Biol. 52, 705 (2003). M. L. Porter, M. Perez-Losada, K. A. Crandall, Mol. Phylogenet. Evol. 37, 355 (2005). A. J. Drummond, S. Y. W. Ho, M. J. Phillips, A. Rambaut, PLoS Biol. 4, e88 (2006). D. E. G. Briggs, in Crustacean Phylogeny, F. R. Schram, Ed. (A. A. Balkema, Rotterdam, The Netherlands, 1983), Vol. 1, pp. 1–22. D. E. G. Briggs, M. D. Sutton, D. J. Siveter, D. J. Siveter, Proc. Roy. Soc. Lond. B 272, 2365 (2005). F. R. Schram, Crustacea (Oxford University Press, New York, 1986), pp. 606.

Sea urchins (Echinoidea) Andrew B. Smith Department of Palaeontology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK ([email protected])

Abstract The Echinoidea (sea urchins) is one of the five classes of the Phylum Echinodermata and includes ~900 living species in 50 families. Their traditional taxonomy, based on skeletal characters, has been largely corroborated by recent molecular phylogenetic analyses with one marked exception: clypeasteroids are not found to be monophyletic. The echinoid timetree deduced from molecular data is largely concordant with the fossil record, placing the basal divergence in the late Paleozoic (265 million years ago, Ma). Echinoids diversified during the Mesozoic (251–66 Ma) and there is a good match between paleontological and molecular estimates of divergence times, with clypeasteroids again proving to be an exception.

The Phylum Echinodermata is a clade of marine invertebrate deuterostomes that includes such well-known animals as the starfishes and sea urchins. All echinoderms possess a calcitic endoskeleton with a distinctive and unique three-dimensional structure, a stereom, and they all undergo an unusual asymmetrical development in which right larval coelomic components are suppressed and lost. There are five living classes of echinoderm, of which the Echinoidea or sea urchins (Fig. 1) is probably the best known and certainly the one that has left the most complete fossil record. Living echinoids have a mesodermal skeleton constructed of 10 columns of plates, all of which bear tubercles and spines. The modern taxonomy of echinoids was established by Mortensen (1), based primarily on the detailed arrangement of plates making up the skeleton. This has the great advantage of allowing fossils to be placed with confidence into any taxonomic scheme constructed for the living species. About 900 living species of echinoids have been described and placed in ~50 families (1, 2), not all of which are considered monophyletic. Here, I review the relationships and divergence times of the major echinoid groups. It has long been recognized that the cidaroids differ in several fundamental ways from other echinoids

(Euechinoidea), and this division is reflected at the subclass level (1, 2). Cidaroids and euechinoids differ in their style of ambulacral plating, and have different jaw apparatus morphologies and musculature. Whereas cidaroids remained rather conservative in their morphology, the euechinoids have given rise to a wide diversity of forms (3), currently classified into 13 orders. There are a number of small, early branching groups but about 80% of the modern diversity lies in two major clades, the Irregularia and the Echinacea. Irregular echinoids are so named because their pentameral symmetry is disrupted by migration of the anus from an apical to a posterior position during ontogeny, and most live infaunally. Echinacea are regular echinoids with a derived lantern morphology and keeled teeth and all are epifaunal. The classification of the Echinacea has been particularly difficult and relies on small differences rarely preserved in fossils (4). The taxonomy of irregular echinoids, on the other hand, has been much less problematic with the long-standing major groups (spatangoids, holasteroids, clypeasteroids, and cassiduloids) readily differentiated

Fig. 1 An echinoid (Arbacia punctulata) from Carrie Bow Cay, Belize, viewed from above. Credit: A. B. Smith.

A. B. Smith. Sea urchins (Echinoidea). Pp. 302–305 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Echinodermata; Echinoidea

303

12

Toxopneustidae Strongylocentrotidae

20

Echinidae

19

7

Echinacea

Genocidaridae 26

Camarodonta

Temnopleuridae 15

Echinometridae Stomopneustidae

11

Arbaciidae

5

Brissidae Paleopneustidae

9

18

Schizasteridae Astriclypeidae

17

8

Mellitidae

14

2

Laganidae 22

13

Echinocyamidae Echinolampadidae

6

10

23

Cassidulidae Arachnoididae

1

Cassiduloida Clypeasteroida

Plexechinidae

Acroechinoidea

Archaeopneustes 25

16

Loveniidae

Irregularia

21

Spatangoida

Spatangidae 24

Echinoneidae Diadematidae 3

Aspidodiadematidae 4

Pedinidae Echinothuriidae

Triassic 250

Jurassic Cretaceous MESOZOIC 200

150

100

Pg Ng CENOZOIC 50

0 Million years ago

Fig. 2 A timetree of sea urchins (Echinoidea). Divergence times are from Table 1. Abbreviations: Ng (Neogene) and Pg (Paleogene).

on morphological grounds (1, 2). The monophyly of the clypeasteroids, a group characterized by the unique synapomorphy of multiple tube feet (and pores) on each ambulacral plate, has never been disputed from morphological grounds (5), though the cassiduloids, from which they emerged, are now recognized to be a paraphyletic grade (6). Molecular phylogenies for the Echinoidea that encompassed a number of different families started to appear from 1992 onward (6–10). These have all been constructed

from nuclear and mitochondrial ribosomal RNA genes (18S, 28S, and 16S rRNA), sometimes with the addition of sequence data from three subunits of mitochondrial cytochrome oxidase genes (COI, COII, and COIII). The first studies achieved only a very sparse taxonomic coverage and failed to find convincing evidence for the cidaroid–euechinoid basal dichotomy (6). By 1995, however, the first analysis that could claim reasonable taxonomic coverage appeared (7) and showed good correspondence with morphology-based phylogenetic trees. In the latest

304


Table 1. Divergence times (Ma) and their confidence/ credibility intervals (CI) among sea urchins (Echinoidea), based on ref. (10). Timetree Node

Time

1

245

2

232

3

221

4

210

5

193

6

181

7

171

8

164

9

160

10

138

11

137

12

111

13

109

14

98

15

97

16

95

17

86

18

72

19

65

20

61

21

54

22

41

23

37

24

36

25

33

26

28

Note: Molecular dates are the means of estimates obtained from the analysis of concatenated 18S rRNA, 28S rRNA , and COII partial gene sequences using different methodologies: LF (Langley–Fitch), NPRS (nonparametric rate smoothing), PL-A (penalized likelihood with additive penalty function, PL-L (penalized likelihood with logarithmic penalty function), and Bayesian.

analysis (10), the molecular phylogeny is now based on gene sequence data from almost 50 taxa, with representatives from 13 of the 14 extant orders. In addition to these studies addressing the higher-level relationships of echinoids, detailed molecular phylogenies have appeared outlining the phylogenetic relationships of specific groups at

genus level (e.g., temnopleuroids (11), spatangoids (12), strongylocentrotids (13)] and species level [e.g. Eucidaris (14), Diadema (15)). From the beginning, phylogenetic studies have often analyzed morphological and molecular data in parallel and in combination, and have used the rich fossil record for dating divergences. The basic taxonomic framework for echinoids established from skeletal morphology has stood up well to this molecular scrutiny (1). Cidaroids consistently turn out to represent the deepest branch in the echinoid tree and echinothurioids the deepest branch on the euechinoid side, exactly as predicted by morphology. The next few branches are very closely spaced and branching order of pedinoids and diadematoids is not clear. There is a monophyletic Irregularia, within which the echinoneid Echinoneus represents the basal branch, and holasteroids and spatangoids are closest relatives. There is one major surprise—molecular data suggest that clypeasteroids are not monophyletic. The two suborders (Clypeasterina and Scutellina) are recognized, but they are not identified as closest relatives. Instead, representatives from two families of cassiduloid are the closest relatives of the Scutellina. The very short branches leading to the cassiduloid taxa suggest that this is not a long-branch attraction problem, and the inferred relationships are robust to addition or removal of taxa. However, it is hard to reconcile this observation with the strong morphological evidence for clypeasteroid monophyly. Only one study has estimated divergence times among echinoid families from molecular data (10). This study examined 26 internal nodes and compared molecular estimates based on ribosomal gene divergence with paleontological estimates (Fig. 2). In order to generate a semilinearized tree one taxon was selected from each family, avoiding extremely long or short terminal and branches. Bayesian and nonparametric rate smoothing semiparametric penalized likelihood methods were all used for estimating divergence times along with the Langley– Fitch strict clock method, and error bars calculated. A selection of taxa from the other echinoderm classes formed the outgroup, with a prior depth of the root node set at 480 My based on the fossil record. Four internal calibrations were set as minimal divergence times from across the tree topology to provide constraints on local rate variation. Molecular estimates of divergence times derived from applying both molecular clock and relaxed molecular clock models are concordant with estimates based on the fossil record for 70% of the nodes. Mismatch is confined to three areas of the tree, the most serious of which

Eukaryota; Metazoa; Echinodermata; Echinoidea

concerns the clypeasteroids, where a late Jurassic divergence (156 ± 24 Ma) for Clypeasterina from Scutellina was predicted from molecular data. In contrast, the fossil record provides no evidence for any clypeasteroid before 60 Ma (Middle Paleocene) (2, 16). The fact that so much of the molecular phylogeny matches what is known from morphology and the fossil record is encouraging, and should allow for a more confident integration of data. For example, the divergence of many of the basal euechinoid clades apparently occurred in the Triassic (251–200 Ma), during the very earliest stages of the breakup of Pangea (17) as marine conditions started to spread over the continental shelves after a major sea-level low stand (18). By comparative analysis it is possible to show that the mismatch between molecular clock and paleontological estimates of divergence increases as the marine rock record deteriorates in quality (17).

Acknowledgment Support was provided by a grant from the Leverhulme Foundation.

References 1.

T. Mortensen, A Monograph of the Echinoidea, 5 volumes (C. A. Reitzel, Copenhagen, 1928–1951).

305

2. A. B. Smith, The Echinoid Directory, http://www.nhm. ac.uk/research-curation/projects/echinoid-directory (The Natural History Museum, London, 2006). 3. A. B. Smith, in Evolving Form and Function—Fossils and Development, D. E. G. Briggs, Ed. (Yale University Press, New Haven, 2005), pp. 181–194. 4. A. B. Smith, Mol. Biol. Evol. 5, 345 (1988). 5. R. Mooi, Paleobiology 16, 25 (1990). 6. A. B. Smith, Paleobiology 27, 392 (2001). 7. A. B. Smith, B. Lafay, R. Christen, Phil. Trans. Roy. Soc. Lond. B 338, 365 (1992). 8. D. T. J. Littlewood, A. B. Smith, Phil. Trans. Roy. Soc. Lond. B 347, 213 (1995). 9. A. B. Smith, D. T. J. Littlewood, G. A. Wray, Phil. Trans. Roy. Soc. Lond. B 349, 11 (1996). 10. A. B. Smith, D. Pisani, J. A. Mackenzie-Dodds, B. Stockley, B. L. Webster, D. T. J. Littlewood, Mol. Biol. Evol. 23, 1832 (2006). 11. C. H. Jeffery, R. B. Emlet, D. T. J. Littlewood, Mol. Phylogenet. Evol. 28, 99 (2003). 12. B. Stockley, A. B. Smith, D. T. J. Littlewood, H. A. Lessios, J. A. MacKenzie-Dodds, Zool. Scripta 34, 447 (2005). 13. Y.-H. Lee, Mol. Biol. Evol. 20, 1211 (2003). 14. H. A. Lessios, B. D. Kessing, D. R. Robertson, G. Pauley, Evolution 53, 806 (1999). 15. H. A. Lessios, B. D. Kessing, J. S. Pearse, Evolution 55, 955 (2001). 16. P. M. Kier, Palaeontology 25, 1 (1982). 17. A. B. Smith, Paleobiology 33, 311( 2007). 18. A. B. Smith, A. J. McGowan. Palaeontology 50, 765 (2007).


VERTEBR ATES


Vertebrates (Vertebrata) S. Blair Hedges Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA ([email protected])

Abstract The vertebrates (~58,000 sp.) comprise a phylum of mostly mobile, predatory animals. The evolution of jaws and limbs were key traits that led to subsequent diversification. Atmospheric oxygen change appears to have played a major role, with an initial rise in the late Precambrian (~580–542 million years ago, Ma) permitting larger body size, followed by two Paleozoic pulses affecting prey. The First Pulse (~430–390 Ma) brought fishes to brackish and freshwater environments where they diversified, with one lineage giving rise to tetrapods. The Second Pulse (~340–250 Ma) led to a Permo-Carboniferous explosion of tetrapods, adapting to diverse terrestrial niches.

The Phylum Vertebrata includes ~58,000 living species in seven evolutionary clades that diverged in the latest Precambrian and in the Paleozoic Era, ~600–360 Ma. Approximately one half of the species (28,183 sp.) are fishes—not an evolutionary group—and the other half (29,638 sp.) are tetrapods, which comprise a monophyletic group. Vertebrates (Fig. 1) are mobile animals that possess a cranium (skull) and, at least ancestrally, a backbone consisting of vertebrae protecting the nerve cord (1–4). The cyclostomes (~85 sp.) are jawless fishes (agnathans) and include lampreys and hagfishes. Chondrichthyans (~1200 sp.) are the cartilaginous fishes and include the sharks, rays, and chimaeras. Actinopterygians (26,890 sp.) are ray-finned fishes and include bichirs, sturgeons, paddlefishes, gars, bowfins, and teleosts, with the latter group comprising nearly all species of actinopterygians. Actinistia (two sp.), which is alternatively called Coelacanthimorpha, includes lobe-finned fishes (coelacanths). Dipnoans (six sp.) are the lungfishes. Lissamphibians (6200 sp.) are the living amphibians and include the frogs and toads, salamanders, and caecilians. Amniota (23,438 sp.) comprises the mammals, tuataras, squamates (lizards, snakes, and amphisbaenians), turtles, crocodilians, and birds). Here, the relationships and divergence times of these major lineages of vertebrates are reviewed.

Vertebrates are treated here as a separate phylum rather than a subphylum of Chordata. The morphological disparity among the chordates (urochordates, cepahalochordates, and vertebrates), and their deep time of separation based on molecular clocks (5) is as great as that among other groups of related animal phyla (e.g., arthropods, tardigrades, and onycophorans). The phylogeny of the lineages covered here is uncontroversial, for the most part. Evidence from nuclear genes and morphology (1, 2, 6, 7) agree in the backbone phylogeny of vertebrates represented by these nested groups: Tetrapoda (Lissamphibia, Amniota), Sarcopterygii (Actinistia, Dipnoi, Tetrapoda), Osteichthyes (Actinopterygii, Sarcopterygii), and Gnathostomata (Chondrichthyes, Osteichthyes). Cyclostomata was originally considered a basal, monophyletic group based on morphology (8), but later morphological analyses placed lampreys as closest relatives of gnathostomes (9–11). However, molecular phylogenies from many genes since the early 1990s have consistently supported cyclostome monophyly (6, 7, 12–17) and therefore this basal branch is a classical example of

Fig. 1 Representative vertebrates. (A) A clownfish, Amphiprion ocellaris, from East Timor (upper left); (B) A lizard, Anolis allisoni, from Cuba (upper right); (C) a frog, Eleutherodactylus portoricensis, from Puerto Rico (lower left), and a white ibis, Eudocimus albus, from the Dominican Republic (lower right). Credits: N. Hobgood (upper left), E. Fernandez (upper right), and A. Sanchez (lower left and right).

S. B. Hedges. Vertebrates (Vertebrata). Pp. 309–314 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Amniota 5

Lissamphibia

4

Tetrapoda


Dipnoi

3

Actinistia

2

Sarcopterygii

310

Actinopterygii Chondrichthyes

1

Cyclostomata Np

Cm O S

PR 600

D

C

P

PALEOZOIC 500

400

Tr

J

K

MESOZOIC 300

200

100


Fig. 2 A timetree of vertebrates. Times of divergence are averages of estimates from different studies listed in Table 1. Abbreviations: C (Carboniferous), Cm (Cambrian), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Np (Neoproterozoic), O (Ordovician), P (Permian), Pg (Paleogene), PR (Proterozoic), S (Silurian), and Tr (Triassic).

apparent conflict between morphological and molecular data. The only currently unresolved portion of the vertebrate backbone tree, from the molecular perspective, involves the relationships among coelacanths, lungfishes, and tetrapods. This question has been pursued vigorously in many morphological and molecular studies over the last several decades, without full resolution, although a close relationship between lungfishes and tetrapods is currently favored among paleontologists (18) and has been the most frequent and recent result in molecular studies (e.g., 8, 19–27). The fossil record of vertebrates is more complete than that of most organisms, primarily because of the durability of their bony endoskeleton (in most groups) which includes a mineral, hydroxyapetite. Fossil vertebrates were important for Darwin in development of his ideas on evolution (28) and even, before that, were used by Cuvier to establish the concept of extinction. Two large compendia (29, 30) provide an overview of the vertebrate fossil literature up to ~1990, and there are several more recent texts (1–3). The earliest vertebrate fossils are of agnathans, resembling cyclostomes, from the Lower Cambrian of China (~525 Ma) (31). The first jawed vertebrates (gnathostomes) to appear in the fossil record are—controversially—conodonts in the late Cambrian (~500 Ma), although probably not closest relatives of any living lineage (11). Alternatively, the earliest jawed vertebrates were acanthodians from the mid-Ordovician (~461 Ma) (32). However, vertebrate diversity was low until the late Silurian and early Devonian (~430–380 Ma), during which time all of the remaining living groups of fishes first appear including chondrichthyans,

actinopterygians, actinistians, and lungfishes (1, 33, 34). One of the best-documented evolutionary transitions is recorded in Devonian fossils (~390–360 Ma), linking one group of lobe-finned fish to tetrapods (8, 24, 35–39). Fossils from the Carboniferous (359–299 Ma) and Permian (~299–251) show a great diversification of landdwelling vertebrates, including the first lissamphibians and amniotes (~330 Ma) (32, 40) as well as many extinct groups. Only two molecular clock studies have been published that have estimated all, or nearly all, of the nodes in the backbone tree of major vertebrate groups. The first involved analyses of nuclear protein-coding genes, 13–107 genes depending on the node, screened for lineage-specific rate variation by relative rate tests, and used two calibration points (41). The resulting time estimates for the split of Lissamphibia and Amniota (360 Ma) and the split of Actinopterygii and Sarcopterygii (450 Ma) were only 20–35 million years older (5–8%) than fossil record estimates for those divergences (32, 42–44). Fewer fossil constraints have been available for the earlier divergences, between Chondrichthyes and Osteichthyes, and between Agnatha and Gnathostomata (45), and therefore the estimates for those splits—528 and 564 Ma—are older than the fossil record suggests. A more recent study used a Bayesian relaxed clock method and 48–325 genes depending on the node (6). Similar times for the Lissamphibia–Amniota (370 Ma), Actinopterygii–Sarcopterygii (476 Ma), and Chondrichthyes–Osteichthyes (525 Ma) divergences were obtained, but an older time (652 Ma) was estimated for the Agnatha–Gnathostomata divergence. Also, the

Eukaryotes; Metazoa; Vertebrata

311

Table 1. Divergence times (Ma) among vertebrates. Timetree Node

Estimates Ref. (6)

Time Time

Ref. (41) CI

Time

Ref. (46) CI

Ref. (47)

Time

CI

Time

CI

1

608

652

742–605

564

710–418

–

–

–

–

2

527

525

580–494

528

639–417

–

–

–

–

3

455

476

494–442

450

520–380

–

–

458

499–421

4

430

430

438–421

–

–

–

–

–

–

5

361

370

–

360

389–331

383

414–352

–

–

Estimates (continued)

Node Ref. (48)

Ref. (49)

Ref. (50)(b)

Ref. (50)(a)

Time

CI

Time

CI

Time

CI

Time

CI

1

–

–

–

–

–

–

–

–

2

–

–

–

–

–

–

–

–

3

–

–

–

–

438

480–412

451

495–413

4

–

–

–

–

–

–

–

–

5

360

373–346

354

367–341

–

–

–

–

Estimates (continued)

Node Ref. (51)

Ref. (52)(a)

Ref. (52)(b)

Time

CI

Time

CI

Time

CI

1

–

–

–

–

–

–

2

–

–

–

–

–

–

3

–

–

–

–

–

–

4

–

–

–

–

–

–

5

356

369–341

353

365–341

354

370–340

Note: See text for details. In the case of two studies (50, 52), each analyzed two different data sets (a) = nucleotides, (b) = amino acids and both estimates are shown here.

three-way Dipnoi–Actinistia–Tetrapoda split was estimated as 430 Ma. Six other studies have estimated single nodes in the timetree of vertebrates (46–52), all concerning the three latest divergences, among osteichthyians (Table 1). The timetree of vertebrates (Fig. 2) reflects the nodal averages for all of these molecular studies and, except for the early divergence of cyclostomes and gnathostomes, is more-or-less uncontroversial. Continental reconstructions for the Paleozoic are not as well resolved as those for later time periods. In general, they show a gradual coalescence of land areas into a supercontinent Pangaea (53, 54). There is no evidence yet that breakup of land areas was responsible for any of the

major divergences under consideration here. Sea-level changes also can cause the separation of evolutionary lineages, but the pattern for the Paleozoic, which shows a high in the Ordovician (488–444 Ma) followed by generally falling levels (55), also does not map directly to the vertebrate timetree (Fig. 2). However, variation in oxygen levels has been invoked as a major driver of animal evolution (56–63). During a surprisingly short interval in the latest Precambrian, 580–542 Ma, oxygen levels increased from 1% to 10% of the present level (64, 65) to nearly 100% of the present level (61, 66). If correct, a spike in oxygen of that magnitude would explain the Cambrian Explosion in the fossil record, reflecting an increase in

312


animal body size and production of hard parts (57–59). Evidence from molecular clocks indicates that this was decoupled from evolutionary divergences among animal phyla which occurred much earlier (67, 68). Atmospheric oxygen levels varied considerably during the Phanerozoic (542–0 Ma), presumably in response to changes in land floras and hence carbon burial, resulting in two major pulses (66). The First Pulse was in the late Silurian and early Devonian (~430–390 Ma), where the level reached a maximum of 25% at 410 Ma (the current level is 21%). This was followed by the Second Pulse in the Carboniferous and Permian (~340–250 Ma), reaching a maximum of 35% at ~270 Ma. Some effects of high oxygen levels are known, at least anecdotally, because insect gigantism is directly associated with the Second Pulse (56, 60). Also at a gross level, the timing of the two pulses coincides with periods diversification of vertebrates (1), suggesting a relationship. Otherwise it is difficult to link specific oxygen levels to favorable or unfavorable effects on vertebrates (62, 63). For example, the lowest level (~12%) during the entire Phanerozoic occurred in the Jurassic (200–146 Ma), yet this is the time when vertebrates were diversifying in body plans and increasing greatly in body size (e.g., dinosaurs), indicating that whatever limitations existed were overcome by the organisms. Nonetheless, the association of oxygen levels and Phanerozoic taxonomic diversity is still compelling and suggests some cause–effect relationship, even if mostly tied to invertebrate prey. Against this backdrop of atmospheric change, the origin of the vertebrate lineage began deep in the Neoproterozoic (~800–700 Ma), according to molecular clocks (5). That the earliest divergence among vertebrates (Fig. 2) occurred in the Precambrian (Neoproterozoic) as well is supported by the presence of diverse lineages of agnathans in the early Cambrian (31), only 20 million years after the Precambrian–Cambrian boundary. Under the interpretation of cyclostome monophyly, the distinction of craniates and vertebrates disappears, and therefore the first vertebrates that existed before the divergence between cyclostomes and gnathostomes are inferred to have had both a cranium and a vertebral column. However, because oxygen levels were very low they were surely small in size and soft-bodied, lacking bone, and initially probably fed on microorganisms. Prediction of the environment of the earliest vertebrates based on kidney structure and function has been debated (69–71), with an estuarine habitat favored most recently (72). In the latest Precambrian (e.g., Edicaran) when oxygen levels rose dramatically, there would have been already

multiple lineages of early vertebrates. One led to the living cyclostomes and another—the gnathostomes— took on the role of a major predator. The interpretation of conodonts as being on the gnathostome lineage (11) helps to fill in the otherwise large gap in the fossil record of gnathostomes forced by the presence of agnathans in the early Cambrian. Still, the absence of vertebrates and many other phyla from the latest Precambrian fossil record remains a problem for this hypothesis. If they were much smaller in size and soft-bodied, their fossils may only be discoverable in the finest-grain sediments representing low-energy environments (e.g., 73). In the Paleozoic, a good case can be made for oxygen as a driver of vertebrate evolution through prey abundance and diversity. The fossil record shows an explosion of fish lineages arising precisely during the First Pulse in the late Silurian and early Devonian. Presumably the pulse was from the land flora, known to be expanding at that time (74). A diversity of aquatic and terrestrial invertebrates also underwent diversification during the interval (63), making coastal and freshwater habitats a new resource niche for small fishes, which in turn were food for larger fishes. The first tetrapods appear in the fossil record later in the Devonian (~360 Ma), between the first and second pulses, and most authors consider the Second Pulse as key for tetrapod evolution (60, 62, 63). This is almost certainly true for the diversification of tetrapods, including the origin of lissamphibians and amniotes. The timetree (Fig. 2) suggests that the phylogenetic divergence of lissamphibians and amniotes occurred at a time (~360 Ma) when the first tetrapods appear in the fossil record. This rapid diversification is supported by the fossil record as well, which shows evidence that those two major tetrapod groups split no more than 25 million years later (44), with subsequent splitting into the lineages leading to caecilians, frogs, and salamanders on the one hand and synapsids and sauropsids on the other (4, 32, 40, 42, 43, 52, 75, 76). The ecological and evolutionary details of this “Permo-Carboniferous Explosion” of tetrapods have yet to be fully understood. However, most of the stage for the conquest of land by tetrapods was likely set during the First Pulse. Shallowwater habitats were teeming with life and the great number of lineages of fishes that appeared at that time, some having adaptations for living in shallow water (e.g., lungs, bony elements in fins) attests to the importance of that habitat. These included two major lineages of living fishes—coelacanths and lungfishes—and other extinct groups, one of which (osteolepiforms) led to tetrapods.

Eukaryotes; Metazoa; Vertebrata

Species transitional between fishes and tetrapods first appear in the fossil record ~385 Ma (37–39), at the end of the First Pulse. Although these transitional forms lacked digits and were adapted to shallow water—not land—they had already started to evolve most of the major tetrapod body plan traits. Whether oxygen did or did not play a major role in the origin of tetrapods, the reason was still probably related to a new prey resource (1) rather than escape from drying ponds (8). In summary, the early history of vertebrates may have extended 100–300 million years into the Precambrian but little is known about those organisms other than that they lived in very low oxygen levels and therefore were necessarily small and soft-bodied. Since the evolution of jaws in the latest Precambrian or Cambrian (according to the molecular timetree), vertebrates have dominated most environments on Earth as the top predators. Variation in oxygen levels during the Phanerozoic affected vertebrate prey and hence vertebrate diversity, and appears to have been responsible for bringing vertebrates onto land. The only remaining unsolved portion of the vertebrate backbone phylogeny, based on molecular data, is the relationships of tetrapods to lungfishes and coelacanths. In this case it is unclear whether resolving that node will answer any major questions not already addressed by fossils (18) and molecular clocks (Fig. 2), which place the divergences close in time. In the future, more sequences and fossils, and more reliable phylogenies incorporating fossil data, will allow for better calibration of molecular clocks and reduced variance of time estimates, and therefore a better resolution of the timescale of vertebrate evolution.

5.

6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21.

22.

Acknowledgments

23.

M. Laurin commented on an earlier draft. Support was provided by the U.S. National Science Foundation and National Aeronautics and Space Administration (NASA Astrobiology Institute).

24. 25.

References 1. M. J. Benton, Vertebrate Palaeontology (Blackwell Science, Oxford, 2005). 2. R. L. Carroll, Patterns and Processes of Vertebrate Evolution (Cambridge University Press, Cambridge, 1997). 3. E. H. Colbert, M. Morales, E. C. Minkoff, Colbert’s Evolution of the Vertebrates (Wiley-Liss, New York, 2001). 4. F. H. Pough, C. M. Janis, J. B. Heiser, Vertebrate Life (Pearson/Prentice Hall, Upper Saddle River, New Jersey, 2005).

26. 27. 28. 29. 30. 31. 32.

313

J. E. Blair, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 223–230. J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 2275 (2005). J. Mallatt, C. J. Winchell, Mol. Phylogenet. Evol. 43, 1005 (2007). A. S. Romer, Vertebrate Paleontology (The University of Chicago Press, Chicago, 1966). J. Mallatt, J. Zool. 204, 169 (1984). P. Janvier, Early Vertebrates (Clarendon Press, Oxford, 1996). P. C. J. Donoghue, P. L. Forey, R. J. Aldridge, Biol. Rev. 75, 191 (2000). D. W. Stock, G. S. Whitt, Science 257, 787 (1992). J. Mallatt, J. Sullivan, Mol. Biol. Evol. 15, 1706 (1998). S. Kuraku, D. Hoshiyama, K. Katoh, H. Suga, T. Miyata, J. Mol. Evol. 49, 729 (1999). S. B. Hedges, in Major Events in Early Vertebrate Evolution: Palaeontology, Phylogeny, Genetics and Development, P. E. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 119–134. C. Delarbre et al., Mol. Phylogenet. Evol. 22, 184 (2002). N. Takezaki, F. Figueroa, Z. Zaleska-Rutczynska, J. Klein, Mol. Biol. Evol. 20, 287 (2003). M. Zhu, X. B. Yu, P. E. Ahlberg, Nature 410, 81 (2001). D. E. Rosen, P. L. Forey, B. G. Gardiner, C. Patterson, Bull. Am. Mus. Nat. Hist. 167, 159 (1981). A. Meyer, A. C. Wilson, J. Mol. Evol. 31, 359 (1990). H. P. Shultze, in Origins of the Higher Groups of Tetrapods, H. P. Shultze, L. Trueb, Eds. (Comstock, Ithaca, New York, 1991), pp. 29–67. M.-M. Chang, in Origins of the Higher Groups of Tetrapods, H. P. Shultze, L. Trueb, Eds. (Comstcok, Ithaca, New York, 1991), pp. 3–28. S. B. Hedges, C. A. Hass, L. R. Maxson, Nature 363, 501 (1993). P. E. Ahlberg, Z. Jonanson, Nature 395, 792 (1998). R. Zardoya, Y. Cao, M. Hasegawa, A. Meyer, Mol. Biol. Evol. 15, 506 (1998). A. Meyer, R. Zardoya, Ann. Rev. Ecol. Evol. Syst. 34, 311 (2003). N. Takezaki, F. Figueroa, Z. Zaleska-Rutczynska, N. Takahata, J. Klein, Mol. Biol. Evol. 21, 1512 (2004). C. Darwin, On the Origin of Species by Means of Natural Selection (John Murray, London, 1859). R. L. Carroll, Vertebrate Paleontology and Evolution (W. H. Freeman & Company, New York, 1988). M. J. Benton, The Fossil Record 2 (Chapman & Hall, London, 1993). D. G. Shu et al., Nature 402, 42 (1999). M. J. Benton, C. J. Donoghue, R. J. Asher, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 35–86.

314


33. M. L. J. Stiassny, E. O. Wiley, G. D. Johnson, M. R. de Carvalho, in Assembling the Tree of Life, J. Cracraft, M. J. Donoghue, Eds. (Oxford University Press, New York, 2004), pp. 410–429. 34. T. Rowe, in Assembling the Tree of Life, J. Cracraft, M. J. Donoghue, Eds. (Oxford University Press, New York, 2004), pp. 384–409. 35. M. I. Coates, J. A. Clack, Nature 352, 234 (1991). 36. J. A. Clack, Gaining Ground: The Origin and Early Evolution of Tetrapods (Indiana University Press, Bloomington, Indiana, 2002). 37. P. E. Ahlberg, J. A. Clack, Nature 440, 747 (2006). 38. E. B. Daeschler, N. H. Shubin, F. A. Jenkins, Nature 440, 757 (2006). 39. N. H. Shubin, E. B. Daeschler, F. A. Jenkins, Nature 440, 764 (2006). 40. J. S. Anderson, R. R. Reisz, D. Scott, N. B. Fröbisch, S. S. Sumida, Nature (Lond.) 453, 515 (2008). 41. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 42. M. Ruta, M. I. Coates, in Telling Evolutionary Time, P. C. J. Donoghue, M. P. Smith, Eds. (CRC Press, London, 2003), pp. 224–262. 43. M. Ruta, M. I. Coates, D. L. J. Quicke, Biol. Rev. 78, 251 (2003). 44. G. Vallin, M. Laurin, J. Vertebr. Paleontol. 24, 56 (2004). 45. P. C. J. Donoghue, M. P. Smith, I. J. Sansom, in Telling the Evolutionary Time: Molecular Clocks and the Fossil Record, P. C. J. Donoghue, M. P. Smith, Eds. (CRC Press, Boca Raton, Florida, 2003), pp. 190–223. 46. H. Mannen, S. S. Li, Mol. Phylogenet. Evol. 13, 144 (1999). 47. J. E. Blair, P. Shah, S. B. Hedges, BMC Bioinformatics 6, 53 (2005). 48. P. Zhang, H. Zhou, Y. Q. Chen, Y. F. Liu, L. H. Qu, Syst. Biol. 54, 391 (2005). 49. P. Zhang et al., Gene 346, 133 (2005). 50. J. G. Inoue, M. Miya, B. Venkatesh, M. Nishida, Gene 349, 227 (2005). 51. S. L. Pereira, A. J. Baker, Mol. Biol. Evol. 23, 1731 (2006). 52. A. F. Hugall, R. Foster, M. S. Y. Lee, Syst. Biol. 56, 543 (2007).

53. 54.

55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68.

69. 70. 71. 72. 73. 74. 75. 76.

A. Hallam, An Outline of Phanerozoic Biogeography (Oxford University Press, New York, 1994). F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds., A Geologic Timescale 2004 (Cambridge University Press, Cambridge, UK, 2005), pp. 610. A. Hallam, Phanerozoic Sea-Level Changes (Columbia University Press, New York, 1992). É. Harlé, A. Harlé, Bull. Soc. Geol. Fr. 4, 118 (1911). J. R. Nursall, Nature 183, 1170 (1959). P. Cloud, Paleobiology 2, 351 (1976). B. Runnegar, J. Geol. Soc. Aust. 29, 395 (1982). J. B. Graham, R. Dudley, N. M. Aguilar, C. Gans, Nature 375, 117 (1995). R. A. Berner, Proc. Natl. Acad. Sci. U.S.A. 96, 10955 (1999). R. A. Berner, J. M. VandenBrooks, P. D. Ward, Science 316, 557 (2007). P. Ward, C. Labandeira, M. Laurin, R. A. Berner, Proc. Natl. Acad. Sci. U.S.A. 103, 16818 (2006). D. Canfield, Ann. Rev. Earth Planet. Sci. 33, 1 (2005). D. E. Canfield, S. W. Poulton, G. M. Narbonne, Science 315, 92 (2007). R. A. Berner, Geochim. Cosmochim. Acta 70, 5653 (2006). G. A. Wray, J. S. Levinton, L. H. Shapiro, Science 274, 568 (1996). S. B. Hedges, F. U. Battistuzzi, J. E. Blair, in Neoproterozoic Geobiology and Paleobiology, S. Xiao, A. J. Kaufman, Eds. (Springer, New York, 2006), pp. 199–229. A. S. Romer, The Vertebrate Body (Saunders, Philadelphia, 1949). H. Smith, Quart. Rev. Biol. 7, 1 (1932). R. W. Griffith, Comp. Biochem. Physiol. 87, 523 (1987). H. Ditrich, Zool. J. Linn. Soc. 150, 435 (2007). S. H. Xiao, Y. Zhang, A. H. Knoll, Nature 391, 553 (1998). P. Kenrick, P. R. Crane, Nature 389, 33 (1997). D. Marjanovic, M. Laurin, Syst. Biol. 56, 369 (2007). S. S. Sumida, K. L. Martin, Eds., Amniote Origins— Completing the Transition to Land, (Academic Press, New York, 1997).

FISHES


Jawless fishes (Cyclostomata) Shigehiro Kurakua,b,*, Kinya G. Otaa, and Shigeru Kuratania a

Laboratory for Evolutionary Morphology, RIKEN Center for Developmental Biology, 2–2-3 Minatojima-minami, Chuo-ku, Kobe 650–0047, Japan; bLehrstuhl für Zoologie und Evolutionsbiologie, Department of Biology, University of Konstanz, Universitätsstrasse 10, D-78457 Konstanz, Germany *To whom correspondence should be addressed (shigehiro.kuraku@ uni-konstanz.de)

Abstract Cyclostomata comprises two families of living jawless fishes: hagfishes (Myxinidae, 44 species) and lampreys (Petromyzonidae, 41 species). Morphological analyses have favored the closer relationship of lampreys to jawed vertebrates (gnathostomes) than to hagfishes. However, most of the recent molecular phylogenetic analyses have supported a hagfish–lamprey relationship. The estimated divergence time for hagfishes and lampreys among several studies averages 482 million years ago (Ma), but varies (520–432 Ma) depending mostly on the assumed timing of the cyclostome–gnathostome divergence. Nonetheless, there is agreement that hagfish and lamprey lineages diverged relatively shortly (within 100 million years) after the divergence of cyclostomes and gnathostomes.

Cyclostomata consists of two extant orders, Myxiniformes and Petromyzoniformes (1). Myxiniformes contains a single family, Myxinidae that includes 44 species in six genera (2) (hagfishes; Fig. 1). Petromyzoniformes also consists of a single family, Petromyzonidae that includes 41 species in six genera (3) (lampreys, Fig. 1). Historically, hagfishes and lampreys have been classified as cyclostomes (“round mouth”) because both have a jawless mouth armed with retractable horny teeth (4). However, some morphological traits in hagfishes (e.g., lack of vertebrae, heart innervation, and eye lens) have been taken to suggest that lampreys are more closely related to jawed vertebrates (gnathostomes) than to hagfishes (5). To settle this controversy, molecular phylogenetic studies have been conducted since the early 1990s using ribosomal DNA genes (6, 7), mitochondrial

genes (8–11), and protein-coding genes of the nuclear genomes (12–16). Here we review the phylogenetic relationships and the divergence time of the two families of Cyclostomata. Most molecular studies to date have concluded that Cyclostomata is a monophyletic group (see 17 and references therein). These findings suggest that many primitive morphological traits found in hagfishes are just due to secondary losses in the hagfish lineage, or that the presumed primitive nature of the hagfish morphology is due to the unavailability of detailed morphological description, as recently exemplified by hagfish embryonic morphology (18). Rediscovery of cyclostome monophyly can be regarded as one of the most outstanding examples in which molecular phylogenetics has resolved controversial phylogenetic relationships among major animal groups (19). With this rediscovery, researchers can now estimate the timing of hagfish–lamprey divergence using molecular sequence information. Phylogenetic studies have revealed that the hagfish lineage exhibits a long branch, suggesting elevation of evolutionary rate at the molecular level in this lineage (e.g., 17). For this reason, application of local clock analysis is expected to improve the precision of divergence time estimation, especially in this animal group. Using a lineage-specific method, the hagfish–lamprey divergence was estimated to be 499 (536–462) Ma in an analysis of seven nuclear genes (20). A subsequent Bayesian analysis of 25 nuclear genes resulted in a slightly older estimate of 520 (596–461) Ma (16). In this study, the minimum time constraint for hagfish–lamprey divergence was incorporated based on fossil-based discoveries of extinct relatives

Fig. 1 Two Japanese lampreys (Lethenteron japonicum) and an inshore hagfish (Eptatretus burgeri, tangled). Credit: S. Kuraku and K. G. Ota.

S. Kuraku, K.G. Ota, and S. Kuratani. Jawless fishes (Cyclostomata). Pp. 317–319 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

318

THE TIMETREE OF LIFE Myxinidae 1

O

Petromyzonidae S

D

C

P

Tr

PALEOZOIC 400

J

K

Pg

MESOZOIC 300

200

CZ

100

0 Million years ago

Fig. 2 A timetree of jawless fishes (Cyclostomata). Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), O (Ordovician), P (Permian), Pg (Paleogene), S (Silurian), and Tr (Triassic).

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) between jawless fishes (Cyclostomata). Timetree Node

1

Estimates Ref. (16)

Time

482.3

Ref. (17)(a)

Ref. (17)(b)

Ref. (20)

Time

CI

Time

CI

Time

CI

Time

CI

520.0

596–461

432.0

473–391

478.0

497–459

499.0

536–462

Note: Node times in the timetree represent the mean of time estimates from different studies. From ref. (17), estimates are presented from (a) nuclear and (b) mitochondrial data.

of modern hagfishes and lampreys (325 Ma; 21 and references therein). In a more recent study, the divergence time of hagfishes and lampreys was estimated using different sets of genes (nuclear genes and mitochondrial genes) and calibration dates (molecules and fossils) (17). Application of molecular estimates (16, 22) as the maximum and minimum time constraints resulted in dates that are similar to the previous estimate (520 Ma; 16) or much more ancient (612 Ma) (data not shown in Table 1; see 17 for details). When the cyclostome–gnathostome split is constrained at 500 Ma, however, the analysis resulted in a 432 (473–391) Ma date based on 10 nuclear genes and 478 (497–459) Ma date based on 12 mitochondrial genes for the hagfish–lamprey divergence (17). Since those studies were completed, the fossil record of lampreys has been extended into the late Devonian (~360 Ma) (23), which is 11% older than the calibration date that was used for the divergence of hagfishes and lampreys (16). While the estimates of hagfish–lamprey divergence times are highly variable, depending on the calibrations used, all available studies unequivocally suggest that the ancestors of Myxinidae (hagfishes) and Petromyzonidae (lampreys) diverged relatively soon (less than 100 million years) after the ancestors of cyclostomes split from the gnathostome lineage.

Acknowledgment Support was provided by Grants-in-Aid from the Ministry of Education, Science and Culture of Japan.

References 1. 2. 3. 4.

5. 6. 7. 8. 9. 10. 11.

M. W. Hardisty, Biology of the Cyclostomes (Chapman & Hall, London, 1979). J. M. Jørgensen, The Biology of Hagfishes, 1st ed. (Chapman & Hall, London; New York, 1998). M. W. Hardisty, I. C. Potter, The Biology of Lampreys (Academic Press, London; New York, 1971). A. M. C. Duméril, Zoologie Analytique, ou Méthode Naturelle de Classification des Animaux, Rendue plus facile a l’Aide de Tableaux Synoptiques (Allais, Paris, 1806). P. Janvier, Early Vertebrates (Clarendon Press; Oxford University Press, Oxford, England, New York, NY, 1996). J. Mallatt, J. Sullivan, Mol. Biol. Evol. 15, 1706 (1998). D. W. Stock, G. S. Whitt, Science 257, 787 (1992). C. Delarbre et al., Mol. Biol. Evol. 17, 519 (2000). C. Delarbre, C. Gallut, V. Barriel, P. Janvier, G. Gachelin, Mol. Phylogenet. Evol. 22, 184 (2002). C. Delarbre, A. S. Rasmussen, U. Arnason, G. Gachelin, J. Mol. Evol. 53, 634 (2001). A. S. Rasmussen, A. Janke, U. Arnason, J. Mol. Evol. 46, 382 (1998).

Eukaryota; Metazoa; Vertebrata; Cyclostomata

12. 13. 14. 15. 16. 17. 18.

F. Delsuc, H. Brinkmann, D. Chourrout, H. Philippe, Nature 439, 965 (2006). N. Takezaki, F. Figueroa, Z. Zaleska-Rutczynska, J. Klein, Mol. Biol. Evol. 20, 287 (2003). S. Kuraku, D. Hoshiyama, K. Katoh, H. Suga, T. Miyata, J. Mol. Evol. 49, 729 (1999). R. F. Furlong, P. W. H. Holland, Zool. Sci. 19, 593 (2002). J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 2275 (2005). S. Kuraku, S. Kuratani, Zool. Sci. 23, 1053 (2006). K. G. Ota, S. Kuraku, S. Kuratani, Nature 446, 672 (2007).

19.

319

A. Meyer, R. Zardoya, Annu. Rev. Ecol. Evol. Syst. 34, 311 (2003). 20. S. B. Hedges, in Major Events in Early Vertebrate Evolution, P. E. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 119–134. 21. P. C. J. Donoghue, M. P. Smith, I. J. Sansom, in Telling the Evolutionary Time, P. C. J. Donoghue, M. P. Smith, Eds. (Taylor & Francis, London, 2003), pp. 190–223. 22. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 23. R. W. Gess, M. I. Coates, B. S. Rubidge, Nature 443, 981 (2006).

Cartilaginous fishes (Chondrichthyes) Matthew P. Heinickea,*, Gavin J. P. Naylor b, and S. Blair Hedgesa a

Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA; bSchool of Computational Biology, Dirac Science Library, Florida State University, Tallahassee, FL 32306, USA *To whom correspondence should be addressed ([email protected])

Abstract Sharks, rays, chimaeras, and relatives (Class Chondrichthyes) comprise an important component of living vertebrate diversity, with two subclasses, 18 orders, ~55 families, and ~1200 species. Recent morphological studies have supported a position for rays deeply nested within sharks. Molecular analyses, however, support a basal divergence between rays and sharks. New molecular timing analyses presented here suggest that the earliest divergences in Chondrichthyes occurred deep in the Paleozoic, 460–300 million years ago (Ma), and that most living families originated before the end of the Cretaceous (66 Ma). If accurate, these dates imply large ghost ranges in the fossil record for many chondrichthyan groups.

Living members of the Subclasses Holocephali (chimaeras, including ratfishes, spookfishes, and rabbitfishes, ~43 sp.) and Elasmobranchii (sharks, rays, skates, sawfishes, and guitarfishes, ~1125 sp.) together comprise the extant representatives of the Class Chondrichthyes (cartilaginous fishes) (1). Holocephali includes only a single living order with three families (1, 2). Elasmobranchs are more diverse, with ~17 orders and ~52 families (there is some disagreement in ordinal and familial limits, especially among rays) (1, 3–5). Chondrichthyans can be differentiated from their closest living relatives, Osteichthyes (bony vertebrates), by possession of a skeleton of prismatic cartilage and internal fertilization via modified male pelvic fins (claspers). Other characters common to the group are possession of placoid (toothlike) scales and, in many lineages, a heterocercal tail fin. While most sharks and chimaeras have a generally cylindrical “fishlike” body form (Fig. 1), some sharks and all batoids (rays, skates, sawfishes, and guitarfishes) are

dorsoventrally flattened and benthic in habit. Although early chondrichthyans included many freshwater forms, living species are overwhelmingly marine in distribution, excluding a few euryhaline sharks and rays and some freshwater stingrays. Here, we review the relationships of the subclasses, orders, and families of cartilaginous fishes. Additionally, molecular divergence times of these groups are estimated from publicly available sequence data and presented. The fossil record of Chondrichthyes has been considered excellent, based largely on rich deposits of dental material (6). The cartilaginous skeleton of Chondrichthyes fossilizes poorly; therefore, skeletal fossil material is much rarer (7). The earliest fossils assigned to Chondrichthyes are from the Silurian (444–416 Ma) (8). Fossils become more common in the Devonian (416–359 Ma), including many representatives of extinct groups. Based on these fossils, the Subclasses Holocephali and Elasmobranchii are estimated to have diverged by 410 Ma (8). Fossil evidence for modern representatives of these subclasses—Suborder Chimaeroidei (chimaeras) and Infraclass Neoselachii (sharks and rays)—does not occur until the Mesozoic (251–66 Ma) (9, 10). Living orders and families can be identified from the Jurassic (200–146 Ma) onward, with fossil evidence of most families before the end of the Mesozoic (7, 10).

Fig. 1 A Great White Shark (Carcharodon carcharias) from near Isla Guadalupe, Mexico. Credit: T. Goss.

M. P. Heinicke, G. J. P. Naylor, and S. B. Hedges. Cartilaginous fishes (Chondrichthyes). Pp. 320–327 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Chondrichthyes

Carcharhinidae

44

Triakidae-1

40

35

Triakidae-2 Leptochariidae

22 21

Pseudotriakidae

25

Proscylliidae

16

Carcharhiniformes

Hemigaleidae

42

Scyliorhinidae-1 Scyliorhinidae-2 Alopiidae

39

Odontaspididae

31

Pseudocarchariidae 29

Lamnidae

36 20

Cetorhinidae

32

Lamniformes

33

13

Galeomorphii

Megachasmidae

Carchariidae 7

Mitsukurinidae

Heterodontidae Somniosidae Oxynotidae 28

3

Dalatiidae Etmopteridae

23

Centrophoridae 12

Squalidae

9

Squatinidae

18

Echinorhinidae

Notorynchidae

43

Hexanchidae

15

Chlamydoselachidae


O

S

D

C

P

PALEOZOIC 400

Fig. 2 Continues

Tr

J

K

MESOZOIC 300

200

100


Hexanchiformes

Pristiophoridae

4

Squalimorphii

Parascylliidae

Squatiniformes Squaliformes

Orectolobidae

Heterodontiformes

Brachaeluridae 27

14

Echinorhiniformes

Hemiscylliidae

19

Orectolobiformes

Stegosomatidae

38 5

Ginglymostomatidae

Pristiophoriformes

45

321

THE TIMETREE OF LIFE (continued on previous page)

Myliobatidae

34

Gymnuridae Dasyatidae-1

2

Dasyatidae-2 30

Urolophidae Hexatrygonidae Plesiobatidae

Myliobatiformes

322

Torpedinidae 26

Narcinidae-2

6

Narcinidae-1 Rajidae Rhinochimaeridae

37

Chimaeridae

17

Callorhinchidae O

S

D

C

P

PALEOZOIC 400

Tr

J

K

MESOZOIC 300

200

100

Chimaeriformes Torpediniformes

Platyrhinidae

Batoidea Holocephali

Rhinobatidae 8

Rhynchobatiformes

Pristidae

Rhinobatiformes

24

1

Pristiformes

Rhynchobatidae

Platyrhiniformes

Rhinidae

41

10

Rajiformes

Urotrygonidae

Rhiniformes

Potamotrygonidae 11


Fig. 2 A timetree of cartilaginous fishes (Chondrichthyes). Divergence times are from Table 1. Galeomorphii, Squalimorphii, and Batoidea comprise the Subclass Elasmobranchii. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), O (Ordovician), P (Permian), Pg (Paleogene), S (Silurian), and Tr (Triassic). Codes

for paraphyletic and/or polyphyletic groups are as follows: Triakidae-1 (Mustelus), Triakidae-2 (Triakis), Scyliorhinidae-1 (Pentanchinae), Scyliorhinidae-2 (Scyliorhininae), Dasyatidae-1 (Dasyatis), Dasyatidae-2 (Himantura), Narcinidae-1 (Narcininae), and Narcinidae-2 (Narkinae).

Division of living cartilaginous fishes into Elasmobranchii and Holocephali is strongly supported by morphological analyses, as is uniting these groups to form Chondrichthyes (6, 9, 11). Within Holocephali, it is believed that Rhinochimaeridae and Chimaeridae form a group to the exclusion of Callorhinchidae (12). The relationships of the more species-rich elasmobranchs are more contentious. Early studies suggested a basal split between sharks and rays (13, 14). In 1992, Shirai published an extensive and influential analysis of morphological variation among sharks and rays in which he proposed a “Hypnosqualean hypothesis” wherein the batoids fall together with the dorsoventrally

compressed sawsharks (Pristiophoriformes) and angel sharks (Squatiniformes) (15). These in turn group with the Orders Squaliformes, Hexanchiformes, and Echinorhiniformes in the Hypnosqualean clade. The Orders Lamniformes, Carcharhiniformes, Orectolobiformes, and Heterodontiformes are grouped as Galea (3, 4). Minor modifications to Shirai’s original 1992 hypothesis of elasmobranch interrelationships were made by de Carvalho in 1996 (4). This hypothesis remains the consensus from morphological data. Because the monophyly of Chondrichthyes and reciprocal monophyly of Elasmobranchii and Holocephali have not been controversial and are supported by


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among cartilaginous fishes (Chondrichthyes), based on analyses presented here. Timetree Node

Time

Estimates This study (a) Time

CI

This study (b)

This study (c)

Time

CI

Time

CI

1

471

471

494–434

486

495–463

436

482–411

2

393

393

431–354

440

471–403

357

402–319

3

350

350

392–309

419

452–380

273

319–235

4

327

327

372–283

392

431–345

256

304–214

5

318

318

359–279

374

414–330

258

301–222

6

291

291

333–250

308

368–248

308

357–262

7

289

289

329–252

344

386–300

234

276–201

8

281

281

324–241

–

–

294

344–249

9

276

276

323–232

329

385–265

219

269–178

10

274

274

318–235

–

–

285

334–240

11

265

265

307–227

283

341–228

273

321–228

12

263

263

311–220

312

370–247

207

256–165

13

259

259

297–226

304

346–262

220

260–190

14

237

237

287–186

280

338–215

–

–

15

236

236

295–183

285

353–202

–

–

16

226

226

261–195

269

310–228

189

227–166

17

220

220

320–125

248

351–128

–

–

18

214

214

269–163

252

324–176

–

–

19

195

195

249–139

231

296–160

–

–

20

185

185

224–148

222

270–175

–

–

21

179

179

210–153

213

251–177

155

194–124

22

173

173

204–149

205

242–171

141

179–111

23

170

170

218–128

190

268–121

142

188–106

24

159

159

219–105

–

–

169

233–110

25

153

153

183–127

183

222–147

114

156–77

26

145

145

205–98

–

–

154

215–101

27

142

142

201–84

169

243–97

–

–

28

135

135

184–93

154

229–91

108

155–73

29

126

126

160–104

167

217–125

120

154–102

30

124

124

159–98

–

–

127

166–99

31

122

122

155–100

140

193–93

–

–

32

122

122

154–101

122

167–100

117

150–100

33

119

119

151–97

120

172–74

117

151–98

34

111

111

139–97

115

148–100

115

149–98

35

110

110

133–96

112

140–96

–

–

36

109

109

142–83

93.4

139–59

109

142–86

37

107

107

182–51

–

–

123

197–65

38

106

106

162–60

127

196–69

–

–

39

104

104

139–73

103

157–58

–

–

323

324



Time

Estimates This study (a) Time

This study (b)

CI

Time

CI

This study (c) Time

CI

40

98.6

98.6

118–90

99

123–90

107

139–90

41

92.4

92.4

150–47

–

–

99.6

164–49

42

91.6

91.6

111–77

91

116–73

94.8

127–73

43

79.7

79.7

150–29

98.6

185–34

–

–

44

72.1

72.1

94–53

70.9

97–50

–

–

45

53.1

53.1

97–23

64.6

119–27

–

–

Note: Node times for the timetree are from the combined analysis of RAG1, 12S, and 16S alignments [shown with CI in column (a)]. Columns (b) and (c) present estimates from RAG1 and 12S/16S analyses, respectively.

numerous morphological characters, molecular studies have not been designed to specifically address these relationships. However, recent molecular studies that have included a broad enough sample of taxa to draw conclusions have supported the monophyly of these groups (16, 17). The interrelationships among the holocephalan families have not yet been addressed with molecular data. However, one mitochondrial gene study, using several holocephalan species as outgroups, has suggested that Rhinochimaeridae is embedded within Chimaeridae (18). Most molecular studies have focused on elasmobranch interrelationships. Studies in the early to mid-1990s included too few taxa or sites to infer strong conclusions (19–21). Since 2003, elasmobranch relationships have been inferred with more comprehensive data sets of both nuclear and mitochondrial data, including most orders and families (7, 17, 18, 22, 23). These studies consistently (but weakly) reject the Hypnosqualea hypothesis, and instead suggest a basal divergence between sharks and batoids. Within the batoids, skates (Rajiformes) appear basal, followed by electric rays (Torpediniformes), then sawfishes and guitarfishes (Pristiformes, Rhinobatiformes), with stingrays (including butterfly, eagle, and manta rays; Myliobatiformes) being the most derived (7, 17, 21, 24). In these studies, most of the batoid orders were represented by only one or a few families, but there are numerous myliobatiform families. Analyses including these families have not found significantly supported relationships, although it appears that the butterfly rays and manta/eagle rays (Gymnuridae and Myliobatidae) form a group (24). No studies have yet

determined the relationships among the families of guitarfishes (Rhinidae, Rhynchobatidae), thornback rays (Platyrhinidae), or panrays (Zanobatidae). Based on analysis of mitochondrial 12S ribosomal RNA (rRNA) gene sequences available in GenBank, however, it appears that Rhinidae and Rhyncobatidae form a sawfish/guitarfish group with Pristidae and Rhinobatidae, while the position of thornback rays remains unresolved (results not shown). Molecular studies of shark orders and families have led to a somewhat better understanding of relationships. The two major groups of sharks, galeomorphs and squalimorphs, are supported in most molecular studies (7, 17, 18, 22, 23, 25). Although morphologically part of Galeomorphii, the horn sharks (Heterodontiformes) are in a basal position in molecular phylogenies, and cluster with both Squalimorphii and Galeomorphii, depending on the data set. Within the Galeomorphii, the orders Lamniformes and Carcharhiniformes are generally recovered as closest relatives (7, 18, 23). In the Squalimorphii, Squatiniformes (angel sharks), and Echinorhiniformes (bramble sharks) are close relatives, while cow sharks (Hexanchiformes) are outside all other squalimorph orders (7, 17, 18, 22, 23, 25). At the family level, the nominal groups Scyliorhinidae and Triakidae are estimated to be paraphyletic (26, 27) while the position of the hammerhead sharks is seen to fall within the Carcharhinidae. Accordingly, they are not considered a distinct family herein (24). Carchariidae and Odontaspididae (often considered a single family) form divergent branches in Lamniformes (7, 28). The interfamilial relationships of Squaliformes remain unexplored.


Until now, no timing analyses have been performed at or above the family level using molecular sequence data. Martin et al. (29) calculated the rate of evolution in sharks for cytochrome b sequences, but did not use this rate to infer times of divergence among different families. Batoid divergence times have been calculated, but only within families (30, 31). However, divergence times of higher chondrichthyan taxa have been inferred using immunological distances (32). These data suggest a very old divergence between sharks and batoids (392 Ma), and show divergences among sharks beginning 300 Ma. Because there is no study reporting molecular divergence times of chondrichthyan families, we report herein the results of an analysis using published sequence data employing methodology described elsewhere (33). Sequence data were obtained from the most comprehensive available studies, using the nuclear protein-coding RAG1 gene and the mitochondrial 12S and 16S rRNA genes (7, 18, 26). Additional 12S and 16S sequences of 15 batoid families were included from GenBank, as only Rajidae and Urolophidae were included in the study of Douady et al. (18). Together these data encompass a patchwork of sequences for 53 of 55 families of Chondrichthyes, excluding only Zanobatidae (panrays) and Rhincodontidae (whale shark). We note that while 53 of 55 families are represented, relatively few families are represented by all three genes, as a consequence of concatenating the data from three different studies with few overlapping taxa. In total, eight batoid families are represented only by 12S sequences, and 15 shark families by only RAG1 sequences, while 17 families include all data and the remaining 13 families include data for two genes. Tree topology was based on the studies that reported the sequences, although branches that are not resolved or conflict among these and other published molecular phylogenies were collapsed to polytomies for the final timetree (Fig. 2). These polytomies mainly affect Squaliformes and the batoid orders, as molecular studies including squaliform families have very short, poorly supported internal branches, and relationships within batoid orders are similarly poorly supported (7, 17, 22, 24). An analysis of batoid 12S sequences used in the timetree did not find any significantly supported relationships within batoid orders (results not shown). For the timetree, a combined analysis of all data was used. Analyses were also performed for the separate RAG1 and 12S/16S data sets (Table 1). An amniote (Homo), amphibian (Xenopus), actinopterygian (Danio), cyclostome (Petromyzon), and echinoderm (Strongylocentrotus) were used as outgroups, but these

325

taxa are not presented in the timetree. A total of 14 minimum (min.) and three maximum (max.) fossil constraints used to calibrate the timetree were obtained from the literature (7, 8, 10, 34). These include the divergence of Centrophoridae from other Squaliformes (min. 89 Ma); the divergence of Squatinidae and Echinorhinidae (min. 151 Ma); the divergence of Hexanchidae and Chlamydoselachidae (min. 176 Ma); the divergence of Triakidae and Carcharhinidae (min. 89 Ma); the divergence of Scyliorhinidae and other Carcharhiniformes (min. 165 Ma); the divergence of Carchariidae and Lamnidae (min. 100 Ma); the divergence of Parascyllidae and other Orectolobiformes (min. 100 Ma); the divergence of Heterodontidae and other sharks (min. 176 Ma); the divergence of Dasyatidae and Myliobatidae (min. 100 Ma); the divergence of Rajidae and other batoids (min. 176 Ma); the divergence of sharks and batoids (min. 190 Ma); the divergence of elasmobranchs and holocephalans (min. 410 Ma, max. 495 Ma); the divergence of amniotes and amphibians (min. 340 Ma, max. 370 Ma); and the divergence of actinopterygians and sarcopterygians (min. 435 Ma, max. 495 Ma). Times of divergence obtained from the separate RAG1 and 12S/16S analyses differ markedly for most comparisons. Of the nodes shared between these two analyses, only the estimates for nodes within Batoidea and Lamniformes, and among derived carcharhiniform families (Carcharhinidae, Hemigaleidae, Triakidae), show noteworthy temporal concordance. In general the RAG1based estimates are much older than those based on 12S/16S sequences (Table 1). In some cases, the discrepancy in age estimates is quite large. For example, RAG1 data result in times more than 100 million years older than 12S and 16S data for divergences among the major chondrichthyan groups (chimaeras, batoids, galeomorph sharks, and squalimorph sharks). This may be caused by the large amount of branch length variation in the RAG1 data set (7), while the 12S and 16S data have relatively less variation (18). Time estimates from the combined analysis, discussed later, are generally between values from the individual analyses. Conclusions based on the combined analysis must be tempered by the knowledge that not all genes are present for all taxa (i.e., a large amount of missing data) and the large differences in times of deep branches obtained with RAG1 as compared to 12S and 16S data. Notwithstanding the discrepancies in age estimates among genes, the timetree (Fig. 2) suggests that holocephalans and elasmobranchs diverged in the Ordovician, 471 (494–434) Ma. Fossil evidence indicates

326


that these subclasses had diverged by at least 410 Ma (8). The living families of Holocephali apparently diverged in the Mesozoic (251–66 Ma). The divergence of sharks and batoids is inferred to have occurred in the Devonian, 393 (431–354) Ma. This date is more than 100 million years older than the first appearance of neoselachian elasmobranchs in the fossil record, and over 200 million years older than unambiguous evidence of modern orders (10). If these estimates are accurate, one must infer a large ghost range in the fossil record for early divergences within modern elasmobranchs. Times obtained with only 12S and 16S data are substantially younger, at 357 (402–319) Ma, but still suggests a large ghost range. Our analyses of the presented molecular data suggest that ordinal divergences were largely completed by the beginning of the Triassic, 251 Ma (whether considering the combined, RAG1, or 12S/16S analyses) and that living families diverged throughout the Mesozoic, but especially during the Cretaceous (146–66 Ma). With the possible exceptions of Ginglymostomatidae (nurse sharks) and Stegostomatidae (zebra shark), all elasmobranch families are estimated to have appeared by the end of the Cretaceous. Because of these apparent ancient divergences, oceanic habits of chondrichthyans, and large differences in time estimates depending upon analysis used, it is difficult to infer the biogeographic history of the living families. Most chondrichthyan families today are cosmopolitan in distribution, or found in widely divergent (i.e., separate ocean basins) areas of suitable habitat. For the many families with pelagic or deep-sea distributions, it may be impossible to infer biogeographic history due to the worldwide nature of their habitats. Extensive plate tectonic activity has contributed to substantial changes in ocean basins since the divergence of most families of Chondrichthyes. Although all ocean floor is geologically young (Mesozoic and Cenozoic), oceans differ in age when considered as bodies of water (aquatic habitat). For example, the Atlantic is relatively young (~150 Ma) compared with the Pacific (35, 36), which may explain why no living families are restricted to the Atlantic. Other families may have their origins in basins that no longer exist. For example, many inshore, benthic families, such as batoids, diverged in the Cretaceous (based on the results of the presented analyses, as well as fossil data) when the sea level was much higher and shallow continental seaways covered large portions of North America and Asia. At the same time, the now-gone Tethys Sea existed between the northern and southern continents (35, 36). These water bodies may have

been the early sites of diversification within batoids and inshore sharks. The timetree (Fig. 2) is compatible with previous interpretations of shark evolution based on the fossil record, including a major radiation of neoselachian sharks in the Jurassic and Cretaceous (200–66 Ma), possibly related to a parallel radiation of prey, actinopterygian fishes (37). In order to better understand the factors leading to diversification in Chondrichthyes, additional fossil (especially skeletal) and paleogeographic data will be needed to complement the emerging molecular phylogenetic data. In addition, more comprehensive molecular data, including nuclear gene loci that exhibit more uniform rates of evolution among lineages, are needed to resolve poorly known parts of the chondrichthyan tree and to estimate better-constrained times of divergence.

Acknowledgments Support was provided by grants from the U.S. National Science Foundation (NSF) to G.J.P.N. and from NSF and the National Aeronautics and Space Administration (NASA Astrobiology Institute) to S.B.H.

References 1. L. J. V. Compagno, D. A. Didier, G. H. Burgess, in Sharks, Rays, and Chimaeras, the Status of the Chondrichthyan Fishes, S. L. Fowler et al., Eds. (IUCN/SSC Shark Specialist Group, Gland, 2005), pp. 4–11. 2. D. A. Didier, in Biology of Sharks and their Relatives, J. C. Carrier, J. A. Musick, M. R. Heithaus, Eds. (CRC Press, Boca Raton, 2004), pp. 115–138. 3. S. Shirai, in Interrelationships of Fishes, M. L. J. Stiassny, L. R. Parenti, G. D. Johnson, Eds. (Academic Press, San Diego, 1996), pp. 9–34. 4. M. R. de Carvalho, in Interrelationships of Fishes, M. L. J. Stiassny, L. R. Parenti, G. D. Johnson, Eds. (Academic Press, San Diego, 1996), pp. 35–62. 5. J. D. McEachran, K. A. Dunn, T. Miyake, in Interrelationships of Fishes, M. L. J. Stiassny, L. R. Parenti, G. D. Johnson, Eds. (Academic Press, San Diego, 1996), pp. 63–84. 6. J. G. Maisey, Zool. J. Linn. Soc. 82, 33 (1984). 7. J. G. Maisey, G. J. P. Naylor, D. J. Ward, in Mesozoic Fishes 3—Systematics, Paleoenvironments and Biodiversity, G. Arriata, A. Tintori, Eds. (Verlag Dr. Friedrich Pfeil, München, 2004), pp. 17–56. 8. M. I. Coates, S. E. Sequeira, in Major Events in Early Vertebrate Evolution: Paleontology, Phylogeny, Genetics and Development, P. E. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 241–262.


9.

10. 11. 12. 13.

14.

15.

16. 17. 18. 19. 20. 21. 22. 23.

E. D. Grogan, R. Lund, in Biology of Sharks and their Relatives, J. C. Carrier, J. A. Musick, M. R. Heithaus, Eds. (CRC Press, Boca Raton, 2004), pp. 3–32. C. J. Underwood, Paleobiology 32, 215 (2006). J. G. Maisey, Cladistics 2, 201 (1986). D. A. Didier, Am. Mus. Nov. 3119, 1 (1995). H. W. Bigelow, W. C. Schroeder, in Fishes of the Western North Atlantic, Part I, J. Tee-Van et al., Eds. (Yale University Press, New Haven, 1948), pp. 59–576. H. W. Bigelow, W. C. Schroeder, in Fishes of the Western North Atlantic, Part II, J. Tee-Van et al., Eds. (Yale University Press, New Haven, 1953), pp. 1–514. S. Shirai, Squalean Phylogeny: A New Framework of “Squaloid” Sharks and Related Taxa (Hokkaido University Press, Sapporo, 1992), pp. 151. U. Arnason, A. Gullberg, A. Janke, Zool. Scripta 30, 249 (2001). J. Mallatt, C. J. Winchell, Mol. Phylogenet. Evol. 43, 1005 (2007). C. J. Douady, M. Dosay, M. S. Shivji, M. J. Stanhope, Mol. Phylogenet. Evol. 26, 215 (2003). G. Bernardi, G. A. Powers, Mol. Phylogenet. Evol. 1, 161 (1992). K. A. Dunn, J. F. Morrissey, Copeia 1995, 526 (1995). T. Kitamura et al., Fish. Sci. 62, 340 (1996). C. J. Winchell, A. P. Martin, J. Mallatt, Mol. Phylogenet. Evol. 31, 214 (2004). G. J. P. Naylor, J. A. Rayburn, O. Fedrigo, A. Lopez, in Reproductive Biology and Phylogeny of Chondrichthyes:

24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

36.

37.

327

Sharks, Batoids, and Chimaeras, W. C. Hamlett, Ed. (Science Publishers, Enfield, 2005), pp. 1–25. K. A. Dunn, J. A. McEachran, R. L. Honeycutt, Mol. Phylogenet. Evol. 27, 259 (2003). B. A. Human, E. P. Owen, L. J. V. Compagno, E. H. Harley, Mol. Phylogenet. Evol. 39, 384 (2006). S. P. Iglesias, G. Lecointre, D. Y. Sellos, Mol. Phylogenet. Evol. 34, 569 (2005). J. A. Lopez, J. A. Ryburn, O. Fedrigo, G. J. P. Naylor, Mol. Phylogenet. Evol. 40, 50 (2006). A. P. Martin, A. T. Pardini, L. R. Noble, C. S. Jones, Mol. Phylogenet. Evol. 23, 205 (2002). A. P. Martin, G. J. P. Naylor, S. R. Palumbi, Nature 357, 153 (1992). N. R. Lovejoy, E. Bermingham, A. P. Martin, Nature 396, 421 (1998). E. Valsecchi et al., J. Evol. Biol. 18, 436 (2005). R. Lawson et al., J. Zool. Lond. 237, 101 (1995). M. P. Heinicke, W. E. Duellman, S. B. Hedges, Proc. Nat. Acad. Sci. U.S.A. 104, 10092 (2007). J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 2275 (2005). A. G. Smith, D. G. Smith, B. M. Funnell, Atlas of Mesozoic and Cenozoic Coastlines (Cambridge University Press, Cambridge, 1994), pp. 99. F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds., A Geologic Time Scale 2004 (Cambridge University Press, Cambridge, 2004), pp. 610. M. J. Benton, Vertebrate Paleontology (Blackwell Science, Oxford, 2000), pp. 452.

Ray-finned fishes (Actinopterygii) Thomas J. Neara,* and Masaki Miyab a

Department of Ecology and Evolutionary Biology & Peabody Museum of Natural History, Yale University, New Haven, CT 06520, USA; bNatural History Museum and Institute, Chiba, 955-2 Aobacho, Chuo-ku, Chiba 260-8682, Japan *To whom correspondence should be addressed (thomas.near@ yale.edu)

Abstract Extant Actinopterygii, or ray-finned fishes, comprise five major clades: Polypteriformes (bichirs), Acipenseriformes (sturgeons and paddlefishes), Lepisosteiformes (gars), Amiiformes (bowfin), and Teleostei, which contains more than 26,890 species. Phylogenetic analyses of morphology and DNA sequence data have typically supported Actinopterygii as an evolutionary group, but have disagreed on the relationships among the major clades. Molecular divergence time estimates indicate that Actinopterygii diversified in the Lower Devonian (416–397 million years ago, Ma), and the major clades had diversified by the end of the Carboniferous (~300 Ma).

Actinopterygii, or ray-finned fishes, are one of the two major lineages of osteichthyan vertebrates, the other being Sarcopterygii (1). There are more than 26,890 species of actinopterygian fishes and the group has diversified into a wide range of marine and freshwater habitats (2). Typically five major clades are recognized in Actinopterygii: Polypteriformes (bichirs), Acipenseriformes (sturgeons and paddlefishes), Lepisosteiformes (gars), Amiiformes (Bowfin), and Teleostei. In this account, we review the evidence presented for the monophyly of Actinopterygii, the phylogenetic hypotheses of relationships among the major actinopterygian clades, and the inferences of divergence times resulting from analyses of DNA sequence data sampled from whole mitochondrial genomes and nuclear genes. The only plausible skepticism regarding the monophyly of actinopterygians was directed specifically at the phylogenetic relationships of Polypteriformes. Since the early part of the twentieth century polypteriforms have been considered actinopterygians (3) however, doubts

have been raised regarding the phylogenetic affinities of polypteriforms within Actinopterygii with an alternative hypothesis that they are more closely related to sarcopterygians (4–6). Phylogenetic analyses of morphological characters have supported the hypothesis that Polypteriformes is most closely related to all other extant actinopterygians (1, 7–15). The phylogenetic position of Polypteriformes consistently inferred from morphological data has also been supported in phylogenetic analyses of nuclear encoded 28S rRNA gene sequences (16, 17), DNA sequences from whole mitochondrial genomes (18, 19), and a combined data analysis of seven singlecopy nuclear genes (20). There is little doubt that the five major actinopterygian clades are each monophyletic (Amiiformes contains only a single extant species, Fig. 1). However, the hypotheses of relationships among these clades have differed dramatically among analyses of both morphological and molecular data sets. These discrepancies involve the relationships of Amiiformes and Lepisosteiformes, and whether Amiiformes, Acipenseriformes, and Lepisosteiformes form an “ancient fish” clade. Perhaps the most problematic aspect of actinopterygian phylogenetics involves the relationship of Lepisosteiformes and Amiiformes. These two lineages were traditionally grouped together along with several extinct lineages in the Holostei (21, 22). Analyses of morphological characters has resulted in an alternative hypothesis that Amiiformes is the closest relative of Teleostei and Lepisosteiformes is most closely related to this clade (23–30). However, other morphological studies (4, 31–33), and molecular phylogenetic analyses of both nuclear and mtDNA gene sequences, have resulted in a monophyletic Holostei (16, 19, 20, 29, 34).

Fig. 1 The only extant amiiform species (Amia calva), from Texas, USA. Credit: B. H. Bauer.

T.J. Near and M. Miya. Ray-fi nned fishes (Actinopterygii). Pp. 328–331 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Actinopterygii

329

Lepisosteiformes 4

Acipenseriformes

3

Amiiformes

2

Teleostei

1

Polypteriformes D

C

P

PALEOZOIC 400

300

Tr

J

K

200

Pg Ng CZ

MESOZOIC 100

0 Million years ago

Fig. 2 A timetree of ray-finned fishes (Actinopterygii). Divergence times are shown in Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

One of the more recent controversies in actinopterygian phylogenetics involves the phylogenetic analysis of whole mitochondrial genome sequences that fi nds the Amiiformes, Lepisosteiformes, and Acipenseriformes form a clade that is most closely related to Teleostei. This has been referred to as the “ancient fish” clade (19). Maximum likelihood Kishino-Hasegawa tree topology tests using the whole mitochondrial genome sequences were unable to reject several more traditional hypotheses of actinopterygian relationships, including Holostei as being most closely related to Teleostei, Amiiformes most closely related to Teleostei, and a polytomy involving Lepisosteiformes, Amiiformes, and Teleostei (19). Inoue et al. (19) cite support for the “ancient fish” clade in a study examining relationships among jawed vertebrates using insertions and deletions of amino acid sites in nuclear genes (35). This study did not support the monophyly of a clade containing Lepisosteiformes, Amiiformes, and Acipenseriformes, but only provided apomorphic character states to distinguish two clades, one containing Amiiformes, Lepisosteiformes, Acipenseriformes, and Teleostei (Actinopteri), and Teleostei as distinct from the Actinopteri (35). The inability of data sets consisting of whole mitochondrial genomes or multiple nuclear genes to reject alternative phylogenetic relationships among the major actinopterygian lineages indicates that much work remains to resolve these relationships (19, 33). Several studies have presented molecular divergence time estimates for the Actinopterygii. Yet, these investigations differ in the type of data sampled, the calibrations used to convert estimated genetic divergence to absolute age estimates, and the methods used to account for rate heterogeneity among lineages. The first set of studies discussed in this review address the age of the split between Actinopterygii and

Sarcopterygii, and these studies attempted to estimate divergence times among all major deuterostome or vertebrate lineages. Pairwise genetic distances of amino acid sequences sampled from 44 genes and calibrated with a single amniote fossil resulted in an age estimate for the split of Actinopterygii and Sarcopterygii at 450 ± 35.5 Ma (36). To account for heterogeneity of molecular evolutionary rates among lineages, Kumar and Hedges (36) excluded genes that exhibited rate heterogeneity. A very similar study that differed by using a combination of fossil and molecular calibrations resulted in a nearly identical age estimate for the Actinopterygii–Sarcopterygii split (37). A more recent study provided an age estimate of 476 Ma with a 95% credibility interval (CI) of (494– 442) for the Actinopterygii–Sarcopterygii split using amino acid sequences from 325 nuclear genes analyzed with a Bayesian local clock method and calibrated with 13 fossils (38). The second group of studies specifically addressed the divergence times among the major extant actinopterygian clades, and provides age estimates for the common ancestor of all living actinopterygians. Inoue et al. (39) presented a time-calibrated phylogeny estimated using amino acid sequences from 26 mitochondrial genomes sampled among Polypteriformes, Acipenseriformes, Lepisosteiformes, Amiiformes, and Teleostei. The timetree is shown in Fig. 2, and the phylogeny depicts a slight alteration of the “ancient fish” clade presented in an earlier study (19). Divergence times were estimated with a partitioned Bayesian strategy using the computer program Multidivtime, and a set of 13 fossil calibrations. Interestingly, the molecular age estimate for the common ancestor of all living Actinopterygii overlaps completely with the confidence interval estimated for the lower bound age of actinopterygians at 425.6 Ma using

330


Table 1. Divergence time estimates (Ma) and their confidence/ credibility intervals (CI) among ray-finned fishes (Actinopterygii). Timetree Node

Estimates Ref. (19)

Time

Ref. (33)

Time

CI

Time

CI

376–446

–

–

1

407.0

407

2

343.0

343

310–381

372

347–391

3

327.0

327

295–366

–

–

4

312.0

312

279–351

–

–

Note: Node times in the timetree are from ref. (19).

a gap analysis of the actinopterygian fossil record (40). Additionally, the Lower Devonian molecular age estimate is consistent with phylogenetic relationships of several extinct Middle and Upper Devonian actinopterygian lineages (e.g., Mimia, Moythomasia, and Kentuckia) that are phylogenetically nested between Polypteriformes and Acipenseriformes. These fossil lineages were not used as calibrations by Inoue et al. (39), but their ages ranging between 345 and 392 Ma are very close to the molecular age estimate for the most recent common ancestor of Actinopterygii (Fig. 2; Table 1). Hurley et al. (33) estimated divergence times among actinopterygian lineages using data from whole mitochondrial genomes and four nuclear genes. The phylogeny generated from a combined data analysis of the four nuclear genes was different from the tree resulting from analysis of mitochondrial genomes (Fig. 2). The phylogeny presented in Hurley et al. (33) did not contain the “ancient fish” clade, but this hypothesis could not be rejected in statistical comparisons of alternative phylogenies. A partitioned Bayesian strategy using the computer program Multidivtime was used to estimate divergence times. Critical information was presented to refine the fossil information used in calibrating the actinopterygian molecular phylogenies. Unfortunately, this study did not sample any polypteriform species, preventing Hurley et al. (33) from estimating the age of the common ancestor of all living Actinopterygii. Analysis of the nuclear genes resulted in a molecular age estimate for the common ancestor of all living actinopterygians exclusive of polypteriforms (Actinopteri) that was similar to the estimate resulting from analysis of whole mitochondrial genomes (Table 1). This was the only node shared between the actinopterygian phylogenies estimated from whole mitochondrial genomes and those inferred from nuclear genes (19, 33).

The molecular divergence times estimated for Actinopterygii support observations from the fossil record that the early diversification of this clade occurred in the Paleozoic (10–12). These age estimates also indicate that the diversification of the major extant actinopterygian clades also occurred in the Paleozoic, with subsequent diversification of major teleost clades occurring in the late Paleozoic and Mesozoic. The molecular age estimate presented for the split of Teleostei from other actinopterygians (Table 1) is more than 100 Ma older than the earliest teleost fossils (e.g., Pholidophorus) that date to the Middle Triassic (26, 41, 42). The earliest fossils of extant teleost lineages (e.g., Elopiformes, Osteoglossomorpha, and Ostariophysi) appear at the Upper Jurassic–Lower Cretaceous boundary (151–140 Ma) (43). Molecular divergence times for Actinopterygii provide a broad temporal perspective to examine macroevolutionary patterns in a clade that contains more than 50% of all extant vertebrate species. We anticipate future investigations of actinopterygian divergence times to utilize the fossil record with an increased sophistication and apply new and developing tools to estimate and correct for rate heterogeneity of molecular evolutionary rates among lineages.

Acknowledgments T.J.N. is supported by the National Science Foundation, and M.M. is supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and Japan Science for the Promotion of Science.

References 1. R. Diogo, The Origin of Higher Clades: Osteology, Myology, Phylogeny, and the Evolution of Bony Fishes and the Rise of Tetrapods (Science Publishers, Enfield, 2007). 2. J. S. Nelson, Fishes of the World, 4th ed. (John Wiley, Hoboken, 2006), pp. 601. 3. E. S. Goodrich, Palaeobiologica 1, 87 (1928). 4. G. J. Nelson, Bull. Amer. Mus. Nat. Hist. 141, 475 (1969). 5. G. J. Nelson, in Interrelationships of Fishes, P. H. Greenwood, R. S. Miles, C. Patterson, Eds. (Academic Press, London, 1973), pp. 333–349. 6. H. L. Jessen, in Interrelationships of Fishes, P. H. Greenwood, R. S. Miles, C. Patterson, Eds. (Academic Press, London, 1973), pp. 227–232. 7. B. G. Gardiner, in Interrelationships of Fishes, P. H. Greenwood, R. S. Miles, C. Patterson, Eds. (Academic Press, London, 1973), pp. 105–135.

Eukaryota; Metazoa; Vertebrata; Actinopterygii

8. D. E. Rosen, P. L. Forey, B. G. Gardiner, C. Patterson, Bull. Amer. Mus. Nat. Hist. 167, 163 (1981). 9. C. Patterson, Am. Zool. 22, 241 (1982). 10. B. G. Gardiner, Bull. Brit. Mus. (Nat. Hist.) Geol. 37, 173 (1984). 11. B. G. Gardiner, B. Schaeffer, Zool. J. Linn. Soc. 97, 135 (1989). 12. B. G. Gardiner, B. Schaeffer, J. A. Masserie, Zool. J. Linn. Soc. 144, 511 (2005). 13. M. I. Coates, Zool. J. Linn. Soc. 122, 27 (1998). 14. M. I. Coates, Phil. Trans. Roy. Soc. B 354, 435 (1999). 15. B. Schaeffer, in Interrelationships of Fishes, P. H. Greenwood, R. S. Miles, C. Patterson, Eds. (Academic Press, London, 1973), pp. 207–226. 16. H. L. V. Le, G. Lecointre, R. Perasso, Mol. Phylogenet. Evol. 2, 31 (1993). 17. R. Zardoya, A. Meyer, in Major Events in Early Vertebrate Evolution, P. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 135–155. 18. K. Noack, R. Zardoya, A. Meyer, Genetics 144, 1165 (1996). 19. J. G. Inoue, M. Miya, K. Tsukamoto, M. Nishida, Mol. Phylogenet. Evol. 26, 110 (2003). 20. K. Kikugawa, et al. BMC Biol. 2, 1 (2004). 21. A. S. Romer, Vertebrate Paleontology, 3rd ed. (University of Chicago Press, Chicago, 1966), pp. 468. 22. M. Jollie, Copeia 1984, 476 (1984). 23. C. Patterson, in Interrelationships of Fishes, P. H. Greenwood, R. S. Miles, C. Patterson, Eds. (Academic Press, London, 1973), pp. 233–305. 24. C. Patterson, in Major Features of Vertebrate Evolution, D. R. Prothero, R. M. Schoch, Eds. (Paleontological Society, Knoxville, 1994), pp. 57–84. 25. G. Arratia, J. Vert. Paleo. 21, 767 (2001). 26. G. Arratia, in Mesozoic Fishes 3: Systematics, Paleoenvironments and Biodiversity, G. Arratia, A. Tintori, Eds. (Verlag Dr. Friedrich Pfeil, Munich, 2004), pp. 279–315. 27. G. Arratia, in Mesozoic Fishes 2—Systematics and Fossil Record, G. Arratia, H.-P. Schultze, Eds. (Verlag Dr. Friedrich Pfeil, Munich, 1999), pp. 265–334.

28. 29.

30. 31. 32.

33. 34. 35. 36. 37.

38. 39. 40.

41. 42. 43.

331

L. Grande, W. E. Bemis, J. Vert. Paleo. 18 (Memoir 4), 1 (1998). B. G. Gardiner, J. G. Maisey, T. J. Littlewood, in Interrelationships of Fishes, M. L. J. Stiassny, L. R. Parenti, G. D. Johnson, Eds. (Academic Press, San Diego, 1996), pp. 117–146. W. E. Bemis, E. K. Findeis, L. Grande, Envir. Biol. Fish. 48, 25 (1997). G. J. Nelson, Ann. N.Y. Acad. Sci. 167, 18 (1969). L. Grande, in Abstracts of the Fourth International Meeting on Mesozoic Fishes—Systematics, Homology, and Nomenclature, F. J. Poyato-Ariza, Ed. (Ediciones Universidad Autónoma et Madrid, Madrid, 2005), pp. 119–121. I. A. Hurley, et al. Proc. Roy. Soc. Lond. B 274, 489 (2007). B. B. Normark, A. R. McCune, R. G. Harrison, Mol. Biol. Evol. 8, 819 (1991). B. Venkatesh, M. V. Erdmann, S. Brenner, Proc. Natl. Acad. Sci. U.S.A. 98, 11382 (September 25, 2001). S. Kumar, S. B. Hedges, Nature (London) 392, 917 (1998). S. B. Hedges, in Major Events in Early Vertebrate Evolution: Paleontology, Phylogeny, Genetics and Development, P. E. Ahlberg, Ed. (Taylor & Francis, London, 2001), pp. 119–134. J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 2275 (2005). J. G. Inoue, M. Miya, B. Venkatesh, M. Nishida, Gene 349, 227 (2005). P. C. J. Donoghue, M. P. Smith, I. J. Sansom, in Telling the Evolutionary Time, P. C. J. Donoghue, M. P. Smith, Eds. (Taylor & Francis, London, 2004), pp. 190–223. C. Patterson, D. E. Rosen, Bull. Amer. Mus. Nat. Hist. 158, 85 (1977). C. Patterson, in Fossil record 2, M. J. Benton, Ed. (Chapman & Hall, London, 1993), pp. 621–656. G. Arratia, in Mesozoic Fishes: Systematics and Paleoecology, G. Arratia, G. Viohl, Eds. (Verlag Dr. Friedrich Pfeil, Munich, 1996), pp. 243–259.

Sturgeons and paddlefishes (Acipenseriformes) Zuogang Penga,d,*, Rui Diogob, Arne Ludwigc, and Shunping Hea a

Institute of Hydrobiology, The Chinese Academy of Sciences, Wuhan, 430072, China; bDepartment of Anthropology, The George Washington University, Washington, DC, 20052, USA; cDepartment of Evolutionary Genetics, Leibniz-Institute for Zoo and Wildlife Research, Alfred-Kowalke-Street 17, 10315 Berlin, Germany; dPresent address: School of Biology, Georgia Institute of Technology, Atlanta, GA 30332, USA *To whom correspondence should be addressed (pengzuogang@ gmail.com)

Abstract The Order Acipenseriformes includes 25 living sturgeon species and two living paddlefish species, which are commonly considered “living fossils.” Phylogenetic analyses have supported two morphological divisions within acipenseriforms, Polyodontidae (paddlefishes) and Acipenseridae (sturgeons). Divergence times from molecular data range from 184 million years ago (Ma) to 114 Ma, although the oldest time is considered to be the most reliable and is in better agreement with the fossil record. The molecular estimates and fossil record suggest that the major lineages of Acipenseriformes diversified in the Jurassic and early Cretaceous (~180–100 Ma), probably associated with continental breakup.

The extant sturgeons (Acipenseridae, containing four genera—Acipenser, Huso, Pseudoscaphirhynchus, and Scaphirhynchus) and paddlefishes (Polyodontidae, containing two monospecific genera—Polyodon and Psephurus) with some extinct families form a monophyletic group of the ray-finned fishes, the Order Acipenseriformes. Sturgeons are diagnosed by presenting five rows of bony scutes or plates on their body, four barbels in front of mouth, and absence of teeth in adults. Paddlefishes are diagnosed by their paddle-like snout, absence of large scutes on their body, and minute barbels on their snout (1). Additionally, Polyodon is best known for its filtered-feeding habit based on numerous thin, elongate gill rakers unique to them among sturgeons and paddlefishes (2). Acipenseriforms only inhabit the Northern Hemisphere, and the present biogeographic distribution of the

extant species of this group reflects ancient relationships among fish faunas of Europe, Asia, and North America. Extant representatives are in two families with six genera and 27 species. Here, we review the relationships and divergence times of the major groups of acipenseriforms (Fig. 1). Until recently, our knowledge of the phylogenetic relationships of sturgeons and paddlefishes was mainly based on anatomical studies (3, 4). Researchers usually agree that the diversification of the living acipenseriforms may go back to the Jurassic, where sturgeons and paddlefishes were already diversified (5). The first comprehensive study (6) using molecular data, partial sequences of the mitochondrial genes cytochrome b (cyt b), 16S rRNA, and 12S rRNA, drew three major conclusions: the Pallid Sturgeon, Scaphirhynchus albus, was suggested as the closest species to all species of Acipenser and Huso; the two Huso species were embedded within Acipenser; and three major clades were proposed. Those clades were Acipenser sturio–Acipenser oxyrinchus, Acipenser schrenckii–Acipenser transmontanus, and all Ponto-Caspian species plus Acipenser dabryanus and Acipenser brevirostrium. However, these conclusions were tentative due to both limited taxa sampling as well as use of relatively short, partly nondiagnostic, gene fragments. More recently, studies using combined DNA data sets (4012 bp) from five (7) mitochondrial genes (cyt b, 12S rRNA, cytochrome c oxidase subunit II, tRNA Asp, and tRNAPhe) and two (8) mitochondrial gene regions (16S rRNA and NADH5) and comprehensive taxonomic coverage resulted in five well-supported conclusions: (i) the two species of paddlefish form a monophyletic clade;

Fig. 1 A juvenile Shortnose Sturgeon from North America, Acipenser brevirostrum. Credit: M. H. Sabaj.

Z. Peng, R. Diogo, A. Ludwig, and S. He. Sturgeons and paddlefishes (Acipenseriformes). Pp. 332–334 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Actinopterygii; Acipenseriformes

333

Acipenseridae 1

Polyodontidae

Early K

Jurassic

Late K

MESOZOIC 150

Paleogene

Ng

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of sturgeons and paddlefishes (Acipenseriformes). Divergence times are shown in Table 1. Abbreviations: Ng (Neogene) and K (Cretaceous).

(ii) Acipenser and Scaphirhynchus form a monophyletic assemblage, the Acipenseridae, the most basal position within them remains unresolved, held either by the genus Scaphirhynchus or by the clade containing the sea sturgeons (A. oxyrinchus and A. sturio); (iii) the two species of Huso are embedded within the genus Acipenser; (iv) there are two monophyletic groups within the Acipenser/ Huso assemblage closely correlated to their geographic distribution—the Atlantic clade and the Pacific clade; and (v) the three species of Pseudoscaphirhynchus are clustered within the Atlantic clade of Acipenser/Huso (7, 9, 10). Only one study has estimated divergence times among major groups of acipenseriforms in a comprehensive manner so far. Based on complete mitochondrial genome data and using the Bayesian relaxed molecular clock method (10), the estimated origin time for Acipenseriformes was at 390 million years ago (Ma) with a 95% credibility interval of 414–362 Ma, and the estimated time for splitting between sturgeons and paddlefishes was at 141 Ma with a 95% credibility interval of 160–132 Ma (Table 1). This latter time estimate was similar to estimates from two data sets presented in another study, 114 Ma and 145 Ma (11). Using more taxa but only one gene (cyt b), the estimate for the sturgeon–paddlefish split was somewhat older, 184 Ma (10). The divergence time of the two families shown in the timetree (Fig. 2), 184 Ma, reflects the recent estimate based on cyt b, which is in better agreement with the fossil record (10). Together with a more detailed timetree of species relationships based on cyt b (10), and the fossil record, this suggests that most of the major splits in Acipenseriformes occurred during the Jurassic and early Cretaceous, ~180–100 million years ago. These splits were most probably related to the continental breakup. Within the Polyodontidae clade the Chinese Swordfish (Pseudoscaphirhynchus gladius), with a limited distribution in Yangtze River, splits with the Mississippi Paddlefish (Pseudoscaphirhynchus spathula), which also has a

limited distribution in the Mississippi–Missouri basin, at ~68 Ma. Within the Acipenseridae, the divergence time between the A. oxyrinchus–A. sturio cluster and the rest of the acipenserids appears as ~172 Ma; the divergence time between Scaphirhynchus and the Acipenser/ Huso assemblage appears as about 151 Ma; the divergence time between the Pacific and the Atlantic clades appears as about 121 Ma (10). Acipenseriformes has existed at least since the early Jurassic (~200 Ma), and all fossil and recent taxa are from the Holarctic biogeographic region (2). The Atlantic and Pacific Oceans seemingly began to open during the Jurassic and have continued opening during the Cretaceous. About 120 million years ago, the Tethys Sea shrank further, eventually becoming the Black, Caspian, and Aral Seas (12). These geological events appear to have played an important role in acipenseriform diversification and evolution (6, 10). In summary, the acipenseriform timetree shows Jurassic to mid-Cretaceous diversification of sturgeons and paddlefishes, which is indirectly supported by fossil evidence (3, 4, 13) and is consistent with continental movements and paleogeography.

Acknowledgment Support was provided by National Natural Science Foundation of China (NSFC) to Z.P. and S.H.

References 1. J. S. Nelson, Fishes of the World, 4th ed. (John Wiley & Sons, New Jersey, 2006). 2. W. Bemis, E. Findeis, L. Grande, Environ Biol. Fish 48, 25 (1997). 3. L. Grande, W. E. Bemis, J. Vertebr. Paleontol. 11, 1 (1991). 4. L. Grande, W. E. Bemis, in Interrelationships of Fishes, M. L. J. Stiassny, L. R. Parenti, G. D. Johnson, Eds. (Academic Press, New York, 1996), pp. 85–115.

334


Table 1. Divergence time estimates (Ma) and their confidence/ credibility intervals (CI) between sturgeons and paddlefishes (Acipenseriformes). Timetree Node

1


Time

184

Ref. (10)(b)

Time

CI

Time

CI

184

200–150

141

160–132

Note: The estimates are from a Bayesian relaxed clock analysis of two data sets: (a) cytochrome b gene (b) complete mitochondrial genome data. The node time in the timetree uses estimate (a).

5. M. L. J. Stiassny, E. O. Wiley, G. D. Johnson, M. R. Carvalho, in Assembling the Tree of Life, M. J. Donaghue, J. Cracraft, Eds. (Oxford University Press, New York, 2004), pp. 200–247.

6. V. J. Birstein, R. DeSalle, Mol. Phylogenet. Evol. 9, 141 (1998). 7. J. Krieger, A. K. Hett, P. A. Fuerst, E. Artykhin, A. Ludwig, J. Appl. Ichthyol. 24 (S1), 36 (2008). 8. V. J. Birstein, P. Doukakis, R. DeSalle, Copeia 2, 287 (2002). 9. A. Ludwig, et al. Genetics 158, 1203 (2001). 10. Z. Peng et al., Mol. Phylogenet. Evol. 42, 854 (2007). 11, J. G. Inoue, M. Miya, B. Venkatesh, M. Nishida, Gene 349, 227 (2005). 12. G. Smith, D. G. Smith, B. M. Funnell, Atlas of Mesozoic and Cenozoic Coastlines (Cambridge University Press, Cambridge, 1994). 13. F. Jin, in Sixth Symposium Mesozoic Terrestrial Ecosystems and Biota, Short Papers, A. Sun, Y. Wang, Eds. (China Ocean Press, Beijing, 1995), pp. 15–22. 14. A. Choudhury, T. A. Dick, J. Biogeogr. 25, 623 (1998).

Teleost fishes (Teleostei) Zuogang Penga,c, Rui Diogob, and Shunping Hea,* a

Institute of Hydrobiology, The Chinese Academy of Sciences, Wuhan, 430072, China; bDepartment of Anthropology, The George Washington University, Washington, DC, 20052, USA; cPresent address: School of Biology, Georgia Institute of Technology, Atlanta, GA 30332, USA *To whom correspondence should be addressed ([email protected])

Abstract Living Teleost fishes (~26,840 sp.) are grouped into 40 orders, comprising the Infraclass Teleostei of the Class Actinopterygii. With few exceptions, morphological and molecular phylogenetic analyses have supported four subdivisions within Teleostei: Osteoglossomorpha, Elopomorpha, Otocephala (= Ostarioclupeomorpha), and Euteleostei. Despite the progress that has been made in recent years for the systematics of certain teleost groups, the large-scale pattern of teleost phylogeny remains open. The teleost timetree shows that the major groups diversified from mid-Paleozoic to early Mesozoic, 400–200 million years ago, most probably before the breakup of the supercontinents.

Teleosts are a modern group of fishes including more than 26,000 species (1), which are grouped into 40 orders. They are typically grouped together with the garfishes (Lepisosteiformes) and Bowfin (Amiifomes) in the Subclass Neopterygii. Teleosts are the most speciesrich and diversified group of all the vertebrates. There are more teleost species than all the other vertebrates combined (2). They dominate in the world’s rivers, lakes, and oceans. There are four subdivisions within extant teleosts: Osteoglossomorpha (e.g., mooneyes and bonytongues), Elopomorpha (e.g., eels, tarpons, and bonefishes), Otocephala (e.g., ostariophysan and clupeomorph teleosts), and Euteleostei (the remaining teleosts, e.g., Argentiniformes, Osmeriformes, Salmoniformes, and Neoteleostei) (3). At least 27 anatomical shared derived traits were found by de Pinna (4) to support the monophyly of the Teleostei. Here, we review the relationships and divergence times of the major groups of teleosts (Fig. 1).

Until recently, the classification of teleosts pioneered by Greenwood et al. (5) and expanded on by Patterson and Rosen (6) has followed the arrangement proposed by Nelson (7) and today is still reflected in fish textbooks and papers. In it, species were placed in four major groups: Osteoglossomorpha, Elopomorpha, Otocephala, and Euteleostei. This division was based on multiple morphological characters and molecular evidence. Based on morphological characters, Osteoglossomorpha was considered as the most plesiomorphic living teleosts by several works (6, 7). However, the anatomical studies of Arratia (8–10) supported that elopomorphs, and not osteoglossomorphs, are the most plesiomorphic extant teleosts. This latter view was supported by the results of the most extensive morphologically based cladistic analysis published so far on osteichthyan higher-level phylogeny, which included 356 osteological and myological characters and 80 terminal taxa, including both extant and fossil species (3). An early molecular phylogeny based on nuclear 28S rDNA (11) supported a close relationship between (Osteoglossomorpha + Elopomorpha) and (Otocephala + Euteleostei). The clade (Otocephala + Euteleostei; Clupeocephala) was subsequently supported by anatomical (3, 9) and molecular data (12, 13). Molecular studies using longer sequences—complete mitochondrial genomes and greater taxonomic coverage (12, 13)—indicate that the Osteoglossomorpha is the

Fig. 1 An armored catfish from South America, Leporacanthicus triactis. Photo credit: M. H. Sabaj.

Z. Peng, R. Diogo, and S. He. Teleost fishes (Teleostei). Pp. 335–338 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

336


Cypriniformes

6

Clupeiformes 7

Salmoniformes

2

Anguilliformes 1

5

Albuliformes

3

Elopiformes Osteoglossidae

D

C

P

PALEOZOIC 300

J

Tr

K

200

Pg Ng CZ

MESOZOIC 100

Elopomorpha

Perciformes 8

Euteleostei

Gonorynchiformes

Osteoglossomorpha

4

Elopocephala

Characiformes

9

Otocephala

Siluriformes 10

0 Million years ago

Fig. 2 A timetree of teleost fishes (Teleostei). Divergence times are shown in Table 1. The branch shown as Characiformes represents the non-monophyletic group of characiforms with species nested with other taxa from the Order Gymnotiformes

as shown in ref. (13). Abbreviations: C (Carboniferous), CZ (Cenozoic), D (Devonian), J (Jurassic), K (Cretaceous), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

most basal extant teleostean group, as first proposed by Patterson and Rosen (6) and subsequently supported by Lauder and Liem (14) and Nelson (7) (Fig. 2). Only two studies have estimated the divergence times among the major lineages of Teleostei in a comprehensive manner. Both studies used the complete mitochondrial genome data and Bayesian method with different sampling and concerns (13, 15). Both studies used calibrations from the teleost fossil record. The relationships obtained in both studies were similar, although most of the dates estimated by Peng et al.’s study were older than those in Inoue et al.’s study (Table 1). For example, divergence between the Osteoglossomorpha and the remaining teleostean groups in Peng et al.’s study was estimated to be Devonian (384 million years ago, Ma), considerably older than the estimates of 285 Ma (data set 1) or 334 Ma (data set 2) and paleontological estimate of early Permian or mid-Carboniferous in Inoue et al.’s study. The timetree of teleosts based on Peng et al. (Fig. 2) shows that most of the major splits in the tree occurred during mid-Paleozoic to early Mesozoic, 400–200 Ma. Most of those divergences took place when all the continents were joined in a single supercontinent, Pangaea. These included the divergence of osteoglossomorphs

and elopocephalans, elopomorphs and clupeocephalans (Euteleostei + Otocephala), and euteleosts and otocephalans. However, the earliest fossil lineages within teleosts recorded so far were dated back to late Triassic–early Jurassic (~210–200 Ma). There is a significant difference between the time of divergence of major teleostean groups obtained from molecular clocks and those indicated by the oldest fossil record of these groups. For example, the divergence time between euteleosts and otocephalans (the origin of Otocephala) was estimated to be of ~307–230 Ma (13, 15). However, the oldest otocephalan fossil discovered so far is from the late Jurassic, ~150 Ma (9). Nevertheless, as explained by Diogo (16, 17), there is strong indirect evidence supporting that the origin of certain otocephalan groups such as catfishes (Siluriformes) is in fact very likely older than the direct evidence provided by the oldest fossils of those groups might indicate (e.g., evidence regarding the geographic distribution of fossil and/or extant taxa, the phylogenetic relationships between taxa from different continents, the fact that some of the oldest fossils discovered so far for a certain group occupy in fact a phylogenetically derived position within that group). For example, although the oldest catfish (siluriform) fossil discovered to date is about 75–72 Ma, a broader analysis of

Eukaryota; Metazoa; Vertebrata; Actinopterygii; Teleostei

337

Table 1. Divergence times and their confidence/credibility intervals (CI) among teleost fishes (Teleostei). Timetree Node

Estimates Ref. (13)

Time Time

Ref. (15)(a) CI

Time

Ref. (15)(b)

CI

Time

CI

1

384.0

384

447–273

285

320–253

334

372–295

2

355.0

355

420–251

265

300–234

315

352–276

3

336.0

336

401–236

–

–

–

–

4

307.0

307

371–215

230

264–200

278

314–241

5

286.0

286

352–198

228

262–199

264

301–227

6

282.0

282

343–197

201

233–172

239

275–204

7

270.0

270

332–188

–

–

–

–

8

264.0

264

327–183

191

221–164

232

267–197

9

251.0

251

311–175

–

–

–

–

10

210.0

210

265–144

–

–

–

–

Note: Node times in the timetree are based on ref. (13). Estimates from ref. (15) are from two different data sets: (a) = Data set 1 and (b) = Data set 2.

the catfish biogeographical distribution, phylogeny, and fossil record points out that by the late Cretaceous these fishes already had a worldwide distribution. This indicates that the origin of Siluriformes very likely occurred much before 75–72 Ma. The paleobiogeographic data on other teleost groups do also provide interesting indirect evidence supporting that those groups might have an older origin than that indicated by a direct and exclusive analysis of their oldest fossil (18, 19). In summary, with respect to the divergence times obtained so far from molecular studies, they indicate that the origin of the major teleostean groups is probably much older than a direct, exclusive analysis of the oldest fossil of each of these groups might suggest. It should also be noted that there possibly still were some Pangean connections between Gondwana and Laurasia in the late Jurassic, and perhaps even in the early Cretaceous (16, 17, 20). If this is so, this would help to explain the Pangean distribution of taxa such as the cypriniforms, characiforms, and siluriforms, which are primary freshwater fishes with relatively few, and phylogenetically rather derived, marine members.

Acknowledgment Support was provided by National Natural Science Foundation of China (NSFC) to Z.P. and S.H.

References 1. J. S. Nelson, Fishes of the World, 4th ed. (John Wiley & Sons, New Jersey, 2006). 2. M. L. J. Stiassny, E. O. Wiley, G. D. Johnson, M. R. Carvalho, in Assembling the Tree of Life, M. J. Donaghue, J. Cracraft, Eds. (Oxford University Press, New York, 2004), pp. 200–247. 3. R. Diogo, in On the Origin and Evolution of HigherClades: Osteology, Myology, Phylogeny and Macroevolution of Bony fishes and the Rise of Tetrapods (Science Publishers, Enfield, 2007). 4. M. C. C. de Pinna, in Interrelationships of Fishes, M. L. J. Stiassny, L. R. Parenti, G. D. Johnson, Eds. (Academic Press, San Diego, 1996), pp. 147–162. 5. P. H. Greenwood, D. E. Rosen, S. H. Weitzman, G. S. Myers, Bull. Amer. Mus. Nat. Hist. 131, 339 (1966). 6. C. Patterson, D. E. Rosen, Bull. Amer. Mus. Nat. Hist. 158, 81 (1977). 7. J. S. Nelson, Fishes of the World, 3rd ed. (John Wiley & Sons, New York, 1994). 8. G. Arratia, in Early Vertebrates and Related Problems in Evolutionary Biology, M. M. Chang, H. Liu, G. R. Zhang, Eds. (Science Press, Beijing, 1991), pp. 249–340. 9. G. Arratia, Palaeo. Ichthyologica 7, 168 (1997). 10. G. Arratia, in Mesozoic Fishes 3—Systematics, Paleoenvironments, and Biodiversity, G. Arratia, A. Tintori, Eds. (Verlag Dr. Friedrich Pfeil, München, 2004), pp. 279–315.

338


11. H. L. Le, G. Lecointre, R. Perasso, Mol. Phylogenet. Evol. 2, 31 (1993). 12. J. G. Inoue, M. Miya, K. Tsukamoto, M. Nishida, Mol. Phylogenet. Evol. 20, 275 (2001). 13. Z. Peng, S. He, J. Wang, W. Wang, R. Diogo, Gene 370, 113 (2006). 14. G. V. Lauder, K. F. Liem, Bull. Mus. Comp. Zool. 150, 95 (1983). 15. J. G. Inoue, M. Miya, B. Venkatesh, M. Nishida, Gene 349, 227 (2005).

16. R. Diogo, Anim. Biol. 54, 331 (2004). 17. R. Diogo, in Adaptations, Homoplasies, Constraints, and Evolutionary Trends: Catfish Morphology, Phylogeny and Evolution, A Case Study on Theoretical Phylogeny and Macroevolution (Science Publishers, Enfield, 2005). 18. F. J. Poyato-Ariza, Palaeo. Ichthyologica 6, 1 (1996). 19. A. Filleul, J. G. Maisey, Am. Mus. Novitates 3455, 1 (2004). 20. J. C. Briggs, J. Biogeogr. 32, 287 (2005).

Notothenioid fishes (Notothenioidei) Thomas J. Near Department of Ecology and Evolutionary Biology & Peabody Museum of Natural History, Yale University, New Haven, CT 06520, USA ([email protected])

Abstract Notothenioids are a clade of acanthomorph teleosts that represent a rare example of adaptive radiation among marine fishes. The notothenioid Antarctic Clade is characterized by extensive morphological and ecological variation and adaptations to avoid freezing in the ice-laden water of Southern Ocean marine habitats. A recent analysis of notothenioid divergence times indicates that the clade dates to the Cretaceous (125 million years ago, Ma), but the Antarctic Clade diversified near the Oligocene–Miocene boundary (23 Ma). These age estimates are consistent with paleogeographic events in the Southern Ocean that drove climate change from temperate to the polar conditions observed today.

Notothenioids represent an adaptive radiation of teleost fishes in the frigid waters of the Southern Ocean surrounding Antarctica (1). Of the ~129 recognized species, 101 are found in marine costal habitats of Antarctica (Fig. 1), and the remaining species are distributed along costal areas of southern South America, the Falkland Islands, southern New Zealand, southern Australia, and Tasmania (2). In addition to a diverse array of adaptations to survive in the freezing Antarctic marine habitats, notothenioids are unique in that they completely dominate the fish fauna of the Southern Ocean. Among benthic fish samples taken on the Antarctic shelf, notothenioids comprise nearly 77% of the species diversity, more than 91% of the species abundance, and ~91% of the biomass (3). A hypothesized key innovation that facilitated the diversification of Antarctic notothenioids is the origin of an antifreeze glycoprotein from a tyripsinogen-like ancestral gene that confers protection from freezing in the subzero Southern Ocean waters (4). The ecological and morphological diversity of Antarctic notothenioids is extensive and it is thought that the clade diversified

and fi lled vacant niches after the onset of polar conditions ~35 Ma (2). The fossil fishes preserved in the Eocene La Meseta Formation on Seymour Island at the tip of the Antarctic Peninsula indicate that before the development of polar conditions the nearshore fish fauna of Antarctica was diverse, cosmopolitan, and not dominated by notothenioids (5). The only documented notothenioid fossil is a well-preserved neurocranium of the extinct species Proeleginops grandeastmanorum from the La Meseta Formation that is dated to ~40 Ma (6–10). Ecologically, Antarctic notothenioids have diversified into both benthic and water column habitats (2). Several lineages are able to utilize water column habitats despite lacking a swim bladder by modification of buoyancy through the reduction of ossification and the evolution of intra- and intermuscular lipid deposits (11, 12). A notable group of notothenioid species is the Channichthyidae, or icefishes (Fig. 1). Species in this clade are also called the “white-blooded” fishes, because of the absence of the oxygen-transporting molecule hemoglobin, which is the result of deleted globin loci possibly initiated by interspecific hybridization and subsequent introgression (13). These are the only vertebrates that exhibit this bizarre phenotype and it is thought that the persistence of this apparently deleterious trait is due to the cold oxygen-saturated water that provides adequate oxygen via passive diff usion into the body (14).

Fig. 1 A 29.2 cm long (standard length) channichthyid notothenioid (Chionodraco myersi: YPM 16533) sampled from the Bransfield Strait. Credit: T. J. Near.

T. J. Near. Notothenioid fishes (Notothenioidei). Pp. 339–343 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

340

THE TIMETREE OF LIFE Channichthyidae 8

Bathydraconidae-2

7

Bathydraconidae-1

6

Artedidraconidae 5

9

4 3

Harpagiferidae Nototheniidae-3 Nototheniidae-2 Nototheniidae-1

2 1

Pseudaphritidae Bovichtidae

Early K

Late K

Paleogene

125

100

Neogene

CENOZOIC

MESOZOIC 75

50

25

0 Million years ago

Fig. 2 A timetree of notothenioid fishes (Notothenioidei). Divergence times are shown in Table 1. Abbreviation: K (Cretaceous).

The monophyly of Notothenioidei has been supported in phylogenetic analyses of morphological characters (15, 16) and DNA sequences from mitochondrial and nuclear genes (17–20). Phylogenetic relationships among lineages within Notothenioidei inferred from analyses of morphological and molecular data sets are generally consistent with traditional taxonomic hypotheses developed during the time of great Antarctic exploration in the early twentieth century. Taxonomically, eight families are recognized in the Notothenioidei and all but the Bathydraconidae (Dragonfishes) and Nototheniidae were resolved as monophyletic groups in molecular phylogenetic analyses (21–28). Monophyly of Nototheniidae was supported in morphological phylogenetic analyses (29), and phylogenetic analyses of complete mtDNA 16S rRNA (23). Other phylogenetic analyses have focused on specific notothenioid subclades, including the Channichthyidae (30–33), Bathydraconidae (22), Artedidraconidae (15), and Nototheniidae (24, 34). One important result from these phylogenetic investigations is the consistent monophyly of the Antarctic Clade (Nototheniidae, Harpagiferidae, Artedidraconidae, Bathydraconidae, and Channichthyidae) that comprises the major lineages of notothenioids that are found in the Southern Ocean south of the Antarctic Polar Front (25, 26). Eleven published studies have presented molecular divergence time estimates for notothenioid clades. The ages estimated from these molecular analyses were quite broad, but fairly similar across studies. One study estimated the divergence time of the entire notothenioid

radiation at 57–45 Ma (28). The range of estimated divergence times for the common ancestor of the Antarctic Clade was 21–5 Ma (4, 21, 26–28, 35). Several of these studies also estimated the divergence times of particular lineages within the Antarctic Clade. Age estimates for the Channichthyidae ranged between 23.3 and 2 Ma (26, 31, 36), and molecular age estimates for the nototheniid clade Trematomus (with and without Trematomus scotti) were 3.4–2.8 Ma (26, 34). Any effort that attempts to estimate divergence times using molecular data will face a specter of uncertainty and ever present confounding variables, and all of these notothenioid molecular divergence time studies suffer minimally from one of three severe methodological problems. First, calibration is a major concern as most estimates of notothenioid divergence times rely on external or “universal” rates of DNA sequence evolution estimated for other clades and applied to notothenioids (4, 27, 31, 34–37). The extreme example of this strategy was the application of the rate of mtDNA evolution in trout and salmon to estimate the age of the Antarctic Clade from pairwise genetic distances calculated from a nuclear-encoded intron (4). This strategy is problematic and ill-advised, because animal mtDNA genes have a much higher nucleotide substitution rate than any sampled nuclear gene, including introns (38). Even when paleogeographic calibrations have been used, they do not represent current hypotheses in the geological literature. For example, the separation of New Zealand from East Gondwana is given as 57 Ma in ref.

Eukaryota; Metazoa; Vertebrata; Actinopterygii; Teleostei; Notothenioidea

341

Table 1. Divergence times (Ma) and their credibility/confidence intervals (CI) among notothenioid fishes (Notothenioidei). Timetree Node

1

Time

125.0

Estimates Ref. (4)

Ref. (21)

Ref. (26)

Ref. (27)

Ref. (28)

Ref. (35)

Time

Time

Time

Time

Time

Time

Time

Ref. (42) CI

–

–

–

–

57–45

–

125.0

129–121

2

47.0

–

–

–

–

–

–

47.0

48.4–45.6

3

24.1

14–5

12–8

21

16–10

15–11

15–7

24.1

24.6–23.6

4

22.4

–

–

–

–

–

–

22.4

22.9–21.9

5

18.9

–

–

–

–

–

–

18.9

19.3–18.5

6

17.0

–

–

–

–

–

–

17.0

17.4–16.6

7

16.1

–

35–15

–

–

–

–

16.1

16.4–15.8

8

15.8

–

–

–

–

–

–

15.8

16.1–15.5

9

12.7

–

–

–

–

–

–

12.7

–

Note: Node times in the timetree are from ref. (42).

(28) and used to calibrate the divergence of Bovichtidae from other notothenioids. However, it appears that 80 Ma is a more appropriate age for this event (39, 40). In another study, the fragmentation of Australia and Antarctica is presented as occurring 38 Ma and used to calibrate the divergence of Pseudaphritidae and all other notothenioids (26). However, a range of 52–35 Ma is the more appropriate age for this paleogeographic event (39, 40). Second, most molecular estimates of divergence times in notothenioids have ignored heterogeneity of molecular evolutionary rates among lineages. However, a few studies have used relative rate tests, where species exhibiting departure from rate heterogeneity were excluded from the analysis (26–28). Relative rate tests are problematic, because they measure the substitution rate in only a small portion of the phylogeny and the statistical significance of relative rate tests must be corrected when multiple tests are used (41). Third, divergence time estimates have often been presented without confidence or credibility intervals. The collective problems exhibited among these notothenioid divergence time estimates were directly addressed in a study that used a fossil calibration and a tree-based method to account for rate heterogeneity (42). A molecular phylogeny and branch lengths were estimated from mtDNA gene sequences sampled from the 12S and 16S rRNA genes using maximum likelihood. The fossil P. grandeastmanorum was used to provide a fi xed minimal age estimate of 40 Ma for the node that represents the most recent common ancestor of Eleginopidae

and the Antarctic Clade. Penalized likelihood was used to correct for molecular evolutionary rate heterogeneity among lineages and confidence intervals were calculated with a bootstrapping method. The time-calibrated phylogeny is presented in Fig. 2 and the divergence times are shown in Table 1. The clade Nototheniidae is not monophyletic in this analysis, and the timetree in Fig. 2 contains three nototheniid clades: clade 1 containing all the species sampled from Dissostichus, Notothenia, Aethotaxis, Lepidonotothen, Patagonotothen, Trematomus, and Indonotothen; clade 2 containing Gobionotothen gibberifrons and Gobionotothen acuta; and clade 3 containing Pleuragramma antarcticum. Bathydraconidae is also not monophyletic and the species are distributed between two clades: clade 1 contains Gymnodraco acuticeps, and clade 2 contains Cygnodraco mawsoni and Parachaenichthys charcoti. Most of the estimated divergence times from this penalized likelihood rate smoothed molecular phylogeny are older than age estimates resulting from analyses of pairwise genetic distances. For example, the molecular divergence time estimate for the common ancestor of Notothenioidei is ~125 Ma, more than double the single previous molecular estimate (28). One study used the fragmentation of Antarctica and Australia as a calibration set at 38 Ma for the common ancestor of Pseudaphritidae and the remaining notothenioids (26); however, the penalized likelihood estimated age for this node is substantially older (Fig. 2; Table 1). Perhaps of greatest interest to comparative biologists is the age of

342


the common ancestor of the Antarctic Clade, since this is the lineage that exhibits adaptations to polar conditions such as the presence of antifreeze glycoproteins. Previous estimates using pairwise genetic distances had estimated this clade diversified between 21 and 5 Ma (4, 35) (Table 1). The penalized likelihood estimate for the age of this clade is older and close to the Oligocene–Miocene boundary at 23 Ma (Fig. 2; Table 1). This indicates that the Antarctic Clade diversified after the development of the unrestricted Antarctic Circumpolar Current and Antarctic Polar Front, which formed after the opening of the Drake Passage between South America and Antarctica in the late Eocene (43). The available divergence time estimates in notothenioids are far from providing the last word on the topic. Molecular age estimates for notothenioids need exploration with data sets that are much larger in terms of the number of genes and species sampled, relative to the partial mtDNA gene sequences used in all previous studies. In addition, the consistency of the P. grandeastmanorum fossil calibration should be assessed in a cross-validation analysis with external fossil calibrations sampled from other acanthomorph teleost clades. These future investigations also need to utilize strategies of divergence time estimation that account for heterogeneity of molecular evolutionary rates, as well as uncertainty in the fossil calibrations. Such analyses will provide reliable credibility intervals for the molecular age estimates. The nonparametric bootstrap procedure used in the penalized likelihood estimate of notothenioid divergence times does not account for uncertainties in the fossil calibrations and likely results in misleadingly narrow confidence intervals (44, 45). Robust divergence time estimates for notothenioids will facilitate investigations of the role of climate change and adaptive evolution in the diversification of this Antarctic adaptive radiation.

Acknowledgment T.J.N. is supported by the U.S. National Science Foundation.

References 1. A. Clarke, I. A. Johnston, Trends Ecol. Evol. 11, 212 (1996). 2. J. T. Eastman, Antarctic Fish Biology: Evolution in a Unique Environment (Academic Press, San Diego, 1993). 3. J. T. Eastman, Polar Biol. 28, 93 (2005). 4. L. Chen, A. L. DeVries, C.-H. Cheng, Proc. Natl Acad. Sci. U.S.A. 94, 3811 (1997).

5. J. T. Eastman, Antarctic Sci. 12, 276 (2000). 6. J. A. Case, in Geology and Paleontology of Seymour Island, Antarctic Peninsula, R. M. Feldman, M. O. Woodburne, Eds. (Geological Society of America, Boulder, 1988), pp. 523–530. 7. J. A. Case, M. O. Woodburne, D. S. Chaney, in Geology and Paleontology of Seymour Island, Antarctic Peninsula, R. M. Feldman, M. O. Woodburne, Eds. (Geological Society of America, Boulder, 1988), pp. 505–521. 8. M. O. Woodburne, W. J. Zinsmeister, J. Paleontol. 58, 913 (1984). 9. J. T. Eastman, L. Grande, Antarctic Sci. 3, 87 (1991). 10. A. V. Balushkin, Vopr. Ikhtiol. 34, 298 (1994). 11. A. L. DeVries, J. T. Eastman, Nature 271, 352 (1978). 12. J. T. Eastman, A. L. DeVries, Copeia 1982, 385 (1982). 13. T. J. Near, S. K. Parker, H. W. Detrich, Mol. Biol. Evol. 23, 2008 (2006). 14. G. di Prisco, E. Cocca, S. K. Parker, H. W. Detrich, Gene 295, 185 (2002). 15. R. R. Eakin, in Antarctic Research Series, Vol. 31, Biology of the Antarctic Seas IX, L. S. Kornicker, Ed. (American Geophysical Union, Washington, 1981), pp. 81–147. 16. P. A. Hastings, in A History and Atlas of the Fishes of the Antarctic Ocean, R. G. Miller, Ed. (Foresta Institute for Ocean and Mountain Studies, Carson City, 1993), pp. 99–107. 17. W.-J. Chen, C. Bonillo, G. Lecointre, Mol. Phylogenet. Evol. 26, 262 (2003). 18. A. Dettaï, G. Lecointre, Antarctic Sci. 16, 71 (2004). 19. A. Dettaï, G. Lecointre, C. R. Biol. 328, 647 (2005). 20. W. L. Smith, M. T. Craig, Copeia 2007, 35 (February 28, 2007). 21. L. Bargelloni, L. Zane, N. Derome, G. Lecointre, T. Patarnello, Antarctic Sci. 12, 259 (2000). 22. N. Derome, W.-J. Chen, A. Dettai, C. Bonillo, G. Lecointre, Mol. Phylogenet. Evol. 24, 139 (2002). 23. T. J. Near, J. J. Pesavento, C. H. C. Cheng, Mol. Phylogenet. Evol. 32, 881 (2004). 24. S. Sanchez, A. Dettai, C. Bonillo, C. Ozouf-Costaz, H. W. Detrich, G. Lecointre, Polar Biol. 30, 155 (2007). 25. T. J. Near, C. H. C. Cheng, Mol. Phylogenet. Evol. 47, 832 (2008). 26. L. Bargelloni, S. Marcato, L. Zane, T. Patarnello, Syst. Biol. 49, 114 (2000). 27. L. Bargelloni, G. Lecointre, in Fishes of Antarctica: A Biological Overview, G. D. Prisco, E. Pisano, A. Clarke, Eds. (Springer-Verlag, Berlin-Heidelberg, 1998), pp. 259–273. 28. L. Bargelloni, T. Patarnello, P. A. Ritchie, B. Battaglia, A. Meyer, in Antarctic Communities, B. Battaglia, J. Valencia, D. W. H. Walton, Eds. (Cambridge University Press, Cambridge, 1997), pp. 45–50. 29. A. V. Balushkin, J. Ichthyol. 40, S74 (2000). 30. T. Iwami, Mem. Nat. Inst. Polar Res. Tokyo 36, 1 (1985).

Eukaryota; Metazoa; Vertebrata; Actinopterygii; Teleostei; Notothenioidea

31.

32. 33. 34.

35.

36. 37.

W.-J. Chen, C. Bonillo, G. Lecointre, in Fishes of Antarctica: A Biological Overview, G. Di Prisco, E. Pisano, A. Clarke, Eds. (Springer-Verlag, Berlin-Heidelberg, 1998), pp. 287–298. T. J. Near, J. J. Pesavento, C.-H. C. Cheng, Mol. Phylogenet. Evol. 28, 87 (2003). O. S. Voskoboinikova, Zool. Zh. 79, 321 (2000). P. A. Ritchie, L. Bargelloni, A. Meyer, J. A. Taylor, J. A. Macdonald, D. M. Lambert, Mol. Phylogenet. Evol. 5, 383 (1996). L. Bargelloni, P. A. Ritchie, T. Patarnello, B. Battaglia, D. M. Lambert, A. Meyer, Mol. Biol. Evol. 11, 854 (1994). J. T. Eastman, A. R. McCune, J. Fish. Biol. 57, 84 (2000). A. Stankovic, K. Spalik, E. Kamler, P. Borsuk, P. Weglenski, Polar Biol. 25, 203 (2002).

38. 39. 40.

41.

42. 43. 44.

45.

343

W.-H. Li, Molecular Evolution (Sinauer, Sunderland, Massachusetts, 1997). I. Sanmartín, F. Ronquist, Syst. Biol. 53, 216 (2004). A. G. Smith, D. G. Smith, B. M. Funnell, Atlas of Mesozoic and Cenozoic coastlines (Cambridge University Press, Cambridge, 2004). M. J. Sanderson, in Molecular Systematics of Plants, II: DNA Sequencing, D. E. Soltis, P. S. Soltis, J. J. Doyle, Eds. (Kluwer Academic Publishers, Amsterdam, 1998), pp. 242–264. T. J. Near, Antarctic Sci. 16, 37 (2004). H. D. Scher, E. E. Martin, Science 312, 428 (2006). J. L. Thorne, H. Kishino, in Statistical Methods in Molecular Evolution, R. Nielsen, Ed. (Springer, New York, 2005), pp. 233–256. Z. Yang, Computational Molecular Evolution (Oxford University Press, Oxford, 2006).

Labyrinth fishes (Anabantoidei) Lukas Rüber Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK ([email protected])

Abstract Labyrinth fishes (Anabantoidei) are grouped into three families (Osphronemidae, Helostomatidae, and Anabantidae) within the teleost Order Perciformes. Recent phylogenetic analyses have resulted in major changes in their classification. The Family Belontiidae is no longer recognized. The pike-head Luciocephalus, previously considered a separate family (Luciocephalidae) and closest relative of the remaining anabantoids, is now placed in a derived position within the Osphronemidae. The anabantoid timetree shows that the three families diverged either in the Middle Eocene (~40 million years ago, Ma) or in the late Cretaceous (~90 Ma) depending on the assignment of the only available anabantoid fossil.

Labyrinth fishes form a natural (monophyletic) group of teleost fishes, the Suborder Anabantoidei. They are arranged into three families: Osphronemidae (~120 species; gouramies, paradise fishes, and fighting fishes; Fig. 1), Helostomatidae (one species, Kissing Gourami), and Anabantidae (28 species; climbing gouramies and bushfishes). Although a comparatively small group, anabantoids exhibit a striking variation in size, ranging from dwarfed forms such as Parosphromenus ornaticauda, with 19 mm standard length, to large forms such as the giant gouramies of the genus Osphronemus, with up to 70 cm standard length (1, 2). A number of species play an important role as food fishes and are widely used in aquaculture, whereas others are important and highly colorful aquarium fishes. Labyrinth fishes show an astonishing diversity in breeding behavior that is rarely found in any other fish group (3, 4). Parental care is dominant and occurs in 16 of the 19 anabantoid genera. Reproductive modes range from free-spawning without parental care to substrate spawning, submerged plant nest building, bubble nesting, and mouthbrooding with parental care. Labyrinth fishes are diagnosed by

the presence of a peculiar organ above the gills (suprabranchial), consisting of a greatly modified upper element of the gill arches (epibranchial 1), which is housed in a cavity above the gills. Both, the wall of the cavity and the modified epibranchial are covered with respiratory epithelium, and assist in accessory air-breathing. The subrabranchial organ is also called labyrinth organ because of its complex folding that greatly increases respiratory surface. Labyrinth fishes are typically grouped together with the snakeheads (Channoidei, Channidae) in the Labyrinthici (5). Here, I review the relationships and divergence times of labyrinth fishes that include the three families: Osphronemidae, Helostomidae, and Anabantidae. Although anabantoids were already recognized as a natural assemblage in the early nineteenth century by Cuvier and Valenciennes (6) their phylogenetic intrarelationships have been highly contentious. Most of the controversy is focused on the relative phylogenetic position of the enigmatic pike-head Luciocephalus pulcher. This highly morphologically derived teleost (7) was originally included in the family Esocidae (Esociformes) by Gray (8), but subsequently considered a member of the Anabantoidei by Bleeker (9, 10). Later on, Berg (11) and Liem (12) rejected a close relationship of the two taxa.

Fig. 1 A fighting fish (Betta channoides) from Borneo. Credit: Z. Hang.

L. Rüber. Labyrinth fishes (Anabantoidei). Pp. 344–347 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Actinopterygii; Anabantoidei

345

Anabantidae 2

Helostomatidae

1

Osphronemidae

Paleogene

Neogene CENOZOIC

50

25

0 Million years ago

Fig. 2 A timetree of labyrinth fishes (Anabantoidei). Divergence times are from Table 1.

However, it is now generally accepted that Luciocephalus belongs to the anabantoids based on several derived morphological characters (13–15). The first phylogenetic hypothesis of anabantoid relationships was proposed by Lauder and Liem (15), who divided the suborder into five families: Luciocephalidae, Anabantidae, Helostomatidae, Osphronemidae, and Belontiidae. They identified Luciocephalidae as the most basal anabantoid family, Anabantidae as the closest relative of Helostomatidae, and Osphronemidae as the closest relative of Belontiidae. However, Britz (13, 14, 16) and Britz et al. (17) revised Lauder and Liem’s hypothesis in essential aspects. Britz (13) showed that Luciocephalus belongs to a monophyletic group, called Osphronemidae (18), that includes Osphronemus, and Liem’s belontiids. The monophyly of Osphronemidae and Anabantidae is well supported by morphological studies (5, 15). The only molecular phylogenetic study addressing anabantoid relationships was based on four mitochondrial genes (cyt b, 12S rRNA, tRNA Val, and 16S rRNA) and one nuclear gene (RAG1) and an extensive taxonomic coverage (4). The monophyly of both Osphronemidae and Anabantidae are well supported (Fig. 2). Based on morphological evidence, Lauder and Liem (15) considered Helostoma temminkii, the only representative of the Helostomatidae, to be the closest relative of the Family Anabantidae. The molecular phylogenetic analyses on the other hand were unable to resolve the relative position of the Helostomatidae with respect to the other two anabantoid families. While the mitochondrial genes indicated that Anabantidae is the closest relative of the group containing Osphronemidae and Helostomidae, RAG1 and a combined nuclear and mitochondrial DNA data set showed that Osphronemidae is the closest relative of the group containing Anabantidae and Helostomidae. Short internal branches connecting the Osphronemidae, Anabantidae, and Helostomatidae along with a wide posterior probability distribution of the root location may account for this lack of resolution (4).

Clearly, further morphological and molecular data studies are needed to rigorously test basal anabantoid relationships. Rüber et al. (4) also estimated divergence times among the anabantoid families based on two separate data sets, using either the four mitochondrial genes or a combined data set consisting of the mitochondrial genes plus the nuclear gene RAG1 and applying a Bayesian approach. The anabantoid fossil record that can be utilized for the calibration of the lineage divergence times is scarce. The only known articulated anabantoid fossil is from the genus Osphronemus (19, 20). It was found in the Sangkarewang Formation (Central Sumatra) dated with palynological data as late Eocene to early Oligocene (37–28.5 Ma) by Barber et al. (21). It is not possible to assign the fossil Osphronemus to any extant species in that genus with certainty, nor is it possible to assign it without doubt to Osphronemus. Therefore, two different age estimates based on different assignments of the fossil Osphronemus were conducted: (a) assignment of the fossil to the most recent common ancestor of Osphronemus and its closest relative Belontia (first calibration) and (b) assignment of the fossil to the most recent common ancestor of Osphronemus (second calibration). Calibration with fossils from early-diverging lineages, based on both, mitochondrial genes and the combined mitochondrial and nuclear gene data set, indicate a Middle Eocene (~40 Ma) origin of anabantoids as well as a Middle Eocene Helostomatidae–Anabantidae split (Table 1). Divergence times based on the combined mitochondrial plus nuclear data resulted in estimates that were on average a few million years younger than those derived from the four mitochondrial genes using the first calibration (Table 1). The second calibration, based on the four mitochondrial genes, on the other hand, resulted in a late Cretaceous age, ~90 Ma, for the origin of anabantoids and the Helostomatidae–Anabantidae split (Table 1).

346


Table 1. Divergence times (Ma) and confidence/credibility intervals (CI) among labyrinth fishes (Anabantoidei). Timetree Node

Time


CI

Ref. (4)(b)

Ref. (4)(c)

Time

CI

Time

CI

1

56.9

41.7

50–35

37.7

44–32

91.3

127–64

2

54.3

39.7

48–33

36.2

42–31

87.0

122–61

Note: Node times in the timetree represent the mean of time estimates from different columns. Times from the analysis of four mitochondrial genes assigning the fossil Osphronemus as a basal branch (plus Belontia) are shown in (a), from the analysis of four mitochondrial genes plus the nuclear gene RAG1 assigning the fossil Osphronemus to a basal branch (plus Belontia) are shown in (b), and from an analysis of four mitochondrial genes assigning the fossil Osphronemus to a living lineage of Osphronemus are shown in (c).

Anabantoids show a disjunct South and Southeast Asian–African distribution that might be indicative of a restricted Gondwana distribution. With both the RAG1 and a combined nuclear and mitochondrial DNA data set used by Rüber et al. (4), a basal split of labyrinth fishes into osphronemids vs. helostomids plus anabantids was favored. The first two families are exclusively found in Asia, whereas the third family is found on both continents, with Anabas (Anabantinae) from Southeast Asia as the closest relative of the remaining anabantids (Ctenopominae), which occur in Africa. Divergence time estimates of an Asian–African separation of these subfamilies ranged from (a) 35.1 (CI 43–28) based on mitochondrial DNA and 30.8 Ma (CI 37–25 Ma) based on a combined mitochondrial DNA plus nuclear RAG1 data set for the first calibration to (b) 77.0 Ma (CI 109–53 Ma) based on mitochondrial DNA for the second calibration. The African–Asian biogeographic split within labyrinth fishes indicates either a divergence at the Eocene–Oligocene (first calibration) or an Upper Cretaceous–Paleocene divergence (second calibration). Thus, the oldest divergence time estimate based on the second calibration is close to the suggested divergence of the Madagascar and Indian continent from Africa at 165–121 Ma. On the other hand, late Mesozoic dispersal from Africa to Asia, or vice versa, via land bridges between Africa, India, and Eurasia cannot be ruled out based on the first calibrations. It has been pointed out repeatedly that the late Mesozoic–early Cenozoic history of the Gondwana breakup is still poorly known and thus allows for wide speculations regarding African–Asian biotic exchanges via land bridges (22, 23). Given the poor fossil record of anabantoids, as well as the uncertainty in the phylogenetic placement of the fossil Osphronemus, it seems premature to draw any firm

conclusion regarding the relative role that drift vicariance and dispersal have played in shaping anabantoid African– Asian distribution (4).

Acknowledgments Support was provided by the Swiss National Science Foundation, the Janggen-Pöhn-Foundation, Switzerland and the Department of Zoology, the Natural History Museum (UK).

References 1. M. Kottelat, Ichthyol. Explor. Freshwat. 2, 273 (1991). 2. T. R. Roberts, Ichthyol. Explor. Freshwat. 2, 351 (1992). 3. L. Rüber, et al. Evolution 58, 799 (2004). 4. L. Rüber, R. Britz, R. Zardoya, Syst. Biol. 55, 374 (2006). 5. R. Britz, in Spezielle Zoologie. Teil 2: Wirbel–oder Schädeltiere, W. Westheide, R. Rieger, Eds. (Spektrum Akademischer Verlag, Heidelberg and Berlin, Germany, 2004). 6. G. Cuvier, A. Valenciennes, Histoire Naturelle des Poissons, Vol. 7 (F. G. Levrault, Paris, 1831). 7. G. V. Lauder, K. F. Liem, Env. Biol. Fish. 6, 257 (1981). 8. L. E. Gray, Zool. Miscell. 1831, 7 (1831). 9. P. Bleeker, Verh. Akad. Amst. 19, 1 (1879). 10. P. Bleeker, Natuurk. Tijdschr. Nederl. Ind. 20, 395 (1859). 11. L. S. Berg, System der Rezenten und Fossilen Fischartigen und Fische (VEB Verlag, Berlin, 1958). 12. K. F. Liem, Illinois Biol. Monogr. 30, 1 (1963). 13. Britz, R. 1995. Zur phylogenetischen Systematik der Anabantoidei (Teleostei, Percomorpha) unter Berücksichtigung der Stellung des Genus Luciocephalus. Morphologische und ethologische Untersuchungen. PhD Thesis, Universität Tübingen, Tübingen, Germany.

Eukaryota; Metazoa; Vertebrata; Actinopterygii; Anabantoidei

14. 15. 16. 17. 18.

R. Britz, Zool. J. Linn. Soc. 112, 491 (1994). G. V. Lauder, K. F. Liem, Bull. Mus. Comp. Zool. 150, 95 (1983). R. Britz, Ichthyol. Explor. Freshwat. 12, 305 (2001). R. Britz, M. Kokoscha, R. Riehl, Jap. J. Ichthyol. 42, 71 (1995). M. Kottelat, T. Whitten, Freshwater Biodiversity in Asia, With Special Reference to Fish, Technical Report 343 (World Bank Technical Paper, 1996). 19. C. Patterson, in The Fossil Record 2, M. J. Benton, Ed. (Chapman & Hall, London, 1993), pp. 621–656.

20.

347

M. Sanders, Verhandelingen van het Geologischmijnbouwkundig Genoostschap voor Nedeland en Koloniën, Geologische Serie 11, 1 (1934). 21. A. J. Barber, M. J. Crow, J. S. Milsom, Eds., Sumatra: Geology, Resources and Tectonics (Geological Society, London, 2005). 22. J. C. Briggs, J. Biogeo. 30, 381 (2003). 23. S. Chatterjee, C. R. Scotese, Proc. Indian Natl. Sci. Acad. 65A, 397 (1999).

Lungfishes (Dipnoi) Matthew P. Heinickea,*, Jennifer M. Sander b, and S. Blair Hedgesa

Lungfishes (Subclass Dipnoi) number only six species in three families, but are an important group of vertebrates because of their close relationship to tetrapods. Phylogenetic analyses of morphological and molecular data agree that African lungfishes (Protopteridae) and South American Lungfish (Lepidosirenidae) are closest relatives. Molecular clock analyses suggest that the divergence of these families from the Ceratodontidae (e.g., Australian Lungfish) occurred in the Permian 277 (321–234) million years ago (Ma). The divergence of South American and African lungfishes was in the early Cretaceous, 120 (165–94) Ma, and was probably related to the breakup of Gondwanaland.

content, not unlike many ray-finned fishes with the ability to breathe air, but does not aestivate. Extant lungfishes are intolerant of marine conditions, and are restricted to freshwater habitats, as were most Mesozoic lungfishes (2). Paleozoic lungfishes included numerous marine representatives, however, and the group may have originally been marine (3). Here, the relationships of the three living families of lungfishes are reviewed and the first estimates of divergence times are presented based on analyses of published sequence data. The fossil record of lungfishes is moderately complete. Tooth plates and scales are well represented, but skeletal material is relatively rare (2, 4). In addition to these remains, fossilized burrows are known (5), some harboring skeletal remains. The earliest fossils of sarcopterygians on the lungfish lineage (rather than tetrapod or coelacanth lineage) are from the Devonian (6). True members of the Subclass Dipnoi also appear in the Devonian, and the peaks of diversity of Dipnoi were in the Devonian and Triassic (251–200 Ma) (6). These early lungfishes represent extinct groups, and the living families appear later in the fossil record. The three extant families are all known from the Cretaceous (146–66 Ma) (as fossils

The six species of lungfish are the living representatives of the Subclass Dipnoi. These species are divided into two suborders (Lepidosirenoidei and Ceratodontoidei) and three families (Lepidosirenidae, Protopteridae, Ceratodontidae) (1). Several additional families are known from fossils extending back to the Devonian (416– 359 Ma). Lungfishes comprise one of three extant groups of Sarcopterygii, along with tetrapods and coelacanths (1). Living species are characterized by stocky, eellike bodies and fleshy fins without spines or rays. The paired pectoral and pelvic fins are paddle-like in the Australian Lungfish (Ceratodontidae) and whiplike in the African and South American lungfishes (Protopteridae and Lepidosirenidae). All extant species of lungfish are obligate air-breathers, and the African and South American lungfishes have the ability to aestivate during periods of drought (for months at a time in the case of African lungfishes) (1). The Australian Lungfish, Neoceratodus (Fig. 1), is a riverine species able to tolerate water with low oxygen

Fig. 1 An Australian Lungfish (Neoceratodus forsteri). Credit: J. Joss.

a

Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA; bDepartment of Evolution, Ecology, and Organismal Biology, Ohio State University, Columbus, OH 43210-1293, USA *To whom correspondence should be addressed ([email protected])

Abstract

M. P. Heinicke, J. M. Sander, and S. B. Hedges. Lungfishes (Dipnoi). Pp. 348–350 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Protopteridae

1

Ceratodontidae

P

Triassic

Jurassic

PZ

Cretaceous

MESOZOIC 250

200

150

Pg

Ng

349

Ceratodontoidei

Lepidosirenidae 2

Lepidosirenoidei

Eukaryota; Metazoa; Vertebrata; Sarcopterygii; Dipnoi

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of lungfishes. Divergence times are from Table 1. Abbreviations: Ng (Neogene), P (Permian), Pg (Paleogene), and PZ (Paleozoic).

of the three modern genera) (7–9). Ceratodontidae has been suggested to extend back to the Triassic (251–200 Ma), depending on how fossil taxa are allocated (6, 7). Species diversity declined in the Cenozoic, although several extinct species, including some in extinct genera, occur as late as the Pleistocene (1.81–0.01 Ma) (7). The phylogenetic relationships of the living lungfish families are not controversial. It is universally accepted that Dipnoi is monophyletic, and that within Dipnoi the African and South American lungfishes (Protopteridae and Lepidosirenidae) are closest relatives. They share numerous anatomical characters, including external larval gills, fin shape, and two-lobed lungs (1), and are in fact often grouped together in the Family Lepidosirenidae. These relationships are strongly supported by both morphological cladistic (10, 11) and molecular studies, including molecular studies employing nuclear or mitochondrial data (12–16). Molecular data also support the monophyly of Protopteridae, the only family that includes more than one living species (17). There have been no published molecular timing analyses among the three families of lungfish. Therefore, we conducted two molecular clock analyses using published sequence data and the Bayesian program Multidivtime (18). One analysis includes all families and uses published RAG1 and RAG2 nucleotide sequences from another study (12). The other analysis includes only Lepidosirenidae and Protopteridae, but uses amino acid data for six genes (RAG1, RAG2, TPI, GAG, ALDc, and GAD65) from three studies (12, 19, 20). Methodology is as described elsewhere (21). Several vertebrate outgroups are included in both analyses for calibration purposes (Mus, Oryctolagus, Homo, Gallus, Xenopus, Danio, and Carcharhinus/Triakis in the two-gene set; Mus, Homo, Gallus, Danio in the six-gene set), although these do not appear in the timetree.

Seven minimum and three maximum constraints were used in the two-gene data set, based on fossil data obtained from the literature (6, 7, 9, 22, 23). These include the divergence of Lepidosirenidae and Protopteridae (minimum, 92.7 Ma); the divergence of Lepidosirenoidei and Ceratodontoidei (minimum, 199 Ma); the divergence of primates and rodents (minimum, 62 Ma); the divergence of mammals and birds (minimum, 312 Ma, maximum, 370 Ma); the divergence of amniotes and amphibians (minimum, 330 Ma, maximum, 370 Ma); the divergence of tetrapods and lungfish (minimum, 404 Ma); and the divergence of ray-finned and lobe-finned fishes (minimum, 416 Ma, maximum, 495 Ma). For the six-gene data set, only the primate/rodent and mammal/ bird divergences were used. The following Multidivtime parameters were employed in both analyses: rttm (450), rttmsd (100), bigtime (3000). For the two-gene analysis, rtrate was set at 0.001, rtratesd at 0.0005, brownmean at 0.0025, and brownsd at 0.0025. For the six-gene analysis, these values were 0.04, 0.04, 0.001, and 0.001, respectively. The results of both analyses for the Protopteridae/ Lepidosirenidae divergence are similar, with confidence intervals that broadly overlap (Table 1). The timetree (Fig. 2) shows that the African and South American lungfishes diverged in the early Cretaceous, 120 (165–94) Ma. This date agrees well with the fossil evidence, as it is not substantially earlier than the earliest fossils of African lungfish that appear beginning in the Cenomanian stage of the Cretaceous (100–93 Ma) (9). The divergence between these two families and Ceratodontidae occurs much earlier, in the Permian, 277 (321–234) Ma. This date also agrees well with the fossil record, as putative ceratodontids are known from the Triassic, and several other Triassic genera of dipnoans are thought to be more closely related to lepidosirenoid lungfishes (2, 6, 7).

350


Table 1. Divergence times (Ma) among lungfishes and their confidence/credibility intervals (CI). Timetree Node

Estimates

Time

(a)

(b)

Time

CI

Time

CI

1

277.0

277

321–234

–

–

2

120.0

114

154–94

126

165–95

Note: Node times in the timetree are averages from two data sets: twogene analysis (a) and six-gene analysis (b).

The divergence of protopterid and lepidosirenid lungfishes has long been suggested to be related to Gondwanan breakup, because these families are restricted to freshwater and have fossil records extending back to the Cretaceous, but restricted to Africa and South America, respectively (2, 24, 25). The timetree supports this hypothesis, as the South Atlantic Ocean opened largely during the Aptian and Albian stages of the Cretaceous, 125–100 Ma (26), the time period during which these families diverged according to the molecular time estimate. The ceratodontids are a much older, and formerly more widespread, group. The divergence of ceratodontid and lepidosirenoid lungfishes was too early (277 Ma) to be explained by continental vicariance, as the continents were joined into the supercontinent Pangaea at this time (2). Further, although now restricted to Australia, fossils referable to ceratodontids have been described from Mesozoic deposits in Africa and South America, indicating a much wider distribution (27, 28). In the future, it is likely that additional fossils, rather than molecular data, will contribute more to elucidating the biogeographic history of the lungfishes, a relict group.

Acknowledgment Support was provided by U.S. National Science Foundation and National Aeronautics and Space Administration (NASA Astrobiology Institute) to S.B.H.

References 1. J. S. Nelson, Fishes of the World, 4th ed. (John Wiley, San Francisco, 2006). 2. L. Cavin, V. Suteethorn, E. Buffetaut, H. Tong, Zool. J. Linn. Soc. 149, 141 (2007).

3. K. S. W. Campbell, R. E. Barwick, J. Morph. Supp. 1, 99 (1986). 4. C. R. Marshall, J. Morph. Supp. 1, 151 (1986). 5. D. S. Berman, J. Paleont. 50, 1034 (1976). 6. R. Cloutier, P. E. Ahlberg in Interrelationships of Fishes, M. L. J. Stiassny, L. R. Parenti, G. D. Johnson, Eds. (Academic Press, San Diego, 1996), pp. 445–479. 7. A. Kemp, J. Paleont. 71, 713 (1997). 8. H. P. Schultze, in Fosiles y Facies de Bolivia. Volumen I—Vertebrados, R. Suarez-Soruco, Ed. (Revista Tecnica de Yacimientos Petroliferous Fiscales Bolivianos 12, Santa Cruz, 1991), pp. 441–448. 9. A. M. Murray, Fish Fisheries 1, 111 (2000). 10. H. P. Schultze, C. R. Marshall, in Palaeontological Studies in Honour of Ken Campbell, P. A. Jell, Ed. (Association of Australasian Palaeontologists, Brisbane, 1993), pp. 211–224. 11. H. P. Schultze, in Mesozoic Fishes 3—Systematics, Paleoenvironments and Biodiversity, G. Arriata, A. Tintori, Eds. (Verlag Dr. Friedrich Pfeil, Munchen, 2004), pp. 463–492. 12. H. Brinkmann, B. Venkatesh, S. Brenner, A. Meyer, Proc. Natl. Acad. Sci. U.S.A. 101, 4900 (2004). 13. H. Brinkmann, A. Denk, J. Zitzler, J. J. Joss, A. Meyer, J. Mol. Evol. 59, 834 (2004). 14. S. B. Hedges, C. A. Hass, L. R. Maxson, Nature 363, 502 (1993). 15. A. Meyer, S. I. Dolven, J. Mol. Evol. 35, 102 (1992). 16. R. Zardoya, A. Meyer, Proc. Natl. Acad. Sci. U.S.A. 93, 5449 (1996). 17. M. Tokita, T. Okamoto, T. Hikida, Mol. Phylogenet. Evol. 35, 281 (2005). 18. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 19. K. Kikugawa et al., BMC Biol. 2, 3 (2004). 20. K. Lariviere et al., Mol. Biol. Evol. 19, 2325 (2002). 21. M. P. Heinicke, W. E. Duellman, S. B. Hedges, Proc. Natl. Acad. Sci. U.S.A. 104, 10092 (2007). 22. M. J. Benton, P. C. J. Donoghue, R. J. Asher, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2008), pp. 35–86. 23. J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 2275 (2005). 24. J. G. Lundberg, in Biological Relationships between Africa and South America, P. Goldblatt, Ed. (Yale University Press, New Haven, 1993), pp. 156–199. 25. M. J. Novacek, L. G. Marshall, Copeia 1976, 1 (1976). 26. A. G. Smith, D. G. Smith, B. M. Funnell, Atlas of Mesozoic and Cenozoic Coastlines (Cambridge University Press, Cambridge, 1994). 27. M. Martin, Neus Jahrb. Geol. Palaontol., Abhandlung 169, 225 (1984). 28. M. Martin, C. R. Acad. Sci. Paris, Sci. Ter. Planets 325, 635 (1997).

AMPHIBIANS


Amphibians (Lissamphibia) David C. Cannatellaa,*, David R. Vieitesb, Peng Zhangb, and Marvalee H. Wakeb, and David B. Wakeb a

Section of Integrative Biology and Texas Memorial Museum, 1 University Station C0930, University of Texas, Austin, TX 78712, USA; b Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Sciences Building, University of California, Berkeley, CA 94720, USA *To whom correspondence should be addressed (catfish@mail. utexas.edu)

Abstract Living amphibians (6449 species) include three distinctive orders: salamanders (Caudata), caecilians (Gymnophiona), and frogs (Anura). Each is supported as monophyletic in molecular phylogenetic analyses, with frogs + salamanders forming the clade Batrachia. Molecular time estimates of the origin of Lissamphibia vary greatly (367 -282 million years ago; latest Devonian to Early Permian), although recent analyses favor the youngest ages. Divergences among the three orders likely occurred during the Permian, 300 -251 million years ago. Debates about the origin, relationships to extinct taxa, and monophyly of living amphibians are ongoing.

Lissamphibia, a subclass of Amphibia, includes all living representatives, which form three clades, frogs (Salientia), salamanders (Caudata), and caecilians (Gymnophiona), each readily recognizable based on their highly distinctive body plans (Fig. 1). Frogs generally have large mouths and bulging eyes, but short vertebral columns and no tail. These squat creatures have powerful hind limbs for jumping. They are the most speciose clade with about 5700 species. Most of the 576 species of living salamanders are more typical-looking tetrapods, with a tail and four legs. Some aquatic or fossorial species have reduced limbs and girdles and elongated trunks. Living caecilians are elongate, limbless, tail-less or nearly so, and have grooved rings encircling the body. A distinctive tentacle anterior to and below the typically inconspicuous eye is used for chemoreception. Although the majority of the 176 species are fossorial, one lineage has invaded aquatic habitats. The phylogenetic relationships among the three lissamphibian orders have been controversial for decades.

A close relationship between caecilians and salamanders (Procera hypothesis) was supported by earlier analyses of mitochondrial and nuclear ribosomal DNA sequences (1–3) and mitochondrial genomes (4). The Procera hypothesis has advantages for interpreting distribution patterns and fossil records of the three orders: frogs are distributed worldwide but salamanders and caecilians have strong Laurasian and Gondwanan distribution patterns, respectively; frog-like fossils can be traced back to the Triassic (~250 Ma) but no salamander or caecilian fossils have been found before the Jurassic (~190 Ma). However, most recent analyses, using larger databases of either nuclear genes, mitochondrial genes (including mitochondrial genomes), or a combination of both, have found frogs and salamanders to be closest relatives, a group called Batrachia; the earlier molecular analyses were misled by poor performance of data or insufficient taxon sampling (5–8). Furthermore, analyses of morphological data, including fossil taxa (9–14), have also found the Batrachia hypothesis to be more strongly supported than the Procera hypothesis, and we follow that conclusion here.

Fig. 1 Representative lissamphibians. Pseudotriton ruber (Plethodontidae), a salamander (upper); Hypsiboas helprini (Hylidae), an anuran (lower left); and Schistometopum thomense (Caeciliidae), a caecilian (lower right). Credits: S. B. Hedges.

D. C. Cannatella, D. R. Vieites, P. Zhang, M. H. Wake, and D. B. Wake. Amphibians (Lissamphibia). Pp. 353–356 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).


Anura 2

Caudata

1

Batrachia

354

Gymnophiona

Permian

Triassic

PZ 300

Jurassic

Cretaceous

200

150

Ng

CENOZOIC

MESOZOIC 250

Pg

100

50

0 Million years ago

Fig. 2 A timetree of amphibians (Lissamphibia). Divergence times are shown in Table 1. Abbreviations: Ng (Neogene), Pg (Paleogene), and PZ (Paleozoic).

The identity of the Paleozoic relatives of modern amphibians is controversial (15). Three hypotheses have been proposed. The Temnospondyl hypothesis suggests that frogs, salamanders, and caecilians form a clade that is nested within dissorophoid temnospondyls. Dissorophoidea (Moscovian of the late Carboniferous to upper portion of the Early Permian) is a large superfamily of temnospondyl amphibians, including some small, paedomorphic forms, such as Doleserpeton, amphibamids, and branchiosaurids, which share many derived features with modern amphibians (9–12, 16). The most recent study found the dissorophoid amphibians Amphibamus and Doleserpeton to be most closely related to modern amphibians (13). In diametric opposition to the earlier is the Lepospondyl hypothesis. Modern amphibians are nested within the lepospondyls; their closest relatives are Lysorophia (Brachydectes), known from the Bashkirian of the late Carboniferous to the upper part of the Early Permian, with the next successive relatives being microsaurs (17, 18). According to this hypothesis, modern amphibians are more closely related to amniotes than they are to temnospondyls. A third view, the Polyphyletic hypothesis, incorporates elements of both of the earlier hypotheses. It argues for polyphyly of modern amphibians with respect to the major Palaeozoic lineages (14). As in the Lepospondyl hypothesis, living caecilians are seen as most closely related to lepospondyls, which are in turn are closer to amniotes than to other modern amphibians. Furthermore, as in the Temnospondyl hypothesis, salamanders and frogs are derived from temnospondyls, specifically the branchiosaurs. This hypothesis was strengthened by the discovery of the Lower Jurassic taxon Eocaecilia, a putative limbed caecilian that may be nested between goniorhynchid microsaurs, such as Rhynchonkos (Lower Permian), and the extant caecilians (19). However, the Eocaecilia–microsaur relationship has

recently been rejected (20). A monophyletic origin of modern amphibians with respect to living amniotes is strongly supported in all molecular phylogenetic studies to date (1–8, 21). Thus, if the lepospondyls are a clade most closely related to amniotes, the Polyphyletic hypothesis of modern amphibians can be indirectly rejected because it requires a closer relationship of caecilians to amniotes than to the Batrachia. Using numbers of nuclear genes and a molecularclock-based method, Kumar and Hedges (22) provided the first time estimate for the origin of lissamphibians at about 360 Ma, using average evolutionary rates of rateconstant genes to apply a clock. More recent analyses use relaxed-clock methods, which allow evolutionary rates to vary among genes and lineages. Zhang et al. (8) analyzed 14 complete lissamphibian mitochondrial genomes and obtained an estimate of the age of the lissamphibian ancestor (337 Ma, Table 1). By comparing the confidence intervals to the temporal distribution of related fossil taxa, the Temnospondyl hypothesis was found to be most compatible with these results. This finding was challenged because of imprecise characterization of the stratigraphic ranges of the groups of Palaeozoic tetrapods, and these estimates would be more compatible with a polyphyletic origin of lissamphibians (23), which is questioned (24). There are few early lissamphibian fossils, the oldest being the proto-frogs Triadobatrachus and Czatkobatrachus, from the Early Triassic (~250 Ma; 25, 26). Dates estimated from molecular data typically are much older than the fossil record suggests. The summary of published dates for the nodes Lissamphibia and Batrachia (Table 1; Fig. 2) indicates substantial lack of consensus. For example, three published dates based on nuclear sequences place the split of Lissamphibia (into Gymnophiona and Batrachia) at 367 Ma (27), 369 Ma (7), and 360 Ma (22), approximately at the

Eukaryota; Metazoa; Vertebrata; Lissamphibia

355

Table 1. Divergence times and credibility/confidence intervals (CI) among amphibians (Lissamphibia). Timetree Node

Time


Ref. (6)(b)

Time

Time

Time

Ref. (7)(a) CI

Time

Ref. (7)(b) CI

Time

Ref. (8) CI

1

294

322

292

368.8

396–344

351.6

370–304

337

353–321

2

264

267–266

267–266

357.8

385–333

332.9

353–289

308

328–289

Timetree Node

Time

Estimates (Continued) Ref. (22) Time

Ref. (24)

Ref. (27)

Time

CI

Time

Ref. (29) CI

Time

CI

1

294

360

282

356–250

367

417–328

294

319–271

2

264

–

254

257–246

357

405–317

264

276–255

Note: Node times in the timetree are from ref. (29). Estimates from ref. (6) are from r8s penalized likelihood analyses of (a) 2613 basepairs and (b) 871 amino acids of the RAG1 gene. Estimates from ref. (7) are from (a) Multidivtime and (b) r8s penalized likelihood analysis of 3747 base pairs from 16S, CXCR4, NCX1, RAG1, and SLC8A3 genes. Estimates from ref. (8) are from Multidivtime analysis of mitochondrial genomes. Estimates from ref. (22) are sequence divergence analysis of multiple nuclear genes. Estimates from ref. (24) are from r8s penalized likelihood analysis of the dataset from ref. (27). Estimates from ref. (27) are from Multidivtime analysis of the 1368 basepairs of RAG1. Estimates from ref. (29) are from BEAST analysis of mitochondrial genomes.

Devonian–Carboniferous boundary, the time of the occurrence of the first tetrapods (Table 1). At the other extreme, a reanalysis (24) of the mitogenome data set (8) reported several dates for Lissamphibia and Batrachia, based on various combinations of topology, analytical parameters in r8s, and use of external (outside of amphibians) or internal (within) calibrations. The youngest of the estimates for lissamphibians (267 Ma), based only on internal calibration points, is just slightly older than the oldest lissamphibian fossils, from the Triassic. Recent molecular studies are reducing the discrepancies between molecular dating and the fossil record. An analysis of RAG1 from all tetrapods (6) estimated ages for Lissamphibia and Batrachia of 322 and 274 Ma, respectively (no internal amphibian calibration dates were used). Another analysis of complete mitochondrial genomes using several internal amphibian calibration points and a “soft” bound calibration strategy (28) produced dates for Lissamphibia and Batrachia of 294 and 264 Ma, respectively (Zhang and Wake, unpublished results). Improved analytical methods (30–32) and larger databases are producing dates more in line with what might be expected from extrapolating from fossils. Thus, the 264 Ma date estimated from molecular data for Batrachia is close to the age (250 Ma) and phylogenetic position of the Lower Triassic fossils Triadobatrachus and Czatkobatrachus. In general, available time tree estimates for the origin of Lissamphibia and Batrachia are controversial, with about

87 and 103 million years difference between the youngest and oldest estimates, respectively (Table 1). Robust time estimation using different methods with multiple markers should clarify how the group originated and evolved. For example, if the split of the caecilians from Batrachia happened after the time the Permian microsaur Rhynchonkos lived, the Polyphyletic hypothesis will be rejected. The most recent analyses with the largest data sets are tending to favor younger times of divergence.

Acknowledgment Support was provided by U.S. National Science Foundation grants to D.C.C., D.B.W., and M.H.W.

References 1. S. B. Hedges, K. D. Moberg, L. R. Maxson, Mol. Biol. Evol. 7, 607 (1990). 2. S. B. Hedges, L. R. Maxson, Herpetol. Monogr. 7, 27 (1993). 3. A. E. Feller, S. B. Hedges, Mol. Phylogenet. Evol. 9, 509 (1998). 4. P. Zhang et al., Mol. Phylogenet. Evol. 28, 620 (2003). 5. R. Zardoya, A. Meyer, Proc. Natl. Acad. Sci. U.S.A. 98, 7380 (2001). 6. A. F. Hugall, R. Foster, M. S. Lee, Syst. Biol. 56, 543 (2007). 7. K. Roelants et al., Proc. Natl. Acad. Sci. U.S.A. 104, 887 (2007).

356


8. P. Zhang et al., Syst. Biol. 54, 391 (2005). 9. A. R. Milner, in The Phylogeny and Classification of the Tetrapods. 1. Amphibians, Reptiles, Birds, M. J. Benton, Ed. (Oxford University Press, Oxford, 1988), pp. 59–102. 10. A. R. Milner, Herpetol. Monogr. 7, 8 (1993). 11. L. Trueb, R. Cloutier, in Origins of the Higher Groups of Tetrapods: Controversy and Consensus, H.-P. Schultze, L. Trueb, Eds. (Cornell University Press, Ithaca, 1991), pp. 233–313. 12. M. Ruta, M. I. Coates, D. L. J. Quicke, Biol. Rev. Cambridge Phil. Soc. 78, 251 (2003). 13. M. Ruta, M. I. Coates, J. Syst. Palaeont. 5, 69 (2007). 14. R. L. Carroll, Zool. J. Linn. Soc. 150 (Suppl. 1), 1 (2007). 15. R. L. Carroll, in Amphibian Biology. Volume 4, H. Heatwole, R. L. Carroll (Surrey Beatty & Sons, Chipping Norton, 2000), pp. 1270–1273. 16. J. R. Bolt, J. Paleontol. 51, 235 (1977). 17. M. Laurin, R. R. Reisz, in Amniote Origins. Completing the Transition to Land, S. S. Sumida, K. L. M. Martin, Eds. (Academic Press, San Diego, 1997), pp. 9–59.

18. G. Vallin, M. Laurin, J. Vert. Paleont. 24, 56 (2004). 19. R. L. Carroll, in Amphibian biology, Vol. 4, H. Heatwole, R. L. Carroll (Surrey Beatty & Sons, Chipping Norton, 2000), pp. 1402–1411. 20. F. A. Jenkins, Jr., D. M. Walsh, R. L. Carroll, Bull. Mus. Comp. Zool. 158, 285 (2007). 21. J. M. Hay et al., Mol. Biol. Evol. 12, 928 (1995). 22. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 23. M. S. Lee, J. S. Anderson, Mol. Phylogenet. Evol. 40, 635 (2006). 24. D. Marjanovic, M. Laurin, Syst. Biol. 56, 369 (2007). 25. J. C. Rage, Z. Rocek, Palaeontogr. Abt. 206, 1 (1989). 26. S. E. Evans, M. Borsuk-Bialynicka, Acta Palaeontol. Polon. 43, 573 (1998). 27. D. San Mauro et al., Am. Nat. 165, 590 (2005). 28. Z. Yang, B. Rannala, Mol. Biol. Evol. 23, 212 (2006). 29. P. Zhang, D. B. Wake, Submitted ms. (2008). 30. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 31. M. J. Sanderson, Bioinformatics 19, 301 (2003). 32. A. J. Drummond, A. Rambaut, BMC Evol. Biol. 7, 214 (2007).

Frogs and toads (Anura) Franky Bossuyt and Kim Roelants Biology Department, Unit of Ecology & Systematics, Amphibian Evolution Lab, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium *To whom correspondence should be addressed (fbossuyt@vub. ac.be)

Abstract Anura (frogs and toads) constitute over 90% of living amphibian diversity. Recent timetree constructions have shown that their diversification was a highly episodic process, with establishment of the major clades in three periods: Triassic (251–200 million years ago, Ma), end of Jurassic to early Cretaceous (~150–100 Ma), and end of Cretaceous to early Paleogene (~70–50 Ma). The early diversification of anurans predated the initial north–south breakup of Pangaea, and resulted in distinct assemblages in both hemispheres. The subsequent radiation of neobatrachian frogs has been largely determined by Gondwanan fragmentation and resulted in recurrent patterns of continent-scale endemism.

Anura (Fig. 1) (“tail-less amphibians”) represent the largest living order of amphibians, and currently include ~5400 described species (1). Most of them undergo the typical amphibious life history and are dependent on the presence of water for their reproduction and development. Multiple lineages, however, show an evolutionary trend toward increased terrestriality in larval or adult frog stages. Despite an evolutionarily conserved body plan, anurans have diversified into a myriad of ecomorphs and have adapted to life in habitats as distinct as rainforest canopies, mangroves, and sand dune burrows. In addition, anurans have attained a subcosmopolitan distribution and are currently only absent in extreme northern latitudes, Antarctica, and most oceanic (noncontinental) islands (2). The independent occupation of similar ecological niches by frog taxa in different geographic regions has resulted in extraordinary cases of evolutionary convergence. The consequent high levels of morphological homoplasy have complicated anuran systematics for decades. However, a major ongoing upsurge

of molecular phylogenetic studies is now leading to an increasingly resolved consensus for the anuran tree. In this chapter, we review the relationships and divergence times among 59 anuran families and argue that their evolutionary history is largely congruent with major geological and environmental changes in Earth’s history. We mostly implement clade and family names derived from the taxonomy recently proposed by Frost et al. (3). However, we believe that evolutionary time is an important parameter in conveying useful comparative information in biological classification (4). We therefore treat Ascaphidae, Discoglossidae, Nasikabatrachidae, some subfamilies in Nobleobatrachia, and all subfamilies of Microhylidae sensu lato as distinct families. The sequence of early divergences in Anura has been the subject of major controversy. Most of the debate focused on the phylogenetic position of archaeobatrachian families (taxa with primitive or transitional characters, covering ~4% of all extant species) with respect to neobatrachian families (“advanced” taxa, covering

Fig. 1 A tree frog (Rhacophorus lateralis) from India. Credit: F. Bossuyt.

F. Bossuyt and K. Roelants. Frogs and toads (Anura). Pp. 357–364 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

358

THE TIMETREE OF LIFE Microhylidae 43

Dyscophidae

37

Kalophrynidae 46

Cophylidae

44

32

Melanobatrachidae Gastrophrynidae

31

Microhyloidea

Asterophryidae 35

Scaphiophrynidae

28

Hoplophrynidae

25

Phrynomeridae

Astylosternidae

26

Hyperoliidae

21

Brevicipitidae 38

Hemisotidae Mantellidae

41

Rhacophoridae

36 16

Dicroglossidae

39

33

Neobatrachia

42

Afrobatrachia

Arthroleptidae

17

Ranidae

Micrixalidae 40

24

11

Ranixalidae

34

Ceratobatrachidae

Natatanura

Nyctibatrachidae 29

Petropedetidae 30

23

Pyxicephalidae

27

Ptychadenidae Sooglossidae 22

Triassic

Jurassic

Cretaceous

MESOZOIC 250

200

Nasikabatrachidae

150

Pg

Ng

Sooglossoidea

Phrynobatrachidae

CENOZOIC 100

50

0 Million years ago

Fig. 2 Continues

the remaining 96% of extant species). Morphological studies have supported diverse paraphyletic arrangements of archaeobatrachian families (5–10). Although Neobatrachia have traditionally been considered to constitute a single-nested clade, analyses of combined larval and adult characters have recently questioned their monophyly (9, 10). Early analyses of ribosomal DNA sequences clustered the archaeobatrachian families in a single clade

(Archaeobatrachia) as the closest relatives of Neobatrachia (11–13). Recent phylogenetic studies, implementing expanded taxon sampling, nuclear and mitochondrial protein-coding DNA sequences, and model-based reconstruction methods, have provided robust support for a paraphyletic arrangement of four major archaeobatrachian lineages: (i) Amphicoela: Ascaphidae + Leiopelmatidae, (ii) Costata (previously known as Discoglossoidea), (iii) Xenoanura (Pipoidea), and (iv)

Eukaryota; Metazoa; Vertebrata; Lissamphibia; Anura

359


Pelodryadidae 56

Phyllomedusidae

52

Brachycephalidae Craugastoridae

55

Eleutherodactylidae

51

Strabomantidae 49

Hylidae 54

Ceratophryidae Telmatobiidae

47

Nobleobatrachia

10

Bufonidae 53

Leptodactylidae

50

Rhinodermatidae

13

Limnodynastidae 20

Rheobatrachidae

19

4

Myobatrachidae

15

Calyptocephalellidae

Myobatrachoidea

Centrolenidae

Neobatrachia

Dendrobatidae

45 48

9

Pelobatidae

3

18

Megophryidae

14

Pelodytidae

8

Scaphiopodidae 2

Pipidae 7

Rhinophrynidae Alytidae 12

1

Discoglossidae

6

Bombinatoridae Ascaphidae 5

Leiopelmatidae

Triassic

Jurassic

Cretaceous

MESOZOIC 250

200

150

Pg

Amphicoela Costata Xenoanura Anomocoela

Heleophrynidae

Ng

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of frogs and toads (Anura). Divergence times are shown in Table 1 Ng (Neogene), Pg (Paleogene), and Tr (Triassic)..

Anomocoela (Pelobatoidea) (3, 14–18). Consistent with most morphological evidence, they supported the basal divergence of Amphicoela, and identified Anomocoela as the closest relative of Neobatrachia. A remaining point of ambiguity is the phylogenetic position of Xenoanura. Molecular studies have variously resolved Xenoanura as the closest relative of Costata (17, 19), of Neobatrachia

(14), of an Anomocoela + Neobatrachia clade (16–18), and of a Costata + Anomocoela + Neobatrachia clade (3, 15). The monophyly of Neobatrachia has always received strong support from molecular data (3, 12–22). Various arrangements of the following four well-supported lineages have been published: (i) Heleophrynidae,

360


Table 1. Divergence times (Ma) and their 95% confidence/credibility intervals (CI) among anurans. Timetree

Estimates Ref. (15)

Ref. (18)

Ref. (26)

Ref. (30)

Ref. (37)

Node

Time

(Fig. 2)

(Ma)

Time

CI

Time

CI

Time

Time

CI

Time

1

243

262

304–223

243

264–217

–

–

–

–

2

234

245

288–204

234

258–210

183

–

–

–

3

229

–

–

229

253–204

–

–

–

–

4

222

216

260–176

222

247–197

–

–

–

–

5

203

237

281–193

203

227–177

172

–

–

–

6

203

199

244–155

203

228–180

132

–

–

–

7

193

–

–

193

218–170

–

–

–

–

8

167

164

208–121

167

191–146

127

–

–

–

9

167

162

199–128

167

186–149

125

–

–

–

10

161

150

186–117

161

180–143

–

–

–

–

11

159

131

167–99

159

178–141

110

–

–

–

12

152

152

199–105

152

179–129

–

–

–

–

13

146

138

172–108

146

164–129

106

–

–

–

14

143

142

187–101

143

166–121

112

–

–

–

15

129

120

154–91

129

147–113

94.4

–

–

–

16

119

99

132–70

119

131–106

–

133

154–115

–

17

117

–

–

117

129–104

–

127

148–110

–

18

115

118

162–77

115

136–93

93.3

–

–

–

19

107

–

–

107

123–89

–

–

–

–

20

102

–

–

102

118–85

–

–

–

–

21

102

–

–

102

115–90

–

107

126–89

–

22

101

–

–

101

121–83

–

–

–

–

23

94.4

–

–

82.5

95–71

–

94.4

113–78

–

24

89.7

–

–

–

–

–

89.7

108–74

–

25

88

–

–

75.6

83–70

–

88

102–77

–

26

87.7

–

–

87.7

100–75

–

96

117–78

–

27

86.3

–

–

–

–

–

86.3

104–71

–

28

83.9

–

–

–

–

–

83.9

97–74

–

29

82.9

–

–

–

–

–

82.9

100–68

–

30

81.1

–

–

72.6

85–62

–

81.1

99–66

–

31

80.5

–

–

–

–

–

80.5

93–72

–

32

79

–

–

–

–

–

79

91–70

–

33

78.1

–

–

–

–

–

78.1

95–64

–

34

77.6

–

–

–

–

–

77.6

95–63

–

35

76.8

–

–

73.0

79–68

–

76.8

89–69

–

36

74.4

–

–

70.3

81–60

–

74.4

91–61

–

37

73.6

–

–

70.2

76–66

–

73.6

85–67

–

38

72.1

–

–

72.1

85–59

–

74.6

94–58

–

39

71.7

–

–

–

–

–

71.7

88–58

–

40

71.2

–

–

–

–

–

71.2

89–56

–


361

Table 1. Continued Timetree

Estimates Ref. (15)

Ref. (18)

Ref. (26)

Ref. (30)

Ref. (37)

Node

Time

(Fig. 2)

(Ma)

Time

CI

Time

CI

Time

Time

CI

41

69.7

–

–

62.8

74–53

–

69.7

86–56

–

42

69.1

–

–

69.1

81–57

–

–

–

–

43

68.7

–

–

67.1

71–66

–

68.7

78–65

–

44

66.8

–

–

–

–

–

66.8

80–56

–

Time

45

63

–

–

63

77–53

–

–

–

–

46

61.8

–

–

–

–

–

61.8

76–49

–

47

61.1

–

–

61.1

76–51

–

–

–

–

48

59.7

–

–

59.7

74–50

–

–

–

–

49

58.4

–

–

58.4

72–49

–

–

–

–

50

57.9

–

–

57.9

72–49

–

–

–

–

51

56.7

–

–

56.7

70–47

–

–

–

–

52

55

–

–

55

68–46

–

–

–

–

53

54.6

–

–

54.6

68–45

–

–

–

–

54

53.2

–

–

53.2

66–44

–

–

–

–

55

46.5

–

–

–

–

–

–

–

46.5

56

43.2

–

–

43.2

55–36

–

–

–

–

Note: Node times in the timetree are based on refs. (18) and (30) for Natatanura and Microhyloidea, because the use of time estimates averaged across all studies would be incompatible with the depicted topology.

(ii) Sooglossoidea: Nasikabatrachidae + Sooglossidae, (iii) Nobleobatrachia (the “Hyloidea” of 14, 15, 21) + Myobatrachoidea + Calyptocephalellidae, and (iv) Ranoides: Afrobatrachia + Microhyloidea + Natatanura. Most studies of the past few years have converged on a basal split between the South African endemic Heleophrynidae and the remaining neobatrachians (3, 14–18). An important addition to the amphibian tree resulted from the discovery of a new frog lineage (Nasikabatrachidae) in the Western Ghats of India (21). All molecular studies have found this lineage to be the closest relative of Sooglossidae, a small family endemic to the Seychelles (3, 18, 21, 22). Despite the absence of derived morphological characters supporting Nobleobatrachia, its monophyly is strongly supported by DNA sequence evidence of different loci, and includes a unique codon insertion in the RAG-1 gene. Within this clade, Leptodactylidae and Hylidae as traditionally defined (1, 2) are now known to be polyphyletic (3, 18, 23, 24) and several of their lineages have been assigned to separate families (3). Most of these studies also identified the ex-leptodactylid genera

Calyptocephalella (= Caudiverbera) and Telmatobufo (both now Calyptocephalellidae) of Chile as the closest relatives of the Australo-Papuan Myobatrachoidea (3, 15, 18, 24). The same analyses recovered this previously unrecognized clade (Australobatrachia) as the closest relatives of Nobleobatrachia. Several controversies however remain in nobleobatrachian phylogeny: first, Leptodactylidae as defined here was not found monophyletic in a recent molecular study (25), although this clade is supported by a unique insertion of two codons in the Ncx1 gene (18). Second, Hemiphractidae were variously found to be monophyletic (26) or polyphyletic (3). Most important, the sequence of rapid diversification in the radiation of nobleobatrachian families can be considered largely unresolved. Ranoides are composed of three highly supported family assemblages: Afrobatrachia, Microhyloidea, and Natatanura. Afrobatrachia represents a well-resolved African endemic clade and has recently been shown to include the Brevicipitidae (27), which were long considered part of the microhyloid clade. Studies of natatanuran phylogeny incorporating both mitochondrial and nuclear genes agreed on the

362


basal divergence of several African lineages and corroborated morphological evidence for a close relationship between Mantellidae and Rhacophoridae (3, 28–30). Recent analyses of Microhyloidea using similar multigene data sets have demonstrated the non-monophyly of at least five out of nine traditionally recognized subfamilies (2). A consensus for early microhyloid relationships is yet unavailable, but at least three studies have provided evidence for a close relationship of Asian Microhylidae and Madagascan Dyscophidae (18, 30, 31). During the past few years, molecular divergence time analyses have resulted in increasingly comprehensive timetrees for Anura (Table 1, Fig. 2). The earliest studies incorporated Thorne et al.’s (32) relaxed molecular clock model (divtime) to date primary divergences in Natatanura, suggesting mid- to late Cretaceous diversification of this clade (33). Subsequent studies implementing different relaxed-clock models, calibration points, and sampling strategies have corroborated the late Cretaceous radiation of both Natatanura and Microhyloidea (29–31, 34). Similar application of Thorne and Kishino’s (35) upgraded model, adapted to accommodate rate variation across multiple loci (Multidivtime), resulted in late Jurassic–early Cretaceous time estimates for the basal divergences of Neobatrachia (based on five genes and five calibration points) (21) and Triassic estimates for basal anuran divergences (based on five genes and seven calibration points) (16). San Mauro et al. (15) constructed the first family-level timetree for amphibians using Multidivtime analyses of RAG1 sequences and nine calibration points. Their analyses were based on broad phylogenetic sampling of frogs and provided confidence intervals for the age of the major anuran clades. An expanded study using penalized likelihood analyses of 84 anuran RAG1 sequences and 11 calibration points (26) produced overall younger time estimates (Table 1) when the Caudata–Anura split was arbitrarily fi xed at 300 Ma. A parallel study, applying both Multidivtime and penalized likelihood analyses using 24 calibration points on a five-gene data set including 120 anuran taxa (18), resulted in divergence time estimates that were very similar to those of San Mauro et al. (15), with strong overlap of 95% credibility intervals. These studies indicate that the evolutionary rise of anuran diversity was a highly episodic process, with the establishment of archaeobatrachian clades in the Triassic–early Jurassic (251–199.6 Ma), of the primary neobatrachian lineages in the late Jurassic–early Cretaceous, of the natatanuran and microhyloid radiations in the late Cretaceous, and of Nobleobatrachia

around the Cretaceous–Paleogene (K-P) boundary (65.5 Ma). Zhang et al. (19), based on multidivtime analyses of mitogenomic data in combination with a single external calibration point, recovered noticeably older age estimates for several nodes, including a Carboniferous– Permian (299 Ma) origin of living anurans and a midCretaceous (99.6 Ma) age for the nobleobatrachian clade. However, because mitochondrial genes evolve much faster than most nuclear genes used in other studies, it is likely that they pose increased risks of mutational saturation and biases in branch length estimation. The rise of living anurans shows strong overlap with major shifts in vertebrate faunal compositions in the late Permian and Triassic. Both the end Permian mass extinction and the Triassic extinction episodes represented severe losses of amphibian diversity (36) and in parallel to amniote groups, anurans may have taken opportunistic advantage of ecological niche vacancy in the redeveloping and increasingly complex vertebrate ecosystems. Molecular divergence time estimates of all studies also imply that the major archaeobatrachian lineages were present on Pangaea before its Jurassic north– south breakup into Laurasia and Gondwana (15, 16, 18). The subsequent formation of distinct anuran faunas in both landmasses is illustrated by three independent divergences between Laurasian and Gondwanan taxon pairs: (i) Ascaphidae of North America vs. Leiopelmatidae of New Zealand, (ii) Rhinophrynidae of North– Central America vs. Pipidae of South America–Africa, and (iii) Anomocoela of North America–Eurasia vs. Neobatrachia, originally a Gondwanan group. These results reinforce the predicted importance of Pangaea breakup in shaping distinct amphibian faunas in both hemispheres (13). The late Jurassic or early Cretaceous divergences of the four major neobatrachian lineages (Fig. 2) constitute a second distinct wave of anuran radiation. Two of the four major lineages are now only represented by few species endemic to restricted geographic regions on different ex-Gondwanan landmasses. Heleophrynidae (167 Ma) consist of six extant species that occur in the mountain ranges of the Cape and Transvaal regions of Repulic of South Africa; Sooglossoidea (159 Ma) consists of only five described species, one of which (Nasikabatrachidae) is endemic to the Indian Western Ghats, and four of which (Sooglossidae) occur only on the Seychelles. The deep split between both families (101 Ma) identifies each of them as relict lineages that testify for a midCretaceous biogeographic link between the Seychelles and the Indian subcontinent. A similar deep-time


intercontinental link is represented by the Cretaceous divergence of the Chilean Calyptocephalellidae and the Australo-Papuan Myobatrachoidea (129 Ma). This suggests that Australobatrachia once had a trans-Gondwanan distribution, ranging from South America over Antarctica to Australia. Although the nobleobatrachian radiation produced approximately half of the currently living anuran species, molecular divergence time estimates suggest that its initial diversification commenced relatively late, near the K-P boundary (15, 18). The rapid establishment of a large number of lineages that currently represent a broad range of ecomorphs (including toads, litter frogs, glass frogs, poison arrow frogs, fossorial frogs, and several lineages of tree frogs) fits the pattern of opportunistic radiation in the aftermath of the K-P extinction episode. Given the neotropical distribution of most nobleobatrachian families, this radiation is likely to have taken place primarily in South America. At least four lineages dispersed to other continents in the Tertiary: one lineage of Eleutherodactylidae reached North America (37, 38), Hylidae and Bufonidae attained widespread distributions, probably by dispersing through North America and Eurasia (39, 40), and the occurrence of Pelodryadidae in the AustraloPapuan realm provides evidence for a Tertiary transAntarctic range extension (15). The prevalence of continent-scale endemism in Ranoides, that is, the clear historical association of families with a single Gondwanan landmass, suggests that continental breakup has played a key role in the distribution of these frogs (29, 30). The late Cretaceous radiations of Natatanura and Microhyloidea indicate that dispersal between Gondwanan landmasses took place at least until the end Cretaceous (30, 31). In addition, comparable temporal and spatial divergence patterns within the microhyloid and natatanuran radiations suggest that the isolation of their daughter lineages on different continents were determined by the same geological events. The disruption of terrestrial passages that persisted between continents long after they started drifting apart may have resulted in parallel instances of vicariance in both clades (30). The late Cretaceous diversification of Natatanura and Microhyloidea imply a survival of multiple lineages across the K-P boundary. Some of the surviving lineages are represented by only few relict species (e.g., Phrynomeridae in Africa, Melanobatrachidae on the Indian subcontinent), but several of their largest families radiated substantially in the Paleogene (e.g., Dicroglossidae,

363

Mantellidae, Rhacophoridae, Ranidae, Asterophryidae, and Microhylidae) (18, 29). The timetree that is emerging from the rapid succession of molecular analyses provides an increasingly detailed temporal framework for anuran evolution. Besides shedding light on historical biogeography, this framework allows us to study patterns and rates of evolutionary change in morphological, ontogenetic, and genomic data. A remaining challenge for future phylogenetic studies is represented by the explosive radiations of Natatanura, Microhyloidea, and Nobleobatrachia, and several of their families. Resolving these radiations will most likely require alternative (and more expensive) strategies, provided by the expanding field of phylogenomics (e.g., using SINES or EST data). The credibility of the anuran timetree is reinforced by the relative consistency among independent studies, despite the use of different data sets, calibration points, and methods. In addition, although molecular time estimates for some anuran nodes are notably older than those derived from the fossil record, there are no major incompatibilities between the two types of data (41). Rather, molecular and fossil analyses can be considered complementary tools to understand the paleobiological processes and events that shaped the present-day anuran diversity.

Acknowledgment Support was provided by the Fonds voor Wetenschappelijk Onderzoek, Vlaanderen en de Vrije Universiteit Brussel.

References 1. AmphibiaWeb, Information on Amphibian Biology and Conservation, http://amphibiaweb.org/ (Berkeley, California, 2008). 2. W. E. Duellman, L. Trueb, Biology of Amphibians (The Johns Hopkins University Press, Baltimore and London, 1994). 3. D. R. Frost et al., Bull. Am. Mus. Nat. Hist. 297, 1 (2006). 4. J. C. Avise, G. C. Johns, Proc. Natl. Acad. Sci. U.S.A. 96, 7358 (1999). 5. J. D. Lynch, in Evolutionary Biology of the Anurans: Contemporary Research on Major Problems, J. L. Vial, Ed. (University of Missouri Press, Columbia, Missouri, 1973), pp. 133–182. 6. L. S. Ford, D. C. Cannatella, Herpetol. Monogr. 7, 94 (1993). 7. A. Haas, J. Zool. Syst. Evol. Res. 35, 179 (1997). 8. M. Maglia, L. A. Pugener, L. Trueb, Am. Zool. 41, 538 (2001).

364


9. A. Haas, Cladistics 19, 23 (2003). 10. L. A. Pugener, A. M. Maglia, L. Trueb, Zool. J. Linn. Soc. 139, 129 (2003). 11. S. B. Hedges, L. R. Maxson, Herpetol. Monogr. 7, 27 (1993). 12. J. M. Hay, I. Ruvinsky, S. B. Hedges, L. R. Maxson, Mol. Biol. Evol. 12, 928 (1995). 13. E. Feller, S. B. Hedges, Mol. Phylogenet. Evol. 9, 509 (1998). 14. S. Hoegg, M. Vences, H. Brinkmann, A. Meyer, Mol. Biol. Evol. 21, 1188 (2004). 15. D. S. San Mauro, M. Vences, M. Alcobendas, R. Zardoya, A. Meyer, Am. Nat. 165, 590 (2005). 16. K. Roelants, F. Bossuyt, Syst. Biol. 54, 111 (2005). 17. C. Gissi, D. San Mauro, G. Pesole, R. Zardoya, Gene 366, 228 (2006). 18. K. Roelants et al., Proc. Natl. Acad. Sci. U.S.A. 104, 887 (2007). 19. P. Zhang, H. Zhou, Y. Q. Chen, Y. F. Liu, L. H. Qu, Syst. Biol. 54, 391 (2005). 20. I. Ruvinsky, L. R. Maxson, Mol. Phylogenet. Evol. 5, 533 (1996). 21. S. D. Biju, F. Bossuyt, Nature 425, 711 (2003). 22. S. K. Dutta, K. Vasudevan, M. S. Chaitra, K. Shankar, R. K. Aggarwal, Curr. Sci. 86, 211 (2004). 23. C. R. Darst, D. C. Cannatella, Mol. Phylogenet. Evol. 31, 462 (2004). 24. J. J. Wiens, J. W. Fetzner, C. L. Parkinson, T. W. Reeder, Syst. Biol. 54, 719 (2005).

25. T. Grant et al., Bull. Am. Mus. Nat. Hist. 299, 1 (2006). 26. J. J. Wiens, Am. Nat. 170, S86 (2007). 27. A. van der Meijden, M. Vences, A. Meyer, Proc. Roy. Soc. Lond. B 271, S378 (2004). 28. A. van der Meijden, M. Vences, S. Hoegg, A. Meyer, Mol. Phylogenet. Evol. 37, 674 (2005). 29. F. Bossuyt, R. M. Brown, D. H. Hillis, D. C. Cannatella, M. C. Milinkovitch, Syst. Biol. 55, 579 (2006). 30. I. Van Bocxlaer, K. Roelants, S. D. Biju, J. Nagaraju, F. Bossuyt, PLoS ONE 1, e74 (2006). 31. A. van der Meijden et al., Mol. Phylogenet. Evol. 44, 1017 (2007). 32. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 33. F. Bossuyt, M. C. Milinkovitch, Science 292, 93 (2001). 34. M. Vences et al., Proc. Biol. Sci. 270, 2435 (2003). 35. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 36. J. Benton, The Fossil Record 2 (Chapman & Hall, London, 1993), pp. 845. 37. M. P. Heinicke, W. E. Duellman, S. B. Hedges, Proc. Natl. Acad. Sci. U.S.A. 104, 10092 (2007). 38. S. B. Hedges, W. E. Duellman, M. P. Heinicke, Zootaxa 1737, 1 (2008). 39. S. A. Smith, P. R. Stephens, J. J. Wiens, Evolution 59, 2433 (2005). 40. J. B. Pramuk, T. Robertson, J. W. Sites, Jr., B. P. Noonan, Global Ecol. Biogeogr. 17, 72 (2008). 41. D. Marjanovic, M. Laurin, Syst. Biol. 56, 369 (2007).

Salamanders (Caudata) David R. Vieites*, Peng Zhang, and David B. Wake Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Sciences Building, University of California, Berkeley, 94720-3160, California, USA *To whom correspondence should be addressed ([email protected])

Abstract Living salamanders (~570 species) are placed in 10 families, comprising the Order Caudata. Their classification is relatively stable, but phylogenetic relationships among families are contentious. Recent molecular phylogenetic analyses have found five major clades. The salamander timetree shows the deepest divergence, between Cryptobranchoidea and all other familes, in the Triassic (251–200 million years ago, Ma) with subsequent diversification occurring in the Jurassic (200–146 Ma) and early Cretaceous (146–100 Ma).

Salamanders form a monophyletic group, constituting one of the three orders of modern amphibians (Lissamphibia), together with frogs and caecilians. Salamanders comprise the second most species-rich order of amphibians (1) and are typically classified in 10 families, with ca. 68% of the species belonging to the Family Plethodontidae. The body plan has remained relatively stable since the Jurassic (2, 3) (Fig. 1), displaying several features that in combination distinguish it from the body plan of other amphibians: presence of a tail both in larval and adult phases, two pairs of limbs of equal size (when present) set perpendicular to the body, presence of teeth on both jaws, presence of ribs on most trunk vertebrae, and absence of several skull bones (4). Here we review the phylogenetic relationships and the divergence times of salamander families. The families are grouped into five suborders: Cryptobranchoidea (Cryptobranchidae and Hynobiidae), Sirenoidea (Sirenidae), Salamandroidea (Salamandridae, Ambystomatidae, Dicamptodontidae), Proteoidea (Proteidae), and Plethodontoidea (Plethodontidae, Rhyacotritonidae, and Amphiumidae). Despite the increasing number of studies and data addressing the phylogeny of salamander families, their relationships are difficult to resolve. Several relationships are consistently recovered

with different data sets, while the positions of others, in particular the sirenids and proteids, have remained contentious. A monophyletic Cryptobranchoidea and a clade consisting of Dicamptodontidae and Ambystomatidae are recovered in every molecular phylogenetic study (Vieites and Wake, submitted; Zhang and Wake, submitted; 2, 5–10). Salamandridae is usually recovered as the closest relative of the clade constituted by dicamptodontids and ambystomatids (Vieites and Wake, submitted; Zhang and Wake, submitted; 2, 5–7, 9, 10). Frost et al. (9) proposed placing Dicamptodontidae in Ambystomatidae because the two families form a clade and each contains only a single living genus, but divergence between the lineages is great and the two are very old (115.8 Ma, Table 1). Furthermore, dicamptodontids have a long and rather rich fossil record so the recognition of only a single family is misleading. The position of Proteidae has been contentious. North American and European species form a clade. Several molecular studies recovered proteids as closest relatives of sirens and nested within the crown (2, 6, 8, 9). Recent molecular phylogenetic studies have found that Proteidae is closest to Salamandroidea (7, 9), although with low statistical support. In contrast, a recent study using one mitochondrial and four nuclear markers found that Proteidae is the closest relative of Plethodontoidea (10). The same relationship was found, with high statistical support, with a data set of 19 nuclear markers as well as with complete mitochondrial genomes (Fig. 2; Vieites and Wake, submitted; Zhang and Wake, submitted). Plethodontoidea is monophyletic and Rhyacotritonidae is closest to a clade constituted by amphiumids and plethodontids (Vieites and Wake, submitted; Zhang and

Fig. 1 A plethodontid salamander (Karsenia koreana) female from the type locality. Credit: D. R. Vieites.

D. R. Vieites, P. Zhang, and D. B. Wake. Salamanders (Caudata). Pp. 365–368 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Amphiumidae

7

Rhyacotritonidae

5

Proteidae 3

Ambystomatidae 9

Dicamptodontidae

4

1

Sirenidae Hynobiidae 6

Tr

Jurassic

Cryptobranchidae

Cretaceous

MESOZOIC 200

150

Pg

Ng

Cryptobranchoidea

Salamandridae

Sirenoidea

2

Proteoidea

Plethodontidae 8

Salamandroidea

THE TIMETREE OF LIFE Plethodontoidea

366

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of salamanders (Caudata). Divergence times are shown in Table 1. Ng (Neogene), Pg (Paleogene), and Tr (Triassic).

Wake, submitted; 7, 9, 10). The large Family Plethodontidae always has been found to be monophyletic (e.g., 11). Multiple nuclear markers and complete mitochondrial genomes found that amphiumids are closest to plethodontids (Vieites and Wake, submitted; Zhang and Wake, submitted; 7, 10). One study that combined analyses of nuclear rRNA and mtDNA data suggested that amphiumids are at the base of the salamander tree, as closest relative of the remaining salamanders (8). The base of the salamander tree has been subject to controversy for decades. The Sirenidae is a small clade restricted to the southeastern United States and extreme northeastern Mexico that has a characteristic morphology. The species are greatly elongated, permanently aquatic, and gilled. They lack hind limbs and have jaws that lack teeth (except on the small coronoid bone on the inner surface of the lower jaw) and are covered by hardened keratinized “beaks.” Their reproduction is external but they differ from all other salamanders in lacking pelvic glands. They were once considered not to be salamanders and placed in a separate Order Trachystomata (12), later becoming the Order Meantes (13). In contrast, a paleontological study found Sirenidae to be a deeply nested clade, closest to Salamandridae (14). All relevant molecular studies have clustered sirenids with other salamanders. Analyses of complete mitochondrial genomes of all families found Sirenidae to be the closest relative of all other salamanders with high statistical support (Zhang and Wake, submitted). One study found them to be part of a basal polytomy

based on partial RAG1 sequences (15). An early study that combined morphological and molecular data suggested that sirenids represent a basal stem group (5). However, another study that combined nuclear RNA (based on the relatively scanty data then available) and morphological data found that Cryptobranchoidea is the closest relative of all other families (2). The same result was found with recent analyses of single (7) and multiple nuclear loci (10) data sets, but without strong statistical support. Bayesian and maximum likelihood analyses of a data set comprising 19 nuclear markers (Vieites and Wake, submitted) (Fig. 2) strongly support Cryptobranchoidea as the first branch, closest to a group including Sirenoidea and the remaining suborders. Cryptobranchoids display several traits that appear to be ancestral. They have external fertilization, like sirenids, and they are the only salamanders known to have a separate angular bone in the lower jaw, and high numbers of microchromosomes. Relationships of the extinct taxa Batrachosauroididae, Prosirenidae, and Scapherpetontidae are unclear because the fossil record is incomplete, but Karauridae is widely accepted as the closet relative of the living and extinct salamanders (14). Salamander fossils are scarce, and few phylogenetic studies have estimated divergence times among all salamander families. San Mauro et al. (15) estimated the divergence of Caudata and Anura at 271 ± 19 million years ago (Ma), based on a partial fragment of RAG1. Using several nuclear and a mitochondrial marker for representatives of all living families,

Eukaryota; Metazoa; Vertebrata; Lissamphibia; Caudata

367

Table 1. Divergence times (Ma) and confidence/credibility intervals among salamanders (Caudata). Timetree Node

Estimates

Time Ref. (10)(a)

Vieites and Wake (submitted)

Ref. (10)(b)

Time

CI

Time

CI

Zhang and Wake (submitted)

Time

CI

Time

CI

1

217.5

248.7

282–220

220.1

247–196

218.3

234–204

183.0

201–167

2

200.1

232.2

266–199

194.8

232–169

202.2

221–184

171.0

186–158

3

181.7

209.8

242–178

176.0

211–149

181.1

200–162

160.0

177–144

4

168.0

198.6

231–165

162.6

196–131

165.8

186–146

145.0

165–125

5

166.2

190.0

221–160

154.0

188–123

169.8

189–151

151.0

168–134

6

157.2

174.0

208–150

145.5

168–146

158.3

166–155

151.0

162–145

7

137.3

156.5

185–127

119.5

147–92

137.3

157–119

136.0

153–118

8

124.4

144.6

175–115

106.0

130–80

123.1

143–104

124.0

143–107

9

115.8

136.4

170–107

107.6

151–80

113.2

134–94

106.0

137–74

Note: Node times in the timetree represent the mean of time estimates from different studies. Divergence times calculated from an analysis of four nuclear and one mitochondrial markers using Bayesian (a) and penalized likelihood (b) methods (10) are shown. In another study (Vieites and Wake, submitted), 19 nuclear markers were used, and in a third study (Zhang and Wake, submitted) complete mitochondrial genomes were analyzed.

Roelants et al. (10) provided a younger estimate, using both Bayesian and penalized likelihood approaches (Table 1). A 19-nuclear-marker study (Vieites and Wake, submitted) using a Bayesian approach and minimum constraints, instead of fi xed calibration points, yielded divergence times that were on average close to the penalized likelihood estimates from Roelants et al. (10). An analysis of complete mitochondrial genomes using a rate-uncorrelated dating technique and a “soft bound” calibration strategy yielded an estimate of ~183 Ma (Zhang and Wake, submitted). A comparison of results for all salamander families, from studies using the same method of divergence time estimation (Table 1), shows discrepancies between the different estimates averaging 25 million years. The nuclear (Vieites and Wake, submitted) and mitochondrial (Zhang and Wake, submitted) data sets provided overlapping divergence time estimates in young nodes, while the mitochondrial data set gave much younger estimates for older nodes. All available data sets suggest that most of the families of salamanders diversified during the Jurassic (Fig. 2). Sirenoidea diverged from the crown group ~200 Ma. Salamandridae split from Ambystomatidae + Dicamptodontidae at about the same time as Proteoidea split from Plethodontoidea, during the mid-Jurassic (Table 1). The estimated time of divergence of hynobiids and cryptobranchids (157.2 Ma, Table 1) is in agreement with the oldest known fossils (3). The divergence time estimates of Ambystomatidae and Dicamptodontidae

are older than the oldest fossil (55.8 Ma, 16), suggesting an mid-Cretaceous split, 115.8 Ma. A similar result was found with respect to the split of Plethodontidae and Amphiumidae (124.4 Ma, Table 1), much older than the oldest amphiumid fossil (65.5 Ma, 17). In summary, the salamander timetree (Fig. 2) suggests that the diversification of extant salamander families happened during the Jurassic to mid-Cretaceous. Cryptobranchoids are known from fossils dating from 155 Ma in northeastern China (3, 18). The earliest sirenoid fossils are from the Cretaceous of Sudan, Africa, ca. 100 Ma (19), and the late Cretaceous, 83 Ma, and Paleocene of western North America (Alberta to Wyoming, 20). The earliest proteoid is from the late Paleocene of North America (20), the earliest plethodontoid is from the late Cretaceous, 66 Ma, of Montana (17), and the earliest salamandroid is from the latest Paleocene, 56 Ma, of Alberta and Montana (14, 16). Extant species of all families except Hynobiidae today occur in North America, and five families (Ambystomatidae, Amphiumidae, Dicamptodontidae, Rhyacotritonidae, and Sirenidae) are restricted to that continent. Both families of the Cryptobranchoidea occur in East Asia, where they cooccur with Salamandridae and Plethodontidae.

Acknowledgment Support was provided by U.S. National Science Foundation Tree of Life program.

368


References 1. Amphibia Web, Information on Amphibian Biology and Conservation, http://amphibiaweb.org/ (Berkeley, California, 2008). 2. K. Gao, N. H. Shubin, Nature 410, 574 (2001). 3. K. Gao, N. H. Shubin, Nature 422, 424 (2003). 4. W. Duellman, L. Trueb, The Biology of Amphibians (Johns Hopkins University Press, Baltimore, 1994), pp. 461–475. 5. A. Larson, W. Dimmick, Herp. Monogr. 7, 77 (1993). 6. A. Larson, D. W. Weisrock, H. H. Kozak, in Reproductive Biology and Phylogeny of Urodela (Amphibia), D. M. Sever, Ed. (NH: Science Publishers, Enfield, 2003), pp. 31–108. 7. J. J. Wiens, R. M. Bonett, P. T. Chippindale, Syst. Biol. 54, 91 (2005). 8. D. W. Weisrock, L. J. Harmon, A. Larson, Syst. Biol. 54, 758 (2005). 9. D. R. Frost et al., Bull. Am. Mus. Nat. Hist. New York 297, 1 (2006).

10. 11 12. 13. 14. 15. 16. 17. 18. 19. 20.

K. Roelants et al., Proc. Natl. Acad. Sci. U.S.A. 104, 887 (2007). R. L. Mueller, J. R. Macey, M. Jaekel, D. B. Wake, J. L. Boore, Proc. Natl. Acad. Sci. U.S.A. 101, 13820 (2004). C. Goin, O. Goin, Introduction to Herpetology, 2nd ed. (Freeman & Co., San Francisco, 1971). C. Goin, O. Goin, G. Zug, Introduction to Herpetology, 3rd ed. (Freeman & Co., San Francisco, 1978). R. Estes, in Handbuch der Paläoherpetologie (Pfeil Verlag, 1981), pp. 1–115. D. San Mauro, M. Vences, M. Alcobendas, R. Zardoya, A. Meyer, Am. Nat. 165, 590 (2005). B. G. Naylor, R. C. Fox, Can. J. Earth Sci. 30, 814 (1993). J. D. Gardner, J. Vert. Paleont. 23, 769 (2003). D. Marjanovic, M. Laurin, Syst. Biol. 56, 369 (2007). S. E. Evans, A. R. Milner, C. Werner, Paleontology 39, 77 (1996) J. A. Holman, Fossil Salamanders of North America (Indiana University Press, Bloomington, 2006).

Caecilians (Gymnophiona) David J. Gower* and Mark Wilkinson Department of Zoology, The Natural History Museum, London SW7 5BD, UK *To whom correspondence should be addressed (d.gower@nhm. ac.uk)

Abstract The ~170 species of caecilians (Gymnophiona) are grouped into three to six families. Analyses of molecular data since 1993 have largely consolidated earlier hypotheses of family relationships inferred from morphology, although Uraeotyphlidae nests within a paraphyletic Ichthyophiidae rather than being Teresomata’s closest relative. Dating analyses conducted thus far broadly agree. Most families diversified by the end of the Jurassic, 146 million years ago (Ma), with Uraeotyphlidae and Typhlonectidae originating from their ichthyophiid and caeciliid ancestors, respectively, by about 100–40 Ma. The Asian Ichthyophiidae and Uraeotyphlidae diverged after the breakup of Gondwana, probably on the Indian subcontinent before its collision with Asia.

Caecilians are a monophyletic group of elongate, snakeor wormlike amphibians lacking all trace of limbs and girdles, and with tails reduced or absent (Fig. 1). They are one of the three orders of the extant Lissamphibia, the Gymnophiona (~naked snakes), and are most likely the closest relatives of the more familiar frogs and salamanders (1, 2). All caecilians possess a distinctive cranial sensory organ, the tentacle, and have a unique dual jaw closing mechanism (3). Males have an eversible cloaca used in copulation, and fertilization is internal (4). Some groups retain the ancestral trait of an aquatic larval stage, but direct development and viviparity are common (5). The skin is externally segmented, with scales present in dermal pockets in many species. Most of the ~170 known species, grouped into three to six families (1, 6), inhabit soils as adults and, associated with burrowing, have reduced visual systems and heavily ossified skulls. The group has a primarily tropical (Gondwanan) distribution. Here we review the inferred phylogenetic relationships and estimated divergence times of the major lineages of caecilians. The classification used here follows the most recent review (6).

Until 1968 only a single family of caecilians was recognized. Taylor (7, 8) provided a four family classification that has been variously modified and extended by subsequent authors to accommodate new information on morphology, alternative hypotheses of phylogeny and differing perspectives on how best to deal with demonstrably paraphyletic taxa (1, 3, 6, 9–11). The first numerical phylogenetic analysis of caecilians (9) investigated intergeneric relationships using morphological (and life history) data. This and other family-level studies based on these initial data (12, 13) yielded a view of the phylogenetic relationships of the major lineages that has, in the main, been corroborated by subsequent molecular and morphological studies. The major exception has been a change in the placement of the Uraeotyphlidae, an Indian endemic that despite many similarities to the Teresomata (scolecomorphids, caeciliids, and typhlonectids) is now placed in the closest relative of the Teresomata, the Diatriata (Uraeotyphlidae + Ichthyophiidae), based on both morphological (9) and molecular (1, 2, 14–16) data (Fig. 2). The first molecular phylogenetic studies used only partial mitochondrial ribosomal genes. Taxonomic

Fig. 1 A caeciliid caecilian amphibian (Herpele squalostoma) from Cameroon. Credit: © 1999 Natural History Museum, London.

D. J. Gower and M. Wilkinson. Caecilians (Gymnophiona). Pp. 369–372 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

370


Scolecomorphidae

2

Ichthyophiidae 5

1

Uraeotyphlidae

Neocaecilia

Caeciliidae 2

3

Diatriata

Caeciliidae 1

4

Teresomata

Typhlonectidae 6

Rhinatrematidae

Triassic

Jurassic

Cretaceous

MESOZOIC 200

150

Pg

Ng

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of caecilians (Gymnophiona). Divergence times are shown in Table 1. The single species (26) of Ichthyophis that is the closest relative of Uraeotyphlus (thus making Ichthyophiidae paraphyletic) has not yet been included in dating analyses, and is ignored here. Codes for paraphyletic and/or

polyphyletic groups are as follows: Caeciliidae-1 (Caecilia + Oscaecilia, Chthonerpeton, and Typhlonectes) and Caeciliidae-2 (Boulengerula and Herpele). Abbreviations: Ng (Neogene) and Pg (Paleogene).

sampling has been increased steadily so as to improve the coverage of families from four to six, and to begin testing their monophyly as well as their interrelationships (11, 14, 17). Substantial expansions of the available molecular data have seen combined analyses of complete mitochondrial genomes and the RAG-1 nuclear gene of representatives of all six families (15) and of concatenations of multiple nuclear and mitochondrial markers as part of large-scale analyses of amphibian interrelationships (1, 2). Sampling at the generic level is not yet complete. We present a consensus view of the phylogenetic relationships of the major lineages of caecilians emerging from morphological and molecular studies (Fig. 2). Monophyly of four (of six families)—Rhinatrematidae, Uraeotyphlidae, Scolecomorphidae, and Typhlonectidae—is well supported by analyses of morphological and/or molecular data. The large and heterogeneous Caeciliidae, and the relatively more uniform Ichthyophiidae, have been convincingly demonstrated to be paraphyletic to Typhlonectidae and Uraeotyphlidae, respectively (1, 2, 9, 10, 16). Interfamilial relationships are generally well-supported by both morphology and molecules, except for the position of the Scolecomorphidae. While some molecular analyses have placed scolecomorphids within the Caeciliidae (1, 17), analyses of morphological data, of complete mitochondrial genomes and RAG-1, and of the most recent concatenated mitochondrial and nuclear markers indicate that Scolecomorphidae is the closest relative of the group containing Caeciliidae and Typhlonectidae (2, 10, 15).

Only a few studies have used molecular data to estimate the age of divergences among caecilian families (2, 14, 18, 19). Only Roelants et al. (2) have estimated dates of divergence for all six nodes in the inter-family tree, and so we focus on that study here (Fig. 2). This study dated a phylogeny of 171 amphibians (24 caecilians) based on ca. 3750 kb of sequences for one mitochondrial and four nuclear genes by using 22 calibrations from both (caecilian and noncaecilian) paleobiogeographic and (noncaecilian) fossil data. Use of two different statistical methods produced similar results (2). Divergences among most caecilian families are estimated to have occurred in the early Mesozoic (251–146 Ma), at least by the end of the Jurassic (146 Ma). The two exceptions are later Mesozoic/early Cenozoic (100–40 Ma) divergences associated with the paraphyly of Ichthyophiidae and Caeciliidae (Fig. 2). Depth of divergence might profitably be employed to determine rank in future revisions of caecilian classification (20). Other molecular dating estimates for divergences of some caecilian families based on smaller taxon and character samplings and using a variety of methods (14, 18, 19) are generally a little older, but they overlap with those from Roelants et al. (2). The main difference is an estimate of 250 Ma for the divergence between Diatriata and Teresomata (19), which is based on a single nonamphibian fossil calibration. Reanalysis of that data set with improved, multiple calibrations generally resulted in substantially younger dates throughout amphibians, although a revised estimate for the Diatriata–Teresomata divergence was not reported (21). Two other studies have

Eukaryota; Metazoa; Vertebrata; Lissamphibia; Gymnophiona

371

Table 1. Divergence times (Ma) and their 95% confidence/credibility intervals (CI) among caecilians (Gymnophiona). Timetree Node

Time


Ref. (2)(b)

Ref. (14)

Ref. (18)

Ref. (19)

Time

CI

Time

CI

Time

Interval

Time

CI

Time

CI

242–192

–

–

214.3

256–177

–

–

1

226.4

226.4

254–197

217.8

2

195.8

195.8

223–168

188.2

214–167

178

278–126

192.4

233–160

250

274–224

3

169.3

169.3

193–146

162.8

185–145

–

–

177.1

218–148

–

–

4

161.8

161.8

185–140

156.5

175–139

–

–

155.2

193–134

–

–

5

75.3

75.3

100–54

74.6

99–55

94

123–72

104.3

151–65

–

–

6

62.0

62.0

83–46

59.6

77–42

–

–

–

–

–

–

Note: Node times in the timetree are from Thorne–Kishino analysis (a) of one mitochondrial and four nuclear genes (~3750 basepairs) for 24 caecilian species as reported in ref. (2), in which results were also reported for penalized likelihood analysis (b) of the same data.

used molecular dating analyses to interpret caecilian evolution. Published substitution rates for amphibian mitochondrial DNA were used to estimate the divergence between Indian and the monophyletic Sri Lankan ichthyophiids at between 9.25 and 26 Ma (22), and relative dating was used to demonstrate that the divergences of three pairs of disjunctly distributed East–West African caeciliids and scolecomorphids were not contemporaneous (23). Currently, the poverty of the “caecilian” fossil record (24) renders it irrelevant to the issue of dating divergences within Gymnophiona, because it comprises only two (ca. 190 and 140 Ma) fossil taxa not assignable to living lineages (= Gymnophiona) and three kinds of fossil vertebrae that may or may not belong to living lineages. Thus, it is not possible to use any currently known “caecilian” fossil to estimate the minimum age of Gymnophiona. As a result, molecular dating estimates have relied on paleogeographic data and noncaecilian fossils for calibration (2, 14, 18, 19). The timetree indicates that multiple lineages of Gymnophiona coexisted with the two fossil (possibly stem-) taxa that do not belong to Gymnophiona, Eocaecilia and Rubricacaecilia. The stegokrotaphic (closed roofed) skull of Eocaecilia has been used to argue that, unlike the frogs and salamanders, and rhinatrematid caecilians (which have open roofed skulls), the ancestral caecilian was stegokrotaphic and therefore had a separate ancestry from the other amphibians (25). However, the long independent histories of Eocaecilia and Gymnophiona, the plausibility of their convergent adaptation to burrowing and independent evolution of stegokrotaphy, and the morphology of rhinatrematids caution against

accepting Eocaecilia as an accurate model for the ancestral caecilian. The timetree is consistent with the hypothesis, based on present-day geographical distributions, that Gymnophiona is primarily a radiation of Gondwana (and the Gondwanan part of Pangea) and that the divergence of the exclusively Asian Ichthyophiidae and Uraeotyphlidae occurred on the Indian plate subsequent to the breakup of Gondwana and before its collision with Laurasian Asia (14). One of the most interesting aspects of caecilian biology is the diversity of reproductive modes within the group. Some caecilians have been recently discovered to have extended parental care in which altricial young feed on a modified, lipid-rich epidermis of their attending mothers (5). Consideration of phylogenetic relationships suggests that this maternal dematophagy may be fairly widespread within Neocaecilia (27). Molecular dates indicate that this highly unusual form of parental care has persisted in multiple lineages for perhaps more than 138 million years.

Acknowledgment H. Taylor photographed the caecilian used in Fig. 1.

References 1. D. R. Frost et al., Bull. Am. Mus. Nat. Hist. 297, 1 (2006). 2. K. Roelants et al., PNAS 104, 887 (2007). 3. R. A. Nussbaum, Occ. Pap. Univ. Mich. Mus. Zool. 683, 1 (1977). 4. D. J. Gower, M. Wilkinson, Bull. Nat. Hist. Mus. (Zool.) 68, 143 (2002). 5. A. Kupfer et al., Nature 13, 440 (2006).

372

6.

7. 8. 9. 10. 11. 12. 13.


M. Wilkinson, R. A. Nussbaum, in Reproductive Biology and Phylogeny of Gymnophiona, J.-M. Exbrayat, Eds. (Science Publishers, Enfield, NH, 2006), pp. 39–78. E. H. Taylor, The Caecilians of the World (University of Kansas Press, 1968). E. H. Taylor, Univ. Kansas Sci. Bull. 10, 297 (1969). R. A. Nussbaum, Occ. Pap. Univ. Mich. Mus. Zool. 687, 1 (1979). M. Wilkinson, R. A. Nussbaum, Copeia 1996, 550 (1996). S. B. Hedges, R. A. Nussbaum, L. R. Maxson, Herpetol. Monogr. 7, 64 (1993). W. E. Duellman, L. Trueb, Biology of Amphibians (John Hopkins University Press, Baltimore, 1994), pp. 696. D. M. Hillis, in Amphibian Cytogenetics and Evolution, D. M. Green, S. K. Sessions, Eds. (Academic Press, San Diego, 1991), pp. 7–31.

14. M. Wilkinson et al., Mol. Phylogenet. Evol. 23, 401 (2002). 15. D. San Mauro et al., Mol. Phylogenet. Evol. 33, 413 (2004). 16. D. J. Gower et al., Proc. Roy. Soc. Lond. B 269, 1563 (2002). 17. M. Wilkinson et al., Afr. J. Herpetol. 52, 83 (2003). 18. D. San Mauro et al., Am. Nat. 165, 590 (2005). 19. P. Zhang et al., Syst. Biol. 54, 391 (2005). 20. J. C. Avise, G. C. Johns, PNAS 96, 7358 (1999). 21. D. Marjanovic, M. Laurin, Syst. Biol. 56, 369 (2007). 22. F. Bossuyt et al., Science 306, 479 (2004). 23. S. P. Loader et al., Biol. Lett. 3, 505 (2007). 24. S. E. Evans, D. Sigogneau-Russell, Palaeontology 44, 259 (2001). 25. R. L. Carroll, Zool. J. Linn. Soc. 150 (s1), 1 (2007). 26. D. J. Gower et al., J. Zool. 272, 266 (2007). 27. M. Wilkinson et al., Biol. Lett. 4, 358 (2008).

AMNIOTES


Amniotes (Amniota) Andrew M. Shedlock* and Scott V. Edwards Department of Organismic and Evolutionary Biology, Museum of Comparative Zoology, 26 Oxford Street, Harvard University, Cambridge, MA 02138, USA *To whom correspondence should be addressed (shedlock@oeb. harvard.edu)

Abstract Amniota is a remarkably diverse clade of tetrapod vertebrates comprising more than 23,000 living species of mammals, non-avian reptiles, and birds adapted to a wide variety of primarily terrestrial habitats. Our most recent common amniote ancestor probably lived ~325 million years ago (Ma), and molecular data suggest that the major lineages of living reptiles arose in the Permian and Triassic (299–200 Ma), when land areas were coalesced into a single supercontinent, Pangaea. Conflicting morphological and molecular results for the origin of turtles has been particularly challenging to resolve, although most recent analyses place turtles with birds and crocodilians.

Amniota is a clade noted for the extraordinary ecological and taxonomic diversity of its more than 23,000 living species, comprising mammals and reptiles, including birds. The taxon is named for the characteristic egg structure in which the amnion membrane forms a fluidfilled cavity that surrounds the developing embryo. The amniote egg is considered one of the key adaptations in the evolution of vertebrate terrestrial life history (Fig. 1). This shared-derived character is generally accepted to be one of several key adaptations for enduring the terrestrial life adopted by our amniote ancestors, thereby facilitating the remarkable evolutionary success of the group. Considerable debate has been focused on precisely when our most recent common amniote ancestor lived (1). Tightly constrained radiometric analysis of fossils places the divergence between the first undisputed synapsids (mammals which have only one temporal fenestration in the skull) and diapsids (reptiles with two temporal fenestrations) at 306.1 ± 8.5 Ma. This dating technique generally exhibits about 1% error (2). The upper bound of this narrow estimate for bird–mammal divergence

overlaps with a widely cited value of 310 Ma, based on the approximate methods of Benton (3) and employed by Kumar and Hedges (4) to calibrate a comprehensive molecular timescale for vertebrate evolution. The amniote timetree considered here includes six terminal taxa: Mammalia (mammals), Sphenodontia (the tuatara), Squamata (lizards and snakes), Testudines (turtles), Crocodylia (alligators and crocodiles), and Aves (birds). A combination of molecular and fossil evidence suggests that the major lineages of living reptiles likely originated in the Permian and Triassic (3–5) although exact divergence time estimates vary among the molecular studies used in the present synthesis. Conflicting phylogenetic results from different morphological and molecular data sets have clouded the picture of amniote macroevolution and have slowed the process of establishing a clear consensus of views between paleontologists and molecular systematists. Synapsids (mammals) are accepted to be the closest relatives of diapsid reptiles. Turtles lack temporal holes in the skull, however, and have been viewed as the only surviving anapsid amniotes. Based on this and other characters, they

Fig. 1 An amniote, the Rough Greensnake (Opheodrys aestivus), hatching from an egg. Photo credit: S. B. Hedges.

A. M. Shedlock and S. V. Edwards. Amniotes (Amniota). Pp. 375–379 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

376


Aves

2

Squamata 3

1

Sphenodontia Mammalia

C

P

Tr

J

PALEOZOIC 300

250

Cretaceous

Pg

150

Ng

CZ

MESOZOIC 200

Lepidosauria

Crocodylia 5

Archosauria

Testudines 4

100

50

0 Million years ago

Fig. 2 A timetree of amniotes. The divergence times are from Table 1. Abbreviations: C (Carboniferous), CZ (Cenozoic), J (Jurassic), Ng (Neogene), P (Permian), Pg (Paleogene), and Tr (Triassic).

are traditionally placed as the closest relatives of all other living reptiles (5, 7, 8). However, a recent morphological reevaluation of skull characters (9), and essentially all molecular genetic studies, indicates that turtles should be nested within diapsids despite their apparent anapsid skull morphology. Based on a growing body of genomic evidence (6, 10–14), it appears that turtles exhibit a secondary loss of skull fenestration or reversal to an ancestral condition. Molecular data sets have not completely clarified the amniote picture, however, since different genes and sampling schemes have in some cases placed turtles within archosaurs, as the closest relative of crocodilians (6, 10, 13, 15), and in other cases, as the closest relative of an archosaur clade that includes crocodilians and birds (11, 12, 14, 16). Statistical evaluation of phylogenetic signal in these studies (10) has revealed that mitochondrial genome data have tended to separate turtles from archosaurs, whereas concatenated nuclear gene sequences under potentially strong selection such as globin genes, lactate dehydrogenase (LDH), and ribosomal RNAs have tended to group turtles with crocodilians. The affi nity of turtles and crocodilians has also been observed in evaluating short DNA word motifs embedded in more than 84 million basepairs (Mbp) of genomic sequence for amniotes (13). The earliest turtles appear ~223 Ma in the fossil record (3), which is within the interval of both older and younger molecular clock estimates for the time of their origin. The most recent comprehensive effort to resolve the relationships of the major groups of amniotes used more than 5100 amino acids from mostly single-copy nuclear DPLA and GAG genes within a maximum likelihood framework, and presented strong statistical support for a close

relationship of crocodilians and birds to the exclusion of turtles (11). Despite considerable debate on the subject, few studies have presented average divergence time estimations across major branches of the amniote tree based on a calibrated clock analysis of numerous genes. Two highly visible studies within the last decade have set the foundation for an ongoing debate about integrating fossils and genes to estimate divergence times among major amniote clades. Kumar and Hedges (4) published their landmark comprehensive molecular timescale for vertebrate evolution based on protein clock calibrations of 658 nuclear genes and 207 vertebrate species. A point estimate of 310 Ma for the bird–mammal divergence, derived from radiometric dating of fossil evidence, was employed to externally calibrate the protein clock. Time estimates based on statistical tests of rate constancy (17, 18) were averaged across multiple genes and taxonomic groups and presented with 95% confidence intervals. Published dates relevant for the amniote timetree included a Permian average estimate for the origin of Lepidosaria at ~276 ± 54.4 Ma, and a bird–crocodilian split at ~222 ± 52.5 Ma. Less than a year later, similar methods were focused specifically on the Reptilia. Hedges and Poling (6) analyzed combinations of 23 nuclear and two mitochondrial genes, including globulins, LDH, rRNAs, alphacrystalline, alpha-enolase, and cyt b, to infer a molecular phylogeny of reptiles. As with previous analyses, the 310 synapsid–diapsid split date was used to anchor the molecular clock. The paper unconventionally joined turtles with crocodilians and suggested a more distant relationship between the tuatara and squamates based on a subset of amino acid data available for Sphenodon.

Eukaryota; Metazoa; Vertebrata; Amniota

377

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among amniotes (Amniota). Timetree Node


Time

Refs. (6, 27)

Refs. (12, 21, 23)

Ref. (15)

Ref. (20)(a)

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

324.5

–

–

323(27)

343–305

326(21)

354–311

–

–

–

–

2

274.9

276

383–169

245

269–221

285

–

250

267–233

285

296–274

3

271.5

–

–

–

–

–

–

–

–

268

280–256

4

230.7

225(25)

238–205

–

–

–

–

152

176–128

265

278–252

5

219.2

222

325–119

258(27)

279–238

196(23)

294–98

214

259–169

201

216–186

1

Timetree

Estimates (Continued) Ref. (22)

Ref. (24)

Time

CI

Time

Time

Time

CI

Time

CI

324.5

–

–

–

–

–

–

–

–

2

274.9

289

302–276

–

294

–

–

–

–

3

271.5

275

292–258

–

–

–

–

–

–

4

230.7

273

291–255

191

278

–

–

–

–

5

219.2

194

218–170

175

254

259

282–236

258

279–238

Node

1

Ref. (20)(b)

Time

Ref. (26)

Ref. (27)

Note: Node times in the timetree represent the mean of time estimates from different studies. Estimates are presented from an analysis of nine nuclear genes examined within a comprehensive analysis of 658 nuclear genes (4); 23 nuclear and nine mitochondrial protein-coding genes (6), 11 nuclear proteincoding genes (23); nuclear LDH-A, LDH-B, and alpha-enolase genes (15); 11 mitochondrial protein-coding genes (24); 12 protein-coding genes, two rRNA genes, and 19 tRNAs from mitochondria (25); 11 protein-coding genes and 19 tRNAs from mitochondria (12), 11 protein-coding genes and unspecified RNA genes from mitochondria (22); 325 protein-coding genes (21), mitochondrial genomes (26, 27), and the nucleotide (a) and amino acid sequences (b) of the nuclear RAG1 gene (20).

Squamates were estimated to have diverged from other reptiles ~245 ± 12.2 Ma, birds from the crocodilian + turtle clade ~228 ± 10.3 Ma, and turtles from crocodilians ~207 ± 20.5 Ma. These results differ from previous average dates published for the vertebrate timescale (4) by suggesting a younger Triassic vs. Permian time frame for the origin of squamate lizards, and a slightly older time for the bird–crocodilian split. These divergence time estimates cannot be reconciled with fossil evidence that provides a minimum age estimate of at least ~223 Ma for the oldest known turtles (3), but suggest that a number of key innovations during amniote evolution occurred within a roughly 40 million period of the Triassic. The study also posed a challenge for paleontologists to reconcile the derived position of turtles among amniotes and set a methodological framework for expanded application of molecular clock analyses to a variety of questions regarding vertebrate evolution. Several additional studies warrant summary here for their contributions to establishing the amniote timetree. The first is specifically aimed at attempting to resolve

the phylogenetic position of turtles (15). The study adds considerable additional LDH-A, LDH-B, and alphaenolase protein sequences to the available data matrix and employs the average distance methods and 310 Ma synapsid–diapsid clock calibration of Kumar and Hedges (4). Strong statistical support was provided for a turtle– crocodilian relationship to the exclusion of birds, with an exceptionally recent Upper Jurassic average divergence time estimate for turtles from crocodilians of only 152 Ma, some 71 million years younger than the oldest fossil turtle. No data were presented for Sphenodon, highlighting the overall lack of published divergence time estimates for this unique “living fossil.” Rest et al. (12) examined 11 protein-coding genes and 19 tRNAs from the mitochondrial genomes of all major amniote lineages, emphasizing the importance of sampling the tuatara to accurately reconstructing reptile phylogeny. The study employed advances in Bayesian model-based methods of inference and resolved turtles as the sister group to archosaurs with 100% statistical support. Adopting a maximum likelihood approach to

378


estimating divergence times under a relaxed molecular clock assumption and multiple fossil calibration points (19), the authors reported a divergence time well into the Permian for the origin of Lepidosauria. Most recently, Hugall et al. (20) analyzed the nuclear gene RAG1 across 88 taxa spanning all major tetrapod clades. The study supported the close relationship of turtles to a monophyletic Archosauria and employed model-based rate-smoothing methods to estimate divergence times in amniotes. Results highlighted an ancient Permian origin for the tuatara estimated at ~268–275 Ma and revealed slower molecular evolutionary rates in archosaurs and especially turtles that tend to underestimate divergence times for these groups without using appropriate fossil calibration points. Comparison with other molecular clock studies underscored the bias toward inflated divergence time estimates created by saturated mitochondrial gene sequence data. Seven other studies also present molecular time estimates for at least one node in the amniote tree of life (21–27). The genes examined in each of these studies are listed in the caption for Table 1 and the average time estimates from all 12 investigations considered here are reflected by the topology in Fig. 2. Based on this literature synthesis, the amniote common ancestor is dated at 325 Ma, with the origin of lepidosaurs taking place in the mid-Permian ~275 Ma, and the divergence of sphenodontids from squamates at 272 Ma, distinguishing the living tuatara as a remarkably ancient evolutionary relict. Turtles are estimated to have diverged from archosaurs in the early Triassic some 231 Ma, whereas crocodylians subsequently split from birds as recently as 219 Ma. It must be noted that confidence intervals on the estimates summarized in Table 1 are not available for several studies (12, 22, 24) or vary widely among studies depending on the number of genes investigated, the nature of the gene sequence data analyzed, and taxonomic sampling (e.g., 4 vs. 20). The results in Fig. 2 should therefore be interpreted cautiously within narrow time frames of amniote evolutionary history. In summary, the amniote timetree paints a mixed portrait of organismal evolution among major tetrapod lineages that underwent a number of key vertebrate innovations adapted for the rigors of terrestrial life. The late Permian mass extinction was due presumably to widespread stressful environmental conditions produced by unfavorable ocean–atmosphere chemical interactions (28). Fossil diversity indicates a protracted biological recovery through the lower Triassic ~250 Ma (29) that falls between molecular time estimates for two major

periods of amniote cladogenesis: (1) the divergence and early diversification of lepidosaurs between ~270 and 275 Ma and (2) the origin of turtles and subsequent divergence of crocodilians and birds between roughly ~230 and 220 Ma. Young amniote lineages would have emerged into a relatively unconstrained evolutionary landscape before extensive Pangean supercontinental breakup. A combination of favorable environmental and biogeographic forces likely facilitated successful tetrapod invasion into a variety of open terrestrial niches during the Triassic, exemplified today by the extraordinary ecological, phenotypic, and taxonomic diversity of living amniotes.

Acknowledgments Support for the preparation of this manuscript was provided by Harvard University and in part by a U.S. National Science Foundation grant to S.V.E.

References 1. S. B. Hedges, S. Kumar, Trends Genet. 20, 242 (2004). 2. J. Remane et. al., Eds., International Stratigraphic Chart (International Union of Geological Sciences: International Commission on Stratigraphy, Paris, 2002). 3. M. J. Benton, Vertebrate Paleontology (Chapman & Hall, New York, 1997). 4. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 5. R. R. Reisz, Trends Ecol. Evol. 12, 218 (1997). 6. S. B. Hedges, L. L. Poling, Science 283, 998 (1999). 7. J. Gauthier, A. G. Kluge, T. Rowe, Cladistics 4, 105 (1988). 8. M. S. Y. Lee, Zool. J. Linn. Soc. 120, 197 (1997). 9. O. Rieppel, R. R. Reisz, Ann. Rev. Ecol. Syst. 30, 1 (1999). 10. Y. Cao, M. D. Sorenson, Y. Kumazawa, D. P. Mindell, M. Hasegawa, Gene 259, 139 (2000). 11. N. Iwabe et al., Mol. Biol. Evol. 22, 810 (2005). 12. J. R. Rest et al., Mol. Phylogenet. Evol. 29, 289 (2003). 13. A. M. Shedlock et al., Proc. Natl. Acad. Sci. U.S.A. 104, 2767 (2007). 14. R. Zardoya, A. Meyer, Proc. Natl. Acad. Sci. U.S.A. 95, 14226 (1998). 15. H. Mannen, S. S.-L. Li, Mol. Phylogenet. Evol. 13, 144 (1999). 16. Y. Kumazawa, M. Nishida, Mol. Biol. Evol. 16, 784 (1999). 17. F. Tajima, Genetics 135, 599 (1993). 18. N. Takezaki, A. Rzhetsky, M. Nei, Mol. Biol. Evol. 12, 823 (1995). 19. M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997). 20. A. F. Hugall, R. Foster, M. S. Lee, Syst. Biol. 56, 543 (2007). 21. J. E. Blair, S. B. Hedges, Mol. Biol. Evol. 22, 2275 (2005).

Eukaryota; Metazoa; Vertebrata; Amniota

22. G. L. Harrison et al., Mol. Biol. Evol. 21, 974 (2004). 23. S. Hughes, D. Zelus, D. Mouchiroud, Mol. Biol. Evol. 16, 1521 (1999). 24. A. Janke, D. Erpenbeck, M. Nielsson, U. Arnason, Proc. Roy. Soc. Lond. B 268, 623 (2001). 25. T. Paton, O. Haddrath, A. J. Baker, Proc. Roy. Soc. Lond. B. 269, 839 (2002).

26.

379

P. Zhang, H. Zhou, Y.-Q. Chen, Y.-F. Liu, L.-H. Qu, Syst. Biol. 54, 391 (2005). 27. S. L. Pereira, A. J. Baker, Mol. Biol. Evol. 23, 1731 (2006). 28. D. H. Erwin, Extinction. How Life on Earth Nearly Ended 250 Million Years Ago (Princeton University Press, Princeton, NJ, 2006). 29. J. L. Payne et al., Science 305, 506 (2004).


REPTILES


Lizards, snakes, and amphisbaenians (Squamata) S. Blair Hedgesa,* and Nicolas Vidala,b a

Department of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA; bUMR 7138, Systématique, Evolution, Adaptation, Département Systématique et Evolution, C.P. 26, Muséum National d’Histoire Naturelle, 43 Rue Cuvier, Paris 75005, France *To whom correspondence should be addressed ([email protected])

Abstract Living species of lizards, snakes, and amphisbaenians (~8200 sp.) are grouped into 58 families within the sauropsid Order Squamata. Recent phylogenetic analyses of nuclear genes have resulted in major changes in their classification. Iguanian lizards, once considered basal in the squamate tree, are now placed in a highly nested position together with snakes and anguimorph lizards. The squamate timetree shows that most major groups diversified in the Jurassic and Cretaceous, 200–66 million years ago (Ma), possibly related to the breakup of supercontinents. In contrast, five of the six families of amphisbaenians are younger, having arisen during the Cenozoic (66–0 Ma).

The lizards, snakes, and amphisbaenians form a monophyletic group of scaly reptiles, the Order Squamata. They are typically grouped together with the tuataras (Order Rhynchocephalia) in the Subclass Lepidosauria. Male squamates have a pair of unique copulatory organs, hemipenes, located in the tail base. Limb reduction or loss has occurred independently in multiple lineages. Nearly 8200 living species of squamates have been described and placed in ~58 families: ~4900 species in 26 families of lizards (Fig. 1), ~200 species in six families of amphisbaenians, and ~3070 species in 26 families of snakes (1, 2). Here, we review the relationships and divergence times of the families of squamates, excluding snakes (Serpentes), which are treated elsewhere (3). The classification of squamates was pioneered by Camp (4) and has, until recently, followed the arrangement proposed by Estes et al. (5). In it, species were placed in two major groups: Iguania (iguanids, agamids,

and chamaeleonids) and Scleroglossa (all other families). This division was based on multiple morphological characters (5) but emphasized tongue morphology and mode of feeding. Iguanians have muscular tongues and use tongue prehension, a feeding mode which is thought to be primitive, whereas scleroglossans have hard tongues and use jaw prehension. Recently, morphological analyses have continued to find support for this conventional classification of squamates (6). Historically, three groups of squamates having limb reduction or loss (snakes, amphisbaenians, and dibamids) have been the most difficult to classify, probably because of their specialization and loss of characters. Nonetheless, all three groups have been placed with scleroglossans in most classifications (7, 8). The long, deeply forked tongue and other characters of snakes have allied them with anguimorph lizards, especially the monitor lizards (Varanidae and Lanthanotidae). The Mesozoic (251–66 Ma) fossil record of squamates is sparse (9). There are no known fossils before the Triassic/ Jurassic boundary (200 Ma), although indirect evidence

Fig. 1 Representative squamate reptiles. (A) An iguanid lizard, Anolis baracoae, from Cuba (upper left); (B) a sphaerodactylid lizard (Sphaerodactylus richardsoni) from Jamaica (upper right); (C) an amphisbaenid amphisbaenian (Amphisbaena bakeri) from Puerto Rico (lower left), and a colubrid snake (Lampropeltis triangulum), from the United States (lower right). Credits: S. B. Hedges.

S. B. Hedges and N. Vidal. Lizards, snakes, and amphisbaenians (Squamata). Pp. 383–389 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

384


Bipedidae Blanidae

14

26

Laterata

Trogonophidae

25

Amphisbaenia

Amphisbaenidae 27

Cadeidae

11

Rhineuridae

Teiidae 19

Gymnophthalmidae

Teiioidea

Lacertidae

6

Anguidae 22 17

Helodermatidae Xenosauridae

13

Lanthanotidae 23

Varanidae

16 3

Toxicofera

Diploglossidae

4

Anguimorpha

Anniellidae

Shinisauridae

7

Chamaeleonidae 20

Agamidae

Gerrhosauridae 24

Cordylidae Sphaerodactylidae

15

Phyllodactylidae Eublepharidae

9

Diplodactylidae 21

Carphodactylidae Pygopodidae Dibamidae

Tr

Jurassic

Cretaceous

MESOZOIC 200

150

Paleogene

Pygopodomorpha

1

Gekkonidae 18

12

Gekkomorpha

Xantusiidae 8

Cordylomorpha

5 2

Scinciformata

Scincidae

Gekkota

Iguanidae 10

Iguania

Serpentes

Ng

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of squamate reptiles (Squamata). Divergence times are shown in Table 1. Abbreviations: Ng (Neogene) and Tr (Triassic).

Eukaryota; Metazoa; Vertebrata; Sauropsida; Squamata

385

Table 1. Divergence times (Ma) and their credibility/confidence intervals (CI) among squamate reptiles (Squamata). Timetree Node

Estimates Ref. (22)

Time

Ref. (27)

Ref. (29)

Ref. (25)

Ref. (24)

Ref. (35)

Ref. (38)

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

Time

CI

209.4

240

251–221

178.7

184–173

–

–

–

–

–

–

–

–

–

2

197.9

225

240–207

178.7

184–173

190.0

204–176

–

–

–

–

–

250–236

278–208

3

188.3

215

230–199

173.9

179–169

176.0

190–162

–

–

–

–

–

–

–

1

4

179.7

191

206–179

168.3

174–163

–

–

–

–

–

–

–

–

–

5

170.5

192

209–176

157.6

167–149

162.0

175–149

–

–

–

–

–

–

–

6

169.3

177

193–164

161.6

168–156

–

–

–

–

–

–

–

–

–

7

166.4

178

194–167

163.1

169–158

158.0

171–145

–

–

–

–

–

–

–

8

161.4

179

197–161

143.8

156–132

–

–

–

–

–

–

–

–

–

9

144.6

111

133–90

86.5

95–78

94.0

104–84

144.6

206–85

–

–

–

196–163

233–130

10

144.2

–

–

146.4

154–139

142.0

155–129

–

–

–

–

–

–

–

11

139.3

152

169–136

126.6

139–115

–

–

–

–

–

–

–

–

–

12

133.9

–

–

–

–

–

–

133.9

197–77

–

–

–

195–142

228–111

13

127.3

142

157–113

118.0

125–111

122.0

137–107

–

–

–

–

–

–

–

14

121.9

138

157–121

119.1

131–107

–

–

–

–

108.5

154–76

–

–

–

15

113.0

–

–

–

–

–

–

113.0

161–65

–

–

–

–

–

16

109.5

–

–

110.0

–

109.0

123–95

–

–

–

–

–

–

–

17

105.0

114

129–106

102.0

100–98

99.0

111–87

–

–

–

–

–

–

–

18

95.7

–

–

–

–

–

–

95.7

116–75

–

–

–

–

–

19

85.5

–

–

99.0

99–99

72.0

80–64

–

–

–

–

–

–

–

20

83.5

–

–

–

–

80.0

90–70

–

–

–

–

87

–

–

21

68.6

–

–

–

–

54.0

61–47

68.6

97–40

–

–

–

–

–

22

68.5

–

–

68.5

76–61

–

–

–

–

–

–

–

–

–

23

65.0

–

–

65.0

65–65

–

–

–

–

–

–

–

–

–

24

60.3

–

–

60.3

71–50

–

–

–

–

–

–

–

–

–

25

49.3

46

59–35

48.6

57–41

–

–

–

–

53.2

74–38

–

–

–

26

40.3

–

–

–

–

–

–

–

–

40.3

58–27

–

–

–

27

38.6

32

43–23

32.9

39–27

–

–

–

–

51.0

69–37

–

–

–

Note: Node times in the timetree represent the mean of time estimates from different studies. Note that the gekkotan nodes in the timetree use only the time estimates from the comprehensive gecko study (25) to maintain the tree topology.

suggests that earliest divergences had already occurred by that time. Except for fragmentary remains in Africa and India, all of the Jurassic (200–146 Ma) fossils are from localities on northern continents (Laurasia). The paucity of Mesozoic fossils in general, and especially of those from Gondwana, has hindered biogeographic reconstructions and phylogenetic implications from the fossil record (9). Initial molecular phylogenies based on subsets of the different families, and those studies using mitochondrial DNA, have tended to give conflicting results (10–17). Phylogenies from mitochondrial DNA usually show

considerable rate variation among branches, particularly long-branch lengths in snakes, and they place snakes outside of all lizards and amphisbaenians (14, 16, 17). The first studies using a nuclear gene (C-mos) and comprehensive taxonomic coverage resolved the monophyly of most families that were examined but not interfamilial relationships (18, 19). In early 2004, the first study using multiple nuclear genes (C-mos and RAG-1) and broad taxonomic coverage (20) found statistical support that snakes are not the closest relatives of varanid lizards, as was generally

386


believed. Other groupings in the tree, although weakly supported, suggested that most of the classical phylogeny of squamates, based on morphology, was incorrect. For example, iguanians appeared in a highly nested position in the tree, together with anguimorph lizards and snakes. Also, amphisbaenians clustered with lacertid lizards, xantusiids clustered with scincids and cordylids, and dibamids appeared as the most basal living branch of the squamate tree. All of this implied major reversals and convergences in the key morphological characters used in squamate classification over the last century (4, 5). Later that year, a second study (21) provided additional support for this new phylogeny of squamates with longer sequences of RAG-1 and greater taxonomic coverage within lizard families. Some of the weakly supported nodes in the previous study now were significant with non-Bayesian methods, although the interrelationships of snakes, anguimorphs, iguanians, and lacertiforms could not be resolved. A new mitochondrial DNA data set of the ND2 gene for the same taxa showed some conflicting results, such as the unorthodox nesting of snakes within Iguania, probably a result of long-branch attraction (21). Subsequently, a study using nine nuclear genes provided further resolution of squamate phylogeny (22). The same groupings defined in the 2004 studies were bolstered, and additional groupings were discovered. A clade of three major squamate groups (iguanians, anguimorphs, and snakes) was defined, and further supported by the discovery of the ability for venom production in all three groups involving a suite of molecular and morphological characters (23). The limbless dibamids were found to be the most basal branch of living squamates, now with significant support. Because this new phylogeny was so different from the classical phylogeny (5), the previous classification based on tongue characters and feeding was abandoned and a new one erected. More recently, new families of amphisbaenians (24) and lizards (25) have been recognized based on molecular phylogenetic analyses, and families once separated from Iguanidae based on morphology are now either not recognized or considered as subfamilies of Iguanidae (26). New morphological characters were identified that were consistent with the molecular phylogeny and used in the new classification (22). Because all squamate families except Dibamidae have a bifurcated tongue, this large clade was named Bifurcata. The presence of one egg tooth (as opposed to two) defines the next most inclusive clade, Unidentata, which excludes dibamids and gekkotans. Scinciformata includes Scincidae, Xantusiidae, Gerrhosauridae, and Cordylidae. Teiformata includes

the Superfamily Teiioidea (Teiidae and Gymnophthalmidae). The venom clade (23) was named Toxicofera and the lacertid–amphisbaenian group was named Lacertibaenia. Lacertibaenians and teiformatans were grouped into Laterata, most of which have tile-like ventral scales. Toxicoferans and lateratans were grouped into Episquamata (“top squamates”). Several groups of families (Amphisbaenia, Iguania, Anguimorpha, Teiioidea, and Gekkota) agree with previous classifications. Within Iguania, the conventional grouping of chamaeleonids and agamids (Acrodonta) is supported (21). However, molecular phylogenetic studies in recent years have resolved more of squamate phylogeny than these large clades. Detailed relationships of families are now well supported (20–22, 24, 25, 27), leading to some recent adjustments in the taxonomy (28). For example, the monophyly of the previously defined Varanoidea (Varanidae, Lanthanotidae, and Helodermatidae) has not been supported by molecular evidence (20–22, 27, 29). Also, the monophyly of the Anguidae has been difficult to obtain because the anguid Subfamily Diploglossinae has a similar level of molecular divergence as the Family Anniellidae (21, 27, 30). Thus, the Anguidae was restricted to the Subfamilies Anguinae and Gerrhonotinae, and Diploglossidae was recognized as a family (28) as has been done in the past (e.g., 31). Two clades of anguimorph families are now defined in molecular analyses that correspond to geography (20–22, 27, 29). The first is a mostly New World (ancestrally North American) clade composed of Anguidae, Anniellidae, Diploglossidae, Helodermatidae, and Xenosauridae. The second is an Old World (ancestrally Asian) clade composed of Lanthanotidae, Shinisauridae, and Varanidae. These clades are so different from previous morphological groupings (e.g., Shinisaurus was usually placed in the Xenosauridae and helodermatids were usually associated with varanids and lanthanotids) that they were given new names: Neoanguimorpha for the New World clade and Paleanguimorpha for the Old World clade (28). Within the Neoanguimorpha, the superfamily Anguioidea was restricted to the three closely related families Anguidae, Anniellidae, and Diploglossidae, with the remaining families placed in their own superfamilies, Helodermatoidea (Helodermatidae) and one newly named, Xenosauroidea (Xenosauridae). Within the Paleoanguimorpha, the Superfamily Varanoidea was restricted to the two closely related families Lanthanotidae and Varanidae, and Shinisauridae was placed in its own superfamily, Shinisauroidea (28). Within geckos (Gekkota), two recent molecular phylogenetic studies (25, 32) have recognized seven families


and defined several well-supported clades, now recognized taxonomically (28): Eublepharoidea (Eublepharidae), Gekkonoidea (Gekkonidae, Phyllodactylidae, and Sphaerodactylidae), and Pygopodoidea (Carphodactylidae, Diplodactylidae, and Pygopodidae). The first two superfamilies were placed in the now redefined Gekkomorpha and the third in the taxon Pygopodomorpha (28). Amphisbaenian relationships are now well supported (24), and likewise their taxonomy has been adjusted (28), with recognition of the Superfamilies Amphisbaenoidea (Amphisbaenidae and Trogonophidae), Rhineuroidea (Rhineuridae), Bipedoidea (Bipedidae), and Blanoidea (Blanidae and Cadeidae). Rhineuroidea was placed in Rhineuriformata and the other three superfamilies in Amphisbaeniformata. The scinciformatan Families Cordylidae and Gerrhosauridae have always been found to be close relatives and were placed in the Superfamily Cordyloidea, and together with Xantusioidea (Xantusiidae) in the taxon Cordylomorpha. Scincomorpha was redefined to include only the Scincoidea with its single family, Scincidae. Finally, within Iguania, the Agamidae and Chamaeleonidae were placed in the Chamaeleonoidea, which in turn was placed in Acrodonta. Iguanidae was placed in Iguanoidea, which in turn was placed in the taxon Neoiguania (28). Few studies have estimated divergence times among squamate families. In a globin gene sequence study using absolute rates of change, the iguanian–varanid split was estimated as 139–86 million years ago (Ma) and the lizard–snake split as 161–92 Ma (33). In mitochondrial DNA studies, the divergence of scincids and iguanians was estimated as 167 Ma (34) and 158 Ma (13), and the split of agamids and chamaeleonids was estimated as 87 Ma (35). However, some of these estimates may have been influenced by known mitochondrial rate differences among squamates (e.g., the fast rate in snakes) and conflict with the squamate fossil record, particularly the earliest anguimorph fossil at 166 Ma (9), in the context of the new phylogeny. Only three studies have estimated divergence times among the major lineages of squamates in a comprehensive manner. The first study (22) used nine nuclear proteincoding genes and a Bayesian method (Fig. 2). The second study (27) used published nuclear RAG-1 sequences (21), and a penalized likelihood method (RAG-1 was one of the nine genes used in the first study). The third study also analyzed published RAG-1 sequences with a penalized likelihood method and five fixed calibrations (29). Yet another study used sequences from mitochondrial genomes and a Bayesian method of time estimation, although fewer taxa

387

were included and the time estimates themselves were not published (17). All studies used calibrations from the tetrapod and squamate fossil record. The relationships obtained in these studies were all similar in general, in that they supported the “new” squamate phylogeny and did not support the basal split between Iguania and Scleroglossa (the conventional morphological classification). However, because the mitochondrial study showed evidence of substantial rate variation among branches (especially snakes), the focus here is on the three studies using nuclear genes and broader taxonomic sampling. The time estimates in the RAG-1 studies tended to be younger than those in the nine-gene study of Vidal and Hedges (Table 1). For example, the basal squamate divergences (nodes 1–2) in the RAG-1 studies were estimated to be 190–178 Ma, considerably later than the nine (nuclear) gene estimate of 240–225 Ma and the mid-Triassic paleontological estimate (9). Hugall et al. (29) attributed the difference to the use of two calibration points out of five used by Vidal and Hedges (22) that they considered problematic (difficult to diagnose taxonomically because of limited material): the earliest anguimorph, Parviraptor (~166 Ma), and the earliest teiid, Ptilotodon (~112 Ma). More recently, Brandley et al. (36) went further and claimed that Vidal and Hedges (22) used incorrect calibrations, causing the difference in time estimates, citing Hugall et al. (29). However, Brandley et al. misinterpreted Hugall et al., because the difference is a matter of opinion, not of correctness. Other paleontologists considered those two fossils to be correctly assigned (9, 37). Nonetheless, these two particular fossils, whether they are correctly assigned or not, are unlikely to explain the difference in molecular time estimates. Wiens et al. (27) and Hugall et al. (29) both estimated the Ptilotodonconstrained node as Jurassic (200–146 Ma), much older than the Cretaceous fossil (112 Ma), and they estimated the Parviraptor-constrained node as 162–160 Ma, almost identical in age to that fossil (166 Ma). Moreover, there are other, uncontested, Middle and Late Jurassic fossils of anguimorphs (9) that would similarly constrain that node if Parviraptor were not used. The difference in molecular time estimates among those studies may be from the use of smaller data sets and different methods (penalized likelihood rate smoothing) used by Wiens et al. (27) and Hugall et al. (29) compared with the larger data set (more genes and sites) and Bayesian method used by Vidal and Hedges (22), or from differences in other calibration points. In addition to those comprehensive studies, three additional timing studies have appeared recently that have

388


focused on smaller clades of squamate families. One study analyzed divergence times among the six families of amphisbaenians using two mitochondrial genes and Bayesian methods (24). They found that five of the six (all except Rhineuridae) were considerably young, having arisen only in the Cenozoic. The other study used five nuclear genes and nonparametric rate smoothing in a diverse sample of geckos, finding deep (Mesozoic) divergences among most of the six families (25). Another gecko study focused mainly on eublepharids but also presented some time estimates among families of squamates, using mitochondrial DNA sequences and a Bayesian method of time estimation (38). They generally found a great range in divergence times for nodes, with most times being older than those estimated in the other studies. The timetree of squamates (Fig. 2) represents a synthesis of these various molecular studies, although emphasizing the three comprehensive studies using nuclear genes. It shows that most of the major splits in the tree occurred during the Jurassic and Cretaceous, 200–66 million years ago. The earliest of those divergences took place when all of the continents were joined in a single supercontinent, Pangaea. These included the divergence of dibamids and bifurcatans, gekkotans and unidentatans, and scinciformatans and episquamatans. Therefore a strong geographic influence in the ancestral distributions of these groups is not expected. As noted, the fossil record is essentially silent on the early biogeographic history of squamates (9). Pangaea broke into Laurasia and Gondwana in the Jurassic, ~170–150 Ma (39, 40). Considering the timetree and confidence intervals (Table 1), a large number of squamate lineages may have split at this time, including the earliest divergences among scinciformatans (scincids, cordyloids, and xantusiids), toxicoferans (snakes, anguimorphs, and iguanians), lateratans (teiioids, lacertids, and amphisbaenians), iguanians (iguanoids and chamaeleonoids), and gekkotans (gekkomorphs and pygopodomorphs). The timetree shows early Cretaceous (146–100 Ma) divergences among several gecko families based on the study of Gamble et al. (25). Concerning the earliest split, between gekkomorphs and pygopodomorphs (146 Ma), two other nuclear gene studies (22, 29) found much younger dates (111–94 Ma) while a recent mitochondrial gene study (38) found older dates (196– 163). Thus there remains considerable uncertainty in the timescale of the gekkotan portion of the timetree. The two major clades of anguimorphs, Neoanguimorpha and Paleoanguimorpha, also diverged in the early Cretaceous (Fig. 2), probably related to the continuing separation of

the continents and rising sea levels, creating (in some cases) inland seas. Divergences among anguioid and varanoid families apparently occurred near the Mesozoic– Cenozoic boundary (66 Ma). The late Cretaceous divergence (86 Ma) between Teiidae and Gymnophthalmidae, two primarily South American families, probably occurred in South America where both groups are distributed, although it is not clear how they came to inhabit that continent. Also, if they were widely distributed on the Africa–South America supercontinent at 105 Ma, it is not clear what became of the teiioids that presumably inhabited Africa after it split from South America (no fossils or living representatives have been discovered in Africa). The divergence of Chamaeleonidae and Agamidae (84 Ma) postdates the breakup of Gondwana and supports oceanic dispersal as a mechanism to explain the origin of chameleons on Madagascar (41). Likewise, continental breakup is unlikely to explain any of the remaining divergences among lizard families, all in the Cenozoic (Fig. 2). According to squamate timetree, one-third (10) of all families of lizards and amphisbaenians diverged within a few million years of the Mesozoic/Cenozoic boundary, including all three pygopodomorph families, the cordyloids, the varanoids, and the anguioids. This suggests a possible relationship with the asteroid impact at 66 Ma and the resulting extinctions and ecological changes, although this was also a time of major global sea level change and increasing connections among continents (39, 40, 42). In the case of the amphisbaenians, at least one transatlantic dispersal event in the Cenozoic explains the origin of New World amphisbaenids, representing one-half of all known species of amphisbaenians (24). Despite the coincidence of early splitting events with the breakup of continents in the Mesozoic, there is yet no clear distributional and fossil evidence to support vicariance as a major mechanism in the early evolution of squamates. Nonetheless, the Laurasian distribution of anguimorphs and the possibly Gondwanan distribution of snakes and iguanians, in the Mesozoic, is intriguing and may reflect early vicariant events (22). More Mesozoic fossil material from the southern continents is needed, along with a better resolved tree of toxicoferans.

Acknowledgments We thank M. P. Heinicke for comments on an earlier version of the manuscript. Support was provided by the U.S. National Science Foundation and National Aeronautics and Space Administration (NASA Astrobiology Institute)


to S.B.H. and by the Service de Systématique Moléculaire du Muséum National d’Histoire Naturelle to N.V.

References 1. N. Vidal, A.-S. Delmas, S. B. Hedges, in Biology of the Boas and Pythons, R. W. Henderson, R. Powell, Eds. (Eagle Mountain Publishing, Eagle Mountain, Utah, 2007), pp. 27–33. 2. N. Vidal et al., C. R. Biologies 330, 182 (2007). 3. N. Vidal, J.-C. Rage, A. Couloux, S. B. Hedges, in The Timetree of Life, S. B. Hedges, S. Kumar, Eds. (Oxford University Press, New York, 2009), pp. 390–397. 4. C. L. Camp, Bull. Am. Mus. Nat. Hist. 48, 289 (1923). 5. R. Estes, K. de Queiroz, J. A. Gauthier, in Phylogenetic Relationships of the Lizard Families, Essays Commemorating Charles L. Camp, R. Estes, G. Pregill, Eds. (Stanford University Press, Stanford, 1988), pp. 119–281. 6. O. Rieppel, Zool. J. Linn. Soc. 152, 131 (2008). 7. E. R. Pianka, L. J. Vitt, Lizards: Windows to the Evolution of Diversity (University of California Press, Berkeley, 2003). 8. F. H. Pough et al., Herpetology, 3rd ed. (Prentice Hall, Upper Saddle River, New Jersey, 2003). 9. S. E. Evans, Biol. Rev. (Camb.) 78, 513 (2003). 10. M. R. J. Forstner, S. K. Davis, E. Arévalo, Mol. Phylogenet. Evol. 4, 93 (1995). 11. J. R. Macey, A. Verma, Mol. Phylogenet. Evol. 7, 272 (1997). 12. K. M. Saint, C. C. Austin, S. C. Donnellan, M. N. Hutchinson, Mol. Phylogenet. Evol. 10, 259 (1998). 13. J. S. Rest et al., Mol. Phylogenet. Evol. 29, 289 (2003). 14. Y. Kumazawa, DNA Res. 11, 137 (2004). 15. A. S. Whiting, A. M. Bauer, J. W. Sites, Mol. Phylogenet. Evol. 29, 582 (2003). 16. K. Zhou, H. Li, A. M. Bauer, J. Feng, Mol. Phylogenet. Evol. 40, 887 (2006). 17. Y. Kumazawa, Gene 388, 19 (2007). 18. D. J. Harris, J. C. Marshall, K. A. Crandall, Amphibia– Reptilia 22, 235 (2001). 19. D. J. Harris, Mol. Phylogenet. Evol. 27, 540 (2003).

389

20. N. Vidal, S. B. Hedges, Proc. Roy. Soc. Lond. B 271, S226 (2004). 21. T. M. Townsend, A. Larson, E. Louis, J. R. Macey, Syst. Biol. 53, 735 (2004). 22. N. Vidal, S. B. Hedges, C. R. Biologies 328, 1000 (2005). 23. B. G. Fry et al., Nature 439, 584 (2006). 24. N. Vidal, A. Azvolinsky, C. Cruaud, S. B. Hedges, Biol. Lett. 4, 115 (2008). 25. T. Gamble, A. M. Bauer, E. Greenbaum, T. R. Jackman, J. Biogeogr. 35, 88 (2008). 26. P. Uetz, The Reptile Database, http://www.reptiledatabase.org/ (Research Center Karlsruhe, Karlsruhe, Germany, 2008). 27. J. J. Wiens, M. C. Brandley, T. W. Reeder, Evolution 60, 123 (2006). 28. N. Vidal, S. B. Hedges, C. R. Biologies in press (2009). 29. A. F. Hugall, R. Foster, M. S. Y. Lee, Syst. Biol. 56, 543 (2007). 30. J. R. Macey et al., Mol. Phylogenet. Evol. 12, 250 (1999). 31. A. E. Greer, J. Herpetol. 25, 166 (1991). 32. T. Gamble, A. M. Bauer, E. Greenbaum, T. R. Jackman, Zool. Scripta 37, 355 (2008). 33. T. Gorr, B. K. Mable, T. Kleinschmidt, J. Mol. Evol. 47, 471 (1998). 34. A. Janke, D. Erpenbeck, M. Nilsson, U. Arnason, Proc. Roy. Soc. Lond. B 268, 623 (2001). 35. S. A. Amer, Y. Kumazawa, Gene 346, 249 (2005). 36. M. C. Brandley, J. P. Huelsenbeck, J. J. Wiens, Evolution 62, 2042 (2008). 37. R. L. Nydam, R. L. Cifelli, J. Vertebr. Paleontol. 22, 286 (2002). 38. P. Jonniaux, Y. Kumazawa, Gene (Amst.) 407, 105 (2008). 39. F. M. Gradstein, J. G. Ogg, A. G. Smith, Eds., A Geologic Timescale 2004 (Cambridge University Press, Cambridge, 2005), pp. 610. 40. A. G. Smith, D. G. Smith, B. M. Funnell, Atlas of Mesozoic and Cenozoic Coastlines (Cambridge University Press, Cambridge, 1994). 41. C. J. Raxworthy, M. R. J. Forstner, R. A. Nussbaum, Nature 415, 784 (2002). 42. S. B. Hedges, P. H. Parker, C. G. Sibley, S. Kumar, Nature 381, 226 (1996).

Snakes (Serpentes) Nicolas Vidala,b,*, Jean-Claude Ragec, Arnaud Coulouxd, and S. Blair Hedgesb a

UMR 7138, Systématique, Evolution, Adaptation, Département Systématique et Evolution, C. P. 26, Muséum National d’Histoire Naturelle, 43 Rue Cuvier, Paris 75005, France; bDepartment of Biology, 208 Mueller Laboratory, Pennsylvania State University, University Park, PA 16802-5301, USA; cUMR 5143, Paléobiodiversité & Paléoenvironnements, Département Histoire de la Terre, C. P. 38, Muséum National d’Histoire Naturelle, 8 rue Buffon, Paris 75005, France; dCentre national de séquençage, Genoscope, 2 rue GastonCrémieux, CP5706, 91057 Evry cedex, France *To whom correspondence should be addressed ([email protected])

Abstract Snakes have a Gondwanan origin and their early evolution occurred mainly on West Gondwana, the supercontinent comprising South America and Africa. New data from nine genes indicate that the divergence of Amerophidia and Afrophidia occurred 106 (116–97) million years ago (Ma), supporting their origin by continental breakup. Most (~85%) living snakes are afrophidians and are globally distributed now, but their initial radiation can be explained by dispersal out of Africa through Laurasia or India. Most basal afrophidian families (Henophidia) diverged in the Cretaceous, 104–70 Ma, while most advanced afrophidian families (Caenophidia), diverged in the early Cenozoic, 63–33 Ma.

Snakes are among the most successful groups of reptiles, numbering about 3070 extant species (1). They are divided into two main groups. The fossorial scolecophidians (~370 sp.) are small snakes with a limited gape size and feed on small prey (mainly ants and termites) on a frequent basis. The alethinophidians, or typical snakes (~2700 sp.), are more ecologically diverse and most species feed on relatively large prey, primarily vertebrates, on an infrequent basis (2, 3). According to most morphological studies, a distinctive evolutionary trend within living snakes is the increase of the gape size from fossorial scolecophidians (Typhlopidae, Leptotyphlopidae, and Anomalepididae) and fossorial alethinophidians (Aniliidae, Cylindrophiidae, Uropeltidae, and Anomochilidae) to ecologically diverse macrostomatan alethinophidian

snakes such as boas, pythons, and caenophidians (advanced snakes) (2, but see 4). Macrostomatans are able to ingest very large prey, often greater in diameter than the snake itself (5), and the monophyly of the macrostomatan condition is supported by several unambiguous shared-derived characters (6). All venomous snakes are found within Caenophidia, which includes the great majority of extant snakes (~2500 sp.) (1). Previously, caenophidians were thought to comprise five families: the aquatic acrochordids, the atractaspidids (now a subfamily; some of them with a frontfanged venom system), the elapids, and the viperids (all of them with a front-fanged venom system), and the large and paraphyletic family Colubridae (now split into eight families), which includes rear-fanged snakes and the vast majority of caenophidians (~1900 sp.) (7–12). Here, the relationships and fossil record of snakes are reviewed and new data from nine nuclear protein-coding genes are analyzed, resulting in a timetree of snake families with new biogeographic implications. Several higher-level snake phylogenies using nuclear genes, including some that incorporated mitochondrial genes, have been published since 2002 (13–21). They

Fig. 1 Typhlops arator from Cuba, Typhlopidae (upper left); Rhinocheilus lecontei from southwestern United States, Colubridae (upper right); Cryptelytrops albolabris, from southeastern Asia, Viperidae (lower left); and Tropidophis feicki from Cuba, Tropidophiidae (lower right). Credits: S. B. Hedges.

N. Vidal, J.-C. Rage, A. Couloux, and S.B. Hedges. Snakes (Serpentes). Pp. 390–397 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Sauropsida; Squamata; Serpentes

391

Dipsadidae 21

Pseudoxenodontidae

20

Colubridae

19 16

Natricidae Elapidae Lamprophiidae

13

Viperidae Pareatidae

11

Xenodermatidae

8

Acrochordidae

Alethinophidia

Homalopsidae

14

Afrophidia

15

18

Pythonidae 17

5

Loxocemidae

12

Xenopeltidae

10

Boidae

7

Bolyeriidae 2

Tropidophiidae 9

Aniliidae

1

Anomalepididae Typhlopidae 3

J

Leptotyphlopidae

Early K

Late K

MESOZOIC 150

100

Paleogene

Scolecophidia

Uropeltidae

6

Amerophidia

4

Ng

CENOZOIC 50

0 Million years ago

Fig. 2 A timetree of snakes. Divergence times are shown in Table 1. Abbreviations: J (Jurassic), Ng (Neogene), and K (Cretaceous).

all agree on the monophyly of alethinophidians, but a striking result is the paraphyly of the macrostomatan condition. The fossorial small-gaped Aniliidae (South American genus Anilius) and the terrestrial large-gaped (macrostomatan) Tropidophiidae (Neotropical genera Trachyboa and Tropidophis) cluster together, and form the most basal alethinophidian lineage (13, 16, 17, 19, 20). The genus Anilius is therefore not closely related to the Asian families formerly placed in “Anilioidea.” We propose that Uropeltoidea Müller be used to describe the monophyletic group (22) that includes Cylindrophiidae, Uropeltidae, and Anomochilidae. Also, we provisionally use the taxon Henophidia Hoffstetter to describe all noncaenophidian Afrophidia, which usually form a monophyletic group in molecular phylogenetic analyses.

The alethinophidians were therefore primitively macrostomatan, and this condition was secondarily lost twice by Aniliidae and Uropeltoidea, in connection with burrowing (13, 17, 20). From a biogeographic point of view, the deep split between the Aniliidae– Tropidophiidae clade, which is of South American origin, and all remaining alethinophidians was recently hypothesized to represent a vicariant event: the separation of South America from Africa in the midCretaceous. Accordingly, those two clades were named Amerophidia and Afrophidia (20). Among alethinophidians, the monophyly of the group including the Pythonidae, Xenopeltidae, and Loxocemidae is found in most molecular studies (13, 15–17, 20), with Loxocemidae as the closest relative to Pythonidae. Another large group

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among snakes (Serpentes). Timetree Node

Time

Estimates Ref. (19)

This study Time

CI

Time

Ref. (56)

Ref. (58)

CI

Time

CI

Ref. (59)

Ref. (60)

Ref. (61)

Ref. (62)

Time

Time

CI

Time

CI

Time

Time

CI

1

159.9

159.9

166–148

–

–

102.3

113–94

–

131.1

138–124

109.0

119–99

–

144.2

–

2

155.6

155.6

164–144

–

–

–

–

–

–

–

–

–

–

–

–

3

151.9

151.9

163–137

–

–

–

–

–

–

–

–

–

–

109.3

–

4

105.8

105.8

116–97

121

129–106

62.1

75–49

–

–

–

50.0

56–44

–

76.1

87–69

5

103.7

103.7

114–95

112

119–99

–

–

–

93.5

–

–

–

–

–

–

6

96.9

96.9

108–87

–

–

–

–

–

–

–

–

–

–

–

–

7

92.0

92.0

102–82

–

–

–

–

–

–

–

–

–

–

–

–

8

90.7

90.7

104–78

–

–

54.9

66–44

–

–

–

–

–

–

58.7

71–54

9

89.1

89.1

100–78

110

123–93

–

–

–

–

–

–

–

–

63.1

–

10

86.3

86.3

96–77

–

–

–

–

–

–

–

–

–

–

–

–

11

82.2

82.2

96–69

–

–

–

–

–

–

–

–

–

–

48.7

–

12

70.1

70.1

81–59

–

–

–

–

–

–

–

–

–

–

51.3

–

13

64.0

64.0

77–52

68

82–53

–

–

–

35.6

–

–

–

–

–

–

14

54.3

54.3

67–43

–

–

38.5

45–33

46.0

–

–

–

–

74.3

–

–

15

49.2

49.2

61–39

–

–

–

–

–

–

–

–

–

69.4

–

–

16

46.3

46.3

58–36

–

–

–

–

42.5

–

–

–

–

–

–

–

17

43.7

43.7

56–33

–

–

–

–

–

–

–

–

–

–

37.1

–

18

41.5

41.5

53–32

–

–

–

–

40.5

–

–

–

–

62.9

34.0

–

19

39.8

39.8

50–31

–

–

–

–

–

–

–

–

–

61.2

38.2

–

20

36.6

36.6

46–28

–

–

–

–

–

–

–

–

–

–

–

–

21

32.9

32.9

43–25

–

–

–

–

–

–

–

–

–

–

–

–

Note: Node times in the timetree are from the new analyses presented here. Other published estimates are also shown for comparison.


includes Calabaria, “boines,” “erycines,” and ungaliophiines (genera Ungaliophis and Exiliboa), with North American erycines and ungaliophiines as closest relatives (13, 17, 19, 20). Unfortunately, several higher-level henophidian relationships are still unresolved (20), a situation contrasting with our better state of knowledge of the interfamilial relationships among caenophidian snakes. As recently as 2007, a study using seven nuclear protein-coding genes (C-mos, RAG1, RAG2, R35, HOXA13, JUN, and AMEL) resolved with strong support the relationships of all families of caenophidians (21). Caenophidians devoid of a front-fanged venom system were traditionally lumped into a large (~1900 sp.) family, “Colubridae,” including several subfamilies. Because this family was shown to be paraphyletic, most of the subfamilies were elevated to a familial rank to reflect their evolutionary distinctiveness, and the name Colubridae was restricted to a less inclusive monophyletic group (21). The caenophidian venom apparatus has experienced extensive evolutionary tinkering throughout its history. All traits, ranging from biochemical (specialization of the venoms) to dentition and glandular morphology, have changed independently, resulting in many kinds of toxins and diverse delivery systems (12, 14, 23). Rearfanged—or more correctly defined, non-front-fanged— caenophidians possess complex venoms containing multiple toxin types, while the front-fanged venom system appeared three times independently: once early in caenophidian evolution with viperids, once within atractaspidines (a lamprophiid subfamily), and once with elapids. Further, a reduction of the venom system is observed in species in which constriction has been secondarily evolved as the preferred method of prey capture or dietary preference has switched from live prey to eggs or to slugs and snails (12, 14, 23). Until now, the most comprehensive study to estimate divergence times among alethinophidian families used five nuclear genes (C-mos, RAG1, BDNF, NT3, ODC) and one mitochondrial gene (cyt b), and a Bayesian method (19). It showed that most interfamilial splits among alethinophidians occurred within the span of 25 million years in the early Cretaceous, 121–98 Ma, suggesting a radiation. Also, it suggested that dispersal and vicariant events associated with the fragmentation of the Gondwanan supercontinent have shaped the global distribution of alethinophidians. In that study (19), a scolecophidian was used as outgroup and the earliest snake divergences were therefore not dated. Also, one

393

caenophidian exemplar (Acrochordidae) was used and interfamilial caenophidian splits were not dated. Divergence times among all major groups of snakes are estimated here using nine nuclear protein-coding genes (C-mos, RAG1, RAG2, R35, HOXA13, BDNF, JUN, AMEL, and NT3). These were sequenced in 49 snake taxa representing all families with the exception of the Xenophidiidae, Anomochilidae, and Cylindrophiidae (Alethinophidia). Tissue samples were obtained from the tissue collections of N. V. and S. B. H. (see 13, 14, 16, 20, 24, 25 for details of the samples used). The taxa included Iguanidae: Cyclura, Helodermatidae: Heloderma, Anomalepididae: Liotyphlops, Typhlopidae: Ramphotyphlops, Typhlops, Leptotyphlopidae: Leptotyphlops, Aniliidae: Anilius, Tropidophiidae: Tropidophis, Trachyboa, Uropeltidae: Rhinophis, Uropeltis, Bolyeriidae: Casarea, Loxocemidae: Loxocemus, Xenopeltidae: Xenopeltis, Pythonidae: Python, Liasis, Apodora, Boidae: Calabaria, Boa, Acrantophis, Candoia, Eryx, Gongylophis, Ungaliophis, Charina, Lichanura, Acrochordidae: Acrochordus, Xenodermatidae: Stoliczkaia, Pareatidae: Aplopeltura, Pareas, Viperidae: Bothriechis, Homalopsidae: Homalopsis, Dipsadidae: Leptodeira, Alsophis, Diadophis, Colubridae: Phyllorhynchus, Hapsidophrys, Calamaria, Grayia, Pseudoxenodontidae: Pseudoxenodon, Natricidae: Xenochrophis, Elapidae: Elapsoidea, Laticauda, Bungarus, Dendroaspis, Micrurus, and Lamprophiidae: Psammophylax, Leioheterodon, Lamprophis, Mehelya, Atractaspis. DNA extraction was performed using the DNeasy Tissue Kit (Qiagen). Amplification and sequencing was performed using sets of primers already described (13, 19, 25). The two strands obtained for each sequence were aligned using the BioEdit Sequence Alignment Editor program (26). The sequences produced for this work have been deposited in GenBank under Accession Numbers FJ433886-FJ434106. Sequence entry and alignment (51 taxa) were performed manually with the MUST2000 soft ware (27). Amino acid properties were used, and ambiguous gaps deleted. This resulted in 561 bp for C-mos, 510 bp for RAG1, 708 bp for RAG2, 708 bp for R35, 408 bp for HOXA13, 669 bp for BDNF, 330 bp for the JUN gene, 378 bp for AMEL, and 519 bp for NT3. In all analyses, remaining gaps were treated as missing data. Phylogenies were constructed using probabilistic approaches, with maximum likelihood (ML) and Bayesian methods of inference. ML analyses were performed with PAUP*4 (28). Bayesian analyses were performed with MrBayes 3.1 (29, 30). For ML methods, an appropriate model of sequence evolution was inferred

394


using ModelTest (31), for both separate and combined analyses. As we used only protein-coding nuclear genes, and because separate analyses showed no significant topological incongruence, we performed combined analyses, which are considered to be our best estimates of phylogeny. For the concatenated data set (4791 sites), the model selected was the TVM+I+G model. For the combined ML analysis, we used heuristic searches, with starting trees obtained by random addition with 100 replicates and nearest-neighbor interchange (NNI) branch swapping. For the bootstrap ML analysis, we performed 1000 replicates (NJ starting tree with NNI branch swapping). Bayesian combined analyses were run with model parameters estimated as part of the Bayesian analyses, with nine partitions corresponding to each gene (GTR model). Bayesian analyses were performed by running 2,000,000 generations in four chains, saving the current tree every 100 generations. The last 18,000 trees were used to construct a 50% majority-rule consensus tree. The choice of calibration points is a crucial step in dating analyses, and we therefore present a brief overview of the snake fossil record. Geologic times and boundaries of periods used here are from a recent update (32). Three localities, or group of localities, may be putatively the oldest snake-bearing site(s): Emery, Utah (Coniophis sp.) (33), In Akhamil, Algeria (Lapparentophis defrennei) (34), and El Kohol, Algeria (an indeterminate lapparentophiid-grade snake and a Serpentes incertae sedis) (35). They apparently all fall in the Albian–Cenomanian interval (112–94 Ma) (5), but an older age (Aptian; 125–112 Ma) cannot be ruled out for In Akhamil (P. Taquet, personal communication). Rage and Richter (36) reported a putative snake from the Barremian (early Cretaceous; 130–125 Ma) of Spain, but it is quite likely a lizard (5). Noonan and Chippindale (37) regarded Dinilysia as the earliest representative of the “Booidea,” but there is no consensus about its phylogenetic relationships. The oldest scolecophidian is from the Paleocene of Hainin, Belgium (early Selandian; 62–59 Ma) (38). However, the fossil record of scolecophidians is poor, which likely results from their small size and fragility of their bones. The oldest alethinophidians are an “acrochordoid” (Nubianophis afaahus, Nigerophiidae), a russellophiid (Krebsophis thobanus), and a caenophidian incertae sedis from Wadi Abu Ashim, Sudan (39), a locality that is regarded as Cenomanian (100–94 Ma). None of the known fossil snakes can be reliably assigned to the Aniliidae (restricted here to Anilius). The extinct Coniophis was referred to the Aniliidae, or to the Uropeltoidea, but its monophyly is doubtful and its phylogenetic position

is unknown. No fossil may be confidently assigned to the following lineages: Tropidophiidae, Uropeltoidea, Xenopeltidae, Loxocemidae, and Xenophidiidae. Concerning Bolyeriidae, only a subfossil is known. The earliest pythonid was reported from the late early Miocene of Europe (40). In Australia, Morelia riversleighensis is present at Riversleigh in levels that may be either late Oligocene or early or middle Miocene (41). Older pythons may also be present in a middle Eocene locality of Europe; but this cannot be confi rmed (42). The earliest Boidae are from the mid-Paleocene (62–59 Ma) of Itaboraí, Brazil (43). These boids are represented by the earliest “boines” (including the extant genus Corallus) and the earliest ungaliophiine. The locality comprises several fissure fi llings; therefore it is difficult to correlate the locality with international stratigraphic charts based on marine beds, but it may be regarded as late Selandian. The fossil genus Helagras has been regarded as an “erycine,” but it cannot be assigned to a taxon within the “Booidea” (40). The earliest known member of the North American erycine clade (Charina/ Lichanura) is Charina prebottae from Wyoming (Aquitanian, 23–20 Ma) (44). The oldest caenophidian fossils are mentioned above under the heading Alethinophidia. They are an “acrochordoid” (N. afaahus, Nigerophiidae), a russellophiid (K. thobanus), and a caenophidian incertae sedis. They come from Wadi Abu Ashim (Cenomanian). The oldest acrochordid is Acrochordus recovered from southern Asia (Aquitanian, 23–20 Ma) (45, 46). No fossil may be assigned to the following families: Xenodermatidae, Pareatidae, Homalopsidae, Pseudoxenodontidae, and Lamprophiidae. The earliest Viperidae are from Germany (earliest Aquitanian, 23–20 Ma) (47). In its present understanding, no fossil may be assigned to the Family Colubridae with certainty. Various fossils were assigned to the genus Coluber, including fossils from the Oligocene. But the referral to Coluber is only symbolic because it is not possible to distinguish this genus from several other genera (that are perhaps not all Colubridae) on the basis of the available material (vertebrae). The oldest Dipsadidae would be Paleoheterodon arcuatus from Sansan, France, implying a dispersal from the New World (early Serravallian, 14–12 Ma) (48). The earliest natricid is Natrix mlynarskii from the early Oligocene (Rupelian, 34–28 Ma) of France (49). The oldest ascertained elapids come from Spain and France (late Burdigalian, 20–16 Ma) (50). However, in Australia, an elapid (close to the hydrophiine Laticauda) was recorded from RSO Site of Godthelp Hill, whose age may be either


latest Oligocene or more probably early Miocene (51). This snake may therefore be the earliest elapid, but this cannot be confirmed. Bayesian timing analyses were conducted with Multidivtime T3 (52, 53). The assumed topology was from the ML analysis, with Heloderma used as outgroup. PAML 3.14 (54) was used to estimate model parameters. Multidivtime requires prior estimates for rttm, rttmsd, bigtime, rtrate, rtratesd, brownmean, and brownsd. We followed recommendations accompanying the soft ware and adjusted the last four priors based on the rttm setting. The prior for the rttm (ingroup root) parameter, which is not a calibration point and does not have a major affect on posteriors, was set at 100 Ma (oldest fossil snake), 166 Ma (oldest anguimorph, ref. 55), and 130 Ma (intermediate). The three rttm resulted in less than 1% difference in time estimates, so the intermediate rttm was used in the primary (“best”) analysis. The prior rttmsd was set at one-half of rttm based on recommendations accompanying the soft ware. Analyses were performed treating the nine-gene data set as one partition and as nine partitions. The average deviation between the unpartitioned and the partitioned analyses is −0.06 Ma, and the partitioned analysis was chosen as our primary analysis. The prior bigtime (a value larger than an expected posterior), which is not a calibration point and has little affect on posteriors, was set at 200 Ma (Triassic–Jurassic boundary). Analyses were run for 1,100,000 generations, with a sample frequency of 100 after a burnin of 100,000 generations. The fossil calibration points used here as minimum dates are the oldest elapid (20.4 Ma), the oldest natricid (28.4 Ma), the oldest Charina (20.4 Ma), the oldest ungaliophiine (58.7 Ma), the oldest pythonid (20.4 Ma), and the oldest caenophidian (93.5 Ma). As the use of the latter calibration has been discussed by Sanders and Lee (56), we performed analyses with and without it. The oldest anguimorph (166 Ma) was used as a maximum date for the snake node. We performed analyses using one additional geological calibration point. Because there is no evidence for continuous emergent land in the Antilles before the late Eocene (57), we assigned this date (37.2 Ma) as a maximum constraint for the split between Trachyboa and Tropidophis (maximum divergence times among species of West Indian Tropidophis are similar to the divergence time of Tropidophis and Trachyboa; S. B. H., unpublished data). To examine the effect of the geologic calibration, we performed analyses with and without it. In all analyses, the posterior times obtained for 18 out of the 21 nodes

395

discussed here fell within the credibility intervals derived from the primary partitioned analysis using all eight original calibrations. The three exceptions are Pythonidae vs. Loxocemidae node (extreme value: 62.2 Ma instead of 43.7 Ma in the primary analysis), the Xenopeltidae vs. Pythonidae/Loxocemidae node (extreme value: 85.9 Ma instead of 70.1 Ma), and the Boidae vs. Xenopeltidae/ Loxocemidae/Pythonidae node (extreme value: 97.5 Ma instead of 86.3 Ma). In any case, these differences do not alter the following results and discussion that are based on the analysis performed with all eight calibration points and rttm set at 130 Ma (Fig. 2). As noted, until now, the only study having estimated divergence times among snake families using several nuclear genes is by Noonan and Chippindale (19). Other studies have used one or two nuclear genes or mitochondrial genes (56, 58–62), and those reported time estimates are presented in Table 1, for comparison. Our ML and Bayesian topologies are virtually identical, differing only in the position of Bolyeriidae and Uropeltidae (in the Bayesian tree, Bolyeriidae and Uropeltidae cluster together and form the sister group to Caenophidia). Whatever the method used, these positions are not supported statistically, and we consider them to be unresolved. Similarly, the paraphyly of Scolecophidia is weakly supported (ML BP: 53%, Bayesian PP: 56%), and we conservatively follow the strong morphological evidence available and consider scolecophidians to be monophyletic (63). The remaining interfamilial relationships confirm previously obtained results (20, 21). The timetree of snakes supports a Gondwanan origin for the group, based on the distribution of the basal lineages (Scolecophidia, Aniliidae, Tropidophiidae, Boidae, Bolyeriidae, and Uropeltoidea, whether the last two lineages are basal to henophidians or to caenophidians). According to the same data, snakes most probably evolved on West Gondwana (South America and Africa), which drifted from East Gondwana from 166 to 116 Ma (64). The earliest divergences among living lineages occurred in the late Jurassic between 152 (163–137) Ma and 156 (164–144) Ma. Among toxicoferans, the relative positions of snakes, anguimorphs, and iguanians are still unresolved, but if the traditional clustering of snakes with anguimorphs (that are of Laurasian origin) is confirmed, it would mean that the Jurassic split (166 Ma) may correspond to the breakup of Pangaea (25). Another major result is the split between the group formed by Aniliidae and Tropidopiidae and all remaining Alethinophidia that is estimated here as 106

396


(97–116) Ma. This date corresponds to the opening of the Atlantic Ocean, and supports the inference that this deep alethinophidian split is a vicariant event (20). Furthermore, this result is in agreement with the fossil record, because the oldest known diverse snake fauna is from Africa (Sudan) (39). Among Henophidia (Afrophidia to the exclusion of Caenophidia), all interfamilial splits except one (the divergence between Pythonidae and Loxocemidae) took place in the Cretaceous between 104 (95–114) Ma and 70 (59–81) Ma. Those dates are similar to those of Noonan and Chippindale (112–98 Ma) (19) and suggest that most interfamilial splits among noncaenophidian alethinophidians (Amerophidia and Henophidia) occurred in the middle to late Cretaceous. Among Caenophidia, all interfamilial splits except the two most basal ones occurred during the Paleogene between 63 (52–77) Ma and 33 (25–43) Ma. In contrast, three recent molecular clock analyses using one or two genes obtained much younger time estimates for most divergences (56, 60, 62). However, the two studies using several nuclear genes, ours and the one of Noonan and Chippindale (19) have similar, older estimates that are probably more reliable. Geological and paleobiogeographical data show that the isolation of Africa was broken intermittently during the Cretaceous by contact with Laurasia. Therefore, the initial radiation and dispersal of Afrophidia can be explained by dispersal out of Africa through Laurasia or India or both (64, 65). In turn, the early biogeographic history of Caenophidia is firmly rooted in Asia based on the successive branching, in a ladder-like fashion (basal to derived) of these Asian or mostly Asian families: acrochordids, xenodermatids, pareatids, viperids (partly Asian), and homalopsids (21). Among Henophidia, the relationships between Bolyeriidae, Uropeltoidea, Boidae, and the Xenopeltidae/Loxocemidae/Pythonidae clade are still not well resolved. Thus, their biogeography is more difficult to interpret and probably involves both dispersal and vicariant events (19).

Acknowledgments Assistance with taxonomic issues was provided by P. David and with paleontological issues by L. P. Bergqvist. DNA samples from Pseudoxenodon bambusicola and Calamaria pavimentata were from R. Lawson. Support was provided by the Service de Systématique moléculaire du Muséum National d’Histoire Naturelle to N.V., by the U.S. National Science Foundation and

National Aeronautics and Space Administration (NASA Astrobiology Institute) to S.B.H., and by the Consortium National de Recherche en Génomique, Genoscope.

References 1. P. Uetz, The Reptile Database, http://www.reptiledatabase.org (Research Center Karlsruhe, Karlsruhe, Germany, 2008). 2. D. Cundall, H. W. Greene, in Feeding, Form, Function, and Evolution in Tetrapod Vertebrates, K. Schwenk, Ed. (Academic Press, San Diego, 2000), pp. 293–333. 3. H. W. Greene, Snakes: The Evolution of Mystery in Nature (University California Press, Berkeley, California, 1997). 4. M. S. Y. Lee, J. D. Scanlon, Biol. Rev. 77, 333 (2002). 5. J.-C. Rage, F. Escuillié, Carnets de Géologie/Notebooks on Geology 2003/01, 1 (2003). 6. O. Rieppel, A. G. Kluge, H. Zaher, J. Vert. Paleontol. 22, 812 (2002). 7. J. E. Cadle, in Snakes: Ecology and Evolutionary Biology, R. A. Seigel, J. T. Collins, S. S. Novak, Eds. (Macmillan Publication, New York, 1987), pp. 77–105. 8. J. E. Cadle, Zool. J. Linn. Soc. 110, 103 (1994). 9. S. B. McDowell, in Snakes: Ecology and Evolutionary Biology, R. A. Seigel, J. T. Collins, S. S. Novak, Eds. (Macmillan Publication, New York., 1987), pp. 3–50. 10. G. Underwood, E. Kochva, Zool. J. Linn. Soc. 107, 3 (1993). 11. H. Zaher, Bull. Am. Mus. Nat. Hist. 240, 1 (1999). 12. N. Vidal, J. Toxicol. Tox. Rev. 21, 21 (2002). 13. N. Vidal, S. B. Hedges, C. R. Biologies 325, 977 (2002). 14. N. Vidal, S. B. Hedges, C. R. Biologies 325, 987 (2002). 15. J. B. Slowinski, R. Lawson, Mol. Phylogenet. Evol. 24, 194 (2002). 16. N. Vidal, S. B. Hedges, Proc. Roy. Soc. Lond. B (Suppl.) 271, S226 (2004). 17. N. Vidal, P. David, Mol. Phylogenet. Evol. 31, 783 (2004). 18. R. Lawson, J. B. Slowinski, B. I. Crother, F. T. Burbrink, Mol. Phylogenet. Evol. 37, 581 (2005). 19. B. P. Noonan, P. T. Chippindale, Mol. Phylogenet. Evol. 40, 347 (2006). 20. N. Vidal, A.-S. Delmas, S. B. Hedges, in Biology of the Boas and Pythons, R. W. Henderson, R. Powell, Eds. (Eagle Mountain Publication, Eagle Mountain, Utah, 2007), pp. 27–33. 21. N. Vidal et al., C. R. Biologies 330, 182 (2007). 22. D. J. Gower, N. Vidal, J. N. Spinks, C. J. McCarthy, J. Zool. Syst. Evol. Res. 43, 315 (2005). 23. B. G. Fry et al., Mol. Cell. Proteomics 7, 215 (2008). 24. N. Vidal, S. G. Kindl, A. Wong, S. B. Hedges, Mol. Phylogenet. Evol. 14, 389 (2000). 25. N. Vidal, S. B. Hedges, C. R. Biologies 328, 1000 (2005). 26. T. A. Hall, Nucleic Acids Symp. Ser. 41, 95 (1999). 27. H. Philippe, Nucleic Acids Res. 21, 5264 (1993).


28.

29. 30. 31. 32.

33. 34. 35. 36. 37. 38.

39. 40.

41. 42. 43. 44. 45.

D. L. Swofford, PAUP*. Phylogenetic Analysis Using Parsimony (* And Other Methods), Version 4.0b10 (Sinauer Associates, Sunderland, Massachusetts, 1998). F. Ronquist, J. P. Huelsenbeck, Bioinformatics 19, 1572 (2003). J. A. A. Nylander, F. Ronquist, J. P. Huelsenbeck, J. L. Nieves-Aldrey, Syst. Biol. 53, 47 (2004). D. Posada, K. A. Crandall, Bioinformatics 14, 817 (1998). F. M. Gradstein, J. G. Ogg, A. G. Smith. A Geologic Timescale 2004 (Cambridge University Press, Cambridge, UK 2005). J. D. Gardner, R. L. Cifelli, 1999. Palaeontology (Special Papers in Palaeontology) 60, 87 (1999). R. Hoffstetter, Bull. Soc. Géol. Fr. 1, 897 (1959). G. Cuny, J. J. Jaeger, M. Mahboubi, J.-C. Rage, 1990. C. R. Acad. Sci. (II) 311, 1267 (1990). J.-C. Rage, A. Richter, Neues Jahrbuch Geologie Paläontol. (Monatshefte) 9, 561 (1994). B. P. Noonan, P. T. Chippindale, Amer. Nat. 168, 730 (2006). A. Folie, Evolution des Amphibiens et Squamates de la Transition Crétacé-Paléogène en Europe: Les Faunes du Maastrichtien du Bassin de Hateg (Roumanie) et du Paléocène du Bassin de Mons (Belgique), Ph.D. Thesis (Université libre de Bruxelles, 2006). J.-C. Rage, C. Werner, Palaeontol. Afr. 35, 85 (1999). Z. Szyndlar, J.-C. Rage, Non-Erycine Booidea from the Oligocene and Miocene of Europe (Institute of Systematics and Evolution of Animals, Cracow, 2003). J. D. Scanlon, 2001. Mem. Assoc. Austral. Palaeontol. 25, 1 (2001). Z. Szyndlar, W. Böhme, Mertensiella 3, 381 (1993). J.-C. Rage, Palaeovertebrata 30, 111 (2001). J. A. Holman, Fossil Snakes of North America (Indiana University Press, Bloomington, Indiana, 2000). J.-C. Rage, L. Ginsburg, in Abstracts, Herpetology ’97, Z. Rocek, S. Hart, Eds. (Third World Congress of Herpetology, Prague, 1997), pp. 167–168.

397

46. J. J. Head, D. M. Mohabey, J. A. Wilson, J. Vert. Paleontol. 27, 720 (2007). 47. Z. Szyndlar, J.-C. Rage, in Biology of the Vipers, G. W. Schuett, M. Höggren, M. E. Douglas, H. W. Greene, Eds. (Eagle Mountain Publishing, Eagle Mountain, Utah, 2002), pp. 419–444. 48. M. Augé, J.-C. Rage, Mém. Mus. Nat. Hist. Natur. 183, 263 (2000). 49. J.-C. Rage, Acta Zool. Cracoviensia 31, 457 (1988). 50. Z. Szyndlar, J.-C. Rage, Amphibia–Reptilia 11, 385 (1990). 51. J. D. Scanlon, M. S. Y. Lee, M. Archer, Geobios 36, 573 (2003). 52. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 53. Z. Yang, A. D. Yoder, Syst. Biol. 52, 705 (2003). 54. Z. Yang, CABIOS 13, 555 (1997). 55. S. E. Evans, Biol. Rev. 78, 513 (2003). 56. K. L. Sanders, M. S. Y. Lee, Mol. Phylogenet. Evol. 46, 1165 (2008). 57. M. A. Iturralde-Vinent, R. D. E. MacPhee, Bull. Am. Mus. Nat. Hist. 238, 1 (1999). 58. Z. T. Nagy et al., Proc. Roy. Soc. Lond. B 270, 2613 (2003). 59. J. J. Wiens, M. C. Brandley, T. W. Reeder, Evolution 60, 114 (2006). 60. A. Hugall, R. Foster, M. S. Y. Lee, Syst. Biol. 56, 543 (2007). 61. W. Wüster et al., Mol. Phylogenet. Evol. 45, 437 (2007). 62. F. T. Burbrink, R. A. Pyron, Syst. Biol. 57, 317 (2008). 63. V. Wallach, The Visceral Anatomy of Blindsnakes and Wormsnakes and its Systematic Implications (Serpentes: Anomalepididae, Typhlopidae, Leptotyphlopidae), Ph.D. Thesis (Northeastern University, Boston, 1998). 64. J. R. Ali, J. C. Aitchison, Earth-Sci. Rev. 88, 145 (2008). 65. E. Gheerbrant, J.-C. Rage, Palaeogeogr. Palaeoclimatol. Palaeoecol. 241, 224 (2006).

Turtles (Testudines) H. Bradley Shaffer Department of Evolution and Ecology, University of California, Davis, CA 95616, USA ([email protected])

Abstract Living turtles and tortoises consist of two major clades (Cryptodira and Pleurodira), 14 families, and ~313 living species. Time-calibrated phylogenetic analyses can provide basic insights into the tempo of turtle evolution. Molecular phylogenetic analyses have confirmed that most families are demonstrably monophyletic, as are Cryptodira and Pleurodira. A time-calibrated analysis of all living families of turtles, which spans ~210 million years, demonstrates that many of the most-endangered clades are over 100 million years old and are represented by one or a few species, while the most species-rich families tend to be relatively young.

As a group, the turtles, tortoises, terrapins, and marine turtles (collectively, the Order Testudines) are one of the most instantly recognizable and well-known clades of non-avian reptiles (Fig. 1). Whether kept as pets, revered in temples, or slaughtered for the marketplace, turtles and tortoises are an integral and sometimes sacred part of many societies. Recent morphological and molecular phylogenetic studies (the latter are primarily mitochondrial) have led to considerable reshuffling of generic boundaries and taxonomic instability as apparently nonmonophyletic groups have been identified and reclassified (1). However, at deeper levels, most researchers agree that the 313 living turtle species are distributed among 14 monophyletic families (2). Although species-level diversity is modest, turtles are found in most major habitats and continents on Earth and many island systems. Marine species (Cheloniidae and Dermochelyidae) may exceed 2 m in total length, and are found worldwide in temperate and tropical oceans. Tortoises (Testudinidae) and some Emydidae and Geoemydidae are exclusively terrestrial; tortoises in particular have invaded many of the world’s deserts and some oceanic islands. Among their many unique biological features, turtles are particularly well known for several unique skeletal features, including the “anapsid” skull condition, the lack of teeth, and the shell. Testudines tend to be

very long-lived, and instances of >100-year-old individuals are known (3). They are also extremely variable in their sex-determination mechanisms, and exhibit both genetic- and temperature-dependent sex determination. Recently, turtles have emerged as one of the most threatened major clades of vertebrates, with 132 of 201 evaluated species listed in the highest categories of endangerment (Extinct, Extinct in the wild, Critically Endangered, Endangered, Vulnerable) by the IUCN (http://www.redlist.org/). In this paper, I review the phylogenetic relationships and molecular divergence times of the families of turtles (Fig. 2). The monophyly of turtles has never been questioned. It is based on the derived characters associated with a shell of ribs fused to overlying dermal bones inside of which lie the girdles (4). For the last 130 years, the living turtles have been divided into two reciprocally monophyletic clades (5). Pleurodira, or the side-necked turtles, retract their necks into the shell by bending the neck in a horizontal plane (6). Living pleurodires have relatively modest diversity, with three families and ~86 species (2), and are restricted to the southern continents of South America, Africa, and Australia/New Guinea, although fossil taxa were more widely distributed (6). Cryptodira, or the hidden-neck turtles, retract their head by bending the neck in the vertical plane. They are more diverse, with 11 families and about 227 species (2). Cryptodires are distributed across all temperate and tropical regions

Fig. 1 A kinosternid species (Sternotherus carinatus) endemic to large rivers in the south-central United States. Photo credit: H. B. Shaffer.

H. B. Shaffer. Turtles (Testudines). Pp. 398–401 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Sauropsida; Testudines

399

Geoemydidae 12

Testudinidae

10

Emydidae

9

Platysternidae Kinosternidae

6

11

Dermatemydidae

8

Chelydridae 7

2

Dermochelyidae 13

Cheloniidae Trionychidae

1

4

Carettochelyidae Chelidae

3

Pelomedusidae 5

Tr

Jurassic

Podocnemidae

Cretaceous MESOZOIC

200

150

Paleogene

Ng

CENOZOIC 100

50

0 Million years ago

Fig. 2 A timetree of turtles (Testudines). Divergence times are show in Table 1. Abbreviations: Ng (Neogene) and Tr (Triassic).

on Earth, although their primary diversity is restricted to northern continents. Morphological hypotheses of the interrelationships of the living turtles, including characters diagnosing all nodes, are provided in several recent treatments (4, 7–9). Several novel conclusions regarding interfamilial relationships of turtles were suggested by these morphological studies, including the close relationships of mud turtles (Kinosternidae) and softshell turtles (Trionychidae), the unexpected monophyly of the New-World snapping turtles (Chelydridae) and the Asian Big-headed Turtle (Platysternidae), and the close relationship of the Pig-nosed Turtle (Carettochelyidae) and Trionychidae. Molecular phylogenetic analyses with broad taxonomic coverage across living turtles began about a decade ago (9), and now include studies with both mitochondrial (9, 10) and nuclear (11–13) sequence data. In addition, several comprehensive analyses of within-family relationships (14–17) have helped to further resolve many aspects of relationships among the living turtles. At least three key results have emerged from this body of molecular work. First, the well-sampled families of turtles have been found to be monophyletic, although many of the contained genera have not. The one possible exception is the largely Old-World pond turtle Family Geoemydidae

(sometimes referred to as Bataguridae). Based on morphological evidence, geoemydids were considered to be paraphyletic with respect to the tortoises (Testudinidae), which were hypothesized to be deeply nested within Geoemydidae (4, 17). While molecular evidence has not supported this hypothesis, it is still an open question whether the diverse Old-World geoemydids and Rhinoclemmys, the single New-World genus, are monophyletic with respect to testudinids (17). Second, the relationships of the monotypic Big-headed Turtle Family Platysternidae and the New-World snapping turtles (Family Chelydridae) are finally becoming clear. Based largely on morphological evidence and limited mtDNA data, the two were considered to be closest relatives (4, 9). However, whole mitochondrial genome data (10), RAG1 nuclear data (12), and combined RAG1 and mtDNA data (12) indicate that Platysternidae is most closely related to Testudinoidea (Emydidae, Geoemydidae, Testudinidae). The placement of Chelydridae remains uncertain—nuclear and mtDNA suggest affinities with the Kinosternidae and their relatives (12), whereas whole mitochondrial data place them as the closest relative of a clade containing Testudinoidea, the marine turtles, and Platysternidae (10). Third, the phylogenetic placement of the softshell turtles (Trionychidae) remains enigmatic.

400


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among turtles (Testudines). Timetree Node

Estimates Ref. (19)

Time Time

Ref. (22) CI

Time

CI

1

207.0

–

–

207

221–193

2

175.0

175

186–164

–

–

3

156.5

176

184–168

137

159–115

4

155.0

155

166–144

–

–

5

124.0

124

135–113

–

–

6

94.0

94

100–88

–

–

7

87.0

87

95–79

–

–

8

85.0

85

91–79

–

–

9

74.0

74

93–55

–

–

10

70.0

70

76–64

–

–

11

65.0

65

–

–

–

12

52.0

52

–

–

–

13

50.0

50

57–43

–

–

Note: The node times in the timetree are average of times from difference studies. The times for Nodes 11 and 12 are based on fossil dates supported by a cross-validation procedure with molecular time estimates.

Molecular data suggest that softshells and their close relative, the Pig-nosed Turtle (Carettochelyidae), are either closest to the remaining Cryptodira (12) or perhaps represent the first split among all living turtles and fall outside of Cryptodira [suggested as an unlikely possibility in (12) and (22)]. In either case, they are not closely related to the mud turtle clade as suggested based on morphological analyses (10–12). Because they have a superb fossil record, turtles have long been recognized as excellent candidates for fossil-calibrated molecular phylogenies. In addition, an intriguing early suggestion, based on within-species phylogeographic estimates, that turtles may be characterized by a slow rate of molecular evolution (18) raised the possibility that molecular clocks may be applicable to very ancient phylogenetic events in turtle history. However, only two papers (9, 19) have used fossils to explore divergence times among all of the major lineages of turtles. In the first (9), the primary emphasis was on the impact of including fossils on phylogeny estimation, rather than time estimation. However, one major conclusion from that work was that the major lineages of cryptodiran turtles all appeared to diverge roughly 100 million years ago (Ma), based on evidence from their earliest fossil representatives.

Near et al. (19) provided estimates of divergence times for all major clades of living Testudines, as well as a new cross-validation method for identifying potential outlier fossils that may be providing inaccurate divergence dates. Using 10 consistent fossil calibration dates and 4691 basepairs of nuclear (RAG 1 exon, R35 intron) and mitochondrial (cyt b) data scored for 23 turtles that include representatives of all living families and subfamilies of turtles (12), Near et al. provided several new insights into the tempo of turtle evolution (Fig. 2, Table 1). Primary among them are that most of the among-family divergences of turtles are ancient, and predate the Cretaceous–Cenozoic boundary (>66 Ma). However, divergences among the most diverse families of living Cryptodira (Testudinidae, 55 species; Emydidae, 48 species; Geoemydidae, 66 species) are much more recent, with most of the within-family diversification occurring within the last 50 million years. Some aspects of this work have recently been criticized (20), including the antiquity of the Cryptodira–Pleurodira split (8, 23), although molecular analyses of the RAG-1 gene across tetrapods provides independent evidence in support of the 210 million year age of this event (22). Thus, the fundamental conclusions of Near et al. (21) appear to remain valid. In particular, several of the most critically endangered families of Testudines have origins in the Jurassic (200–146 Ma) or Cretaceous (146–66 Ma) and are currently represented by one or a few living species. These lineages contain a great deal of unique phylogenetic history, and are important candidates for conservation action.

Acknowledgments Support was provided by U.S. National Science Foundation, and the Agricultural Experiment Station and Center for Population Biology at the University of California, Davis. P. Meylan, P. Spinks, and R. Thomson critically reviewed the manuscript.

References 1.

H. B. Shaffer, N. N. FitzSimmons, A. Georges, A. G. J. Rhodin (Eds.), Defining Turtle Diversity: Proceedings of a Workshop on Genetics, Ethics, and Taxonomy of Freshwater Turtles and Tortoises (Chelonian Research Foundation, Lunenburg, Massachusetts, 2007). 2. Turtle Taxonomy Working Group [J. W. Bickham, J. B. Iverson, J. W. Parham, H. D. Philippen, A. G. J. Rhodin, H. B. Shaffer, P. Q. Spinks, P. P. van Dijk], in Defining Turtle Diversity: Proceedings of a Workshop on

Eukaryota; Metazoa; Vertebrata; Sauropsida; Testudines

3. 4.

5. 6. 7. 8. 9. 10. 11.

Genetics, Ethics, and Taxonomy of Freshwater Turtles and Tortoises, H. B. Shaffer, A. Georges, N. FitzSimmons, A. G. J. Rhodin, Eds. (Chelonian Research Foundation, Lunenburg, Massachusetts, 2007), pp. 173–199. J. W. Gibbons, Bioscience 37, 262 (1987). E. S. Gaff ney, P. A. Meylan, in Phylogeny and Classification of the Tetrapods, M. J. Benton, Ed. (Clarendon Press, Oxford, 1988), Vol. 1, pp. 157–219. E. D. Cope, Proc. Amer. Assoc. Adv. Sci. 1871, 194 (1871). E. S. Gaff ney, H. Y. Tong, P. A. Meylan, Bull. Am. Mus. Nat. Hist. 300, 1 (2006). E. S. Gaff ney, P. A. Meylan, A. R. Wyss, Cladistics 7, 313 (1991). W. G. Joyce, Bull. Peabody Mus. Nat. Hist. 48, 3 (2007). H. B. Shaffer, P. Meylan, M. L. McKnight, Syst. Biol. 46, 235 (1997). J. F. Parham, C. R. Feldman, J. L. Boore, BMC Evol. Biol. 6 (2006). M. K. Fujita, T. N. Engstrom, D. E. Starkey, H. B. Shaffer, Mol. Phylogenet. Evol. 31, 1031 (2004).

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

401

J. G. Krenz, G. J. P. Naylor, H. B. Shaffer, F. J. Janzen, Mol. Phylogenet. Evol. 37, 178 (2005). T. Sasaki et al., Mol. Biol. Evol. 21, 705 (2004). T. N. Engstrom, H. B. Shaffer, W. P. McCord, Syst. Biol. 53, 693 (2004). C. R. Feldman, J. F. Parham, Mol. Phylogenet. Evol. 22, 388 (2002). M. Le, C. J. Raxworthy, W. P. McCord, L. Mertz, Mol. Phylogenet. Evol. 40, 517 (2006). P. Q. Spinks, H. B. Shaffer, J. B. Iverson, W. P. McCord, Mol. Phylogenet. Evol. 32, 164 (2004). J. C. Avise, B. W. Bowen, T. Lamb, A. B. Meylan, E. Bermingham, Mol. Biol. Evol. 9, 457 (1992). T. J. Near, P. A. Meylan, H. B. Shaffer, Amer. Nat. 165, 137 (2005). J. F. Parham, R. B. Irmis, Amer. Nat. 171, 132 (2008). T. J. Near, P. A. Meylan, H. B. Shaffer, Amer. Nat. 171, 137 (2008). A. F.Hugall, R. Foster, M. S. Y. Lee, Syst. Biol. 56, 543 (2007). J. Sterli, Biol. Lett. 4, 286–289 (2008).

Crocodylians (Crocodylia) Christopher A. Brochu Department of Geoscience, University of Iowa, Iowa City, IA 52242, USA ([email protected]).

Abstract Crocodylia (23 sp.) includes the living alligators and caimans (Alligatoridae), crocodiles (Crocodylidae), and gharials (Gavialidae). Relatives of Alligatoridae and possibly Gavialidae first appear in the early Campanian of the late Cretaceous (~80 million years ago, Ma), but some molecular estimates place the earliest split within Crocodylia before 150 Ma. Estimating divergences within Crocodylia is complicated by unresolved conflict over how living and extinct gharials are related to alligatorids and crocodylids. If Gavialidae and Crocodylidae are close relatives, their divergence could be anywhere between 20 and 80 Ma.

Crocodylia includes the alligators, caimans, crocodiles, and gharials found throughout the world’s tropics (Fig. 1). Twenty-three living species are currently recognized (1), though some probably represent cryptic species complexes (2–4). The fossil record of the group extends to the early part of the Campanian (84–71 Ma) and includes over 150 known species, with many more awaiting description (5). They are semiaquatic ambush predators and include the largest living reptiles. Some of these species are used in the exotic leather industry and, as such, are important economic resources for impoverished nations; others are critically endangered. Crocodylians are central to research in developmental biology, osmoregulation, cardiophysiology, paleoclimatology, sex determination, population genetics, paleobiogeography, functional morphology, and reptile genomics. Their dense fossil record, with first appearance data throughout the clade’s stratigraphic range, gives us an excellent opportunity to empirically test methods used to estimate divergence times from molecular data (6). In this paper, I discuss divergence times within Crocodylia based both on the fossil record and on the nucleotide sequence data. Virtually all data agree on the monophyly of Alligatoridae, including the two living alligators (Alligator) and six or more living caiman species. Among caimans, the

dwarf or smooth-fronted caimans (Paleosuchus) are basal to other members of the group. There is consensus that 11 species of crocodile (Crocodylus) form a clade, with a 12th—the African Slender-snouted Crocodile (Mecistops cataphractus)—being basal to either Crocodylus or the African dwarf crocodiles (Osteolaemus). The Indonesian False Gharial (Tomistoma schlegelii) is universally seen as being closer to crocodiles than to alligators. These groups—Alligatoridae and Crocodylidae—belong to more inclusive groups (Alligatoroidea and Crocodyloidea, respectively) that include extinct relatives of the “families” (5). Relationships among derived caimans and within Crocodylus are unclear, but this reflects a lack of resolution in most data sets, probably as a result of the recency of their divergences (7, 8). The only real controversy involves the Indian Gharial, G. gangeticus. Morphological data strongly support a distant relationship and comparatively ancient divergence (Mesozoic, minimally 80 Ma) between Gavialis and other living crocodylians. Tomistoma, based on these data, joins Crocodylus, Mecistops, and Osteolaemus within Crocodylidae. Molecular data sets usually support a close relationship between Gavialis and Tomistoma and a much more recent divergence between them. In this case, Gavialis and Tomistoma would form a monophyletic Gavialidae and extant Crocodylidae would be limited to Osteolaemus, Mecistops, and Crocodylus (5, 9–15).

Fig. 1 The Gharial (Gavialis gangeticus) from the Indian subcontinent. Credit: C. A. Brochu.

C. A. Brochu. Crocodylians (Crocodylia). Pp. 402–406 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Sauropsida; Crocodylia

403

Gavialidae 2

Crocodylidae

1

Alligatoridae

Cretaceous

Paleogene

MESOZOIC 100

Neogene

CENOZOIC 0 Million years ago

50

Fig. 2 A timetree of crocodylians (Crocodylia). Divergence times are shown in Table 1.

This complicates efforts to use internal calibration points for the group and compare molecular divergence times with the fossil record. Divergence estimates between Tomistoma and Gavialis make little sense when their fossil relatives cannot be arranged on the tree. There are what appear to be robust calibration points within the clade, but some of these (e.g., the Tomistoma–Crocodylus split) make little sense when the fossils bracketing the divergence point exclude lineages that molecular data argue should be included. Several first appearances within Crocodylia are robust. The oldest alligatorid, Navajosuchus mooki, is from the lowermost Paleocene (66–62 Ma). The lineage including Alligator was generally localized in North America and Eurasia at a time when non-marine vertebrate sampling is good. Caimans also first appear in the Paleocene (66–56 Ma), albeit with a spottier fossil record. Minimal morphological divergence between the oldest known alligatorids and their closest extinct relatives suggests that the last common ancestor of alligators and caimans lived at or near the Cretaceous–Paleogene boundary, approximately 66 Ma (16, 17). Slightly older calibrations used in some analyses (15, 18) are based on arbitrary extensions of the fossil date (17) and may be close to the origination time, but this is difficult to test. The tomistomine–crocodyline split can be placed minimally in the Ypresian stage of the Eocene (56–49 Ma) based on the tomistomine Kentisuchus and the crocodyline Kambara. The degree of disparity among early crocodylids and close relatives is minimal (19–21). Several Campanian (84–71 Ma) alligatoroids are known (22–25), and the oldest crocodyloid is from the Maastrichtian (71–66 Ma) (5). Older crocodyloid fossils have been reported (26), but these are based on fragmentary material that cannot be reliably assigned to Crocodylia. The basal-most alligatoroids and crocodyloids are morphologically very similar, and the fit between stratigraphic and phylogenetic occurrence is

good (5). The inference is that their divergence is probably not much earlier than their first fossil appearances. But although the fossil record of late Jurassic (161–146 Ma) and early Cretaceous (146–100 Ma) crocodyliforms is excellent, the record through the middle Cretaceous, especially of close relatives of Crocodylia, is much less complete (27). Phylogenetic uncertainty over Gavialis and its putative fossil relatives also complicates the situation. These dates are consistent with molecular clock estimates based on distance data (28, 29). Gavialis is the only exception—most (though not all) molecular data posit a Cenozoic (66–0 Ma) divergence between them (Table 1, Fig. 2). This differs from most prominent conflicts between molecular and morphological data; in most cases, molecular data suggest substantially older divergences than the fossil record suggests. In this case, molecular estimates are tens of millions of years younger than the earliest known fossils. For this reason, the phylogenetic identity of Cretaceous and Paleocene gavialoids is controversial (9, 30). Application of quartet dating to several mitochondrial genes showed a strong relationship between the ages of the internal calibrations used and the resulting estimate of divergence time between alligatorids and crocodylids. In all cases, two internal calibration points were used: one within Alligatoridae and another within Crocodylidae. Estimates based on post-Eocene calibrations are uniformly younger than those in which one or both calibration points was of Eocene age or older. Estimates based on two calibrations of very different age—one within the past 30 million years and one older than 50 million years—are usually close to the Campanian first appearance datum (6, 31). Recent studies based on mitogenomic data resulted in widely different divergence estimates for the same nodes. The first (14) used nonparametric rate smoothing (NPRS), penalized likelihood (PL), and Bayesian

404


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) within Crocodylia. Timetree Node

Time


Ref. (6)(b)

Ref. (6)(c)

Time

Time

Time

Time

Ref. (14)(a) CI

Time

Ref. (14)(b) CI

Time

Ref. (15)(a) CI

1

102.6

41.0

71.0

107.0

137.0

144–130

164.0

184–144

101.0

104–98

2

63.8

–

–

–

74.0

80–68

85.0

101–69

47.0

50–44

Timetree Node

Estimates (Continued) Ref. (15)(b)

Time

Ref. (15)(c)

Ref. (16)(a)

Ref. (16)(b)

Ref. (16)(c)

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

1

102.6

97.0

102–92

–

106–100

33.0

39–27

42.0

48–36

55.0

76–34

2

63.8

49.0

54–44

49.0

53–45

–

–

–

–

–

–

Timetree Node

Estimates (Continued) Ref. (16)(d)

Time

Refs. (10, 32) (a)

Refs. (10, 32) (b)

Time

CI

Time

CI

Time

CI

1

102.6

78.0

99–57

112.0

116–110

147.0

155–142

2

63.8

–

–

–

–

73.0

78–71

Note: Node times in the timetree represent the mean of time estimates from different studies. In ref. (6), time estimates were generated from the quartet analysis of four mitochondrial genes with both calibrations from the Neogene (a), with one Neogene and one Paleogene calibration (b), and with calibrations both from the Paleogene (c). In ref. (14), NPRS (a) and Bayesian (b) analyses of mitogenomic amino acid sequences were conducted. In ref. (15), estimates from PL (a), NPRS (b), and Bayesian (c) analyses on mitogenomic amino acid sequences using an internal calibration are shown; additional analyses using nucleotide sequences or excluding the internal calibration are very similar. In ref. (16), PL estimates from nucleotide (a, b) and amino acid (c, d) alignments and based on one (a, c) or five (b, d) calibration points are shown. Refs. (13, 42) report NPRS studies of two nuclear and four mitochondrial genes where Gavialis was excluded (a) and included (b).

methods. It estimated a gavialid–crocodylid divergence near fossil predictions (albeit with a different topology), and the Gavialis–Tomistoma divergence (36–48 Ma) postdates current fossil evidence, but other estimates were substantially older (Table 1), including a late Jurassic (161–146 Ma) alligatorid–crocodylid split. The second study (15) was similar to its predecessor, but with improved taxon sampling and an internal calibration point (alligator–caiman) for some analyses. Recovered dates were much younger than in the preceding mitogenomic analysis, including a Gavialis– Tomistoma divergence of 28–22 Ma (Table 1). The alligatorid–crocodylid divergences were slightly older than fossil first appearances (~100 Ma). These two studies were operationally very similar. They used similar mitogenomic data and dating methods. Outgroup sampling was nearly identical, and they used the same external calibration points. Estimates in the second study (15) did not change appreciably

when the alligator–caiman calibration was excluded. The most significant difference appears to be the lower bound of one of the external calibration points (marsupial–eutherian), which was cut from 174 Ma (14) to 138 Ma (15). This may have inflated the divergence estimates in the earlier study and reinforces the importance of calibration choice in molecular divergence time estimation. Dates reported from RAG1 data using PL (18) appear anomalous at first. As with the mitogenomic studies, a relationship between estimate and calibration choice was noted; but aligned nucleotides put the alligatorid– crocodylid split between 42 and 33 Ma (Table 1) and the alligator–caiman split between 21 and 17 Ma. Fossil first appearances are two to three times older. But when analyzed as amino acid sequences, 95% confidence intervals around dates estimated for the same divergence points either include, or come close to including, first appearances from fossils (Table 1). This is in contrast to the

Eukaryota; Metazoa; Vertebrata; Sauropsida; Crocodylia

second mitogenomic study (15), which found no significant difference between estimates from nucleotide and amino acid sequences. Additional NPRS estimates were obtained for the alligatorid–crocodylid split using a data set (12 taxa, 3667 basepairs) combining nuclear (RAG1 and c-mos) and mitochondrial (12s, 16s, cytochrome b, tRNAglu with a flanking portion of nd6) genes (10, 13, 32) that can be analyzed using the same model (HKY + G + I). The dates were obtained using r8s, version 1.71 (33). Two sets of dates were obtained. The first considered trees in which Gavialis was excluded, and only the alligatorid–crocodylid divergence time was estimated. The second was based on trees including Gavialis as the closest relative of Tomistoma, and two dates—the basal split within Crocodylia and the crocodylid–gavialid split—were estimated. In the first case, four internal constraints were used: Alligator mississippiensis–Alligator sinensis (20 Ma), Alligator–caiman (64 Ma), Tomistoma– Crocodylus (54 Ma), and Osteolaemus–Crocodylus (30 Ma). When Gavialis was included in the analysis, the Tomistoma–Crocodylus calibration was not used. The calibrations were used as upper temporal constraints rather than fi xed points. The resulting estimates for the alligatorid–crocodylid split depend on whether Gavialis is included. Trees excluding Gavialis put the alligatorid–crocodylid split at ~112 Ma, and trees including it put the same split at ~147 Ma. There are several factors that might explain this disparity, including asymmetry in the distribution of calibrations. Alligatorids often have longer branches and higher rates of evolution than crocodylids, and relative rate tests often reject a single-rate model for Crocodylia (9, 11, 31). This is true for the data set analyzed here. Because the Crocodylus–Tomistoma calibration point was not used when Gavialis was included, the resulting trees relied more heavily on calibrations from among alligatorids. If we take the more ancient mitogenomic estimates literally, there are surprisingly few stratigraphic implications. The fossil record, read literally, shows no drop in diversity at the Cretaceous–Paleogene boundary (34, 35). Molecular divergence estimates would draw multiple extant lineages back to the Cretaceous, indicating increased survivorship across a boundary that, for the group in question, already shows a high level of survivorship. They would, however, diminish the apparent drop in crocodylian diversity during the late Eocene and Oligocene, followed by an increase in diversity during the early Miocene (27).

405

The most important implication comes from younger rather than older molecular dates. Many putative gavialoids and tomistomines predate molecular estimates of their divergence (15, 29, 36) by tens of millions of years. We continue to reevaluate these fossils, but for now they continue to support a minimum divergence of 80 Ma between Gavialis and Tomistoma, even if they are constrained as closest living relatives (13, 37). Extension of basal divergences to the late Jurassic or early Cretaceous brings them within the time frame of Gondwanan breakup. It would be tempting to argue that early divergences lend support to a vicariance model for crocodylian historical biogeography. However, the biogeographic distribution of crocodylians, with or without fossils, does not match a vicariant pattern (38). Most extant lineages are fully capable of withstanding exposure to salt water (39), and whether one relies on the preferred morphological or molecular tree, branching order is inconsistent with plate tectonic history. Dispersal remains the best explanation for the distribution of most crocodylian clades.

Acknowledgments I thank A. Janke, L. Densmore, J. Gatesy, P. S. White, R. Willis, D. Ray, and R. McAliley for access to sequence data. This work was supported by the U.S. National Science Foundation.

References 1. J. P. Ross, Crocodiles—Status Survey and Conservation Action Plan, 2nd ed. (International Union for the Conservation of Nature, Gland, Switzerland, 1998). 2. G. Amato, J. Gatesy, in Molecular Ecology and Evolution: Approaches and Applications, B. Schierwater, B. Streit, G. P. Wagner, R. DeSalle, Eds. (Birkhäuser Verlag, Basel, 1994), pp. 215–226. 3. D. A. Ray, P. S. White, H. V. Duong, T. Cullen, L. D. Densmore, in Crocodilian Biology and Evolution, G. Grigg, F. Seebacher, C. E. Franklin, Eds. (Surrey Beatty and Sons, Sydney, 2001), pp. 58–63. 4. A. Schmitz et al., C. R. Palevol. 2, 703 (2003). 5. C. A. Brochu, Ann. Rev. Earth and Planet. Sci. 31, 357 (2003). 6. C. A. Brochu, Evolution 58, 1375 (2004). 7. C. A. Brochu, L. D. Densmore, in Crocodilian Biology and Evolution, G. Grigg, F. Seebacher, C. E. Franklin, Eds. (Surrey Beatty and Sons, Sydney, 2001), pp. 3–8. 8. L. R. McAliley et al., Mol. Phylogenet. Evol. 39, 16 (2006). 9. J. Harshman, C. J. Huddleston, J. P. Bollback, T. J. Parsons, M. J. Braun, Syst. Biol. 52, 386 (2003).

406

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.


R. E. Willis, L. R. McAliley, E. D. Neeley, L. D. Densmore, Mol. Phylogenet. Evol. 43, 787 (2007). C. A. Brochu, Syst. Biol. 46, 479 (1997). S. W. Salisbury, P. M. A. Willis, Alcheringa 20, 179 (1996). J. Gatesy, G. Amato, M. Norell, R. DeSalle, C. Hayashi, Syst. Biol. 52, 403 (2003). A. Janke, A. Gullberg, S. Hughes, R. K. Aggarwal, U. Arnason, J. Mol. Evol. 61, 620 (2005). J. Roos, R. K. Aggarwal, A. Janke, Mol. Phylogenet. Evol. 45, 663 (2007). C. A. Brochu, J. Vert. Paleo. 24, 856 (2004). J. Müller, R. R. Reisz, BioEssays 27, 1069 (2005). A. F. Hugall, R. Foster, M. S. Y. Lee, Syst. Biol. 56, 543 (2007). P. M. A. Willis, R. E. Molnar, J. D. Scanlon, Kaupia 3, 27 (1993). C. A. Brochu, Palaeontology 50, 917 (2007). P. Piras, M. Delfino, L. del Favero, T. Kotsakis, Acta Pal. Pol. 52, 315 (2007). J. E. Martin, J. Vert. Paleontol. 27, 362 (2007). T. E. Williamson, J. Vert. Paleontol. 16, 421 (1996). X.-C. Wu, A. P. Russell, D. B. Brinkman, Can. J. Earth Sci. 38, 1665 (2001). A. K. Martin, C. J. H. Hartnady, J. Geophys. Res. 91, 4767 (1986).

26. M. J. Benton, in The Fossil Record 2, M. J. Benton, Ed. (Chapman & Hall, London, 1993), pp. 681–715. 27. P. J. Markwick, Paleobiology 24, 470 (1998). 28. L. D. Densmore, in Evolutionary Biology, M. K. Hecht, B. Wallace, G. H. Prance, Eds. (Plenum Press, New York, 1983), Vol. 16, pp. 397–465. 29. C. A. Hass, M. A. Hoff man, L. D. Densmore, L. R. Maxson, Mol. Phylogenet. Evol. 1, 193 (1992). 30. C. A. Brochu, J. Vert. Paleontol. 24, 610 (2004). 31. C. A. Brochu, J. Paleontol. 78, 7 (2004). 32. P. S. White, L. D. Densmore, in Crocodilian Biology and Evolution, G. Grigg, F. Seebacher, C. E. Franklin, Eds. (Surrey Beatty and Sons, Sydney, 2001), pp. 29–37. 33. M. J. Sanderson , r8s (Analysis of Rates of Evolution), version 1.71 (University of California at Davis, Davis, CA, 2006). 34. P. J. Markwick, Geological Society of America Abstracts with Programs, A (1994). 35. D. Vasse, S. Hua, Oryctos 1, 65 (1998). 36. L. D. Densmore, H. C. Dessauer, Comp. Biochem. Physiol. 77B, 715 (1984). 37. C. A. Brochu, J. Vert. Paleontol. 20, 1 (2000). 38. C. A. Brochu, in Crocodilian Biology and Evolution, G. Grigg, F. Seebacher, C. E. Franklin, Eds. (Surrey Beatty and Sons, Sydney, 2001), pp. 9–28. 39. L. E. Taplin, G. C. Grigg, Amer. Zool. 29, 885 (1989).

BIRDS


Birds (Aves) Marcel van Tuinen Department of Biology and Marine Biology, 601 South College Road, University of North Carolina at Wilmington, Wilmington, NC 284035915, USA ([email protected])

Abstract Living birds (~9500 species) are grouped into 20–28 orders, comprising the Subclass Neornithes of the Class Aves. With few exceptions, molecular phylogenetic analyses have supported two morphological divisions within Neornithes, Paleognathae (ratites and tinamous) and Neognathae (all other living birds). Within Neognathae, there is universal support for the recognition of two superorders: Galloanserae (landfowl and waterfowl) and Neoaves (all other neognath orders). The neornithine timetree shows a Paleognathae–Neognathae split at ~120 million years ago (Ma) and the Galloanserae–Neoaves split at ~105 Ma, both possibly related to continental breakup.

Living birds are grouped in the Subclass Neornithes, and currently divided into three Superorders: the Paleognathae (ratites and tinamous), Galloanserae (waterfowl and gamefowl), and Neoaves (all other birds; Fig. 1). Based on morphological classification, Neornithes and archaic birds are grouped together in the Class Aves, a subgroup of theropod dinosaurs. The evolution of archaic avians to neornithine birds shows progressive loss of teeth, reduction in tail length, and modification of feathers and limbs for powered flight. These combined characteristics are not yet seen in the oldest lineage of birds, Archaeopteryx. However, several nonmodern birds also displayed remarkable flight adaptations, particularly the Cretaceous enantiornithines. Between 9000 and 10,000 living species of birds have been described, and they have been placed in 20–28 orders (1, 2). Here, the relationships and divergence times of the three superorders of living birds are reviewed. Until recently, the classification of neornithine birds pioneered by Huxley (3) and expanded by Fürbringer (4) has followed the arrangement proposed by Wetmore (5). This classification is still reflected in published bird guides today. In it, species were placed in two major groups: Paleognathae (ratites and tinamous) and

Neognathae (all other orders). These two major divisions reflected differences in jaw morphology and flight capabilities (3). Paleognaths have primitive jaws reminiscent of non-avian theropods, while neognaths possess modern jaws with adaptations reflecting diverse feeding modes and postcranial modifications related to lifestyle and mode of locomotion. Early molecular phylogenies based on immunological (6) and DNA–DNA hybridization (7) distances supported the grouping of paleognath birds, but they did not support the classical definition of Neognathae. Instead of finding penguins, loons, and grebes to be among the earliest diverging neognaths, those molecular studies identified the waterfowl (Anseriformes) and gamefowl (Galliformes) as closest relatives and forming a group (Galloanserae) separate from other neognaths (Neoaves). However, the position of Galloanserae with respect to Neoaves and Paleognathae was not yet firmly established. DNA sequence studies over the last decade have addressed the relationships of avian orders and definition of superorders. Initial studies using small subsets of taxa gave conflicting results. An analysis of some short sequences of a nuclear gene (α-crystallin) in five species joined Galliformes and Anseriformes, with a paleognath (tinamou) basal (8). However, initial analyses of complete mitochondrial genomes in a small selection of avian

Fig. 1 A Rock Pigeon (Columba livia). Credit: M. Peck.

M. van Tuinen. Birds (Aves). Pp. 409–411 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).


Neoaves 2

Galloanserae

1

Paleognathae

Early K

Late K

Paleogene

MESOZOIC 100

Neognathae

410

Neogene

CENOZOIC 0 Million years ago

50

Fig. 2 A timetree of bird superorders. Divergence time estimates are from Table 1. Abbreviation: K (Cretaceous).

orders surprisingly placed songbirds (Passeriformes) in a basal position among living birds, with ratites nested within Neognathae (9–11). Studies using longer sequences from nuclear genes and greater taxonomic coverage (12, 13) provided the strongest support yet for the three superordinal groupings initially defined by immunological distances (6). Galloanserae was found to be the closest relative of Neoaves, together forming Neognathae. Subsequent studies based on one or several nuclear genes (14–19) and mitochondrial genomes (20–25) continued to support this superordinal arrangement, now reflected in taxonomic updates (26). A variety of studies have estimated divergence times among neornithine orders, in particular those covering the early history of modern birds. Based on 19 clocklike nuclear genes, the average divergence time between chicken, goose, Ostrich, and pigeon was estimated at 97 ± 12 million years ago (Ma), with mitochondrial DNA estimates ranging between 68 and 131 Ma (27). Subsequently, a quartet-dating study (28) on the 12S rRNA and c-mos genes from several orders pushed the average neognath– paleognath split deeper into the Cretaceous (135 Ma), and supported Cretaceous ordinal origins as well. Analyses of complete mitochondrial genomes and broad taxon sampling yielded mid-Cretaceous time estimates for the Paleognathae–Neognathae split (120–110 Ma) and the Galloanserae–Neoaves split (100 Ma) (21, 23). Other mitochondrial studies have used a Bayesian approach and minimum constraints instead of fi xed calibration points, which yielded divergence times that were on average 20 My older (22, 25). These early divergence time estimates do not have direct support from the neornithine fossil record. However, the presence of derived waterbird fossils (23, 29, 30) from the latest Cretaceous (66 Ma) and the Paleocene (62–55 Ma) implicate Cretaceous ages for the deepest nodes on the neornithine timetree.

The first study that estimated superordinal divergence times with complete ordinal representation used a concatenated nonprotein-coding portion of the mitochondrial genome (two rRNAs, three tRNAs) and a lineage-specific method (31). Those time estimates agreed closely with estimates reported in two of three mitochondrial DNA studies (21, 23) (Table 1). The only neornithine timetree based on multiple nuclear genes (18) and comprehensive ordinal sampling supported a similar divergence time for the Galloanserae–Neoaves split (95 Ma). In summary, the neornithine timetree (Fig. 2) shows mid-Cretaceous diversification of superorders, which is indirectly supported by fossil evidence (18, 23, 29, 30) and is consistent with continental breakup and paleogeography (27, 32).

Acknowledgments Support was provided by Netherlands Organization for Scientific Research (NWO) and the Epply foundation.

References 1. D. P. Mindell, J. W. Brown, Neornithes: Modern birds, http://tolweb.org/Neornithes/15834/2005.12.14 (University of Arizona, Tucson, 2008). 2. C. G. Sibley, B. L. Monroe, Distribution and Taxonomy of the Birds of the World (Yale University Press, New Haven, 1993). 3. T. H. Huxley, Proc. Roy. Soc. Lond. 1867, 415 (1867). 4. M. Fürbringer, Untersuchungen zur Morphologie und Systematik der Vögel (Von Holkema, Amsterdam, 1888). 5. A. Wetmore, Smithsonian Misc. Coll. 139, 1 (1960). 6. C. Y. K. Ho, E. M. Prager, A. C. Wilson, D. T. Osuga, R. E. Feeney, J. Mol. Evol. 8, 271 (1976). 7. C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds: A Study in Molecular Evolution (Yale University Press, New Haven, 1990). 8. G.-J. Caspers, D. Uit deWeert, J. Wattell, W. W. de Jong, Mol. Phylogenet. Evol. 7, 185 (1997).

Eukaryota; Metazoa; Vertebrata; Aves

411

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among birds (Aves). Timetree Node

Time

Estimates Ref. (18)

Ref. (21)

Time

Time

CI

Ref. (22)

Ref. (23)

Time

CI

Time

CI

Ref. (24)

Ref. (25)

Time

Time

CI

Ref. (29) Time

CI

1

119.0

–

111.0

125–96

139

154–126

110.6

125–96

101

133.2

149–115

119.0

129–108

2

105.0

95.0

99.0

112–87

122

135–110

99.1

112–87

90.0

126.0

140–112

104.0

110–99

Note: Node times in the timetree represent the mean of time estimates from different studies. Divergence times were estimated from an analysis of ribosomal mitochondrial genes (29), protein-coding mitochondrial genes (25), complete mitochondrial genomes (21–24), and five nuclear genes (18).

9.

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

D. P. Mindell, M. D. Sorenson, C. J. Huddleston, H. C. Miranda, A. Knight, S. J. Sawchuk, T. Yuri, Avian Molecular Evolution and Systematics (Academic Press, San Diego, 1997), pp. 213–247. P. J. Waddell, Y. Cao, M. Hasegawa, D. P. Mindell, Syst. Biol. 48, 119 (1999). A. Harlid, U. Arnason, Proc. Roy. Soc. Lond. B. Bio. 266, 305 (1999). J. G. Groth, G. F. Barrowclough, Mol. Phylogenet. Evol. 12, 115 (1999). M. van Tuinen, C. G. Sibley, S. B. Hedges, Mol. Biol. Evol. 17, 451 (2000). J. Garcia-Moreno, M. D. Sorenson, D. P. Mindell, J. Mol. Evol. 57, 27 (2003). U. S. Johannson, P G. P. Ericson, J. Avian Biol. 34, 185 (2003). J. Cracraft, 14 coauthors, Assembling the Tree of Life (Oxford University Press, New York, 2004), pp. 468–489. A. L. Chubb, Mol. Phylogenet. Evol. 30, 140 (2004). P. G. P. Ericson et al., Biol. Lett. 2, 543 (2006). M. G. Fain, P. Houde, Evolution 58, 2558 (2004). E. L. Braun, R. T. Kimball, Syst. Biol. 51, 614 (2002).

21. 22. 23.

24. 25. 26. 27. 28. 29. 30. 31. 32.

T. Paton, O. Haddrath, A. J. Baker, Proc. Roy. Soc. Lond. B. Bio. 269, 839 (2002). S. L. Pereira, A. J. Baker. Mol. Biol. Evol. 23, 1731 (2006). K. E. Slack, C. M. Jones, T. Ando, G. L. Harrison, E. Fordyce, U. Arnason, D. Penny, Mol. Biol. Evol. 23, 1144 (2006). G. L. Harrison et al., Mol. Phylogenet. Evol. 21, 974 (2004). J. W. Brown, J. S. Rest, J. Garcia-Moreno, M. D. Sorenson, D. P. Mindell, BMC Biol. 6, 6 (2008). R. C. Banks, C. Cicero, J. L. Dunn, A. W. Kratter, Auk 120, 923 (2003). S. B. Hedges, P. H. Parker, C. G. Sibley, S. Kumar, Nature 381, 226 (1996). A. Cooper, D. Penny, Science 275, 1109 (1997). G. J. Dyke, M. van Tuinen, Zool. J. Linn. Soc. 141, 153 (2004). J. A. Clarke, C. Tambussi, J. Noriega, G. Erickson, R. Ketcham, Nature 433, 305 (2005). M. van Tuinen, S. B. Hedges, Mol. Biol. Evol. 18, 206 (2001). J. Cracraft, Proc. Roy. Soc. Lond. B. Bio. 268, 459 (2001).

Ratites and tinamous (Paleognathae) Allan J. Baker a,b,* and Sérgio L. Pereiraa a

Department of Natural History, Royal Ontario Museum, 100 Queen’s Park Crescent, Toronto, ON, Canada; bDepartment of Ecology and Evolutionary Biology, University of Toronto, Toronto, ON, Canada *To whom correspondence should be addressed (allanb@rom. on.ca)

Abstract The Paleognathae is a monophyletic clade containing ~32 species and 12 genera of ratites and 46 species and nine genera of tinamous. With the exception of nuclear genes, there is strong molecular and morphological support for the close relationship of ratites and tinamous. Molecular time estimates with multiple fossil calibrations indicate that all six families originated in the Cretaceous (146–66 million years ago, Ma). The radiation of modern genera and species occurred from the Oligocene–Miocene boundary (23 Ma) to the Pleistocene (1.8–0.012 Ma). The basal splits within the ratites are approximately coincident with the breakup of Gondwana, suggesting that the different lineages rafted on continental landmasses to their present locations.

The Superorder Paleognathae consists of the flightless ratites and the volant tinamous. It is the closest relative of the remaining birds in the Superorder Neognathae. Ratites are named for their raft-like (ratis) sternum that lacks a keel, whereas the tinamous have a keeled sternum. Tinamous share with the ratites a complex bone structure in the roof of the mouth termed a paleognathous palate. The ratites include the Emu and four species of cassowaries in Australia and New Guinea (1), five species of kiwis (2) and possibly 14 extinct species of moas from New Zealand (3), two species of rheas from South America, the Ostrich (Fig. 1) now restricted to Africa but once more widely distributed across Europe and Asia, and five extinct species of elephant birds from Madagascar. Forty-six species of tinamous occur in South and Central America (1). Eight fossil species possessing a paleognathous palate occur from the late Paleocene (66–56 Ma) to the middle Eocene (~40 Ma) of the Northern Hemisphere (4). However, they appear to be a paraphyletic assemblage, and have been placed basal to

the ratites (5). Here, we review the phylogenetic relationships and divergence times of the extant clades of ratites, the extinct moas and the tinamous. Longstanding debates about whether the paleognaths are monophyletic or polyphyletic were not settled until phylogenetic analyses were conducted on morphological characters (6–9), transferrins (10), chromosomes (11, 12), α-crystallin A sequences (13, 14), DNA–DNA hybridization data (15, 16), and DNA sequences (e.g., 17–21). However, relationships among paleognaths are still not resolved, with a recent morphological tree based on 2954 characters placing kiwis (Apterygidae) as the closest relatives of the rest of the ratites (9), in agreement with other morphological studies using smaller data sets (6–8, 22). DNA sequence trees place kiwis in a derived clade with the Emus and Cassowaries (Casuariiformes) (19–21, 23, 24). The conflict between morphological and molecular

Fig. 1 An Ostrich (Struthio camelus), Family Struthionidae. Credit: J. Cracraft.

A. J. Baker and S. L. Pereira. Ratites and tinamous (Paleognathae). Pp. 412–414 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Aves; Paleognathae

413

Apterygidae (Apterygiformes) 5

Casuariidae (Casuariiformes)

4

Struthionidae (Struthioniformes)

3

Rheidae (Rheiformes)

2

Dinornithidae (Dinornithiformes)

1

Tinamidae (Tinamiformes) Late K

Paleogene

MESOZOIC 100

75

Neogene

CENOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of ratites and tinamous (Paleognathae). Divergence times are shown in Table 1. Abbreviation: K (Cretaceous).

phylogenies lies in where to place the root of the tree (8). The extinct moas are not recovered as the closest relatives to kiwis (19, 20) in the molecular trees, contrary to the morphological trees where they are closest relatives. The rheas (Rheiformes) and ostrich (Struthioniformes) also exchange places in different molecular trees (8, 19, 20). Nuclear gene sequences often place tinamous (Tinamiformes) within the ratites (25), possibly due to stochastic sorting of gene lineages across short basal branches that prevent recovery of the species tree (26). The Paleognathae timetree is based on the most recent analyses that include multiple fossil calibrations and allow for different rates of evolution in different branches of the tree (21, 27, 28) based on partial or complete mitochondrial genomes (Fig. 2). The basal split between ratites and tinamous is estimated to have occurred ~108 Ma. Moas diverged from the lineage leading to the other modern ratites ~95 Ma, followed by the rheas about 87 Ma, and ostrich 78 Ma. The Emu and cassowary lineage split from the kiwi lineage ~77 Ma, Emu and cassowaries diverged ~41 Ma, the rhea genera diverged about 14 Ma, and the moa and kiwi lineages diversified within the last 18–4 million years. Other estimates within Paleognathae have suggested more recent divergence times (18–31), but these were obtained using single anchor-points and methods of molecular dating that did not account for uncertainty in fossil ages or variable rates of molecular evolution. With the possible exception of the Ostrich, which may have walked to Africa following the rafting of the India–Madagascar plate to Asia, ratite divergence times fit the vicariance biogeography hypothesis. These large flightless birds probably rafted to their current geographic locations on the landmasses resulting from the

fragmentation of the supercontinent Gondwana (21, 27). Early molecular dating studies that did not account properly for phylogenetic and fossil calibration uncertainties resulted in much younger divergence times (17, 18, 30, 31). This in turn led to the alternative hypothesis that flighted ancestors (e.g., lithornids) of crown-group taxa dispersed after the fragmentation of Gondwana (4) and that descendant lineages became secondarily flightless on separate landmasses in the southern hemisphere. New molecular dates (27, 28) based on multiple fossil anchor-points suggest instead that a single loss of flight in modern ratites in Gondwana is highly likely. The tinamou lineage probably originated in the South American portion of Gondwana and never dispersed beyond the Americas. Diversification of moas in New Zealand has been attributed to earth history events and global cooling that fragmented ranges and promoted allopatric divergence of lineages (3).

Acknowledgments Support for this work was provided by NSERC, U.S. National Science Foundation, and the ROM Foundation.

References 1. S. J. J. F. Davies, Ratites and tinamous, Tinamidae, Rheidae, Dromaiidae, Casuariidae, Apterygidae, Struthionidae (Oxford University Press, New York, 2002). 2. M. L. Burbidge, R. M. Colbourne, H. A. Robertson, A. J. Baker, Cons. Genet. 4, 167 (2003). 3. A. J. Baker, L. Huynen, O. Haddrath, C. D. Millar, D. M. Lambert, Proc. Natl. Acad. Sci. U.S.A. 102, 8257 (2005).

414


Table 1. Divergence times (Ma) and credibility/confidence intervals (CI) among ratites and tinamous (Paleognathae). Timetree Node

1

Estimates Ref. (18)

Time

96.7

Ref. (19)

Ref. (20)

Ref. (21)

Ref. (27)

Ref. (28)

Ref. (29)

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

88.9

104–74

92.2

111–73

–

–

105.0

137–88

113

127–100

105.9

128–83

75.2

85–65

2

91.5

–

–

78.9

94–64

–

–

96.0

134–87

99.7

112–88

91.5

116–70

–

–

3

80.6

79.5

94–55

69.3

82–57

89.1

94–84

89.0

127–83

92.2

104–81

81.5

106–59

63.7

75–55

4

75.2

79.5

94–55

65.3

72–58

75.5

78–73

84.0

121–82

84.9

97–74

67.3

92–42

70.1

79–60

5

67.9

55.6

68–43

62.4

69–55

68

72–65

81.0

116–76

76.8

88–66

74.6

100–52

56.7

64–49

Note: Node times in the timetree represent the mean of time estimates from refs. (27, 29). Divergence times were estimated from analyses of complete mitochondrial genomes (19–21, 27, 28) and analysis of partial mitochondrial DNA sequences (18, 29).

4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14. 15.

16.

17. 18.

P. W. Houde, Publ. Nuttall. Club. 22, 1 (1988). G. J. Dyke, M. van Tuinen, Zool. J. Linn. Soc. 141, 153 (2004). J. Cracraft, Ibis 116, 494 (1974). A. H. Bledsoe, Ann. Carnegie Mus. 57, 73 (1988). K. Lee, J. Feinstein, J. Cracraft, Avian Molecular Evolution and Systematics (Academic Press, San Diego, 1997), pp. 213–247. B. C. Livezey, R. L. Zusi, Zool. J. Linn. Soc. 149, 1 (2007). E. M. Prager, A. C. Wilson, D. T. Osuga, R. E. Feeney, J. Mol. Evol. 8, 283 (1976). L. E. M. De Boer, Nature 287, 84 (1980). R. S. De Kloet. S. R. De Kloet, Genetica 119, 333 (2003). S. O. Stapel, J. A. M. Leunissen, M. Versteeg, J. Wattel, W. W. de Jong, Nature 311, 257 (1984). G. Casper, J. Wattel, W. W. de Jong, Mol. Biol. Evol. 11, 711 (1994). C. G. Sibley, J. E. Ahlquist, Evolution today: Proceedings of the Second International Congress for Systematic Evolutionary Biology (Hunt Institute Botanic Document, Pittsburgh, 1981). C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds: A Study in Molecular Evolution (Yale University Press, New Haven, 1990). A. Cooper, Avian Molecular Evolution and Systematics (Academic Press, San Diego, 1997), pp. 345. M. van Tuinen, S. B. Hedges, Mol. Biol. Evol. 18, 206 (2001).

19. 20. 21. 22. 23.

24.

25. 26. 27. 28. 29. 30.

31.

O. Haddrath, A. J. Baker, Proc. Roy. Soc. Lond. B 268, (2001). A. Cooper, C. Lalueza-Fox, S. Anderson, A. Rambaut, Nature 409, 704 (2001). T. Paton, O. Haddrath, A. J. Baker, Proc. Roy. Soc. Lond. B 269, 839 (2002). D. K. Zelenitsky, S. P. Modesto, Can. J. Zool. 81, 962 (2003). G. L. Harrison, P. A. McLenachan, M. J. Phillips, K. E. Slack, A. Cooper, D. Penny, Mol. Biol. Evol. 21, 974 (2004). K. E. Slack, C. M. Jones, T. Ando, G. L. Harrison, E. Fordyce, U. Árnason, D. Penny, Mol. Biol. Evol. 23, 1144 (2006). J. L. Chojnowski, R. T. Kimball, E. L. Braun, Gene 410, 89 (2008). J. H. Degnan, N. A. Rosenberg, PLoS Genet. 2, 762 (2006). S. L. Pereira, A. J. Baker, Mol. Biol. Evol. 23, 1731 (2006). K. E. Slack et al., Mol. Biol. Evol. 23, 1144 (2006). J. W. Brown, J. S. Rest, J. Garcia-Moreno, M. D. Sorenson, D. P. Mindell, BMC Biol. 6, 6 (2008). A. Cooper, C. Mourer-Chauvire, G. K. Chambers, A. Von Haeseler, A. Wilson, S. Pääbo, Proc. Natl. Acad. Sci. U.S.A. 89, 8741 (1992). A. Härlid, A. Janke, U. Árnason, J. Mol. Evol. 46, 669 (1998).

Waterfowl and gamefowl (Galloanserae) Sérgio L. Pereiraa,* and Allan J. Bakera,b a

Department of Natural History, Royal Ontario Museum, 100 Queen’s Park Crescent, Toronto, ON, Canada; bDepartment of Ecology and Evolutionary Biology, University of Toronto, Toronto, ON, Canada *To whom correspondence should be addressed (sergio.pereira@ utoronto.ca)

Abstract The Galloanserae is a monophyletic group containing 442 species and 129 genera of Anseriformes (waterfowl) and Galliformes (gamefowl). The close relationship of these two orders and their placement as the closest relative of Neoaves are well supported. Molecular time estimates and the fossil record place the radiation of living Galloanserae in the late Cretaceous (~90 million years ago, Ma) and the radiation of modern genera and species in the Cenozoic (66–0 Ma). The basal split within Galloanserae is coincident with the breakup of Gondwana, and later diversification occurred following subsequent dispersal to other continental masses in the Eocene (56–34 Ma).

Galloanserae is an ancient clade of birds including the Orders Anseriformes (waterfowl; Fig. 1) and Galliformes (gamefowl). Waterfowl are strong flyers, and most members are also good swimmers. There are about 48 extant genera and 161 species distributed in four families (1). The Family Anhimidae includes three species of South American screamers, named after loud calls emitted when threatened. The monotypic Anseranatidae includes the Australian Magpe Goose, which is unusual among waterfowl because it is the least aquatic species, has partial molt of flight feathers and copulates on land instead of water. The Family Dendrocygnidae groups all eight species of tropical and subtropical whistling ducks in only one genus. Whistling ducks have long necks and legs, and longer hind toes than most ducks. Unlike other families in the order, the Anatidae is more diverse, with 45 genera and about 150 species of geese and swans (Subfamily Anserinae) and true ducks (Subfamily Anatinae). They have fully webbed feet and a worldwide distribution. The family can be divided in two subfamilies: Anserinae (geese and swans) and Anatinae (true ducks).

The Order Galliformes is more diverse than Anseriformes. There are about 281 species and 81 genera of gamefowl (1). Galliformes is classified into five families. The Megapodiidae includes 21 species of megapodes distributed in six genera found in the Australasian region. The family includes the scrub fowl, brush-turkeys, and Mallee Fowl, also known collectively as mound-builders because of their habit of burying their eggs under mounds of decaying vegetation. The Cracidae is a Neotropical group of 10 genera and about 50 species of forest-dwelling birds, including curassows, guans, and chachalacas. Cracids have blunt wings and long broad tails, and may have brightly colored ceres, dewlaps, horns, brown, gray or black plumage, and some have bright red or blue bills and legs. Numididae includes four genera and six species of guinefowl found in sub-Saharan Africa. Guineafowl have most of the head and neck unfeathered, brightly colored wattles, combs or crest, and a large bill. The Family Odontophoridae harbors about 32 species of new world quails, which are medium-sized birds, with short, powerful wings, “toothed” bill, and lack of tarsal spurs. The remaining members of the Galliformes are placed in the Family Phasianidae, which is the most diverse within Galliformes. It includes chicken, grouse, partridges,

Fig. 1 A Mute Swan (Cygnus olor), Family Anatidae, from Lake Ontario, Toronto. Credit: S. L. Pereira.

S. L. Pereira and A. J. Baker. Waterfowl and gamefowl (Galloanserae). Pp. 415–418 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

416


Phasianidae

7

Numididae

5

Cracidae

4

Galliformes

Odontophoridae 8

Anatidae

1

6

Dendrocygnidae

3

Anseranatidae

2

Anhimidae Cretaceous

Paleogene

MESOZOIC 100

Anseriformes

Megapodiidae

Neogene

CENOZOIC 50

0 Million years ago

Fig. 2 A timetree of waterfowl and gamefowl (Galloanserae). Divergence times are shown in Table 1.

pheasants, and turkeys. In general, phasianid birds are medium-sized to large birds, with short, rounded wings, short toes with blunt claws, and raised hallux. Many subdivisions within the Phasianidae have been proposed, with some subfamilies sometimes considered as separate families (e.g., turkeys and grouse placed in the Families Meleagrididae and Tetraonidae, respectively). Although a Cenozoic origin for Galloanserae has been long hypothesized (e.g., 2), based on fragmentary and incomplete specimens, definitive evidence was found only recently (3). Vegavis iaii, the oldest known anseriform fossil from the Maastrichtian stage of the late Cretaceous, is closely related to the lineage of ducks and geese. This finding implies that modern anseriform families, and hence their closest living relative, the Galliformes, were already independent lineages in the late Cretaceous (3). Other fossils of extinct lineages with phylogenetic affinities to extant Galloanserae are known from the Eocene, and modern genera from both orders only show up in the fossil record in the Neogene (23–0 Ma) (2, 3). Here, the systematic and phylogenetic relationships of Galloanserae are reviewed briefly, and estimates of the divergence times for major lineages of Anseriformes and Galliformes derived from molecular data are summarized (Fig. 2). The first suggestion of a close relationship between Anseriformes and Galliformes can be dated back to about 140 years ago (4), and constantly reaffirmed since then by phylogenetic analyses of nonmolecular (e.g., 5, 6) and molecular data (e.g., 7–9). Despite sporadic disputes (e.g., 10), an analysis of 2954 morphological characters scored across 150 extant birds provided compelling evidence

for the monophyly of Galloanserae and indicated that Neoaves is its closest relative (11). This analysis identified nine derived morphological characters supporting Galloanserae as a natural (monophyletic) group (11). Morphological (6, 11, 12) and molecular data (9, 13) have provided strong evidence for family relationships within Anseriformes, placing Anhimidae as the closest relative of Anseranatidae and the remaining families, Anatidae and Dendrocygnidae as each other’s closest relative (Fig. 2). Despite considerable effort to establish the phylogenetic relationships for some groups, such as geese and swans (14) and some ducks (15–17), based on molecular and nonmolecular characters, phylogenetic relationships at the tribal and generic levels are yet not fully resolved within the Subfamily Anatinae (12, 13, 15, 18). This is likely due to insufficient character sampling, leading to poor overall resolution across tribes and genera. The phylogenetic relationships of Galliformes have received considerably more attention than Anseriformes. It is widely accepted that Megapodiidae is the closest relative of all galliform families, followed by Cracidae (11, 19–21). This contrasts with earlier results from DNA–DNA hybridization studies that placed Cracidae and Megapodiidae as reciprocally monophyletic groups (the Craciformes), which in turn were considered the closest relative of remaining galliform families (22). Additionally, there is considerable debate whether Odontophoridae or Numididae is closest to Phasianidae (reviewed in 11, 19, 20). Combined analyses of morphological and behavioral data and DNA sequences (19), or

Eukaryota; Metazoa; Vertebrata; Aves; Galloanserae

417

Table 1. Molecular time estimates (Ma) and their confidence/credibility intervals (CI) among waterfowl and gamefowl (Galloanserae). Timetree Node

Estimates Ref. (8)

Time

Ref. (19)

Ref. (21)

Ref. (35)

Time

CI

Time

CI

Time

CI

Ref. (41)

Time

CI

Time

1

106.9

101.0

113–92

–

–

105.0

135–84

114.6

130–97

–

2

97.9

–

–

–

–

91.0

119–72

104.7

121–86

–

3

97.2

92.1

103–82

–

–

–

–

102.3

119–83

–

4

95.4

86.6

96–79

107.0

122–91

95.6

112–77

103.1

120–85

84.9

5

88.8

–

–

92.8

107–79

88.5

113–71

95.9

114–78

77.8

6

63.2

–

–

–

–

–

–

63.2

84–43

–

7

59.8

52.4

59–48

60.2

71–53

64.1

85–53

72

90–53

50.3

8

62.4

–

–

55.5

66–50

68.8

92–55

66.1

85–48

59.3

Note: Node times in the timetree represent the mean of time estimates from different studies. Data analyzed were complete mitochondrial genomes (8), partial mitochondrial sequences (21, 35, 41), and combined analysis of nuclear and mitochondrial genes (19).

separate analyses of morphological (20) or molecular (23) data, however, seem to point out to a closer relationship between Odontophoridae and Phasianidae (Fig. 2). Additionally, the phylogenetic relationships within Phasianidae are still unsettled (19). Many phylogenetic hypotheses at the genus and species level have also been proposed for many groups, including megapodes (24), cracids (25–27), grouse (28, 29), and some clades within Phasianidae (28, 30–32). In general, the phylogenetic hypotheses based on DNA sequences have provided stronger support than analyses using morphological data for relationships within Anseriformes (12, 13) and Galliformes (19–21, 33), regardless of the taxonomic level under scrutiny. The molecular time estimates within Galloanserae are largely congruent among most studies (8, 19, 21, 34, 35), with appreciable overlap in age uncertainties measured by credible or confidence intervals (Table 1). Estimates suggesting a post-Cretaceous radiation of Galloanserae families have been obtained in three studies that used a single calibration point to calibrate the molecular clock (33, 36, 37), and in a fourth study (38) that used some inappropriate fossil constraints (34), and imposed a maximum age for the root of the timetree (i.e., split between Paleognathae from other birds, including Galloanserae) at 95 Ma, without considering published molecular estimates for an earlier age (e.g., 8, 34, 39). On the other hand, molecular time estimates derived from complete or partial sequences of the mitochondrial genome using multiple fossil constraints and a Bayesian

approach that accounts for uncertainties in both phylogenetics and fossil constraints (8, 21, 35) have placed the split between Anseriformes and Galliformes at the end of the early Cretaceous (Fig. 1). Similar conclusions have been reached using nuclear genes evolving at a constant rate among vertebrates (39, 40) and DNA–DNA hybridization studies (22). The fossil record supports the molecular time estimates (3). The radiation of living Anseriformes is dated at ~97 Ma (or earlier according to a DNA–DNA hybridization study; 36), with the split of Anhimidae from other Anseriformes. The split of Anseranatidae from other anseriforms also occurred early in the history of Anseriformes around the same time (Table 1). The only divergence time available for the split between Dendrocygnidae and Anatidae based on DNA–DNA hybridization (36) places their split at 63.2 Ma. Based on mitochondrial genomes and multiple fossil constraints, the radiation of extant Galliformes is contemporaneous with that of crown Anseriformes, and was placed at ~95 Ma (Table 1) with the separation of the Megapodiidae from the other families (Fig. 2). However, estimates based on DNA–DNA hybridization (36) and combined analysis of mitochondrial and nuclear DNA sequences (27), both using single-point calibrations for the molecular clock, suggest a much younger date (76–59 Ma). The split between Cracidae and the remaining Galliformes occurred at ~88 Ma, also before the Mesozoic–Cenozoic transition. The molecular time estimates for Numididae and Odontophoridae are

418


incongruent across studies (Table 1). As pointed out earlier, the placement of Odontophoridae and Numididae relative to Phasianidae has been contentious (19, 20, 41). Strong support for the close relationship between Odontophoridae and Phasianidae has been achieved only recently with increased character and taxonomic sampling (11, 19). Hence, we have chosen time estimates for nodes 7 and 8 (Fig. 2) from ref. (19). Some studies have shown that the radiation of galliform genera occurred mostly during the Eocene and Oligocene, and speciation occurred mostly in the Miocene to Pleistocene (25–27, 30, 42). The radiation of living Anseriformes and Galliformes as estimated from molecular data (Table 1) and supported by some fossil evidence (3) is consistent with a vicariant mode of diversification due to the breakup of Gondwana during the Mesozoic (251–66 Ma), and continental drift throughout the Cenozoic (66–0 Ma). The first two cladogenic events within both orders isolated the South American and Australian waterfowl and gamefowl from their relatives now found in other Gondwanan and Laurasian landmasses. Dispersal to these other landmasses likely started during the Eocene (56–34 Ma) (19, 21), and favored the diversification of dispersing lineages as they colonized new continents and adapted to new ecological niches throughout the Cenozoic.

Acknowledgments Support was provided by Canada Natural Sciences and Engineering Research Council, Royal Ontario Museum Foundation, and U.S. National Science Foundation.

References 1. C. G. Sibley, B. L. Monroe, Jr., Distribution and Taxonomy of Birds of the World (Yale University Press, New Haven, 1990). 2. G. J. Dyke, M. van Tuinen, Zool. J. Linn. Soc. 141, 153 (2004). 3. J. A. Clarke, C. P. Tambussi, J. I. Noriega, G. M. Erickson, R. A. Ketcham, Nature 433, 305 (2005). 4. T. H. Huxley, Proc. Zool. Soc. Lond. 1867, 415 (1867). 5. J. Cracraft, Auk 98, 681 (1981). 6. G. Mayr, J. Clarke, Cladistics 19, 527 (2003). 7. J. G. Groth, G. F. Barrowclough, Mol. Phylogenet. Evol. 12, 115 (1999). 8. S. L. Pereira, A. J. Baker, Mol. Biol. Evol. 23, 1731 (2006).

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

G. C. Gibb, O. Kardailsky, R. T. Kimball, E. L. Braun, D. Penny, Mol. Biol. Evol. 24, 269 (2007). P. G. Ericson, J. Avian Biol. 27, 195 (1996). B. C. Livezey, R. L. Zusi, Zool. J. Linn. Soc. 149, 1 (2007). B. C. Livezey, Ann. Carnegie Mus. 66, 457 (1997). C. Donne-Gousse, V. Laudet, C. Hanni, Mol. Phylogenet. Evol. 23, 339 (2002). B. C. Livezey, Syst. Biol. 45, 415 (1996). K. P. Johnson, M. D. Sorenson, Auk 116, 792 (1999). B. C. Livezey, Condor 97, 233 (1995). B. C. Livezey, Wilson Bull. 107, 214 (1995). M. Sraml, L. Christidis, S. Easteal, P. Horn, C. Collet, Aust. J. Zool. 44, 47 (1996). T. M. Crowe et al., Cladistics 22, 495 (2006). G. J. Dyke, B. E. Gulas, T. M. Crowe, Zool. J. Linn. Soc. 137, 227 (2003). S. L. Pereira, A. J. Baker, Mol. Phylogenet. Evol. 38, 499 (2006). C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds: A Study in Molecular Evolution (Yale University Press, New Haven, 1990). D. E. Dimcheff, S. V. Drovetski, D. P. Mindell, Mol. Phylogenet. Evol. 24, 203 (2002). S. M. Birks, S. V. Edwards, Mol. Phylogenet. Evol. 23, 408 (2002). E. T. Grau, S. L. Pereira, L. F. Silveira, E. Hofl ing, A. Wajntal, Mol. Phylogenet. Evol. 35, 637 (2005). S. L. Pereira, A. J. Baker, Auk 121, 682 (2004). S. L. Pereira, A. J. Baker, A. Wajntal, Syst. Biol. 51, 946 (2002). V. Lucchini, J. Hoglund, S. Klaus, J. Swenson, E. Randi, Mol. Phylogenet. Evol. 20, 149 (2001). S. V. Drovetski, J. Biogeogr. 30, 1173 (2003). E. Randi et al., Mol. Phylogenet. Evol. 19, 187 (2001). E. Randi et al., Auk 117 (2000). R. T. Kimball, E. L. Braun, P. W. Zwartjes, T. M. Crowe, J. D. Ligon, Mol. Phylogenet. Evol. 11, 38 (1999). G. L. Harrison et al., Mol. Biol. Evol. 21, 974 (2004). J. W. Brown, R. B. Payne, D. P. Mindell, Biol. Lett. 3, 257 (2007). J. W. Brown, J. S. Rest, J. Garcia-Moreno, M. D. Sorenson, D. P. Mindell, BMC Biol. 6, 6 (2008). M. van Tuinen, T. A. Stidham, E. A. Hadly, Hist. Biol. 18, 205 (2006). K. E. Slack et al., Mol. Biol. Evol. 23, 1144 (2006). P. G. P. Ericson et al., Biol. Lett. 2, 543 (2006). M. van Tuinen, S. B. Hedges, Mol. Biol. Evol. 18, 206 (2001). S. Kumar, S. B. Hedges, Nature 392, 917 (1998). M. van Tuinen, G. J. Dyke, Mol. Phylogenet. Evol. 30, 74 (2004). R. T. Kimball, E. L. Braun, J. D. Ligon, Proc. Biol. Sci. 264, 1517 (1997).

Advanced birds (Neoaves) Marcel van Tuinen Department of Biology and Marine Biology, 601 South College Road, University of North Carolina at Wilmington, Wilmington, NC 28403-5915, USA ([email protected])

Abstract Neoaves, the largest superorder of living birds (Neornithes), consists of 16–24 orders and ~9000 species. Although recent progress has been made, available molecular data continue to show remarkable lack of phylogenetic resolution and the basal splits within Neoaves are still uncertain. The neoavian timetree shows an initial divergence at ~95 million years ago (Ma) followed by Cretaceous (87–75 Ma) diversification of southern hemispheric orders and younger times for northern and aquatic orders (Paleogene, 65–30 Ma). The timetree thus implicates possible roles for continental breakup (Cretaceous) and climate (Paleogene) in the diversification of advanced birds.

Neoavian birds are included in the Subclass Neornithes, and most broadly can be defined as those orders with advanced flight capabilities. Such birds include the arboreal songbirds, cuckoos, parrots, and woodpeckers, the nocturnal owls and nightjars, aerial fliers such as swifts and hummingbirds, as well as several other distinct lineages including pigeons, raptors, shorebirds, wading birds, and marine birds (Fig. 1). Excluded from Neoaves are the primitive flightless paleognath landbirds (ratites), the primitive paleognath and neognath volant landbirds (tinamous and landfowl), and primitive neognath waterbirds (waterfowl). Approximately 9000 living species of Neoaves have been described and placed in 16–24 orders. Traditionally, neognath orders with limited aerial flight capability (penguins, loons, grebes) were thought to be primitive and placed ahead of other bird orders in classifications (1–3). The unit Neoaves was first described from molecular data from the 1970s and 1980s (4, 5), and its monophyly continued to gain support from DNA sequence data (6–11). Although initial complete mitochondrial genome data supported neoavian paraphyly with passerines as the earliest diverging modern birds,

this result is now considered an artifact of limited taxon sampling (7, 15). Phylogenetic resolution among the main divergences within Neoaves continues to remain a major hurdle (10), with most neoavian orders appearing to have diverged in close succession. This “neoavian comb” (10) on the one hand has been interpreted as evidence for a real (hard) polytomy among most, if not all, neoavian orders (12), indicating a rapid evolutionary radiation. Others have maintained that additional taxonomic and nucleotide sampling will provide added resolution (10). At present, three molecular studies (7, 13, 14) exist that have included a combination of both complete neoavian ordinal sampling and nuclear gene sequencing. Two of these studies were based on nuclear gene

Fig. 1 A Great Blue Heron (Ardea herodias), Order Ciconiiformes, from Canada. Credit: M. Peck.

M. van Tuinen. Advanced birds (Neoaves). Pp. 419–422 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

420

THE TIMETREE OF LIFE Phoenicopteriformes 5

Podicipediformes Gaviiformes

4

Procellariiformes Sphenisciformes Apodiformes

3

Trochiliformes

2

'Caprimulgiformes' 'Pelecaniformes' 'Ciconiformes' Charadriiformes 'Gruiformes'

1

'Falconiformes' Piciformes Coraciiformes Passeriformes Psittaciformes Coliiformes Trogoniformes Strigiformes Musophagiformes Cuculiformes Columbiformes Cretaceous

Paleogene

100

75

Neogene

CENOZOIC

MESOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of advanced birds (Neoaves). Divergence times are from Table 1.

sequences of single genes (18S rRNA or beta-fibrinogen intron) with or without additional mitochondrial DNA sequences. Both studies differed considerably in phylogenetic resolution among neoavian orders. The study larger in nucleotide sampling (7) showed limited resolution across Neoaves. Instead, the second study (14) based on intron seven of the beta-fibrinogen gene was more extensive in taxon sampling and claimed support for a broad division within neoavian birds, that are between two newly defined groups, Metaves and Coronaves. Several traditional orders were broken up by this new classification, which suggested a provocative scenario for convergent evolution in morphology and ecology across many extant bird orders. For example, Metaves included some

enigmatic gruiform families while the better-known gruiform cranes and rails were placed in Coronaves. The enigmatic Hoatzin was allocated to Metaves, while their supposed cuculiform and musophagiform relatives were placed in Coronaves, also clustering among several waterbird orders. Finally, the pelecaniform tropicbirds were placed with flamingos, grebes, sandgrouse, swifts, hummingbirds, and caprimulgiform birds in Metaves while other pelecaniforms, shorebirds, and owls were allocated to Coronaves. The most recent nuclear study (13) was based on five nuclear genes and extensive taxon sampling across every order. These new sequence data combined with the published beta-fibrinogen sequences also supported

Eukaryota; Metazoa; Vertebrata; Aves; Neoaves

421

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among advanced birds (Neoaves). Timetree Node

Time

Estimates Ref. (13)

Ref. (18)

Ref. (20)

Ref. (21)

Ref. (23)

Ref. (24)

Ref. (25)

Time

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

97.3

95

–

–

118.5

132–105

77

83–72

103.8

111–97

89.3

108–70

100

123–83

2

72.7

62

–

–

–

–

67

69–65

–

–

–

–

89

112–73

3

66.7

45

–

–

80.5

100–61

61

63–60

67.9

73–63

75.7

91–60

70

88–55

4

61.8

58

71.1

76–66

–

–

–

–

41.2

44–38

–

–

77

97–67

5

53.0

31

–

–

52

70–30

–

–

–

–

–

–

76

96–61

1

Note: Node times in the timetree represent the mean of time estimates from different studies. Divergence times were obtained from an analysis of ribosomal mitochondrial genes (24), partial (20), or complete (16, 18, 19) mitochondrial genomes, two nuclear introns (21), five nuclear genes (13, 25), or DNA–DNA hybridization distances (23).

the distinction between Metaves and Coronaves, but only when the fibrinogen sequence data were included. Expanded mitochondrial studies (15–20), and nuclear studies (21) based on incomplete sampling among neoavian orders also fail to show support for Metaves or Coronaves. Thus, these studies exemplify the current status of neoavian molecular systematics. Molecular sequence studies (including expanded mitochondrial gene studies, 15–20) have been unable to provide solid support for the resolution of major ordinal groupings within Neoaves. Nonetheless, there has been consensus support for a grouping of flamingos (Ciconiiformes, Phoenicopteridae) and grebes (Podicipediformes, Podicipedidae) (10, 11, 13, 19, 20, 22); for joining hummingbirds (Trochiliformes) with swifts (Apodiformes) and imbedding this grouping within a paraphyletic Caprimulgiformes (10, 13, 20, 21) and for joining penguins (Sphenisciformes), tubenosed birds (Procellariiformes), and loons (Gaviiformes) together (5, 10, 13, 22, 23). A variety of studies have estimated divergence times among neoavian orders (Table 1). The first study that estimated divergence times from a complete ordinal data set used a concatenated nonprotein-coding portion of the mitochondrial genome (2 rRNAs, 3 tRNAs) and a lineage-specific method (24). The timing of the neoavian divergences agreed closely with that reported in two of three mitogenomic studies (16, 18). This study also confirmed a rapid Cretaceous radiation involving the origination of all neoavian orders starting at 90 Ma. Application of the same calibration method on a taxonomically much larger DNA–DNA hybridization data set suggested a rapid diversification of Gondwanan orders (85–75 Ma) following the origin of Neoaves (104 Ma), but

younger ages for Laurasian orders (75–60 Ma), and the entirely Paleogene origin (55–30 Ma) of several traditional waterbird orders (23). A second timetree (13) of Neoaves based on multiple (five) nuclear genes and comprehensive ordinal sampling was constrained on several nodes internal to Neoaves and used two separate rate-smoothing methods. Results from one of the smoothing approaches (penalized likelihood) agreed well with previous timing results. An alternative timetree was presented that showed considerably younger origination times for neognath orders, and this timetree was preferred because of the better agreement with the fossil record. However, others have used the same and additional data to argue for older Cretaceous ages of many neoavian groups and pointed to an average age of 110 Ma for the first divergence among extant Neoaves (20, 25). Resolution of which of the two timetrees is more accurate is essential to understanding the evolutionary tempo and mode of Neoaves. One timetree scenario (Fig. 2) shows consistency across all molecular data sets in mid-Cretaceous ordinal origination and superordinal diversification. It correlates well with the timing of Gondwanan biogeography, which has direct fossil support for Cretaceous origination of galloanserine orders and superordinal diversification, but it also imposes considerable fossil gaps for many lineages. The second scenario shows better consistency between nuclear DNA sequence and fossil divergence times and suggests that neoavian orders originated rapidly and diversified in the Paleogene. Similar discrepancy exists in the available time estimates for the origin of Caprimulgiformes, Apodiformes, Procellariiformes, and Podicipediformes (Table 1). It is

422


yet unclear whether any of these orders originated in the Cretaceous or Paleogene. A case in point concerns the origin of Apodiformes, which was based on six time estimates from mtrRNA, mtDNA, nuclear exons, and nuclear introns. While BEAST and PATHd8, respectively, provide the maximum (81 Ma) and minimum (45 Ma) estimates for Apodiformes, all confidence/credibility intervals (CI) overlap with the Cretaceous–Paleogene boundary. Without consistency in calibration, gene sampling and timing methodology, the source of the varying time estimates among studies remains as yet unclear. Resolution will likely come from additional nuclear gene sequences, and more Cretaceous and Paleocene fossil material.

Acknowledgments Support was provided by Netherlands Organization for Scientific Research (NWO) and the Epply foundation.

References 1. T. H. Huxley, Proc. Roy. Soc. Lond. 1867, 415 (1867). 2. M. Fürbringer, Untersuchungen zur Morphologie und Systematik der Vögel (Von Holkema, Amsterdam, 1888). 3. A. Wetmore, Smithsonian Misc. Coll. 139, 1 (1960). 4. C. Y. K. Ho, E. M. Prager, A. C. Wilson, D. T. Osuga, R. E. Feeney, J. Mol. Evol. 8, 271 (1976). 5. C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds: A Study in Molecular Evolution (Yale University Press, New Haven, 1990).

6. J. G. Groth, G. F. Barrowclough, Mol. Phylogenet. Evol. 12, 115 (1999). 7. M. van Tuinen, C. G. Sibley, S. B. Hedges, Mol. Biol. Evol. 17, 451 (2000). 8. J. Garcia-Moreno, M. D. Sorenson, D. P. Mindell, J. Mol. Evol. 57, 27 (2003). 9. U. S. Johannson, P G. P. Ericson, J. Avian Biol. 34, 185 (2003). 10. J. Cracraft et al., Assembling the Tree of Life (Oxford University Press, New York, 2004), pp. 468–489. 11. A. L. Chubb, Mol. Phylogenet. Evol. 30, 140 (2004). 12. S. Poe, A. L. Chubb, Evolution 58, 404 (2004). 13. P. G. P. Ericson et al. Biol. Lett. 4, 543 (2006). 14. M. G. Fain, P. Houde, Evolution 58, 2558 (2004). 15. E. L. Braun, R. T. Kimball, Syst. Biol. 51, 614 (2002). 16. T. Paton, O. Haddrath, A. J. Baker, Proc. Roy. Soc. Lond. B Bio. 269, 839 (2002). 17. S. L. Pereira, A. J. Baker, Mol. Biol. Evol. 23, 1731 (2006). 18. K. E. Slack, C. M. Jones, T. Ando, G. L. Harrison, E. Fordyce, U. Arnason, D. Penny, Mol. Biol. Evol. 23, 1144 (2006). 19. M. Morgan-Richards et al. BMC Evol. Biol. 8, 20 (2008). 20. J. W. Brown, J. S. Rest, J. Garcia-Moreno, M. D. Sorenson, D. P. Mindell, BMC Biol. 6, 6 (2008). 21. J. L. Chojnowski, R. T. Kimball, E. L. Braun, Gene 410, 89 (2008). 22. M. van Tuinen, D. B. Butvill, J. A. W. Kirsch, S. B. Hedges. Proc. Roy. Soc. Lond. B Bio. 268, 1345 (2001). 23 M. van Tuinen, T. A. Stidham, E. A. Hadly, Hist. Biol. 18, 205 (2006). 24. M. van Tuinen, S. B. Hedges, Mol. Biol. Evol. 18, 206 (2001). 25. J. W. Brown, R. B. Payne, D. P. Mindell, Biol. Lett. 3, 257 (2007).

Passerine birds (Passeriformes) Joel Cracrafta,* and F. Keith Barker b a

Department of Ornithology, American Museum of Natural History, Central Park West at 79th St., New York, NY 10024, USA; bBell Museum of Natural History, University of Minnesota, 1987 Upper Buford Circle, St. Paul, MN 55108, USA *To whom correspondence should be addressed ([email protected])

Abstract The Order Passeriformes is the largest clade of its rank in birds, encompassing from roughly 40–100 families depending on the classification. In recent years molecular systematic data have greatly clarified interfamilial relationships, although many nodes remain poorly supported and it is clear that numerous traditional families are not monophyletic. Passeriformes is an old group, and most molecular dating studies estimate its age of origin to be late Cretaceous (100–66 million years ago, Ma) on Gondwana, with early lineages being partitioned among New Zealand, Australasia, and South America. The major lineages arose in the Paleogene (66–23 Ma) and then diversified in the Neogene (23–0 Ma).

The Passeriformes constitutes the largest order within modern birds (Neornithes) and includes the suboscines (Tyranni; Old World and New World lineages) and the Passeri (oscine songbirds). Passeriforms represent about 60% of extant avian diversity and estimates range from about 5700 biological species (1) to perhaps 15,000 or more diagnosably distinct taxa, or phylogenetic species (Fig. 1). Over the years, systematists have recognized a variable number of family-ranked taxa, from 44 to 96 (1–4). New molecular sequencing studies are discovering that a substantial number of conventional “families” are not monophyletic, and accordingly it is certain passeriform classification will be a subject of great interest in the coming years. Here, we review recent advances in passerine phylogenetics and use molecular sequence data to estimate the temporal pattern of diversification within the group. It has been accepted for some time now that birds are divided into the paleognaths (tinamous and ratites) and neognaths (all other birds), with the latter being subdivided into the Galloanserae (duck-like and chicken-like

birds) and its closest relative, the Neoaves. Relationships within Neoaves have been controversial and difficult to resolve (5–9). Many workers have placed passeriforms close to the so-called “higher land birds,” particularly the monophyletic Piciformes and taxa of the non-monophyletic “Coraciiformes” (6–11), but some molecular analyses, including DNA hybridization (5) and whole mitochondrial genomes (12), have placed them deeper toward the base of the Neoaves and even— because of a very small taxon sample and a spurious root—at the base of all birds (13). The monophyly of the passeriforms has never been seriously questioned (14). Morphologists of the nineteenth century used syringeal characters to establish the major oscine/suboscine divisions, and through the twentieth-century systematists (2, 15, 16) carried on the process of clustering groups of families together, primarily on the basis of overall similarity and geography and without much new character data. Although the suboscine/ oscine division has been widely accepted, the placement of the New Zealand wrens (Acanthisittidae) relative to

Fig. 1 A Wood Thrush (Hylocichla mustelina), Family Turdidae, from North America. Credit: J. Cracraft.

J. Cracraft and F. Barker. Passerine birds (Passeriformes). Pp. 423–431 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

424


Aegithalidae 54

Hirundinidae Cisticolidae

49 71

Zosteropidae

61 37 50

Sylviidae

Sylvoidea

Timaliidae

53

Pycnonotidae

28

Alaudidae Stenostiridae 36

Paridae Promeropidae

82

Passeri

Cardinalidae 83

Thraupidae Emberizidae

81

Icteridae 30

Parulidae Fringillidae

74

Motacillidae

72

Passeridae

63

Passeroidea

80

85

Passerida

23

Ploceidae

59

Prunellidae

41

Irenidae 39

Dicaeidae 55

Nectariniidae

27 (continued on next page)

Cretaceous

Paleogene CENOZOIC

MESOZOIC 75

Neogene

50

25

0 Million years ago

Fig. 2 Continues

those two lineages was uncertain before the emergence of molecular data (17). DNA hybridization distances (5) suggested that Acanthisitta was at the base of the passeriforms but that position was never tested rigorously with outgroup comparisons. Recent DNA sequencing studies, however, have strongly supported acanthisittids as the closest relatives of the suboscines + oscines (18–20). This finding has been crucial as it has provided the basis for calibrating divergence times within the passeriforms, as discussed later. The division between the Old World suboscines (Eurylaimides) and their New World counterpart (Tyrannides + Furnariides) is well supported (18–23). With ~1200 species, the New World suboscines are the most

diverse lineage of South American birds. They are divided into two major subclades (18–20, 24), the Tyrannides (flycatchers, cotingas, and manakins; Tyranni of 17) and the Furnariides (ovenbirds, woodcreepers, antbirds, tapaculos, antpittas, and allies; Furnarii of 17 and the Furnariida + Thamnophilida of 1). Because of limited taxon sampling in currently published results, many relationships within both of these groups are subject to uncertainty. The overwhelming taxonomic and morphological diversity of oscines (Passeri) has long presented a challenge to understanding their relationships, but in the last few years DNA sequence studies have begun to reveal the overall phylogenetic structure of this group, even though

Eukaryota; Metazoa; Vertebrata; Aves; Passeriformes

425


Bombycillidae 68

Ptilogonatidae

19

Turdidae 65

Cinclidae

56

Mimidae 78

17

Sturnidae Regulidae

31

Passerida

34 32

Muscicapoidea

Muscicapidae

58

Certhiidae 48

Sittidae

42

Polioptilidae 64

Troglodytidae Petroicidae Picathartidae

70

Passeri

Aegithinidae

Prionopidae 76

Vangidae

57

Platysteiridae

Malaconotines

Malaconotidae

69

Artamidae 79

Corvoidea

Cracticidae Rhipiduridae Corcoracidae

77 15

Melampittidae

75

Paradisaeidae

73

Grallinidae 66

84

60

Monarchidae

Corvines

62 46 43

Laniidae 67

Corvidae

44

Dicruridae


Cretaceous

Paleogene

75

Neogene

CENOZOIC

MESOZOIC 50

25

0 Million years ago

Fig. 2 Continues

there remain numerous poorly supported nodes across the tree. Sequences from RAG1 and RAG2 nuclear loci for a large taxon sample, including all but one passeriform family (18, 19), have established that the diverse Australasian “Corvida” that was proposed on the basis of DNA hybridization distances (5) is paraphyletic, and

that some of these lineages are, in fact, successive closest relatives of all remaining oscines. Importantly, however, a monophyletic assemblage (Corvoidea) within the “corvidans” has been found to be the closest relative of all remaining oscines (roughly speaking, Sibley and Monroe’s (1) clade Passerida). Within the latter, at least

426


Oriolidae 52

Paramythiidae

40 13

Colluricinclidae

45 47

Pachycephalidae Corvoidea

Falcunculidae Campephagidae 35

Vireonidae

24

14 16

Daphoenosittidae Passeri

Cnemophilidae 26

8

Callaeatidae

22

Melanocharitidae Orthonychidae

Meliphagidae 25

Pardalotidae

20

Maluridae

4

Meliphagoidea

Pomatostomidae 6

Ptilonorhynchidae 11

Climacteridae Menuridae Dendrocolaptidae 51

Rhinocryptidae

12

Conopophagidae

10

Thamnophilidae

Tityridae

33

1

Cotingidae 29

3

Pipridae Philepittidae

9

Eurylaimidae

7

Pittidae Acanthisittidae Cretaceous

Paleogene CENOZOIC

MESOZOIC 75

Neogene

50

25

0 Million years ago

Fig. 2 A timetree of passerine birds (Passeriformes). Divergence times are shown in Table 1.

Eurylaimides

38

Tyrannides

Tyrannidae

5

Tyranni

Formicariidae 18

2

Furnariides

Furnariidae

21


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among passerine birds (Passeriformes). Timetree Node

Estimates Ref. (12)

Time

Time

Ref. (19)(a)

CI

Time

Ref. (19)(b)

CI

Time

Ref. (38)

CI

Time

CI

1

82.0

95.3

107–85

82.0

–

82.0

–

82

–

2

77.1

91.8

103–81

77.4

81–73

76.8

80–72

–

–

3

69.7

–

–

70.5

78–65

68.8

73–60.9

–

–

4

63.9

–

–

64.7

70–60

63.1

67–57

–

–

5

63.7

–

–

64.6

75–55

62.7

67–56

–

–

6

60.5

–

–

61.3

72–54

59.7

66–53

–

–

7

56.3

–

–

57.3

67–47

55.2

60–47

–

–

8

53.0

–

–

54.6

67–49

51.4

60–47

–

–

9

51.9

–

–

54.4

–

49.3

–

–

–

10

51.2

–

–

53.1

–

49.3

–

–

–

11

51.0

–

–

53.4

70–48

48.5

56–42

–

–

12

50.4

–

–

52.3

–

48.5

–

–

–

13

48.3

–

–

49.7

–

46.9

–

–

–

14

47.9

46.7

54–40

48.5

63–43

47.2

55–41

46.7

54–40

15

47.2

–

–

48.5

–

45.8

–

–

–

16

46.4

–

–

47.1

62–42

45.6

62–39

–

–

17

46.2

–

–

46.7

53–42

45.7

50–40

44.5

52–41

18

46.0

–

–

47.7

–

44.2

–

–

–

19

45.0

–

–

45.3

51–39

44.6

49–39

–

–

20

44.9

–

–

45.8

58–40

44.0

52–37

–

–

21

43.9

–

–

45.6

–

42.2

–

–

–

22

41.1

–

–

42.7

–

39.5

–

–

–

23

40.8

–

–

40.7

54–15

40.8

49–36

–

–

24

39.5

–

–

40.4

45–35

38.6

42–32

–

–

25

39.3

–

–

41.1

–

37.5

–

–

–

26

39.3

–

–

41.0

–

37.6

–

–

–

27

39.2

–

–

38.2

47–14

40.2

48–35

–

–

28

39.0

–

–

38.8

54–15

39.2

54–15

37.5

42–32

29

38.6

–

–

40.5

–

36.6

–

–

–

30

38.5

–

–

38.5

53–14

38.4

53–14

38.1

43–33

31

38.1

–

–

37.9

–

38.2

–

–

–

32

37.8

–

–

37.7

51–13

37.9

50–13

–

–

33

37.4

–

–

39.2

–

35.6

–

–

–

34

37.3

–

–

37.1

–

37.5

–

–

–

35

36.9

–

–

38.5

–

35.3

–

–

–

36

36.9

–

–

36.5

–

37.2

–

–

–

37

35.1

–

–

34.9

51–13

35.2

51–13

33.8

38–29

38

35.1

–

–

36.9

–

33.2

–

–

–

39

35.0

–

–

35.1

–

34.9

–

–

–

427

428



Estimates Ref. (12)

Time

Time

Ref. (19)(a)

CI

Time

Ref. (19)(b)

CI

Time

CI

Ref. (38) Time

CI

40

34.5

–

–

36.0

–

33.0

–

–

–

41

34.2

–

–

34.2

–

34.1

–

–

–

42

34.0

–

–

34.0

41–27

33.9

38–30

–

–

43

33.4

–

–

34.9

49–28

31.9

39–26

–

–

44

32.7

–

–

34.1

–

31.3

–

–

–

45

32.5

–

–

39.3

–

25.7

–

–

–

47

31.9

–

–

33.4

–

30.4

–

–

–

46

31.9

–

–

33.2

–

30.6

–

–

–

48

31.7

–

–

32.0

–

31.3

–

–

–

49

31.5

–

–

31.0

–

32.0

–

–

–

50

31.1

–

–

30.8

48–12

31.3

48–12

–

–

51

30.3

–

–

31.7

–

28.8

–

–

–

52

30.0

–

–

31.2

–

28.7

–

–

–

53

29.4

–

–

28.7

–

30.0

–

–

–

54

29.4

–

–

29.0

–

29.7

–

–

–

55

29.1

–

–

29.3

–

28.8

–

–

–

56

28.6

–

–

27.4

–

29.7

–

–

–

57

28.5

–

–

29.2

35–23

27.7

31–21

23.1

30–21

58

26.9

–

–

25.6

–

28.1

–

–

–

59

26.6

–

–

26.0

–

27.1

–

–

–

60

26.1

–

–

26.8

–

25.4

–

–

–

61

26.0

–

–

25.4

–

26.6

–

–

–

62

25.6

–

–

26.2

–

24.9

–

–

–

63

25.0

–

–

24.3

–

25.7

–

–

–

64

24.6

–

–

24.8

–

24.4

–

–

–

65

24.3

–

–

23.0

–

25.6

–

–

–

66

24.3

–

–

24.8

–

23.7

–

–

–

67

24.2

–

–

24.5

38–19

23.9

38–17

–

–

68

24.2

–

–

24.4

–

24.0

–

–

–

69

24.0

–

–

25.1

–

22.9

–

–

–

70

23.2

–

–

24.1

–

22.2

–

–

–

71

23.1

–

–

22.5

–

23.7

–

–

–

72

23.1

–

–

22.2

–

23.9

–

–

–

73

22.9

–

–

23.4

–

22.3

–

–

–

74

22.1

–

–

21.2

–

23.0

–

–

–

75

22.0

–

–

22.4

–

21.5

–

–

–

76

21.2

–

–

22.3

–

20.1

–

–

–

77

21.0

–

–

21.3

–

20.6

–

–

–

78

20.8

–

–

20.3

26–15

21.2

26–20

–

–

79

20.3

–

–

21.5

–

19.1

–

–

–


429


Estimates Ref. (12)

Time

Time

Ref. (19)(a)

CI

Time

Ref. (19)(b)

CI

Time

CI

Ref. (38) Time

CI

80

20.0

–

–

18.9

–

21.0

–

–

–

81

16.7

–

–

15.5

–

17.8

–

–

–

82

15.2

–

–

14.1

–

16.3

–

–

–

83

12.6

–

–

11.6

–

13.6

–

–

–

84

11.5

–

–

10.8

–

12.2

–

–

–

85

11.3

–

–

10.3

–

12.2

–

–

–

Note: Node times in the timetree represent the mean of time estimates from the study with the most complete taxon sampling (19). Results from nonparametric rate smoothing (a) and penalized likelihood (b) analysis of the RAG1 and RAG2 nuclear genes are shown from ref. (19). In ref. (38), the same two genes and nonparametric rate smoothing is used to estimate divergence times.

three major groups have been delineated—Passeroidea, Muscicapoidea, and Sylvioidea—but due to their short internodal distances from one another and from several problematic groups, their taxonomic boundaries and relationships are still in flux. The largest taxon sample to date for the corvoids uncovered two large clades (19). One clade, corresponding to Sibley and Ahlquist’s Tribe Corvini, includes the crows and their relatives—the shrikes (Laniidae), monarch flycatchers (Monarchidae, Grallinidae), the birds of paradise (Paradisaeidae) and their relatives (Melampittidae, Corcoracidae), as well as the rhipidurine flycatchers. Their closest relative is a group of shrike-like birds with taxa in Africa (malaconotid bush shrikes, prionopid helmet shrikes, among others), Madagascar (vanga shrikes), and Southeast Asia (aegithinids, among others), all corresponding to Sibley and Ahlquist’s Tribe Malaconotini. Relationships among these taxa are still unsettled (25, 26). There are a number of other corvoid lineages that are more distantly related to these two groups, although their phylogenetic placement generally lacks strong support. The RAG1 and RAG2 data set (19) provided an outline of sylvioid relationships based on a small “family-level” sampling, but recently the group has been sampled more broadly using cytochrome b and myoglobin intron II (27). Unfortunately, the markers used in the latter study have limitations when resolving deep branches and many nodes remain poorly resolved. Moreover, there are some significant topological incongruences in the two studies that will need further data to resolve.

The Passeroidea comprises the New World nineprimaried passerines (buntings, cardinals, warblers, tanagers, and blackbirds) as well as Old World finches, sparrows, wagtails, accentors, sunbirds and flowerpeckers, leafbirds, and sugarbirds (18, 19, 28). Within the New World passeroids RAG1 and RAG2 (19) and a diverse array of nuclear and mitochondrial loci (28) support a close relationship of the icterids + parulids with the thraupids + cardinalids + emberizids. These data also are in general agreement about the relationships of Old World lineages to this New World clade. Of particular significance is the finding that the closest relative of all these passeroid lineages appears to be members of an African clade including the sugarbirds (Promeropidae) and the enigmatic “babblers” Arcanator and Modulatrix. The final major clade of Passerida, and the probable closest relative of the passeroids, is the Muscicapoidea. Nuclear RAG gene sequences support the inclusion of the thrushes (Turdidae), muscicapid flycatchers, dippers (Cinclidae), starlings (Sturnidae), mockingbirds and thrashers (Mimidae), and possibly the more distantly related waxwings and silky-flycatchers (Bombicillidae), kinglets (Regulidae), and a monophyletic lineage of wrens (Troglodytidae), gnatcatchers (Polioptilidae), nuthatches (Sittidae), and creepers (Certhiidae) (18, 19, 29). Nodes at the base of the muscicapoids, however, are not well supported, hence the monophyly and relative arrangements of the major clades need further analysis. Age estimates for Passeriformes are logically linked to our understanding of the age of modern birds as a whole and where passeriforms might fit within the avian

430


Tree of Life. Unfortunately, there is no clear understanding about the closest relative of passeriforms, although they probably lie with taxa traditionally classified in the “Coraciiformes” and Piciformes. Those few studies that have attempted to estimate divergence times across the avian tree have all placed the divergence of passeriforms from other orders in the late Cretaceous (8, 12, 30–33). Although there is consistency in these age estimates— that is, before the Cretaceous–Paleogene (K-P) boundary (65 Ma) and within the late Cretaceous (100–65 Ma)— the data and methods vary significantly and the taxon sampling for all these studies was not adequate for examining divergences within passerines. There are currently no passerine fossils that are useful for internal calibration of the passeriform tree. The oldest passeriform (34) is from Australia and of early Eocene in age (~55 Ma), but its phylogenetic relationships are uncertain. As a consequence, there have been several other approaches to calibrating the passerine tree. Van Tuinen and Hedges (31) employed a distant external calibration based on a general vertebrate tree. Barker et al. (19) used a geological vicariance event, the separation of New Zealand (Acanthisitta) from Gondwana to calibrate the base of passeriforms at ~82 Ma. Pereira and Baker (12) estimated divergence times for major groups of birds using 35 complete mitochondrial genomes for 35 species of birds and 13 vertebrate outgroups, with times of splitting for the vertebrate outgroups and five nodes inside modern nonpasseriform birds being calibrated with fossils. In that study, Bayesian analysis of the data placed the origin of modern birds at 139 Ma and the separation between passeriforms and other Neoaves at 108 Ma (in their small sample, passeriforms were resolved as the closest relative of other neoavians). Within passeriforms they dated the divergence of Acanthisitta and other passeriforms at 95.3 Ma, although the lower CI bound included the Barker et al. (19) vicariance age. The only study to estimate the entire passeriform timetree was constructed using nonparametric rate smoothing and penalized likelihood analysis (35, 36) of combined RAG1 and RAG2 genes, which, as noted, was calibrated with the New Zealand vicariance age of 82 Ma (19; Fig. 2). As an independent check on that calibration point, pairwise corrected distances of cytochrome b, converted to time using a commonly applied passerine evolutionary rate (37), yielded a divergence age for the Acanthisitta vs. other passerines at around 87 Ma (19), which is within the confidence interval of the whole mtDNA clock (12).

Assuming the 82 Ma calibration, the suboscines and oscines split (node 2) around 77 Ma, well before the K-P extinction event (19). Other basal splits took place just prior to the K-P boundary (Old and New World suboscines, see also 23), or more or less contemporaneously with it. Thus, the diversification into the major lineages of the New World suboscines, as well as the earliest branches within oscines, took place around the K-P boundary. A sampling of well-supported groups on the passerine timetree gives an overall picture of the temporal pattern of passerine evolution (Fig. 2). Within the New World suboscines, there are two major clades (Fig. 2). The first is the Furnariides—including the antpittas, antbirds, and ovenbirds—which arose around 55 Ma. The other major lineage, the highly speciose Tyrannides, on the other hand, started to diversify substantially later in time, around 38–42 Ma. Numerous early lineages of oscines diversified between 65 and 50 Ma. Thus, all four primary oscine clades—corvoids, muscicapoids, passeroids, and sylvioids—began radiating at roughly the same time (47–38 Ma), although the radiation of the corvoids was likely initiated slightly earlier than the others. Within each of these clades some groups attained very high diversity within relatively short periods of time. The most striking is the New World passeroid radiation that began ~20 Ma and resulted in over 300 genera and 1500 species, most of which apparently diversified within the last 10–15 million years. The highly diverse groups of the Old World muscicapoids and sylvioids began radiating somewhat earlier than the New World passeroids. There has been very little additional analysis of the oscine timetree. Several previous studies (38, 39) employed the same RAG1 and RAG2 data, the Acanthisitta calibration and rate smoothing methods as did Barker et al. (19), hence it is unsurprising that all three studies are consistent with respect to the timetree. Only one other study (25) dated a significant portion of the passerine tree, in this instance within malaconotine corvoids. Employing two nuclear introns (myoglobin intron-2 and GAPDH intron-11) as well as sequence from the mtDNA gene ND2, they calibrated the suboscine–oscine split based on previous work (19, 31, 40), allowing it to vary between 77 and 71 Ma, and estimated splits within malaconotines using Bayesian methods (41). These workers found that malaconotines began radiating ~37.7 Ma using all data and ~38.9 Ma using the nuclear introns alone. These dates are significantly older, by ~9–12 Ma, than those presented in Fig. 2. In parallel with comments about the age of passerines as


a whole (see earlier), discrepancies in the estimated age of the malaconotines are likely attributable to different data sets and rate estimation methods. Our current understanding of passerine phylogeny has led to the strong inference that passeriforms had their origin on Gondwana and that oscines in particular arose in Australasia (18–20, 40). By 47–40 Ma passeridans had reached Laurasia, and corvoids also reached the Asian mainland by about 40 Ma (Fig. 2). Both groups diversified across Laurasia and subsequently invaded the Southern Hemisphere (South America and Africa), as well as reinvaded Australia. Within corvoids, vireos reached the New World around 28–20 Ma (19, 39) and crows/jays did so ~17–14 Ma. There were multiple invasions of passeridans into the New World starting with the wrens at 34 Ma, the mimids at 22–20 Ma, and the emberizines at 22–20 Ma (19). The latter clade includes finches, warblers, blackbirds, and tanagers, which together are the dominant oscines of the New World, especially the Latin American tropics.

Acknowledgments Support was provided by U.S. National Science Foundation and by the American Museum of Natural History through the L.J. and L.C. Sanford Funds, Lewis B. and Dorothy Cullman Program for Molecular Systematics Studies, and the Sackler Institute of Comparative Genomics.

References 1. C. G. Sibley, B. L. Monroe, Jr., Distribution and Taxonomy of Birds of the World (Yale University Press, New Haven, 1990). 2. E. Mayr, D. Amadon, Amer. Mus. Novitates 1496, 1 (1951). 3. A. Wetmore, Smithson. Misc. Coll. 139 (11), 1 (1960). 4. E. C. Dickinson (Ed.), The Howard and Moore Complete Checklist of the Birds of the World (Christopher Helm, London, 2003). 5. C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds (Yale University Press, New Haven, 1990). 6. J. Cracraft et al., in Assembling the Tree of Life, J. Cracraft and M. J. Donoghue, Eds. (Oxford University Press, New York, 2004), pp. 468–489. 7. M. G. Fain, P. Houde, Evolution 58, 2558 (2004). 8. P. G. P. Ericson et al., Biol. Lett. 2, 543 (2006). 9. B. C. Livezey, R. L. Zusi, Zool. J. Linn. Soc. 149, 1 (2007). 10. J. Cracraft, in The Phylogeny and Classification of the Tetrapods, M. J. Benton, Ed. (Oxford University Press, Oxford, 1988), pp. 339–361.

431

11. G. Mayr, J. A. Clarke, Cladistics 19, 527 (2003). 12. S. Pereira, A. J. Baker, Mol. Biol. Evol. 23, 1731 (2006). 13. D. P. Mindell, M. D. Sorenson, D. E. Dimcheff, M. Hasegawa, J. C. Ast, T. Yuri, Syst. Biol. 48, 138 (1999). 14. R. J. Raikow, Auk 99, 431 (1982). 15. J. Delacour, C. Vaurie, Contrib. Sci. Los Angeles Co. Mus. 16, 1 (1957). 16. D. Amadon, Proc. Zool. Soc. Calcutta, Mookerjee Memor. Vol., 259 (1957). 17. R. J. Raikow, Ornithol. Monogr. 41, 1 (1987). 18. F. K. Barker, G. F. Barrowclough, J. G. Groth, Proc. Roy. Soc. Lond. 269, 295 (2002). 19. F. K. Barker, A. Cibois, P. Schikler, J. Feinstein, J. Cracraft , Proc. Natl. Acad. Sci. U.S.A. 101, 11040 (2004). 20. P. G. P. Ericson et al., Proc. Roy. Soc. Lond. 269, 235 (2002). 21. M. Irestedt, U. S. Johansson, T. J. Parsons, P. G. P. Ericson, J. Avian Biol. 32, 15 (2001). 22. M. Irestedt, J. L. Ohlson, D. Zuccon, M. Källersjö, P. G. P. Ericson, Zool. Scripta 35, 567 (2006). 23. Moyle, R. G., R. T. Chesser, R. O. Prum, P. Schikler, J. Cracraft, Amer. Mus. Novitates 3544, 1 (2006) 24. R. T. Chesser, Mol. Phylogenet. Evol. 32, 11 (2004). 25. J. Fuchs, J. Fjeldså, E. Pasquet, Zool. Scripta 35, 375 (2006). 26. R. G. Moyle, J. Cracraft, M. Lakim, J. Nais, F. H. Sheldon, Mol. Phylogenet. Evol. 39, 893 (2006). 27. P. Alström, P. G. P. Ericson, U. Olsson, P. Sundberg, Mol. Phylogenet. Evol. 38, 381 (2006). 28. P. G. P. Ericson, U. S. Johansson, Mol. Phylogenet. Evol. 29, 126 (2003). 29. A. Cibois, A., J. Cracraft, Mol. Phylogenet. Evol. 32, 264 (2004). 30. Cooper, A., D. Penny, Science 275, 1109 (1997). 31. M. van Tuinen, S. B. Hedges, Mol. Biol. Evol. 18, 206 (2001). 32. K. E. Slack et al., Mol. Biol. Evol., 23, 1144 (2006). 33. J. W. Brown, R. B. Payne, D. P. Mindell, Biol. Lett. 3, 257 (2007). 34. W. Boles, Nature 374, 6517 (1995). 35. M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997). 36. M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002). 37. R. C. Fleischer, C. E. McIntosh, C. L. Tarr, Mol. Ecol. 7, 533 (1998). 38. P. Beresford, F. K. Barker, P. G. Ryan, T. M. Crowe, Proc. Roy. Soc. Lond. 272, 849 (2005). 39. S. Reddy, J. Cracraft, Mol. Phylogenet. Evol. 40, 1352 (2007). 40. P. G. P. Ericson, M. Irestedt, U. S. Johansson, J. Avian Biol. 34, 3 (2003). 41. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998).

Shorebirds (Charadriiformes) Allan J. Baker a,b,* and Sérgio L. Pereiraa a

Department of Natural History, Royal Ontario Museum, 100 Queen’s Park Crescent, Toronto, ON, Canada; bDepartment of Ecology and Evolutionary Biology, University of Toronto, Toronto, ON, Canada *To whom correspondence should be addressed (allanb@rom. on.ca)

Abstract The Order Charadriiformes is a major clade of shorebirds and consists of 19 families, ~90 genera, and 366 species. In DNA sequence trees three major clades representing suborders are recognizable: the Scolopaci (sandpipers and allies), Lari (gulls and allies, plus buttonquail), which form a group, and the Charadrii (plovers and allies). The latest divergence time estimates, incorporating uncertainties in phylogenetic trees and fossil dates, suggest that the Charadriiformes originated in the Cretaceous, about 93 million years ago (Ma), and that charadriiform families radiated mostly in the late Cretaceous (93–66 Ma) but also in the Paleogene (66–23 Ma).

Shorebirds are a diverse cosmopolitan group that forms the monophyletic Order Charadriiformes. They represent one of the largest clades in birds with 366 species classified traditionally in 19 families (1). The order can be divided into three major clades (2–4): Scolopaci (sandpipers, jacanas, painted snipes, seedsnipes, and Plains-wanderer), Lari (Crab Plover, coursers and pratincoles, gulls, terns, skimmers, and alcids), and Charadrii (plovers, oystercatchers, Ibisbill, stilts, avocets, sheathbills, and Magellanic Plover) (Fig. 1). Here we review the phylogenetic relationships and estimates of divergence times of the families of shorebirds. Relationships among taxa and even which taxa should be included in the Charadriiformes has proved difficult in the past, due in part to convergent acquisition of traits involved in exploitation of a range of ecological niches (4). Phylogenies derived from an extensive data set of morphological and skeletal characters (5) lack resolution at various nodes, and the topologies change depending on how characters are coded and which ones are excluded from analyses (5–8). In addition to their incongruence,

these trees place the Alcidae as the closest relative of all other Charadriiformes. The three major clades of shorebirds were first detected in DNA–DNA hybridization studies (9), in which Charadrii and Lari (including Alcidae) formed a group, closest to Scolopaci, and the buttonquails (Turnicidae), traditionally classified in their own Order Turniciformes, were closest to the rest of the Lari. Phylogenetic analysis of a very large data set of 2954 morphological characters also led to this topology (10), but placed the Plains-wanderer (Pedionimidae), jacanas (Jacanidae), and painted snipes (Rostratulidae) outside the Scolopaci, and buttonquails as the closest relative to a group including shorebirds and rails (part of Gruiformes). DNA-sequencing studies subsequently altered this arrangement by joining the Lari and Scolopaci as closest relatives (2–4, 11, 12). Additionally, DNA sequence-based phylogenies placed buttonquails as the closest relatives to Lari (2–4, 12). Recent phylogenies of 90 recognized genera of charadriiforms showed that plovers and noddies are not monophyletic assemblages, and the enigmatic Egyptian Plover Pluvianus is not a plover (2, 12), and hence deserves to be placed in its own separate Family Pluvianidae. Similarly, gray and golden plovers Pluvialis

Fig. 1 A Red Knot (Calidris canutus), Family Scolopacidae, from Delaware Bay, New Jersey, USA. Credit: M. Peck.

A. J. Baker and S. L. Pereira. Shorebirds (Charadriiformes). Pp. 432–435 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Aves; Charadriiformes

433

Laridae 17

Rhynchopidae

15

Alcidae 13

7

Lari

Sternidae

10

Stercorariidae

5

Glareolidae Turnicidae 14

Rostratulidae

12

Thinocoridae 16

8

Pedionomidae

Scolopaci

Jacanidae

3

Scolopacidae Haematopodidae

1

19

Ibidorhynchidae

18

Pluvialidae

9

Charadriidae

4

Pluvianidae 2

Charadrii

Recurvirostridae

11

Pluvianellidae 20

Chionidae

6

Burhinidae Late K

Paleogene

MESOZOIC 75

Neogene

CENOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of shorebirds (Charadriiformes). Divergence times are from Table 1. Abbreviation: K (Cretaceous).

are also considered here in their own family Pluvialidae as they form a distinct lineage more closely related to oystercatchers (Haematopidae), Ibisbill (Ibidorhynchidae), and stilts and avocets (Recurvirostridae) than to typical plovers (Charadriidae) (2, 12). A multiple gene phylogeny of the terns (13) found 11 clades that were each classified in separate genera, expanding on the seven or 10 genera recognized previously (14, 15). Parallel evolution and retention of ancestral morphological states was inferred when they were mapped on a multigene phylogeny of the shanks (16). The phylogenetic affinities of sandgrouse (Pterocles) have been debated over a century (10). Morphological characters suggest a closer relationship between sandgrouse and pigeons (Columbiformes) (10), and DNA studies do not support a close relationship between Pterocles and shorebirds (2, 4, 12).

Estimates of divergence time within shorebirds (Table 1) based on analyses of mitochondrial DNA sequences (2, 4, 17–20) agree that the group originated in the Cretaceous, and some lineages have survived the Cretaceous–Paleogene mass extinction. However, other studies that included fewer shorebird lineages, and used different time constraints based on the fossil record or other molecular time estimates (21, 22) have suggested younger ages for the origin of Charadriiformes. We opted to show a timetree (Fig. 2) based in the most comprehensive study for the group, which includes members of all families and 93% of all genera, and used 14 fossil dates as internal constraints within Charadriiformes to properly account for uncertainties in fossil ages (2). Additionally, the minimum age of the root was taken from a study of vertebrates in which additional external fossil ages

434


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among shorebirds (Charadriiformes). Timetree Node

Estimates Ref. (2)

Time

Ref. (4)

Time

CI

Time

Ref. (18) CI

Time

CI

1

93.1

93.1

102–85

–

–

81.7

94–69

2

88.6

88.6

98–80

65.6

79–51

75.6

88–62

3

88.5

88.5

98–80

75.8

80–66

–

–

4

85.1

85.1

95–76

–

–

–

–

5

84.7

84.7

94–76

71.8

78–61

–

–

6

79.9

79.9

90–70

54.0

67–42

–

–

7

77.6

77.6

87–69

45.2

56–37

64.5

78–51

8

73.4

73.4

83–64

63.8

70–57

63.4

77–48

9

69.0

69.0

79–60

32.8

44–22

62.4

77–47

10

68.0

68.0

77–60

36.0

44–25

51.8

67–38

11

66.1

66.1

76–57

–

–

–

–

12

65.2

65.2

75–56

52.1

60–45

49.5

64–32

13

62.9

62.9

72–55

–

–

45.6

59–31

14

60.1

60.1

70–51

45.8

54–38

–

–

15

59.3

59.3

68–51

24.4

33–17

–

–

16

55.3

55.3

66–46

52.1

60–45

–

–

17

55.2

55.2

64–47

22.7

33–16

–

–

18

49.2

49.2

59–40

13.9

22–7

45.4

64–30

19

45.3

45.3

56–36

19.1

28–12

–

–

20

28.3

28.3

37–21

19.0

29–12

–

–

Note: Node times in the timetree are from ref. (2). Ref. (4) provides results from a combined analysis of the nuclear RAG1 gene and three mitochondrial genes, and ref. (18) provides results from the analyses of partial mtDNA genome sequences.

had been employed (19). The common ancestor of the Charadriiformes is estimated to have occurred between 85 and 102 Ma, with a mean estimate of 93 Ma. The much older age of the common ancestor than estimated in (4) is because the age of the root was previously fi xed at 78 Ma, which in turn was based on a molecular age calculated from mtDNA genome sequences (23). Similarly, in an admittedly conservative approach in which calibrations of 62 Ma for the divergence of storks and penguins and 66 Ma for the origin of the Anseriformes (24–26) were used as minimal fossil anchor-points, the Charadriiformes was estimated to have originated about 69 Ma (22). The three suborders of Charadriiformes diverged in the late Cretaceous 98–79 Ma, with a mean estimate of ~88 Ma. Fourteen ancestors of extant lineages were estimated to have occurred before the Mesozoic–Cenozoic

boundary (66 Ma). Ten existing families and the two monogeneric lineages that need to be elevated to family status (Pluvialidae and Pluvianidae) also predated this boundary (Fig. 2). Diversification of genera, however, predominantly occurred after the asteroid impact, from the Paleocene to the mid-Miocene (66–15 Ma) (4). The survival of so many lineages dating from an origin in the late Cretaceous (22, 23, 27, 28) may explain why it has been historically difficult to determine the limits of the Charadriiformes, as many ancient forms have disparate morphologies as they adapted to a wide range of ecological niches over vast timescales.

Acknowledgments Support was provided by Canada Natural Sciences and Engineering Research Council, Royal Ontario Museum

Eukaryota; Metazoa; Vertebrata; Aves; Charadriiformes

Foundation, and U.S. National Science Foundation grants to A.J.B.

References 1.

2. 3. 4. 5. 6. 7. 8. 9.

10. 11. 12. 13.

J. F. Clements, Birds of the World: A Checklist, 5th ed. (Ibis Publishing Company, Temecula, California, 2000). A. J. Baker, S. L. Pereira, T. A. Paton, Biol. Lett. 3, 205 (2007). T. A. Paton, A. J. Baker, Mol. Phylogenet. Evol. 39, 657 (2006). T. A. Paton, A. J. Baker, J. G. Groth, G. F. Barrowclough, Mol. Phylogenet. Evol. 29, 268 (2003). J. G. Strauch, Trans. Zool. Soc. Lond. 34, 263 (1978). M. Bjorklund, Auk 111, 825 (1994). M. F. Mickevich, L. R. Parenti, Syst. Zool. 29, 108 (1980). P. C. Chu, Condor 97, 174 (1995). C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds: A Study in Molecular Evolution (Yale University Press, New Haven, 1990). B. C. Livezey, R. L. Zusi, Zool. J. Linn. Soc. 149, 1 (2007). P. G. Ericson, I. Envall, M. Irestedt, J. A. Norman, BMC Evol. Biol. 3, 16 (2003). M. G. Fain, P. Houde, BMC Evol. Biol. 7, 35 (2007). E. S. Bridge, A. Jones, A. J. Baker, Mol. Phylogenet. Evol. 35, 459 (2005).

435

14. C. G. Sibley, B. L. Monroe, Distribution and Taxonomy of the Birds of the World (Yale University Press, New Haven, 1993). 15. M. J. Gochfeld, J. Burger, I. C. T. Nisbet, in The Birds of North America, No. 370, A. Poole, F. Gill, Eds. (The Birds of North America Inc., Philadelphia, 1998), pp. 1–32. 16. S. L. Pereira, A. J. Baker, Condor 107, 514 (2005). 17. J. W. Brown, R. B. Payne, D. P. Mindell, Biol. Lett. 3, 257 (2007). 18. J. W. Brown, J. S. Rest, J. Garcia-Moreno, M. D. Sorenson, D. P. Mindell, BMC Biol. 6, 6 (2008). 19. S. L. Pereira, A. J. Baker. Mol. Biol. Evol. 23, 1731 (2006). 20. S. L. Pereira, A. J. Baker, Mol. Phylogenet. Evol. 46, 430 (2008). 21. P. G. P. Ericson, C. L. Anderson, T. Britton, A. Elzanowski, U. S. Johansson et al. Biol. Lett. 4, 543 (2006). 22. K. E. Slack, C. M. Jones, T. Ando, G. L. Harrison, R. E. Fordyce, U. Arnason, D. Penny, Mol. Biol. Evol. 23, 1144 (2006). 23. T. A. Paton, O. Haddrath, A. J. Baker, Proc. Roy. Soc. B 269, 839 (2002). 24. E. N. Kurochin, G. J. Dyke, A. A. Karhu, Am. Mus. Novitat. 3386, 1 (2002). 25. G. L. Harrison et al. Mol. Biol. Evol. 21, 974 (2004). 26. J. A. Clarke, C. P. Tambussi, J. I. Noriega, G. M. Erickson, R. A. Ketchman, Nature 433, 305 (2005). 27. A. Cooper, D. Penny, Science 275, 1109 (1997). 28. S. B. Hedges, P. H. Parker, C. G. Sibley, S. Kumar, Nature 381, 226 (1996).

Diurnal birds of prey (Falconiformes) Joseph W. Browna,* and David P. Mindella,b a

Department of Ecology and Evolutionary Biology & University of Michigan Museum of Zoology, 1109 Geddes Road, University of Michigan, Ann Arbor, MI 48109-1079, USA; bCurrent address: California Academy of Sciences, 55 Concourse Drive Golden Gate Park, San Francisco, CA 94118, USA *To whom correspondence should be addressed (josephwb@ umich.edu)

Abstract Diurnal birds of prey (~313 species) are traditionally grouped into five families, constituting the neoavian Order Falconiformes. No consensus has been reached as to whether the group is natural because of uncertainty concerning inclusion of the falcons (Falconidae) and the New World vultures (Cathartidae). However, a clade of “core falconiforms” is supported which includes Sagittariidae (Secretary Bird) and closely related families Pandionidae (Osprey) and Accipitridae (hawks, eagles, kites, and Old World vultures). The Falconiformes timetree suggests that “core falconiforms” diverged in the early Paleogene about 62 million years ago (Ma), but that Cathartidae and Falconidae originated in late Cretaceous 76 Ma.

The diurnal birds of prey constitute the Order Falconiformes, and are generally classified into five reciprocally monophyletic families (1): Cathartidae (New World vultures, seven species; North and South America), Sagittariidae (Secretary Bird, one species; Africa), Pandionidae (Osprey, one species; cosmopolitan), Accipitridae (hawks, eagles, kites, and Old World vultures, ~240 species; cosmopolitan) (Fig. 1), and Falconidae (falcons and caracaras, ~64 species; cosmopolitan). Falconiform taxa are generally characterized by morphological adaptations to predation, be it active hunting (hooked bills and strong talons) or eating carrion (long necks and unfeathered heads), although extensive specialization exists in the order. Here, the relationships and divergence times of the families of Falconiformes are reviewed. Among the traditional avian orders, the controversy currently surrounding the status of Falconiformes as a natural (monophyletic) group is eclipsed only by that of Pelecaniformes (tropicbirds, boobies and gannets,

cormorants and shags, anhingas, pelicans, and frigatebirds) and Caprimulgiformes (nightbirds). Falconiformes has been variously considered monophyletic, polyphyletic, and paraphyletic (2). The basis for this debate involves both the possible inclusion of traditionally nonfalconiform taxa (owls) into Falconiformes, and the possible exclusion of families traditionally thought to belong to the order. A close affinity between the diurnal (Falconiformes) and predominantly nocturnal (owls; Order Strigiformes) birds of prey was hypothesized as early as Linnaeus (3), who placed both (among others) in his Order Accipitres. This scheme was refuted by the influential classifications of Fürbringer (4) and Gadow (5), who found the two groups to be only distantly related, and most subsequent taxonomic treatments have followed suit (1). Cracraft (6) provocatively diverged from this practice by proposing a classification scheme where a monophyletic owl clade is nested among falconiform families, rendering Falconiformes paraphyletic. However, this classification has been criticized (7), and the few subsequent morphological studies (8, 9) recovering this

Fig. 1 A Rough-legged Hawk (Buteo lagopus), Family Accipitridae, from Nunavut, Canada. Credit: G. Court.

J. W. Brown and D. P. Mindell. Diurnal birds of prey (Falconiformes). Pp. 436–439 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Aves; Falconiformes

437

Accipitridae 3

Pandionidae

2

Sagittariidae

1

Falconidae Cathartidae Late K

Paleogene CENOZOIC

MZ 75

Neogene

50

25

0 Million years ago

Fig. 2 A timetree of diurnal birds of prey (Falconiformes). Divergence times are from Table 1. Abbreviations: MZ (Mesozoic) and K (Cretaceous).

arrangement have had only weak statistical support. The recent comprehensive morphological analysis of Neornithes (10) instead strongly supports a close relationship for Falconiformes and Strigiformes, forming the proposed Superorder “Falconimorphae.” Published molecular genetic studies with broad taxon and character sampling fail to ally falconiform and strigiform taxa in any arrangement (2, 11–14). Regardless of whether or not “Falconimorphae” proves to be a natural grouping, it appears we may safely exclude strigiforms in our discussion of the tempo of falconiform diversification. Although the monophyly of the traditional Order Falconiformes has been supported (10, 15), diverse data sets have suggested it as polyphyletic. Karyological (16), morphological (17), and mitochondrial (mt) gene order data (14) have repeatedly emphasized the marked heterogeneity of Falconiformes relative to that found in other traditional avian orders, and have called to question whether such heterogeneity could arise in a natural group. At the extreme (17), Falconiformes has been regarded as an artificial aggregation of four separate (and possibly unrelated) orders: Sagittariiformes, Cathartiformes, Falconiformes, and Accipitriformes (possibly including Pandion). However, phenetic dissimilarities cannot establish the case for polyphyly, and subsequent authors studying the affinities of falconiform and non-falconiform taxa have localized taxonomic uncertainty to individual falconiform families. Chief among the taxa thought not to belong to Falconiformes is the Family Cathartidae. Th is family is generally regarded as being the most distinct falconiform lineage, and recent mtDNA data (14) show that cathartids have a different and less-derived gene order than other falconiform taxa sampled. Both morphological (18) and DNA–DNA hybridization (2) data have suggested an alliance between Cathartidae and storks

(Family Ciconiidae, Order Ciconiiformes), an arrangement first suggested a century earlier (19). Although subsequent research generally supported the separation of Cathartidae from other falconiform taxa, neither morphological (17) nor genetic (20, 21) studies aimed at discerning cathartid affinities have succeeded in recovering the Cathartidae–Ciconiidae pairing. An early mtDNA study (22) repeatedly cited in support of a Cathartidae–Ciconiidae relationship included erroneous sequences (23, 24), and has long since been retracted. In morphological studies that recover a monophyletic Falconiformes (10, 15), Cathartidae is found to be the basal-most lineage in the clade. Of particular note, there is also no support for a close relationship between New World (Cathartidae) and Old World (Accipitridae) vultures, and thus ecological similarities between them provide a striking example of evolutionary convergence. Lineages represented by a single living species have often been difficult to classify in ornithology, presumably because long branches (time) and/or extreme ecological specialization confound the identification of homologous character states. Two falconiform families, Pandionidae and Sagittariidae, are in this category and each has experienced some minor taxonomic turbulence. However, no character data have convincingly excluded them from Falconiformes or placed them elsewhere. Some morphological similarities between the secretary bird and seriemas (Family Cariamidae, Order Gruiformes) have been used to suggest a gruiform ancestry for Sagittarius (19), but this has not been supported in recent analyses. Gadow (5) appears to have been the first to recognize the distinctiveness of the piscivorous osprey (Pandion haliaetus) from accipitrid taxa by placing the former in a separate Family Pandionidae. Recent classifications variously consider the osprey as either the basal-most “extreme” member of Accipitridae, or the closest relative

438


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among falconiform birds. Timetree Node

Time

Estimates Ref. (2)

Ref. (11)

Time

Time

Ref. (12)

Ref. (29)

Ref. (32)

Time

CI

Time

CI

Time

CI

1

76.3

71.1

72.8

81.1

94–67

91.3

113–76

65.1

72–58

2

61.8

46.8

50.4

73.1

87–59

77.0

96–63

–

–

3

49.9

35.1

42

56.8

70–37

65.5

83–52

–

–

Note: Node times in the timetree represent the mean of time estimates. When multiple time estimates were available from the same study, then the mean of reported times and CIs is used as the representative estimate. For the unresolved Node 1, the representative estimates presented are averages of the divergence of Falconidae and Cathartidae from the remaining falconiform families. Results in ref. (2) are derived from DNA–DNA hybridization data; divergence times for Nodes 1 and 3 were not published in the original study, but are derived from melting temperatures presented there, as well as the same calibration factor used to estimate the divergence time for Node 2. The estimate from ref. (32) is derived from complete mtDNA genome sequences and employing a Bayesian autocorrelated model of rate evolution (only Falconidae and Accipitridae were sampled). The estimate presented from ref. (11) is derived from an analysis of five nuclear genes using two different rate-smoothing dating methods: closest-relative smoothing and ancestor-descendant smoothing. Ref. (29) reports a reanalysis of the data from ref. (33) using the same tree topology, but improved fossil calibrations and a dating method that employs a Bayesian autocorrelated model of rate evolution. Ref. (3) constitutes an average estimate from analyses of ~5 kb of mtDNA under eight combinations of different dating methods (n = 5: ancestor-descendent rate smoothing, closest-relative rate smoothing, Bayesian autocorrelated model of rate evolution, overdispersed clock, and Bayesian non-autocorrelated model of rate evolution) and tree topologies (n = 3).

of that family. Regardless, this difference is semantic, and does not influence evolutionary interpretations. Curiously, the namesake family of the order, Falconidae, may also not be closely related to any of the remaining falconiform taxa. Nuclear DNA studies support this notion (11, 13), but do not give a consistent indication as to where Falconidae lies within the neoavian tree. The inconsistent placement of Falconidae in mtDNA studies has been hypothesized to be a result of insufficient taxon sampling (25). However, broad taxonomic sampling (11, 12) and strategically sampled mitochondrial genome sequences (14) have failed to support a close relationship between Falconidae and other falconiform families. A recent morphological study (10) that recovers a monophyletic Falconiformes unites Pandionidae with Falconidae (rather than with Accipitridae); however, this hypothesis is novel and unsupported elsewhere. Notably, given the rules of taxonomic precedence the ordinal name “Falconiformes” must include the Family Falconidae; if falcons are demonstrably shown to be unrelated to the remaining falconiform families, then the grouping of these latter families will be raised to the rank of a novel order. In summary, at present there is no overwhelming evidence that Falconiformes is monophyletic, but also no convincing evidence for an alternative phylogenetic placement of the five traditional falconiform families. Present knowledge therefore suggests refraining from taxonomic modifications until new data are examined. However it seems prudent at this juncture to focus

discussion of divergence times to the “core falconiforms” family set of ((Accipitridae, Pandionidae), Sagittariidae) that is supported as a clade by most studies (2, 11, 13, 15, 26), and consider the relationships of Falconidae and Cathartidae to the “core falconiforms” as unresolved (Fig. 2). Few molecular studies have yet estimated divergence times among all traditional falconiform families (Table 1). Early DNA–DNA hybridization analyses assuming the average rate of change in genome-wide DNA–DNA hybridization analyses was 4.7 million years per degree (centigrade) of DNA–DNA melting temperature lowered (2) suggested that the living lineages of the “core falconiforms” originated in the Eocene about 47 Ma, a scenario supported by a strict interpretation of the fossil record (27). However, doubts surrounding the validity of the vicariance event dating used in calibrating the DNA–DNA hybridization analyses (28), together with the assumption of rate constancy across the entire avian tree, renders this time estimate suspect. Surprisingly, a recent study of five nuclear genes with broad taxonomic sampling (11) yielded even younger estimates in some divergence time analyses. However, this study was found to contain several problems surrounding divergence time estimation (29). Most important in the context of the present chapter, the inferred age of Pandionidae at 29 Ma using a closest-relative rate smoothing procedure (11) significantly postdates the oldest known fossil of that taxon at 37 Ma (30), a fossil that was supposedly used as a minimum age constraint in the dating

Eukaryota; Metazoa; Vertebrata; Aves; Falconiformes

analyses. A reanalysis of these data (29) using a Bayesian modeling of rate evolution together with improved fossil constraints yielded much older divergence time estimates, with Cathartidae and Falconidae diverging from the remaining falconiform lineages at about 91 Ma, followed by the divergence of Sagittariidae at 77 Ma and the Accipitridae–Pandionidae split at 65 Ma (Fig. 2). These nuclear DNA results agree closely to those generated from various dating analyses of ~5 kb of mtDNA for 135 avian taxa (12), despite considering very different tree topologies outside of “core falconiforms.” In the latter study, different assumptions about how substitution rate variation evolves produced a range of falconiform divergence time estimates, the Accipitridae–Pandionidae split ranging from 69 to 50 Ma, the origin of Sagittariidae ranging from 91 to 61 Ma, and the divergence of Cathartidae and Falconidae ranging from 105 to 61 Ma. A study of complete mtDNA genomes employing a Bayesian modeling of rate evolution yielded a slightly younger (but overlapping) mean estimate of about 65 Ma for the first divergence within the traditional Falconiformes; however, sparse taxon sampling likely contributed to this younger estimate (31). The generally close agreement of date estimates across genomes, tree topologies, and dating methods lends credence to the ancient origin of falconiform taxa. However, interpretation of the fossil record in light of these molecular date estimates requires the existence of extensive “ghost lineages” beyond the oldest falconiform fossil, belonging to a living lineage, at ~37 Ma (30).

Acknowledgments J.W.B. thanks R. Pollard for assistance. Support was provided by the University of Michigan to J.W.B. and US National Science Foundation to D.P.M.

References 1. A. Wetmore, Smithson. Misc. Coll. 139, 1 (1960). 2. C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds (Yale University Press, London, 1990). 3. C. Linnaeus, Systema Naturae per Regna Tria Naturae, Secundum Classes, Ordines, Genera, Species cum Characteribus Differentiis, Synonymis, Locis, 10th ed. (Laurentii Salvii, Holmiae, 1758).

439

4. M. Fürbringer, Untersuchungen zur Morphologie und Systematik der Vögel (von Holkema, Amsterdam, 1888). 5. H. Gadow, in Klassen und Ordnungen des Thier-Reichs, H. G. Bronn, Ed. (C. F. Winter, Leipzig, 1893), pp. 259–270. 6. J. Cracraft, Auk 98, 681 (1981). 7. S. L. Olson, Auk 99, 733 (1982). 8. G. Mayr, J. Clarke, Cladistics 19, 527 (2003). 9. M. C. McKitrick, Misc. Publ., Mus. Zool., Univ. Mich. 179, 1 (1991). 10. B. C. Livezey, R. L. Zusi, Zool. J. Linn. Soc. 149, 1 (2007). 11. P. G. P. Ericson et al., Biol. Lett. 4, 543 (2006). 12. J. W. Brown, J. S. Rest, J. García-Moreno, M. D. Sorenson, D. P. Mindell, BMC Biol. 6 (2008). 13. M. G. Fain, P. Houde, Evolution 58, 2558 (2004). 14. G. C. Gibb, O. Kardailsky, R. T. Kimball, E. L. Braun, D. Penny, Mol. Biol. Evol. 24, 269 (2007). 15. C. S. Griffiths, Auk 111, 787 (1994). 16. L. E. M. de Boer, Genetica 46, 77 (1976). 17. M. T. Jollie, Evol. Theory 1, 285; 2, 115; 3, 1 (1976–1977). 18. J. D. Ligon, Occasional Papers/University of Michigan, Museum of Zoology 651, 1 (1967). 19. A. H. Garrod, Proc. Zool. Soc. Lond. 1874, 339 (1874). 20. M. Wink, Zeitschrift der Naturforschenden 50, 868 (1995). 21. I. Seibold, A. J. Helbig, Phil. Trans. Roy. Soc. Lond. B Biol. Sci. 350, 163 (1995). 22. J. C. Avise, W. S. Nelson, C. G. Sibley, Proc. Natl. Acad. Sci. U.S.A. 91, 5173 (1994). 23. S. J. Hackett, C. S. Griffiths, G. P. Bates, N. L. Klein, Mol. Phylogenet. Evol. 4, 350 (1995). 24. A. J. Helbig, I. Seibold, Mol. Phylogenet. Evol. 6, 315 (1996). 25. K. E. Slack, A. Janke, D. Penny, U. Arnason, Gene 302, 43 (2003). 26. H. R. L. Lerner, D. P. Mindell, Mol. Phylogenet. Evol. 37, 327 (2005). 27. A. Feduccia, Trends Ecol. Evol. 18, 172 (2003). 28. M. van Tuinen, C. G. Sibley, S. B. Hedges, Mol. Biol. Evol. 15, 370 (1998). 29. J. W. Brown, R. B. Payne, D. P. Mindell, Biol. Lett. 3, 257 (2007). 30. C. J. O. Harrison, C. A. Walker, Zool. J. Linn. Soc. 59, 323 (1976). 31. H. P. Linder, C. R. Hardy, F. Rutschmann, Mol. Phylogenet. Evol. 35, 569 (2005). 32. K. E. Slack et al., Mol. Biol. Evol. 23, 1144 (2006). 33. P. G. P. Ericson et al., Biol. Lett. 4, 543 (2006)

Cranes, rails, and allies (Gruiformes) Peter Houde Department of Biology, New Mexico State University, Box 30001 MSC 3AF, Las Cruces, NM 88003-8001, USA ([email protected])

Abstract The cranes, rails, and allies (Order Gruiformes) form a morphologically eclectic group of bird families typified by poor species diversity and disjunct distributions. Molecular data indicate that Gruiformes is not a natural group, but that it includes a evolutionary clade of six “core gruiform” families (Suborder Grues) and a separate pair of closely related families (Suborder Eurypygae). The basal split of Grues into rail-like and crane-like lineages (Ralloidea and Gruoidea, respectively) occurred sometime near the Mesozoic– Cenozoic boundary (66 million years ago, Ma), possibly on the southern continents. Interfamilial diversification within each of the ralloids, gruoids, and Eurypygae occurred within the Paleogene (66–23 Ma).

The avian Order Gruiformes, as traditionally conceived, consists of as many as a dozen families of extant or recently extinct birds and nearly as many more families known only from fossils. The Family Gruidae is represented by 15 species of cranes, which are found worldwide except in Antarctica and South America. The nearly cosmopolitan Family Rallidae comprises some 135–142 species of rails, crakes, coots, moorhens, gallinules, and flufftails (1, 2). Cranes and rails comprise the genetically closely related and relatively morphologically homogeneous Suborder Grues (3), which in turn is divided into the stocky rail-like Superfamily Ralloidea (rails, finfoots, adzebills) and the lanky crane-like Superfamily Gruoidea (cranes, Limpkin, trumpeters). Ralloids and gruoids represent extremes along a continuum of size from 139 to 1524 mm in length (4). Most are drab somber birds of wetlands or aquatic habitats, although gallinules are colorful and many others have pigmented patches of skin on the head (Fig. 1). The most primitive members of Grues prefer forested habitats, but a few live even in arid savannahs. Grues are typified by dense plumage, relatively long necks, large feet and/or long legs, wings of low aspect ratio, short tails, and insensitive bills of short to medium length, but most

of these features are subject to allometric scaling. Cranes are exceptional migrators. While most rails are generally more sedentary, they are nevertheless good dispersers. Many have secondarily evolved flightlessness after colonizing remote oceanic islands. Other members of the Grues are nonmigratory. They include the finfoots and sungrebe (Heliornithidae), with three species in as many genera that are distributed pantropically and disjunctly. Finfoots are foot-propelled swimmers of rivers and lakes. Their toes, like those of coots, are lobate rather than palmate. Adzebills (Aptornithidae) include two recently extinct species of flightless, turkey-sized, rail-like birds from New Zealand. Other extant Grues resemble small cranes or are morphologically intermediate between cranes and rails, and are exclusively neotropical. They include three species in one genus of forest-dwelling trumpeters (Psophiidae) and the monotypic Limpkin (Aramidae) of both forested and open wetlands. No fossils of reliably identifiable extant core gruiform families are known to predate the Oligocene, but crane- or raillike fossils date to the early Eocene. Many other extant families have been considered to be Gruiformes, including buttonquails, Australian Plains-wanderer, seriemas, mesites, and bustards. Most are monotypic or nearly so, are morphologically highly

Fig. 1 A Red-chested Flufftail (Sarothrura rufa), Family Rallidae. Photo credit: M. Ford.

P. Houde. Cranes, rails, and allies (Gruiformes). Pp. 440–444 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Aptornithidae Rallidae

4

1

Heliornithidae Rhynochetidae

6

Late K

Paleogene

MESOZOIC 75

Eurypygidae

Grues

Psophiidae 2

441

Eurypygae

Aramidae

3

Ralloidea

Gruidae

5

Gruoidea

Eukaryota; Metazoa; Vertebrata; Aves; Gruiformes

Neogene

CENOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of cranes, rails, and allies (Gruiformes). Only the monophyletic Suborders Grues and Eurypygae are included. Divergence times are shown in Table 1. Abbreviation: K (Cretaceous).

diverged from one another, and have restricted and often disjunct southern distributions. For example, the monotypic family pair of Kagu (Rhynochetidae) and Sunbittern (Eurypygidae) is endemic to New Caledonia and Amazonia, respectively. This has led some to hypothesize a pre-Cenozoic divergence for Gruiformes, while others have questioned its monophyly. This chapter reviews the relationships and divergence times of the families of Grues, and of Kagu and Sunbittern because they are believed by some to have close relationships to some members of Grues. Other putative members of the traditionally conceived Order Gruiformes are disregarded because their most recent divergences are among non-gruiform taxa. Members of Gruiformes were first united on morphological gestalt in the mid-1800s (5), where they have remained in most morphology-based classifications until the present (1, 6, 7). Nevertheless, their monophyly has been questioned on the basis of morphological criteria (8, 9) and they have been the subject of few interfamilial or interordinal molecular genetic studies (3, 5, 10–22). The first attempts to address gruiform relationships using molecular genetic data were DNA hybridization trials (5, 10, 12, 21, 22) that neither included mesites nor incorporated a sufficient diversity of outgroups to test gruiform monophyly. Rather, where these were combined with nongruiforms, the data were primarily intraordinal fitted to an interordinal supertree (5). Extraordinarily large DNA hybridization distances suggested that buttonquails were likely not Gruiformes (5), but their correct placement among shorebirds (Charadriiformes) was determined only later with multiple sequence data (14–16, 23, 24).

The Australian Plains-wanderer was first recognized as a member of Charadriiformes primarily on the basis of osteological characters (8). This conclusion has since been confirmed by numerous molecular studies (5, 14–16, 23, 24), but some morphologists argue that Gruiformes is paraphyletic to Charadriiformes (7). There is irreconcilable disagreement between morphological and molecular studies about whether the remaining gruiforms are monophyletic. Analyses of small-subunit ribosomal mtDNA (12S) suggested that Kagu, Sunbittern, seriemas, bustards, buttonquails, and mesites were all more distant from the Grues clade than charadriiform and ciconiiform outgroups (13). Only two published molecular studies have sampled nonpasserine neoavian families fairly comprehensively (14, 15). These, based on one and five nuclear loci, respectively, and another study combining mitochondrial and nuclear loci addressing Charadriiformes (16), strongly support a polyphyletic origin of Gruiformes. The only monophyletic groups among putative Gruiformes recovered by these studies are Kagu and Sunbittern within the proposed basal clade of Neoaves, Metaves, and among the families of Grues in the other proposed basal clade of Neoaves, Coronaves. In all, these studies found no fewer than six independent lineages of traditional Gruiformes even though the closest relatives of mesites, bustards, and seriemas remain poorly resolved. Grues alone were studied in greater detail with mitochondrial and three nuclear loci (3). Limpkin and cranes are the closest relatives and their group is in turn the closest relative of the trumpeters. Rails appear to be paraphyletic to both finfoots and adzebills, with flufftails (Sarothrura) closest

442


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among cranes, rails, and allies (Gruiformes). Timetree Node

Estimates Ref. (3)

Time Time

Ref. (15)(a)

Ref. (15)(b)

CI

Time

Time

Ref. (31) Time

CI

1

87.4

–

–

74.5

88.2

99.5

123–83

2

64.5

73.2

107–50

45.1

69.0

70.5

89–58

3

56.9

66.4

98–45

32.0

63.8

65.4

83–53

4

44.5

42.6

66–27

45.1

46.9

43.2

57–33

5

37.8

48.5

74–31

17.4

39.4

45.9

61–35

6

36.8

–

–

31.4

38.1

41

56–30

Note: Node times in the timetree represent the mean of time estimates from different studies. Three and five nuclear genes were analyzed using Multidivtime in refs. (3) and (31), respectively. In ref. (15), five nuclear genes were analyzed using (a) PATHd8 and (b) r8s programs.

to finfoots and adzebills the closest relatives of all other rails (Fig. 2). Rates of molecular evolution are heterogeneous across all families of Grues and loci yet tested (3, 20, 22). A relaxed Bayesian clock was used in the Multidivtime program (25, 26) to estimate divergences within Grues based on intron 5 of alcohol dehydrogenase-1 (ADH-1), intron 7 of beta-fibrinogen (FGB), and introns 3–5 and exons 3–4 of glyceraldehyde-3-phosphate dehydrogenase (GAPD-H), both with and without an upper bound set for the basal divergence of Gruidae at 25–20 Ma (3). Lower-bound calibrations internal to Gruidae were liberally construed from the fossil record (20, 22), and the lower bound on the Aramidae–Gruidae divergence was based on an Oligocene fossil (28 Ma) identified as a Limpkin (9). A fossil shorebird (23) was used as a lowerbound calibration external to Gruiformes, and analyses were conducted both with and without the constraints of internal and external upper bounds. The basal divergence within Grues was estimated to be 51 Ma with upper bounds enforced in ref. (3). Without upper bounds, the Aramidae–Gruidae divergence was estimated at 48.5 Ma, the Psophiidae–Aramidae divergence at 66.4 Ma, the Heliornithidae–Rallidae divergence at 42.6 Ma, and the Gruoidea–Ralloidea divergence at 73.2 Ma (Table 1). Limited data (12S and ADH-1 only) precluded the estimation of the Aptornithidae–Rallidae divergence, but it is bracketed by the Heliornithidae– Rallidae divergence (excluding Sarothrura). Ericson et al. (15) estimated divergences across Neoaves using exon 3 of c-myc protooncogene (c-myc), recombination activating gene-1 (RAG1), intron 2 of

myoglobin (myo), intron 7 of FGB, and introns 6–7 and exon 6 of ornithine decarboxylase (ODC). They fi xed the age of the hummingbird-swift divergence at 47.5 Ma and used some 22 diverse fossils as minimum age constraints internal to Neoaves to calibrate their timetree. Two lower-bound constraints were internal to Grues, including a 14 Ma fossil sungrebe (27) and a 34–30 Ma fossil gruoid (28). Using PATHd8 (29) and penalized likelihood (PL) in the r8s program (30), respectively, they estimated the Aramidae–Gruidae divergence at 17.4 and 39.4 Ma, the Psophiidae–Aramidae divergence at 32.0 and 63.8 Ma, the Rallidae–Heliornithidae divergence at 45.1 and 46.9 Ma, the Gruoidea–Ralloidea divergence at 45.1 and 69.0 Ma, the Eurypygidae–Rhynochetidae divergence at 31.4 and 38.1 Ma, and the Eurypygae–Grues (vis á vis Metaves–Coronaves) divergence at 74.5 and 88.2 Ma (Table 1). Brown et al. (31) criticized Ericson et al.’s calibration as being generally too young, taking issue with the appropriateness of PATHd8 and some fossil calibrations; but see (32). They reanalyzed a modified DNA sequence alignment, using some alternative fossil calibrations and parametric Bayesian modeling of evolutionary rate in Multidivtime (25). They estimated the Aramidae– Gruidae divergence at 45.9 Ma, the Psophiidae–Aramidae divergence at 65.4 Ma, the Rallidae–Heliornithidae divergence at 43.2 Ma, the Gruoidea–Ralloidea divergence at 70.5 Ma, the Eurypygidae–Rhynochetidae divergence at 41.0 Ma, and the Eurypygae–Grues divergence at 99.5 Ma (Table 1). The traditionally conceived Order Gruiformes has been said to “exhibit strong Gondwanan distribution

Eukaryota; Metazoa; Vertebrata; Aves; Gruiformes

patterns” (33) because its greatest familial diversity is in southern landmasses. Kagu, adzebills, and Sunbittern have occupied center stage, since the former two are flightless or nearly so. However, the notion that adzebills are the closest relatives of Kagu (1) is contradicted by all available molecular data (13). Ironically, adzebills were described as being a species of rail as early as 1866 (34). Constraints of divergence times of basal ralloids suggest that adzebills diverged from rails in the middle Eocene (more recently than 42.6 Ma). The familiar pattern of rail dispersal to oceanic refugia and their subsequent loss of flight renders the molecular hypothesis for the origin of adzebills biogeographically and evolutionarily plausible (13). Kagu and Sunbittern clearly are closest relatives, but their divergence occurred in the late Eocene (56–34 Ma) or early Oligocene (34–23 Ma). Fossils of Kagu-like or Sunbittern-like birds are well known from early to mid-Eocene deposits of North America and Europe (1, 35). Thus, the extant disjunct distribution of Kagu and Sunbittern simply may reflect the deterioration of tropical forests in the Northern Hemisphere during the Oligocene (36, 37). The average of estimates (64.5 Ma) of the basal divergence of Grues is roughly coeval with the Cretaceous– Paleogene boundary, with estimates and errors ranging 107–45 Ma. The most primitive ralloids (i.e., flufftails, adzebills) are distributed in Africa, Madagascar, and New Zealand and the most primitive gruoids (i.e., trumpeters, Limpkin) are neotropical. Thus, the basal dichotomy of Grues may bear a Gondwanan signature. Alternatively, the generally forest-dwelling most primitive members of Grues may have retreated to southern tropical refugia in response to Oligocene global cooling. Indeed, fossils of basal gruoids are known from Oligocene (but not later) deposits of North America and Europe (9, 28). Interfamilial diversification within each of the ralloids and gruoids appears to have occurred in the early Paleogene (66–23 Ma), with considerable intercontinental dispersal by rails, finfoots, and cranes. The divergence of finfoots from flufftail rails and their Paleogene dispersal through Asia to the Americas is well-documented phylogeographically, by molecular time estimates, and by the fossil record (3, 27).

Acknowledgments I thank C. L. Anderson, J. W. Brown, A. Cooper, P. G. P. Ericson, and M. G. Fain for assistance. This work was supported by the U.S. National Science Foundation.

443

References 1. B. C. Livezey, Phil. Trans. Roy. Soc. Lond. B 353, 2077 (1998). 2. C. G. Sibley, B. L. Monroe, Jr., Distribution and Taxonomy of Birds of the World (Yale University Press, New Haven, 1990). 3. M. G. Fain, C. Krajewski, P. Houde, Mol. Phylogenet. Evol. 43, 515 (2007). 4. J. Van Tyne, A. J. Berger, Fundamentals of Ornithology (Dover Publications, New York, 1971). 5. C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds: A Study in Molecular Evolution (Yale University Press, New Haven, 1990). 6. J. Cracraft et al., in Assembling the Tree of Life, J. Cracraft, M. J. Donoghue, Eds. (Oxford University Press, New York, 2004), pp. 468–489. 7. B. C. Livezey, R. L. Zusi, Zool. J. Linn. Soc. 149, 1 (2007). 8. S. L. Olson, D. W. Steadman, Smithson. Contrib. Zool. 337, 1 (1981). 9. S. L. Olson, Avian Biol. 8, 79 (1985). 10. C. G. Sibley, J. E. Ahlquist, P. deBenedictis, J. Yamashina Inst. Ornithol. 25, 1 (1993). 11. P. Houde, Cladistics 10, 1 (1994). 12. P. Houde, F. H. Sheldon, M. Kreitman, J. Mol. Evol. 40, 678 (1995). 13. P. Houde, A. Cooper, E. Leslie, A. E. Strand, G. A. Montaño, in Avian Molecular Evolution and Systematics, D. P. Mindell, Ed. (Academic Press, San Diego, 1997), pp. 121–158. 14. M. G. Fain, P. Houde, Evolution 58, 2558 (2004). 15. P. G. P. Ericson et al., Biol. Lett. 543 (2006). 16. M. G. Fain, P. Houde, BMC Evol. Biol. 7, 35 (2007). 17. C. Pitra, D. Lieckfeldt, S. Frahnert, J. Fickel, Mol. Phylogenet. Evol. 23, 63 (2002). 18. C. Krajewski, M. G. Fain, L. Buckley, D. G. King, Mol. Phylogenet. Evol. 13, 302 (1999). 19. C. Krajewski, J. W. Fetzner, Jr., Auk 111, 351 (1994). 20. C. Krajewski, D. G. King, Mol. Biol. Evol. 13, 21 (1996). 21. C. Krajewski, Biochem. Genet. 27, 131 (1989). 22. C. Krajewski, Mol. Biol. Evol. 7, 65 (1990). 23. T. A. Paton, A. J. Baker, J. G. Groth, G. F. Barrowclough, Mol. Phylogenet. Evol. 29, 268 (2003). 24. T. A. Paton, A. J. Baker, Mol. Phylogenet. Evol. 39, 657 (2006). 25. J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). 26. B. M. Wiegmann, D. K. Yeates, J. L. Thorne, H. Kishino, Syst. Biol. 52, 745 (2003). 27. S. L. Olson, Proc. Biol. Soc. Wash. 116, 732 (2003). 28. G. Mayr, Naturwissenschaften 92 (2005). 29. T. Britton, C. L. Anderson, D. Jaquet, S. Lundqvist, K. Bremer, http://www.math.su.se/PATHd8 (2006). 30. M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002).

444


31. J. W. Brown, R. B. Payne, D. P. Mindell, Biol. Lett. 3, 257 (2007). 32. P. G. P. Ericson, C. L. Anderson, G. Mayr, Biol. Lett. 3, 260 (2007). 33. J. Cracraft, Proc. Roy. Soc. Lond. B 268, 459 (2001). 34. A. Hamilton, Trans. New Zealand Inst. I. Zoology 24, 175–184 (1891).

35. A. Hesse, J. Ornithol. 129, 83 (1988). 36. J. A. Wolfe, in Eocene–Oligocene Climatic and Biotic Evolution, D. R. Prothero, W. A. Berggren, Eds. (Princeton University Press, Princeton, NJ, 1992), pp. 421–436. 37. J. Zachos, M. Pagani, L. Sloan, E. Thomas, K. Billups, Science 292, 686 (2001).

Woodpeckers, toucans, barbets, and allies (Piciformes) William S. Moorea,* and Kathleen J. Migliab a

Department of Biological Sciences, Wayne State University, Detroit, MI 48202, USA; bDuke University Medical Center, Molecular Genetics & Microbiology, 210 Jones Building, Box 3020 Durham, NC 27710, USA *To whom correspondence should be addressed (wmoore@biology. biosci.wayne.edu)

Abstract

and divergence times of the Order Piciformes and its constituent clades to the level of families. As many as eight nominal families have been included in the Order Piciformes under various classifications: Picidae (wrynecks, piculets, and woodpeckers; ~28 genera, 216 species), Indicatoridae (honeyguides; ~4 genera, 17 species), Megalaimidae (Asian barbets; ~3 genera, 26 species), Lybiidae (African barbets; ~7 genera, 42 species), Capitonidae (New World barbets; ~2 genera, 13 species), Ramphastidae (toucans; ~7 genera, 36 species,

The avian Order Piciformes comprises two major lineages, the Pici and Galbulae, which diverged as early as 70 million years ago (Ma). The jacamars and puffbirds (Galbulae) also diverged relatively early, ~53 Ma. Later diversification of the Pici gave rise to six additional families, beginning ~44–38 Ma. Molecular clock estimates for the origins of piciform clades, some estimated here, are consistent with the chronology of the sparse fossil record. With the exception of the woodpeckers, species of several piciform families were abundant on the northern continents during the Paleogene (66–23 Ma) but are now restricted to the tropics.

The Order Piciformes is a diverse assemblage of bird species that vary greatly in size, appearance, distribution, ecology, and life history. To the extent one can make descriptive generalizations, they tend to be stocky, brightly colored birds with disproportionately large bills—taken to an extreme in the toucans—and arboreal habits (Fig. 1). Their distributions are restricted to the Asian, African, and New World tropics, with the exception of the woodpeckers, which collectively have a more expansive distribution that includes the Old and New World temperate regions. Most species are insectivorous but many eat fruit at least occasionally and the barbets are primarily frugivorous. Cavity nesting is pervasive in the order, as is particularly well known for the woodpeckers, which have adaptations of the bill, skull, and associated musculature and enervation that renders them extraordinarily effective at excavating nest cavities in wood. The honeyguides are nest parasites, but parasitize only cavity-nesting species. Here we review the relationships

Fig. 1 A Red-headed Woodpecker (Melanerpes erythrocephalus), Family Picidae, from North America. Photo credit: R. Moul.

W. S. Moore and K. J. Miglia. Woodpeckers, toucans, barbets, and allies (Piciformes). Pp. 445–450 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

446

THE TIMETREE OF LIFE Ramphastidae 6

Capitonidae Lybiidae

4

Megalaimidae

3

Pici

7

Indicatoridae 5

Picidae Bucconidae

2

Galbulidae Paleogene

Galbulae

1

Neogene CENOZOIC

50

25

0 Million years ago

Fig. 2 The timetree of woodpeckers, toucans, barbets, and allies (Piciformes). Divergence times are shown in Table 1.

including two species of Semnornis), Galbulidae (jacamars; ~5 genera, 18 species), and Bucconidae (puffbirds; ~12 genera, 35 species). Collectively, the species assigned to these families group into two clades that diverged relatively early in the diversification of Neoaves (1, 2). One clade comprises the Galbulidae and Bucconidae and the other comprises the Picidae, Indicatoridae, and barbets and toucans, regardless of how the barbets and toucans are assigned to nominal families. These two ancient lineages are considered distinct orders, Galbuliformes and Piciformes, in some classifications, but suborders, Pici and Galbulae, of Piciformes in others. Regardless of the nomenclature, the evidence is strong that the Pici and Galbulae are both monophyletic. Until very recently, however, it was less certain whether the Pici and Galbulae were closest relatives, but three recent studies based on DNA sequence data from nuclear-encoded genes strongly support this hypothesis (1, 2, 10). Thus, it is reasonable to recognize the Order Piciformes comprising two major lineages, the Pici and Galbulae, totaling ~403 species. Using Peters (3) classification as a reference list for species usually included in an order called Piciformes, the history of systematic groupings of those species into families, superfamilies, and suborders is a kaleidoscope through the nineteenth and much of the twentieth centuries. The early systematic history has been thoroughly reviewed (1,4–6). Considerable stability of inferred relationships was established with the cladistic studies of Swierczewski and Raikow (5), and Simpson and Cracraft (6) based on myological and osteological characters. These studies reached identical conclusions regarding relationships among piciform families, which

were subsequently corroborated in a study by Lanyon and Zink (7) based on protein electromorph characters, although the latter study did not include a honeyguide (Indicatoridae). The tree hypothesized in these studies is ((((Indicatoridae, Picidae), (Ramphastidae, Capitonidae)), (Bucconidae, Galbulidae)), Outgroup). Although these studies produced congruent results, some intraordinal details remained to be determined, and monophyly of the order as well as identification of its closest relative remained contentious. Olson (8), Burton (9), and Sibley and Ahlquist (4) argued that Piciformes is polyphyletic with the Galbulae (Bucconidae and Galbulidae) related to the coraciiforms. Lanyon and Zink (7) were not able to test the hypothesis of monophyly of Piciformes directly, but they noted that their distance data supported a closer relationship of Galbulae to a coraciiform (Momotus) than to other piciforms, which suggests paraphyly. However, recent DNA sequence-based studies consistently support monophyly of Piciformes at statistically significant levels (1, 2, 10). Johansson and Ericson’s (1) study, based on exon segments from the nuclear RAG1 and c-myc genes and intron II from the myoglobin gene totaling 3400 nucleotides, was specifically designed to test the close relationship of Pici and Galbulae and thus monophyly of Piciformes. Combining all sequences from the three genes, monophyly was supported by maximum parsimony (90% bootstrap) and Bayesian (100% posterior probability) analyses. The maximum likelihood analysis based on the “Early Bird” data set also inferred a monophyletic Piciformes with 100% BS support (2). The “Early Bird” data set is more comprehensive and includes

Eukaryota; Metazoa; Vertebrata; Aves; Piciformes

447

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among woodpeckers, toucans, barbets, and allies (Piciformes). Timetree Node

Time

Estimates Ref. (10) Time

Ref. (17) Time

Ref. (22) CI

Time

CI

1

61.8

53

70.5

89–57

93.6

107–80

2

55.0

53

56.9

74–44

73.6

92–53

3

40.9

43.6

38.1

51–28

73.2

89–59

4

31.5

32.5

30.5

42–21

–

–

5

29.9

30.9

28.8

40–20

–

–

6

24.6

24.6

24.6

–

50.9

65–36

7

13.4

13.4

13.4

–

48

63–34

Note: Node times in the timetree represent the mean of time estimates from two studies (10, 17). PATHd8 analysis of DNA sequences from five nuclear genes was conducted in ref. (10). Ref. (17) presents a reanalysis of the same data using the Bayesian program Multidivtime. Some divergence times were estimated here, which is detailed in the text.

~32,000 nucleotides from 19 nuclear gene regions and 169 species. Thus the inferred descent of species, listed by Peters, from a common ancestor is probably true; that is, Piciformes is monophyletic and, thus, a valid taxon. With regard to relationships within and among families, the greatest uncertainty has involved the barbets and toucans. Traditionally the barbets, which occur in the African, Asian, and New World tropics, were considered a single family, the Capitonidae, and the toucans, which are restricted to the New World, were put in a separate family, the Ramphastidae (3). A series of recent studies, however, have been consistent in showing that the traditional Capitonidae is paraphyletic with a close relationship between the New World barbets and toucans and that the Asian and African barbets, each a distinct clade, are more distantly related. These recent studies are based on anatomical (9, 11) as well as molecular characters (4, 12–14). The most definitive study is Moyle’s (14) based on a data set that combines 1045 nucleotides from the mitochondrial encoded cyt b gene with 938 nucleotides from intron 7 of the nuclear-encoded β-fibrinogen gene (β-fibint7). Combining the rapidly evolving cyt b sequences with the more slowly evolving β-fibint7 sequences produced a data set with phylogenetically informative characters at low levels of homoplasy across the full time spectrum of barbet evolution. Moyle also had a relatively complete representation of taxa. The combination of genes evolving at appropriate rates, analytical methods based on detailed nucleotide substitution models (maximum

likelihood and Bayesian) and a dense taxon sample enabled him to resolve nearly complete relationships among species comprising the traditional Capitonidae and Ramphastidae, including enigmatic basal lineages and/or convergent lineages such as Gymnobucco, Calorhamphus, Semnornis, and Trachyphonus. Moyle’s analyses strongly support the existence of three geographically defined clades: Asian barbets, African barbets, and New World barbets, the last one including the toucans. This is consistent with the earlier finding of Sibley and Ahlquist (4), and their recognition of three families is warranted: Lybiidae (African), Megalaimidae (Asian), and Capitonidae (New World). Recognition of the Family Ramphastidae also seems warranted, but the placement of the toucan-barbet, Semnornis ramphastinus, then becomes an issue. This species, which appears very much like what one might imagine as the common ancestor of toucans and New World barbets, is not strongly supported as the closest relative of either. In Moyle’s model-based analyses, Semnornis was closest to the toucans but the ML bootstrap and estimated Bayesian posterior probabilities supporting this node were only 54% and 67%, respectively. The morphological intermediacy of the toucan-barbet, of course, is not the cause of uncertainty in the molecular analyses, but rather the internode that represents the common ancestor of Semnornis, with either the toucans or New World barbets, is short and deep in the tree. Regardless of the exact placement of Semnornis, it, along with the toucans and New World barbets, comprise a clade supported by

448


100% ML bootstrap and posterior probabilities (14). For simplicity of presentation, we will consider the toucans (including Semnornis) a family, Ramphastidae, and similarly we consider the New World barbets (Capitonidae), Asian barbets (Megalaimidae), and African barbets (Lybiidae) as individual families. Monophyly of the other families and their inferred relationships have been more certain. Bucconidae (puffbirds) and Galbulidae (jacamars) are clearly monophyletic and related (4–7, 15). The DNA–DNA hybridization data support the close relationship of Galbulidae and Bucconidae but indicate that the divergence between these two lineages is ancient (4). The Picidae includes the wrynecks, piculets, and woodpeckers, each comprising a subfamily, and is closest to the Indicatoridae (5, 6, 16). In summary, the Order Piciformes as defined here is monophyletic, comprising eight families (Fig. 2). The closest relative of the Piciformes is probably a clade comprising a subset of species traditionally assigned to the Order Coraciiformes (2). Thus, Coraciiformes is paraphyletic but Piciformes is not. The deepest split within Piciformes separates the Suborder Pici from Galbulae. The Galbulae soon bifurcated to give rise to two families, the Bucconidae and Galbulidae. In the Pici, the first bifurcation gave rise to two clades, one comprising the Indicatoridae and Picidae and the other the geographical clades of barbets plus toucans. Within the latter clade, the Asian barbet lineage (Megalaimidae) is basal and the next split gave rise to the common ancestor of African barbets (Lybiidae) and New World barbets plus toucans. Finally, the toucan lineage (Ramphastidae) diverged from the New World barbets (Capitonidae). To date, no study including a molecular clock analysis has focused specifically on Piciformes. However, the broad-based phylogenetic study of 75 families representing essentially all of Neoaves by Ericson et al. (10), and the ensuing comment (17) and reply (18) include molecular clock estimates for five of the seven piciform nodes (Table 1); the remaining two nodes can be estimated from relevant sequences archived in the National Center for Biological Information (NCBI) nucleotide data base. Ericson et al. (10) determined sequences totaling 5007 nucleotides for five nuclear gene regions: c-myc (exon 3), RAG-1, myoglobin (intron 2), β-fibrinogen (intron 7, β-fibint7), and ornithine decarboxylase (introns 6 and 7, and exon 7). They used two computer programs with distinct rate smoothing algorithms, PATHd8 (19) and r8s, a penalized likelihood (PL) method (20), to estimate divergence dates. Calibration was based on 22 fossils mapped onto the Bayesian tree estimated by Ericson et al. (10). A 47.5 million-year-old fossil representing the

hummingbird stem was used as a fi xed calibration point and the remaining 21 fossils established minimum ages for the ancestral lineages they were thought to represent. The PL dates averaged older than the corresponding PATHd8 dates. Moreover, the authors described a “ghost range” in the PL analysis where the origins of lineages represented by fossils averaged 21 million years older than the fossils themselves. Ericson et al. (10) thought the PATHd8 dates to be more reliable, but pointed out that the systematic disparity between the ages of nodes estimated by PATHd8 and PL left the answer to the question of whether diversification of Neoaves came before or after the Mesozoic/Cenozoic boundary (66 Ma) ambiguous, but they presented only the estimates for the PATHd8 analysis (10). Brown et al. (17) reanalyzed Ericson et al.’s (10) data using their revised fossil calibrations, a somewhat different sequence alignment, and a Bayesian methodology (Multidivtime, 21). The contention and disparate results between Brown et al.’s comment (17) and Ericson et al.’s (10) initial study and reply (18) stem from some misunderstanding of details in the original paper (18) but also from different fossil calibrations and analytical methods. The last two illustrate how important these factors are in estimating divergence times based on molecular clocks. These estimates are summarized in Table 1. In addition, we estimated times for the divergence of Ramphastidae, Capitonidae, and Lybiidae as follows (Fig. 2). We downloaded 63 β-fibint7 sequences from 57 piciform species representing all eight families. The sequences were aligned and average genetic distances computed using the MEGA soft ware (Tajima-Nei distance with gamma parameter = 1.0) for all clades shown in Table 1. Two regression lines were then determined for the β-fibint7 distances as functions of the estimated ages inferred using (a) PATHd8 (10) and (b) Multidivtime (17) for all other nodes in Fig. 2. The slopes of the regression lines were estimated such that each line was forced through the origin (i.e., genetic distance equals zero at the time the lineages diverged). The slopes of the leastsquares regression lines are the substitution rates as estimated by each of the two methods. These rates were then interpolated to estimate the ages of the divergence of the two nodes in the ((Ramphastidae, Capitonidae), Lybiida) tree using the genetic distances for β-fibint7. The regression line for the Multidivtime data was a remarkably good fit with no apparent deviation from linearity or outliers (slope = 0.00456 nucleotide substitutions/million years). By this method, the estimated age for the split between Asian barbets (Megalaimidae) and the clade of African and New World barbets (common

Eukaryota; Metazoa; Vertebrata; Aves; Piciformes

ancestor of Ramphastidae, Capitonidae, and Lybiida) is 24 and 13.4 Ma for the split between Capitonidae and Ramphastidae. The regression analysis based on the PATHd8 method is slightly more complex, because an aspect of the PATHd8 program is that it collapses internodes that are short or where there is uncertain phylogenetic resolution (10). Thus, the two oldest splits in Fig. 2 were collapsed to the same level and both dated at 53.0 Ma (Table 1) in Ericson et al.’s (10) PATHd8 analysis. It was apparent from the fitted regression line that the common ancestor of all piciforms is a salient outlier and its age is underestimated by the PATHd8 analysis. Discarding this point and recalculating the slope of the regression line gives, remarkably, a value identical to that fitted to the Multidivtime slope. Thus, the ages estimated for the Ramphastidae, Capitonidae, and Lybiida divergences by the two methods are identical. A better estimate for the common ancestor of Piciformes, based directly on PATHd8 calculated from the regression equation, is 70.8 Ma. Coincidental with the final revision of this paper, Brown et al. (22) published an additional temporal analysis of Aves based on 4594 bp of mtDNA sequence from 135 avian species (ND1, ND2, 12S rRNA, and nine tRNA genes). Although this study focused on diversification of major lineages in relation to the Mesozoic/Cenozoic boundary, estimated dates and confidence intervals for five of the seven piciform nodes can be extracted from their timetree. These have been added to Table 1. Disparities between the ages of nodes based on the mtDNA analysis and the earlier nuclear gene analyses, including those reported in the earlier Brown et al.’s (17) paper, are striking with the five mtDNA-based estimates ranging 1.3–3.6 times older. A thorough exploration of the cause of this apparent bias is beyond the scope of this paper; however, some tentative inferences are evident: The fossil calibrations used in the mtDNA study are nearly identical to those used in the nuclear gene study of Brown et al. (17), which was a modification of the set used by Ericson et al. (10). This suggests that differences in calibration are not the cause of the disparity. Differences in estimation methods suggest a second potential cause. Both sets of nodal times were estimated by Bayesian methods, but the nuclear gene estimates were computed by the program package Multidivtime; whereas, the mitochondrial-gene estimates were computed by BEAST. Brown et al. (22) tested several computational methods on the mtDNA data set: in general the BEAST estimates are higher than the Multidivtime estimates, but not as consistently, or of nearly the magnitude, as the differences between the mitochondrial and

449

nuclear gene estimates. Thus, different estimation procedures might contribute a small amount to the discrepancy, but it is unlikely to be the major factor. This leaves the differences between the data sets, mitochondrial vs. nuclear gene sequences, as a likely cause. Based on simulation studies, Moore et al. (23) showed that the mitochondrial encoded cyt b gene would not perform well as a molecular clock for birds beyond ~10 Ma, whereas the nuclear-encoded β-fibrinogen intron 7 would perform well even to estimate time nodes older than 60 Ma. The rapid evolution of mtDNA and saturation by multiple substitutions leads to underestimation of the substitution rate and over estimation of the age of nodes. For this reason, we did not include the time estimates based on the mtDNA in calculating the average age of piciform nodes, although we did include the estimates in Table 1. It is clear that much work remains to resolve sources of uncertainty and that more attention needs to be paid to differences in the “clocking” accuracy of different genes at different time depths. With the exception of the woodpeckers, the distributions of modern piciform species are restricted to tropical regions of both the New and Old Worlds. This might suggest that their distributions resulted from the breakup of Gondwanaland and the tectonic rifting of the southern continents. The African and South American barbets are closest relatives (14), for example, and their divergence may be attributed to the separation of South America from Africa ~100 Ma (24). However, the timetree dates are at odds with this hypothesis because the inferred dates for all nodes are much too recent; specifically, the split between African and Asian barbets occurred only 24.6 Ma. The piciforms as a whole are relatively weak fliers and dispersers. The Asian barbets, for example, have not crossed Wallace’s Line, and only three species of woodpeckers have crossed the line—and then barely. Trans-Atlantic rafting of terrestrial vertebrates appears to have occurred in rare instances (25) but would seem highly improbable in the case of barbets because they have high metabolic rates and it is doubtful that even a large oceanic raft could support their ecological needs for a sustained journey. Nonetheless, the hypothesis of direct dispersal of an ancestral barbet from Africa to South America via rafting cannot be absolutely rejected. A more plausible hypothesis is that the ancestral piciform species dispersed across Beringia and that they were once distributed broadly across the temperate regions of Eurasia and North America, but these ancestral forms were subsequently extinguished from the northern continents. This is exactly the pattern observed in the fossil record, albeit a sparse fossil record. Moreover, dates

450


associated with various fossil piciforms accord well with the molecular clock estimates for the chronology of nodes. The fossil record shows that early piciforms occurred in both the New and Old World temperate zones in the Lower Eocene (~54 Ma) and perhaps the Paleocene (>54 Ma). Fossils assigned to the extinct piciform Family Gracilitarsidae have been described for the Lower Eocene of Europe, North America, and the Paleocene of Brazil (26). Middle Eocene to Lower Oligocene European fossils assigned to the extinct Family Sylphornithidae exhibit combinations of characters representative of the two major lineages, the Pici and Galbulae, of the Piciformes, and a cladistic analysis based on osteological characters established monophyly of a group comprising Gracilitarsidae, Sylphornithidae, and the crown Piciformes (26). The oldest fossil representative of the Pici dates to the early Oligocene, ~34–30 Ma (27). These fossil dates are reasonably good matches to those inferred in the timetree. The timetree indicates that the ancestral piciform lineage bifurcated ~70.5 Ma to give rise to the Pici and Galbulae lineages. One would expect to find the earliest fossils for the lineage leading to living Piciformes and its two daughter lineages in the Paleocene, Eocene, and Oligocene, which is born out by the fossil record. The inferred bifurcation establishing the Pici and Galbulae predates the Paleocene slightly, but the confidence interval (89–57 Ma) is large and includes much of the Paleocene.

Acknowledgments We thank P. Ericson and J. Brown for information. This work was supported in part by the U.S. National Science Foundation.

References 1. U. S. Johansson, P. G. P. Ericson, J. Avian Biol. 34, 185 (2003). 2. S. J. Hackett et al., Science, 320, 1763 (2008).

3. 4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

20. 21. 22. 23.

24. 25. 26. 27.

J. L. Peters, Check-List of Birds of the World, Vol VI (Harvard University Press, Cambridge, 1948). C. G. Sibley, J. E. Ahlquist, Phylogeny and Classification of Birds: A Study in Molecular Evolution (Yale University Press, New Haven, London, 1990). E. V. Swierczewski, R. J. Raikow, Auk 98, 466 (1981). S. F. Simpson, J. Cracraft, Auk 98, 481 (1981). S. M. Lanyon, R. M. Zink, Auk 104, 724 (1987). S. L. Olson, Auk 100, 126 (1983). P. J. K. Burton, Bull. Brit. Mus. Nat. Hist. Zool. 47, 331 (1984). P. G. P. Ericson et al., Biol. Lett. 2, 543 (2006). R. O. Prum, Zool. J. Linn. Soc. 92, 3131 (1988). S. M. Lanyon, J. G. Hall, Auk 111, 389 (1994). F. K. Barker, S. M. Lanyon, Mol. Phylogenet. Evol. 15, 215 (2000). R. G. Moyle, Mol. Phylogenet. Evol. 30, 187 (2004). E. Hofling, H. M. F. Alvarenga, Zool. Anzeiger 240, 196 (2001). D. M. Webb, W. S. Moore, Mol. Phylogenet. Evol. 36, 233 (2005). J. W. Brown, R. B. Payne, D. P. Mindell, Biol. Lett. 3, 257 (2007). P. G. P. Ericson, C. L. Anderson, G. Mayr, Biol. Lett. 3, 260 (2007). T. Britton, C. Anderson, D. Jacquet, S. Lundqvist, K. Bremer, PATHd8: A Program for Phylogenetic Dating of Large Trees without a Molecular Clock, www.math.su.se/ PATHd8/ (Stockholm University, Stockholm, Sweden, 2007). M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002). J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). J. W. Brown, J. S. Rest, J. Garcia-Moreno, M. Sorenson, D. P. Mindell, BMC Biol. 6, 6 (2008). W. S. Moore, S. M. Smith, T. Prychitko, in Proceedings of 22nd International Ornithological Congress, N. Adams, R. Slotow, Eds. (University of Natal, Durbin, 1999), pp. 745–753. B. C. Storey, Nature 377, 301 (1995). N. Vidal, A. Azvolinsky, C. Cruaud, S. B. Hedges. Biol. Lett. 4, 115 (2007). G. Mayr, Biol. Rev. 80, 515 (2005). G. Mayr, Auk 122, 1055 (2005).

Owls (Strigiformes) Joseph W. Browna,* and David P. Mindella,b a

Department of Ecology and Evolutionary Biology & University of Michigan Museum of Zoology, 1109 Geddes Road, University of Michigan, Ann Arbor, MI 48109-1079, USA; bCurrent address: California Academy of Sciences, 55 Concourse Drive Golden Gate Park, San Francisco, CA 94118, USA *To whom correspondence should be addressed (josephwb@ umich.edu)

Abstract Approximately 202 species of owls comprise the distinctive cosmopolitan neoavian Order Strigiformes. All morphological and genetic studies agree that the order is a natural group. Two families are recognized: Tytonidae (barn owls and bay owls) and Strigidae (typical owls). The strigiform timetree shows that these families are ancient, having diverged in the late Cretaceous 71 million years ago (Ma).

Owls (Order Strigiformes) are grouped into two cosmopolitan families: the species-rich Strigidae (typical owls, ~187 species; Fig. 1) and the relatively depauperate Tytonidae (barn owls and bay owls, ~15 species). Owls are broadly characterized by adaptations to predation (strong zygodactyl feet, raptorial bill and talons, and soft-fringed edges of some flight feathers enabling quiet flight) and adaptations to a predominantly nocturnal or crepuscular lifestyle (large eyes and highly developed auditory system, facilitated by feathers arranged in a distinctive “facial disc”). Here, we review the relationships and divergence times of the strigiform families. Owls form a morphologically homogeneous group that is easily distinguishable from other avian orders. Since the earliest classifications there has been no question that owls form a natural group (1). Recent studies of DNA–DNA hybridization data (1), mitochondrial (mt) (2), nuclear (2–4), and combined (2) DNA sequences, and morphology (5–7) support the monophyletic status of this large avian order. Equally supported is the division of owls into two families, first identified ~160 years ago (8). In addition to the character data establishing monophyly of each family, karyological (9), allozyme (10), and mtDNA restriction fragment (11) data reveal a deep split between the two families. Bay owls (Genus Phodilus) are

in some respects intermediate between Strigidae and the barn owls (Genus Tyto), which underlies early confusion of their taxonomic placement (1). However, recent results strongly support a close relationship between Phodilus and Tyto (2, 5, 9) and thus their placement in the Family Tytonidae. Taxonomy below the family level is currently in a state of flux, partly because the same morphological regularity that clearly delineates groups at higher taxonomic levels hinders phylogenetic classification within the group itself. This is particularly problematic within the speciesrich Family Strigidae, where a comprehensive phylogenetic analysis is long overdue. This is evident from the number of recognized strigid taxa, which increased by 57 species over an 8-year period (12, 13), and which is currently a matter of considerable speculation given recent rediscoveries of taxa previously thought to be extinct (14, 15). Furthermore, a seemingly steady stream of newly described species (16–21), aided in large part by vocalization data, have contributed to uncertainty in the breadth and phylogenetic classification of Strigidae. Relationships of the owls to other avian orders is presently unclear. Historically, Strigiformes has been linked to either the nocturnal nightbirds (Order Caprimulgiformes) (1, 22, 23) or the diurnal raptors (Order Falconiformes) (5). Owls share several morphological characteristics that are separately mirrored in these two orders: the adaptations to a raptorial lifestyle (strong bill and feet) resemble those found in the falconiforms, whereas apparent adaptations to a nocturnal existence

Fig. 1 A Great Grey Owl (Strix nebulosa), Family Strigidae, from Coronado, Alberta. Credit: G. Court.

J. W. Brown and D. P. Mindell. Owls (Strigiformes). Pp. 451–453 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

452

THE TIMETREE OF LIFE Tytonidae

1

K

Strigidae Paleogene

Neogene

CENOZOIC

MZ 50

25

0 Million years ago

Fig. 2 The timetree of owls (Strigiformes). Divergence times are shown in Table 1. Abbreviations: MZ (Mesozoic) and K (Cretaceous).

(large eyes, and soft cryptic plumage used as camouflage from hunting diurnal raptors) are present in the caprimulgiforms. Although recent morphological studies tend to group owls with the diurnal raptors (5, 24–26) (forming the proposed Superorder “Falconimorphae”), molecular genetic studies fail to consistently group owls with any specific order (1, 3, 4, 27, 28). Nevertheless, recent nuclear DNA analyses decisively separate Caprimulgiformes from the two traditional raptorial orders with phylogenetically informative indels (3) and strong statistical support (4, 29). In summary, the morphological and molecular genetic data at hand prohibit uniting all three avian orders. In light of this, some of the defining characteristics of owls have clearly evolved convergently in other lineages, almost certainly with respect to Caprimulgiformes (29), and possibly with Falconiformes (30). Several molecular studies focused on divergence time estimation in birds have included representatives from both strigiform families (Table 1). Early estimates for the timing of divergence between Tytonidae and Strigidae based on allozyme (10) and mtDNA restriction fragment (11) distances yielded similar early Oligocene estimates, 34 Ma and 30–28 Ma, respectively. However, time calibrations used in these two studies are suspect, and subsequent studies with broad species sampling tend to generate much older divergence time estimates. For example, although the age of this divergence was not estimated explicitly by Sibley and Ahlquist (1), the average rate of change in genome-wide DNA–DNA hybridization analyses was 4.5 million years per degree (centigrade) of DNA–DNA melting temperature lowered (1) suggests a much more ancient estimate of ~61 Ma. Recent analyses of five nuclear genes for 87 taxa have generated a range of divergence time estimates when using alternative molecular clock methodologies (4, 31) (Table 1). Original analyses (4) of this data matrix using two different rate smoothing methods recovered dates that ranged from the mid-Eocene (40 Ma) to the late Cretaceous (66 Ma). A reanalysis of these data (31) using

a Bayesian modeling of rate evolution, together with improved fossil constraints, supported the Cretaceous (144–66 Ma) origin of the strigiform families, with the split dated at about 73 Ma (Fig. 2). A variety of analyses making different assumptions about how substitution rate variation evolves in a matrix of ~5 kb of mtDNA for 135 avian taxa (27) generally yielded internally congruent late Cretaceous average age estimates for this node: ancestor-descendant rate smoothing, 95–88 Ma; closestrelative rate smoothing, 89–79 Ma; Bayesian autocorrelated model of rate evolution, 87–82 Ma; overdispersed clock, 92 Ma; Bayesian non-autocorrelated model of rate evolution with unfi xed topology, 84 Ma (Table 1). Although the consensus across these most recent studies supports a late Cretaceous divergence of the two strigiform families, it is evident that the age of this node cannot yet be estimated precisely (Table 1). In a family as large as Strigidae, the sparse taxon sampling used in published dating analyses may have biased subsequent age estimates if the sampled taxa were not representative of their respective families, or if a significant node-density effect was present (32). In addition to the problem of sampling dissimilar species, differences between mt (27) and nuclear DNA (31) estimates may be a result of nucleotide substitution saturation in mtDNA sequences. In contrast to dates inferred from DNA sequence data, fossil gap analysis (33) supports a Cenozoic (44.5 Ma date for the haplorhine node.

34). Among the Malagasy primates, studies yielded an average estimate of 52.2 Ma for the deepest divergence of Daubentoniidae from the remaining lemur families. The deepest node within the remaining four families of lemurs is dated to 37.1 Ma, also based on three studies. Two of these converged at similar dates using different calibration points and data sets [40.9 Ma (18, 35) and 42.3 Ma (10)]. The divergence of Loridae and Galagonidae is estimated at 34.2 Ma. This result is particularly interesting because the aforementioned fossils (Karanisia and Saharagalago) belong to the Loriformes and are dated to ~37 Ma (33, 34). This is the only example where fossil evidence for a taxon antedates the mean given in Table 1. However, this divergence may have occurred earlier than 34.2 Ma. Only three of the studies estimated a date for this node and one used a calibration point for the strepsirrhine/haplorrhine divergence of 63.0 Ma (3), which may be too young. Indeed, although Yoder and Yang (10) calibrated the divergence of Loridae and Galagonidae at 42.0–38.0 Ma based on this fossil evidence, in an earlier study where this divergence was not constrained, they

estimated the node at 40.5 Ma (8). Based on all of the evidence, this divergence date is likely to be closer to ~40 Ma. The earliest divergence within haplorrhines is estimated to be 71.1 Ma. The oldest members of Tarsiidae, Tarsius eocaenus and Xanthorhysis, are dated to the middle Eocene, ~45 Ma, while the earliest stem anthropoid, Eosimias, is similar in age (36, 37). Therefore, the molecular estimate for the divergence date of this node is at least 20 Ma older than the earliest fossil evidence. This suggests that the diversification of primates belonging to living groups may have occurred substantially within the Cretaceous, opening a range of interesting biogeographical questions (38). Based on eight studies, anthropoids are estimated to have split into Platyrrhini and Catarrhini at 44.2 Ma. The earliest catarrhine is Catopithecus (35–34 Ma) (33) and the earliest platyrrhine is Branisella (26–25 Ma) (39). The oldest members of the living platyrrhine radiation come from the early Miocene of South America (6) and a date of 21.4 Ma for the divergence of the living families of

Eukaryota; Metazoa; Vertebrata; Mammalia; Primates

platyrrhines fits well with these fossil data. The estimate for the split of platyrrhines from catarrhines at ~44 Ma and the diversification of the living platyrrhine families at ~21 Ma indicates that the ancestor of New World monkeys migrated to South America sometime during this period. This is particularly interesting, as South America was an island continent during this time. While it is unresolved whether platyrrhine primates derive from North America or Africa, current evidence favors the latter (6). The divergence of the catarrhine groups Hominoidea and Cercopithecidae is estimated to have occurred 29.6 Ma. The oldest cercopithecoids are Victoriapithecus and Prohylobates, dated from the early to middle Miocene, 19–12 Ma (40), while the oldest hominoids are Proconsul and Morotopithecus, both dated to ~20 Ma (41, 42). The fossil record of Africa between 29 and 21 Ma contains few good localities preserving primates. A single hominoid-like fossil (Kamoyapithecus) is dated 27.8–23.9 Ma (43), but the affinities linking this specimen to apes are not diagnostic, making its relevance unclear. Within hominoids, Hylobatidae is estimated to have diverged from Hominidae at 18.8 Ma. The molecular estimates reported here are generally concordant with each other and for the most part are consistent with a timetree based on fossils. However, there are some nodes where molecular estimates greatly exceed the earliest fossil evidence of a taxon. For example, both crown primates and haplorrhines are estimated to originate before the Mesozoic–Cenozoic boundary, while hominoids and cercopithecoids are estimated to have diverged in the earlier part of the Oligocene (~30 Ma). In these instances, molecular estimates are anywhere from 20% to 30% older than those derived from fossil evidence. There are several nonexclusive and potentially overlapping reasons for this result. Critical portions of the fossil record may be undersampled (44), taxa may lack recognizable characters near their origins (45), and homoplasy may make diagnostic traits unreliable (e.g., 46). Alternatively, current molecular estimation techniques may not accurately model important sources or patterns of rate variance that differentially affect these particular nodes (e.g., 47, 48). A combination of paleontological fieldwork targeted on underrepresented portions of the fossil record and empirical research on rates of molecular evolution will help to resolve these areas of contention.

References

Acknowledgment

29. 30.

We thank E. Seiffert for comments.

1.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

485

D. T. Rasmussen, in The Primate Fossil Record, W. Hartwig, Ed. (Cambridge University Press, Cambridge, 2002), pp. 203–220. C. P. Groves, Primate Taxonomy (Smithsonian Institution Press, Washington, DC, 2001). M. Goodman et al., Mol. Phylogenet. Evol. 9, 585 (1998). A. Purvis, Philos. Trans. Roy. Soc. Lond. B Biol. Sci. 348, 405 (1995). J. Shoshani, C. P. Groves, E. L. Simons, G. F. Gunnell, Mol. Phylogenet. Evol. 5, 102 (1996). J. G. Fleagle, Primate Adaptation and Evolution, 2nd ed. (Academic Press, San Diego, 1999). M. DelPero, L. Pozzi, J. C. Masters, Folia Primatol. (Basel) 77, 434 (2006). Z. Yang, A. Yoder, Syst. Biol. 52, 705 (2003). A. D. Yoder, M. Cartmill, M. Ruvolo, K. Smith, R. Vilgalys, Proc. Natl. Acad. Sci. U.S.A. 93, 5122 (1996). A. D. Yoder, Z. Yang, Mol. Ecol. 13, 757 (2004). J. Schmitz, M. Ohme, H. Zischler, Genetics 157, 777 (2001). J. Schmitz, C. Roos, H. Zischler, Cytogenet. Genome Res. 108, 26 (2005). J. C. Opazo, D. E. Wildman, T. Prychitko, R. M. Johnson, M. Goodman, Mol. Phylogenet. Evol. 40, 274 (2006). D. A. Ray et al., Mol. Phylogenet. Evol. 35, 117 (2005). V. M. Sarich, A. C. Wilson, Science 158, 1200 (1967). O. R. Bininda-Emonds et al., Nature 446, 507 (2007). E. Eizirik, W. J. Murphy, M. S. Springer, S. J. O’Brien, in Anthropoid Origins: New Visions. C. F. Ross, R. F. Kay, Eds. (Kluwer Academic/Plenum Publishers, New York, 2004), pp. 45–64. M. E. Steiper, N. M. Young, Mol. Phylogenet. Evol. 41, 384 (2006). R. L. Raaum, K. N. Sterner, C. M. Noviello, C. B. Stewart, T. R. Disotell, J. Hum. Evol. 48, 237 (2005). M. Goodman, J. Barnabas, G. Matsuda, G. W. Moore, Nature 233, 604 (1971). W. H. Li, M. Tanimura, Nature 326, 93 (1987). M. E. Steiper, N. M. Young, T. Y. Sukarna, Proc. Natl. Acad. Sci. U.S.A. 101, 17021 (2004). S. Yi, D. L. Ellsworth, W. H. Li, Mol. Biol. Evol. 19, 2191 (2002). S. Kumar, S. B. Hedges, Nature 392, 917 (1998). J. L. Thorne, H. Kishino, Syst. Biol. 51, 689 (2002). J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). D. T. Rasmussen, in Handbook of Paleoanthropology, W. Henke, I. Tattersall, Eds. (Springer, Berlin, 2007). M. Hasegawa, J. L. Thorne, H. Kishino, Genes Genet. Syst. 78, 267 (2003). C. Poux et al., Syst. Biol. 54, 719 (2005). M. S. Springer, W. J. Murphy, E. Eizirik, S. J. O’Brien, Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003).

486


31. C. Poux, E. J. Douzery, Am. J. Phys. Anthropol. 124, 1 (2004). 32. S. Tavaré, C. R. Marshall, O. Will, C. Soligo, R. D. Martin, Nature 416, 726 (2002). 33. E. R. Seiffert, Proc. Natl. Acad. Sci. U.S.A. 103, 5000 (2006). 34. E. R. Seiffert, E. L. Simons, Y. Attia, Nature 422, 421 (2003). 35. F. C. Chen, W. H. Li, Am. J. Hum. Genet. 68, 444 (2001). 36. K. C. Beard, T. Qi, M. R. Dawson, B. Wang, C. Li, Nature 368, 604 (1994). 37. K. C. Beard, Bulletin of the Carnegie Museum of Natural History 34, 260 (1998). 38. R. D. Martin, C. Soligo, S. Tavaré, Folia Primatol. 78, 277 (2007). 39. B. J. MacFadden, J. Hum. Evol. 19, 7 (1990). 40. B. R. Benefit, M. L. McCrossin, in The Primate Fossil Record. W. Hartwig, Ed. (Cambridge University Press, Cambridge, 2002), pp. 203–220.

41. W. W. Bishop, J. A. Miller, F. J. Fitch, Am. J. Sci. 267, 669 (1969). 42. D. L. Gebo et al., Science 276, 401 (1997). 43. M. G. Leakey, P. S. Ungar, A. Walker, J. Hum. Evol. 28, 519 (1995). 44. R. D. Martin, Nature 363, 223 (1993). 45. S. A. Magallón, Int. J. Plant Sci. 165, S7 (2004). 46. E. R. Seiffert, E. L. Simons, C. V. M. Simons, in Anthropoid Origins: New Visions, C. F. Ross, R. F. Kay, Eds. (Kluwer Academic/Plenum Publishers, New York, 2004), pp. 157–181. 47. N. Elango, J. W. Thomas, S. V. Yi, Proc. Natl. Acad. Sci. U.S.A. 103, 1370 (2006). 48. S. H. Kim, N. Elango, C. Warden, E. Vigoda, S. V. Yi, PLoS Genet. 2, e163 (2006). 49. W. J. Murphy et al., Nature 409, 614 (2001). 50. C. Poux, P. Chevret, D. Huchon, W. W. de Jong, E. J. Douzery, Syst. Biol. 55, 228 (2006).

Pikas, hares, and rabbits (Lagomorpha) Conrad A. Matthee Evolutionary Genomics Group, Department of Zoology, University of Stellenbosch, Stellenbosch, 7602 South Africa ([email protected])

Abstract The pikas, hares, and rabbits (~90 species) are grouped into two families, Ochotonidae and Leporidae, within the mammalian Order Lagomorpha. Based on several molecular studies, the estimated time of divergence of these two families has varied between 30.4 and 51.0 million years ago (Ma), with an average of 40 Ma. Climatic shifts during the Eocene–Oligocene transition may have been responsible for this divergence.

The pikas, hares, and rabbits (Fig. 1) are united into the monophyletic Order Lagomorpha based on the presence of a second peg-like upper incisor and foot morphology showing a unique calcaneal canal running diagonally through the lagomorph calcaneus (1). Based on the latter observation, which is also present in lagomorph fossils, it has been proposed that the order has had a long evolutionary separation from other related mammalian orders such as the Rodentia (1–3). The Lagomorpha currently comprises two extant families, the Ochotonidae (pikas, 30 species and one genus) and Leporidae (rabbits and hares, 60 species and 11 genera) (4). The monotypic pikas have a northern hemisphere distribution with most taxa confined to Asia, where the uplifting of the Tibet Plateau probably played a major role in the diversification of taxa (5). The 11 genera of rabbits and hares have a nearly global distribution (Holarctic, Ethiopian excluding Madagascar, northern Neotropical and Oriental) and their diversification is more than likely due to a series of dispersal and vicariance events that could be correlated to the formation and disappearances of intercontinental landbridges (6). From a paleontological perspective, evidence suggests either an Asian or North American origin for the order, but the pinpointing of the exact location is hampered by a scattered fossil record for the early and middle Eocene (7). It has been postulated that the first expansion of Leporidae occurred in North America during the Miocene (7), a notion supported by at least two recent fossil discoveries

(8, 9), whereas the Ochotonidae probably originated in Asia. Here I review the available information on the timing of the Ochotonidae and Leporidae divergence. The fossil record of the Lagomorpha dates back to at least 45 Ma (1, 7) and the oldest fossil member of the Ochotonidae is known from the early Oligocene of Mongolia ~33–32 million years ago (10). Erbajeva (11) proposed that during this same period the Leporidae and the Ochotonidae started to diverge from each other and although this date falls within the 40–30 million years suggested by Dawson (7), it is more recent than the 50 million years proposed by Benton (12). Analyses of 58 presumptive structural allozyme loci estimated the time of divergence between the Leporidae and the Ochotonidae at 37.5 Ma (13). Using seven gene regions (mtDNA and nuclear introns), the age of the divergence among the families was estimated at 31.7–29.0 Ma in Bayesian and maximum likelihood analyses that used multiple calibration points and allowed for heterogeneity in molecular evolutionary parameters among genes (6). The 95% credibility interval in this study was estimated

Fig. 1 A Riverine Rabbit (Bunolagus monticularis) from South Africa. Credit: T. Camacho, Science Photo Library/Images of Africa/Okapia.

C. A. Matthee. Pikas, hares, and rabbits (Lagomorpha). Pp. 487–489 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

488


Ochotonidae 1

Leporidae Paleogene

Neogene CENOZOIC

40

30

20

0 Million years ago

10

Fig. 2 A timetree of pikas, hares, and rabbits (Lagomorpha). Divergence times are shown in Table 1.

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) between lagomorph families. Timetree Node

1

Time

39.6

Estimates Ref. (6)

Ref. (13)

Ref. (14)

Refs. (15, 16)

Time

CI

Time

CI

Time

CI

Time

CI

30.4

39–22

37.5

–

38.9

39–36

51.0

67–35

Note: Node times in the timetree represent the mean of time estimates.

to be 39.3–22.4 Ma. Using the same methodology and sequences derived from the mtDNA genome only, the date was estimated at 38.9 Ma, ranging from 39.9 to 36.3 Ma (14). Using a larger data set that underpinned a broader study dealing with the diversification of Placentalia, the date for the leporid/ochotonid split was estimated at 51.0 Ma, with a large credibility interval of ±16 million years (15, 16). The discrepancy in these molecular dates is likely due to the inconsistencies in calibration points (constraints). The Springer et al. (15) study used a single Euarchontoglires calibration point (mouse and rat) which was repeated in the Murphy et al. (16) study, while the Matthee et al. (6) investigation focused specifically on leporid evolution used six time constraints within Glires. The Horner et al. (14) study used four lagomorph constraints. Although Matthee et al. (6) allowed different loci to have different patterns of evolutionary change, the divergence time among the Leporidae and Ochotonidae was constrained to be within 40–20 Ma (6). Horner et al. (14) likewise biased the estimate by constraining the divergence of the lagomorph families to between 40 and 35 Ma. The exact date of the split between the Ochotonidae and the Leporidae remains controversial (Table 1), but consensus suggests that the two lagomorph families diverged at some point between the late Eocene (~45 Ma) and early Oligocene (~30 Ma) (Fig. 2). The

Eocene–Oligocene boundary (~34 Ma) is characterized by sudden climatic shifts in global temperatures (a warming trend was immediately followed by cooling events) (17) culminating in large changes in mammalian diversity (18). It is probable that the Eocene–Oligocene transition also contributed toward the diversification of the Lagomorpha. Furthermore it is believed that the pikas followed a diversification trend throughout the Oligocene where they subsequently spread from Asia to North America (and also later to Africa) to become most numerous during the Miocene (25 fossil genera) (10). The Leporidae, on the other hand, did not show a similar diversification trend until the middle Miocene (~15 Ma) at which point the Ochotonidae decreased in diversity and the Leporidae began an increase in diversity (6).

Acknowledgments Comments were provided by T. Robinson and support was provided by the South African National Research Foundation.

References 1. A. R. Bleefeld, W. J. Bock, Acta Palaeontol. Pol. 47, 181 (2002). 2. J. Meng, Bull. Am. Mus. Nat. Hist. 285, 93 (2004). 3. R. J. Asher et al., Science 5712, 1091 (2005).

Eukaryota; Metazoa; Vertebrata; Mammalia; Lagomorpha

4. R. S. Hoff mann, A. T. Smith, in Mammal Species of the World: a Taxonomic and Geographic Reference, D. E. Wilson, D. M. Reeder, Eds. (John Hopkins University Press, Baltimore, 2005), pp. 185–211. 5. N. Yu, C. Zheng, Z. Feng, Acta Theriol. Sin. 12, 255 (1992). 6. C. A. Matthee et al., Syst. Biol. 53, 433 (2004). 7. M. R. Dawson, in Proceedings of the World Lagomorph Conference, K. Myers, C. D. MacInnes, Eds. (University of Guelph Press, Ontario, 1981), pp. 1–8. 8. J. A. White, J. Vert. Paleont. 1, 67 (1991). 9. M. R. Voorhies, C. L. Timperley, J. Vert. Paleont. 17, 725 (1997). 10. B. P. Kraatz, J. Vert. Paleont. 22, 76 (2002).

489

11. M. A. Erbajeva, in Rodent and Lagomorph Families of Asian Origins and Diversification, Y. Tomida, C.-K. Li, T. Setoguchi, Eds. (National Science Museum Monographs, No. 8, Tokyo, 1994), pp. 1–13. 12. M. J. Benton, in The Fossil Record 2 (Chapman & Hall, London, 1993), pp. 845. 13. M. Grillitsch et al., Acta Theriol. 37, 1 (1992). 14. D. S. Horner et al., BMC Evol. Biol. 7, 16 (2007). 15. M. S. Springer et al., Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003). 16. W. J. Murphy et al., Genome Res. 17, 413 (2007). 17. E. A. Bestland, J. Sediment Res. 67, 840 (1997). 18. R. A. Kerr, Science 257, 1622 (1992).

Rodents (Rodentia)

The Order Rodentia is the most diverse group of mammals, represented by 34 extant families. Recent molecular data have helped to clarify relationships among these families, and provide a framework for the higher-level classification of rodents. Molecular time estimates among rodent families have been made in several studies, permitting a timetree of rodent evolution to be constructed. The timetree shows a partitioning of rodent families into three major clades and reveals three periods of diversification: The late Cretaceous, 88–66 million years ago (Ma), the Paleocene to early Eocene, 60–55 Ma, and the late Oligocene to early Miocene, 25–15 Ma.

to either features of the zygomasseteric system (configuration of the infraorbital foramen and placement of the masseter muscles in the jaw) or the angle of the lower jaw relative to the plane of the incisors (15, 16). Tullberg’s classification (15) identifies two suborders, Sciurognathi and Hystricognathi, based on the angle of the lower jaw, and the hystricognathous condition supports the monophyly of a clade containing phiomorph and caviomorph rodents (Fig. 2). Nevertheless, the phylogenetic distribution of features of the infraorbital foramen reveals evidence of parallel changes within Rodentia (17–19). Detailed molecular phylogenetic studies contribute greatly to the resolution of the rodent evolutionary tree, and areas of congruence among the majority of these studies provide strong support for many interfamilial relationships among rodents (Fig. 2). For instance, two major monophyletic groups, Hystricomorpha and a squirrel-like clade, are supported by nuclear (2, 19–22)

The Order Rodentia (represented by squirrels, mice, rats, guinea pigs, and others) contains 42% (2277) of all species and 39% (481) of all genera described for the Class Mammalia (1) (Fig. 1). If one includes the recently discovered extant Family Diatomyidae (2), the order contains 34 families (1). All early classifications of mammals (3) and more recent cladistic analysis of morphological variation (craniodental, postcranial, fetal membrane) (4) support a monophyletic Rodentia. Although several early molecular phylogenetic studies questioned the validity of a monophyletic Rodentia (5, 6), more recent molecular data, based on either nuclear DNA (7–9) or a more thorough analysis of sequences of whole mitochondrial genomes (10), reveal support in favor of monophyly. In this review I will address phylogenetic relationships and divergence times for families within Rodentia. Even though the assignment of species and genera of rodents to specific families is well established (1), the derivation of a family-level phylogeny for rodents is more challenging (11–14). This problem has resulted in a variety of classifications that vary in terms of recognized suborders and the morphological features used to define monophyletic groups (1, 11, 15–17). Two of the major characteristics used to diagnose suborders relate

Fig. 1 A capybara (Hydrochaeris hydrochaeris), Family Caviidae. Credit: R. L. Honeycutt.

Rodney L. Honeycutt Division of Natural Science, Pepperdine University, 24255 Pacific Coast Highway Malibu, CA 90263-4321, USA (rodney.honeycutt@ pepperdine.edu)

Abstract

R. Honeycutt. Rodents (Rodentia). Pp. 490–494 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Mammalia; Rodentia

491

Dasyproctidae

18

20 22

Agoutidae Caviidae

27

Chinchillidae Abrocomidae

17

Octodontidae 26 13

28

Ctenomyidae

25

11

Bathyergidae 15

Petromuridae 21

Thryonomyidae

7

Phiomorpha

Capromyidae 29 Echimyidae 30 Myocastoridae

Hystricomorpha

Dinomyidae

16

Caviomorpha

Erethizontidae

Hystricidae Diatomyidae 12

Ctenodactylidae Anomaluridae

2

10

Peditidae

4

Dipodidae 6 14

3

Nesomyidae 23 24

1

Cricetidae Muridae

Mouse-like

Spalacidae

Castoridae 5

Geomyidae 19

Gliridae 8

Sciuridae 9

Late K

Aplodontidae

Paleogene

MESOZOIC 75

Squirrel-like

Heteromyidae

Neogene

CENOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of rodents (Rodentia). Divergence times are shown in Table 1. The muroid Families Calomyscidae (no divergence time data) and Platcanthomyidae (recently extinct) were not included. Abbreviation: K (Cretaceous).

and mitochondrial (18, 24, 25) sequences, and some nuclear markers (2, 7, 21), including the presence/ absence of retrotransposon insertion loci (9), support a third mouse-like clade. Relationships among these three

clades are somewhat more tenuous, but many molecular studies place the squirrel-like clade at the base of the phylogeny (7, 8, 20, 21, 23). Recent molecular studies also help resolve the placement of several historically

492


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among rodents (Rodentia). Timetree Node

Estimates Ref. (2)

Time

Ref. (21)

Ref. (24)

Ref. (27)

Ref. (28)

Ref. (30)

Ref. (31)

Time

CI

Time

CI

Time

Time

CI

Time

CI

Time

CI

Time

1

92.9

–

–

96.9

105–85

–

–

–

–

–

–

–

–

2

82.8

72.1

80–64

93.5

106–77

–

–

–

–

–

–

–

–

3

78.9

67.8

76–60

90

101–68

–

–

–

–

–

–

–

–

4

76.0

66.5

75–58

85.4

101–65

–

–

–

–

–

–

–

–

5

66.4

63.3

72–55

84.9

93–66

50.9

–

–

–

–

–

–

–

6

66.0

65.4

73–57

87

82–53

45.7

–

–

–

–

–

–

–

7

64.2

61.3

63–54

83.7

97–67

47.6

–

–

–

–

–

–

–

8

62.4

57.3

65–50

78.8

90–62

51.2

–

–

–

–

–

–

–

9

51.5

50.8

59–43

65.3

82–53

38.3

–

–

–

–

–

–

–

10

48.3

56.8

65–49

–

–

39.7

–

–

–

–

–

–

–

11

47.8

45.4

51–40

43.1

51–29

–

55

63–46

–

–

–

–

–

12

44.3

44.3

51–38

–

–

–

–

–

–

–

–

–

13

41.7

42.7

47–37

38.1

54–16

–

49

–

–

36.6

39–34

–

54–43

14

39.0

–

–

–

–

–

–

–

–

–

–

–

39

15

38.0

38

43–33

–

–

–

45

48–41

–

–

30.5

34–27

–

16

35.3

34.3

37–30

–

–

–

–

–

37.7

40–35

33.8

36–32

–

17

32.2

33.3

36–29

32.8

47–24

30.4

–

–

–

–

–

–

–

18

31.7

31.7

35–27

–

–

–

–

–

–

–

–

–

–

19

28.4

27.3

33–22

–

–

–

27

29–25

–

–

31.4

33–29

–

20

26.0

–

–

–

–

–

24

26–21

–

–

27.9

30–26

–

21

25.8

24.5

43–20

–

–

–

27

28–25

–

–

–

–

–

22

25.6

24.1

28–20

–

–

–

–

–

–

–

26.5

29–24

–

23

25.5

–

–

–

–

–

–

–

–

–

–

–

25.5

24

24.2

–

–

–

–

–

–

–

–

–

–

–

24.2

25

21.3

18.6

22–15

–

–

–

–

–

27.4

32–23

17.5

20–15

–

26

21.0

–

–

–

–

–

–

–

–

–

20.6

23–18

–

27

20.2

21.4

26–18

–

–

–

–

–

–

–

19.1

22–16

–

28

17.3

–

–

–

–

–

–

–

19.6

23–17

15

17–13

–

29

10.7

12.4

15–10

–

–

–

9

10–7

–

–

–

–

–

30

7.3

–

–

–

–

–

6

7–5

–

–

8.6

10–7

–

Note: Node times in the timetree represent the mean of time estimates from different studies. In ref. (30), confidence interval is based on estimates of standard deviations provided. In ref. (27), divergence time represents an average of two estimates of the same gene ( VWF) derived from nucleotide and amino acid sequences, and the CI refers to the range of these estimates. In ref. (21), divergence time represents an average of the GHR and BRCA1 genes; only values derived from the rate-smoothing method were used; CI represents the combine rate of values for both genes. In ref. (28), divergence time represents an average of estimates derived from nuclear (GHR) and mitochondrial (12S rRNA) combined genes, and CI is the range of values. In ref. (31), estimates for Node 14 based on GHR gene, optimized with rate smoothing, and dates for Nodes 23 and 24 represented values from a concatenation of all genes and optimization with the penalized likelihood method.

problematic taxa, including the Families Pedetidae (Springhaas), Anomaluridae (scaly-tailed squirrel), Geomyidae/Heteromyidae (pocket gopher and pocket mouse), and Castoridae (beaver) (2, 7, 9, 21, 24).

Tullberg’s (15) Suborder Hystricognathi, characterized by a hystricomorphous zygomasseteric system and a hystricognathous lower jaw, is strongly supported by molecular studies (2, 9, 18, 20, 21, 25). In addition,

Eukaryota; Metazoa; Vertebrata; Mammalia; Rodentia

monophyly of the South American Caviomorpha (guinea pigs and relatives) is strongly supported, suggesting a single invasion from African ancestors, represented today by the Phiomorpha (2, 7, 9, 20, 26). The placement of Old World porcupines (Hystricidae) is more controversial, yet the majority of data, based on nuclear and mitochondrial sequences, suggests a basal position for the family (2, 20, 21, 23, 27). Relationships among families within the Caviomorpha are well resolved with a combination of nuclear and mitochondrial sequences, and with few exceptions (e.g., clade containing Chinchilidae and Dinomyidea, Fig. 2) these results are similar to groups previously defined based on morphology (22, 23, 25, 27–29). The current classification of rodents assigns the 34 families to five suborders (1), and molecular data support the monophyly of these five suborders: Sciuromorpha, Castorimorpha, Myomorpha, Anomaluromorpha, and Hystricomorpha (Fig. 2). Three of these suborders comprise the mouse-like clade. As indicated earlier, placement of some of these major groups (e.g., basal squirrel-like clade and the clade uniting Muroidea, Dipodidae, Peditidae, and Anomaluridae, Fig.2) is not well resolved at this time. Divergence times for the rodent evolutionary tree (Fig. 2, Table 1) represent a compilation from several molecular studies based on the following combinations of genes: (a) portions of four nuclear protein-coding genes (ADRA2B, GHR, IRBP, VWF) and two mitochondrial genes (cytochrome b and 12S rRNA) (2); (b) two mitochondrial genes (Cyt b and 12S rRNA) (24); (c) a nuclear gene (GHR) (20); (d) two nuclear genes (GHR, BRCA1) (21); (e) one nuclear gene (VWF) (22, 27); (f) one nuclear (GHR) and one mitochondrial gene (12S rRNA) (28, 30); (g) four nuclear genes (GHR, BRCA1, RAG1, c-myc) (31); and (h) whole mitochondrial genomes (32). Calibration points and methods used to estimate divergence times vary across these studies. Nevertheless, all of these studies test for uniform rates before applying a molecular clock, and most provide dates estimated by methods that correct for rate heterogeneity across lineages (33–35). The beginning of the rodent radiations (Table 1) represents an average of two estimates: 96.9 Ma (21) and 88.8 ± 4.3 Ma (32). The average date of 92.9 Ma obtained from these two estimates is larger than the basal diversification date of 85.3 Ma based on a recent supertree for extant mammals (36), yet is considerably less than the estimate of 110 Ma provided by Kumar and Hedges (37) for the split between the two major groups of rodents. All of these molecular dates are considerably older than the first appearance of rodents in the fossil record 60–55 Ma (38).

493

The timetree does provide a framework for testing hypotheses related to the distribution and diversification of rodents. The phylogeny in combination with estimates of divergence times (Fig. 2 and Table 1) reveals three major diversifications of rodents, one occurring in the late Cretaceous and early Paleocene, another in the Paleocene to early Eocene, and another that involved the diversification of African phiomorphs and South American caviomorphs in the late Oligocene to early Miocene. This pattern is similar to a recent proposal for mammalian radiations that suggests an early origin for many lineages followed by an increase in diversification during the Eocene and Oligocene (36). Many rodent families in the timetree also reveal long terminal branches and shorter internodes, features that make the resolution of the phylogeny more challenging. The timetree also provides insight into the biogeography of hystricognath rodents. For instance, caviomorph rodents first appear in the fossil record of South America 37–31 Ma (39) and the oldest phiomorphs in Africa date to 37–34 Ma (38), and these dates are similar to those indicated in Table 1. Consequently, these dates are congruent with the hypothesis that the invasion of South America by African hystricognath ancestors involved overwater waif dispersal across an ~1700 km expanse of the Atlantic subsequent to the separation of these two continents at a much earlier date. The classification of rodents has vacillated for well over a century, as a result of a lack of clear understanding of interfamilial relationships. Morphological comparisons of both extant and extinct forms have resulted in a host of phylogenetic hypotheses regarding not only the number of suborders but also the placement of several problematic taxa that are sometimes left as status undetermined in many classifications. Recent molecular studies contribute to the resolution of many problems, and they indicate that traditional morphological features used in earlier classifications have evolved independently in many cases. The timetree in Fig. 2 represents an interpretive framework for testing many hypotheses pertaining to behavioral and ecological evolution. Just as important is the observation that rates of molecular evolution vary across lineages and genes, thus influencing estimates of divergence times. Many of the estimates presented in Table 1 attempt to minimize the effects of rate heterogeneity by the inclusion of multiple calibration points and various types of rate-smoothing methods. The accuracy of these approaches is still debatable, and rodents provide an excellent model for investigating methods for deriving a molecular clock. In addition, rodents are diverse in terms of metabolic rates and body

494


size, and there is at least some information that suggests these features may be correlated with differences in rates across groups of rodents (29). Clearly, this is an area that needs further investigation.

Acknowledgments I thank U.S. National Science Foundation for supporting my research on rodent phylogenetics, and I thank my former graduate students, D. Rowe, M. Allard, M. Nedbal, C. Ingram, and L. Frabotta, for their collaborations on several studies related to the rodent tree of life.

References 1. G. G. Musser, M. D. Carleton, in Mammal Species of the World: A Taxonomic and Geographic Reference, 3rd ed., D. E. Wilson, D. M. Reeder, Eds. (Johns Hopkins University Press, Baltimore, Maryland, 2005), pp. 745–752. 2. D. Huchon et al., Proc. Natl. Acad. Sci. U.S.A. 104, 7495 (2007). 3. G. G. Simpson, Am. Mus. Nat. Hist. Bull. 85, 1 (1945). 4. W. P. Luckett, J.-L. Hartenberger, J. Mammal. Evol. 1, 127 (1993). 5. D. Graur, W. A. Hide, and W.-H. Li, Nature 351, 649 (1991). 6. A. M. D’Erchia, C. Gissi, G. Pesole, C. Saccone, U. Arnason, Nature 381, 597(1996). 7. W. J. Murphy et al., Nature 409, 614 (2001). 8. W. J. Murphy et al., Science 294, 2348 (2001). 9. A. Farwick et al., Syst. Biol. 55, 936 (2006). 10. K. M. Kjer, R. L. Honeycutt, BMC Evol. Biol. 7, 8 (2007). 11. A. E. Wood, J. Mammal. 36, 165 (1955). 12. A. E. Wood, Evolution 19, 115 (1965). 13. J.-L. Hartenberger, in Evolutionary Relationships among Rodents: A Multidisciplinary Analysis, W. P. Luckett, J.-L. Hartenberger, Eds. (Plenum Press, New York, 1985), pp. 59–81. 14. J. J. Jaeger, in The Phylogeny and Classification of the Tetrapods. Vol. 2, Mammals, M. J. Benton, Ed. (Clarendon Press, New York, 1988), pp. 177–199. 15. T. Tullberg, Nova Acta Regiae Societatis Scientarium Upsaliensis 18, 1 (1899).

16. J. K. Brandt, Mémoires de l’Academie Imperiale des Sciences de St. Petersbourg 69, 1 (1855). 17. R. L. Honeycutt, L. J. Frabotta, D. L. Rowe, in Rodent Societies: An Ecological and Evolutionary Perspective, J. O. Wolff and P. W. Sherman, Eds. (The University of Chicago Press, Chicago, Illinois, 2007), pp. 8–23. 18. M. A. Nedbal, R. L. Honeycutt, D. A. Schlitter, J. Mammal. Evol. 3, 201 (1996). 19. L. Marivaux, M. Vianey-Liaud, J.-J. Jaeger, Zool. J. Linn. Soc. 142, 105 (2004). 20. R. M. Adkins, E. L. Gelke, D. Rowe, R. L. Honeycutt, Mol. Biol. Evol. 18, 777 (2001). 21. R. M. Adkins, A. H. Walton, R. L. Honeycutt, Mol. Phylogenet. Evol. 26, 409 (2003). 22. D. Huchon, F. M. Catzeflis, E. J. P. Douzery, Proc. Roy. Soc. Lond. B 276, 393 (2000). 23. D. Huchon et al., Mol. Biol. Evol. 19, 1053 (2002). 24. C. Montgelard, S. Bentz, C. Tirard, O. Verneau, F. M. Catzeflis, Mol. Phylogenet. Evol. 22, 220 (2002). 25. D. L. Rowe, Molecular Phylogenetics and Evolution of Hystricognath Rodents, Ph.D. Dissertation (Texas A&M University, College Station, 2002), pp. 183. 26. M. A. Nedbal, M. W. Allard, R. L. Honeycutt, Mol. Phylogenet. Evol. 3, 206 (1994). 27. D. Huchon, E. J. P. Douzery, Mol. Phylogenet. Evol. 20, 238 (2001). 28. R. L. Honeycutt, D. L. Rowe, M. H. Gallardo, Mol. Phylogenet. Evol. 26, 476 (2003). 29. D. L. Rowe, R. L. Honeycutt, Mol. Biol. Evol. 19, 263 (2002). 30. J. C. Opazo, Mol. Phylogenet. Evol. 37, 932 (2005). 31. S. J. Steppan, R. M. Adkins, J. Anderson, Syst. Biol. 53, 533 (2004). 32. Y. Cao, M. Fujiwara, M. Nikaido, N. Okada, M. Hasegawa, Gene 259, 149 (2000). 33. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 34. M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002). 35. A. D. Yoder, Z. Yang, Mol. Biol. Evol. 17, 1081 (2000). 36. O. R. P. Bininda-Emonds et al., Nature 446, 507 (2007). 37. S. Kumar, S. B. Hedges, Nature 392, 917 (1998). 38. J.-L. Hartenberger, C. R. Acad. Sci. Paris Sci. Terre Planètes 326, 439 (1998). 39. A. R. Wyss et al., Nature 365, 434 (1993).

Hedgehogs, shrews, moles, and solenodons (Eulipotyphla) Christophe J. Douadya,b,c* and Emmanuel J. P. Douzeryd,e a

Université de Lyon, F-69622, Lyon, France; bUniversité Lyon 1, F-69622 Villeurbanne, France; cLaboratoire d'Ecologie des Hydrosystèmes Fluviaux (UMR CNRS 5023), F-69622 Villeurbanne, France; dUniversité Montpellier II, Place Eugène Bataillon, 34095 Montpellier Cedex 05, France; eCNRS, Institut des Sciences de l'Evolution (UMR 5554), CC064-Place Eugène Bataillon, 34095 Montpellier Cedex 05, France *To whom correspondence should be addressed ([email protected])

Abstract Hedgehogs, shrews, moles, and solenodons (~450 sp.) are grouped into four to five families within the mammalian Order Eulipotyphla. Molecular phylogenetic analyses have resulted in major changes in their classification. Former allies have been excluded from what was previously called Lipotyphla or Insectivora. Hedgehogs are considered closest relatives of shrews, with solenodons as the most basal offshoot. The Eulipotyphla timetree shows that the major groups diversified ~80 million years ago (Ma) in the late Cretaceous. Events that led to the mass extinction at the end of the Cretaceous period (66 Ma) might have been instrumental in separating the ancestral shrews and hedgehogs.

Hedgehogs, shrews (Fig. 1), moles, and solenodons form a single, natural group of small mammals, the Order Eulipotyphla. The 452 currently recognized species (1) belong to four living families: Erinaceidae (hedgehogs; two subfamilies, 10 genera, and 24 species), Soricidae (shrews; three subfamilies, 26 genera, and 376 species), Talpidae (moles; three subfamilies, 17 genera, and 39 species), Solenodontidae (solenodons; one genus and four species), and the recently extinct Nesophontidae (West Indian shrews; one genus and nine species). Shared morphological characters include a simple hindgut without a caecum, typically long narrow snouts, and reduced to absent eyes. However, the lack of unique derived characters has convinced many zoologists that they resemble the basic stock that gave rise to most eutherian lineages. Here, we review the relationships and divergence times

of these four living families. We begin by first placing them in the context of what has long been considered as a taxonomic wastebasket, the Insectivora. Under its broadest meaning the former Order Insectivora sensu Wagner encompassed 10 distinct families: Eulipotyphla plus Tenrecidae (tenrecs), Chrysochloridae (golden moles), Macroscelidae (elephant shrews), Tupaiidae (tree shrews), and Cynocephalidae (flying lemurs). However, since Wagner (2), the taxonomic content of the order has gradually decreased. In one of the first attempts to accommodate heterogeneity within Insectivora, Haeckel (3) proposed to split insectivores into two suborders, Menotyphla for insectivores with a caecum (elephant shrews, tree shrews, and flying lemurs) and Lipotyphla for insectivores without a caecum (Eulipotyphla, tenrecs and golden moles). However, evidence arguing against the Menotyphla concept accumulated and these three families were consecutively placed in their own orders. First Leche (4) removed flying lemurs in 1885 but they were not assigned to their own order until 1945 (5). Then Butler (6, 7) assigned ordinal status to both elephant shrews and tree shrews, in 1956 and 1972, respectively. Contrary to this gradual sundering of Menotyphla, morphological studies never challenged the reality of the Suborder Lipotyphla. Regarding

Fig. 1 A Water Shrew (Neomys fodiens), Family Soricidae. Credit: P. Vogel.

C. J. Douady and E. J. P. Douzery. Hedgehogs, shrews, moles, and solenodons (Eulipotyphla). Pp. 495–498 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

496

THE TIMETREE OF LIFE Soricidae 3

Erinaceidae

2

Talpidae

1

Solenodontidae Late K

Paleogene

MZ 75

Neogene

CENOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of hedgehogs, shrews, moles, and solenodons (Eulipotyphla). Divergence times are from Table 1. Abbreviations: MZ (Mesozoic) and K (Cretaceous).

this last group, only the interfamilial and interordinal relationships were questioned (e.g., 8, 9). Unfortunately, and as suggested earlier, the morphological “primitiveness” of this group (i.e., they resemble the “undifferentiated eutherian,” 10), which makes them key taxa in understanding mammalian body plan evolution, has also made difficult attempts at deriving their evolutionary history using morphological data. Without surprise, it appears that molecular approaches have significantly transformed our perception of the group. One of the first successes of molecular tools applied to insectivore-like taxa was the corroboration of the close affinity of flying lemurs (Dermoptera), tree shrews (Scandentia), and Primates (e.g., 11–13). Unexpectedly, de Jong et al. (14) placed elephant shrews within the African supraordinal clade, which is known as Afrotheria (15). This position for elephant shrews has now been corroborated by numerous other molecular studies (e.g., 16–20) and by some paleontological evidence (21). Moreover, Douzery and Catzeflis (22) suggested an association between chrysochlorids and afrotherians, which were represented in their study by a golden mole (Amblysomus) and a hyrax (Procavia), respectively. More recently, several significant improvements in understanding lipotyphlan relationships were obtained with expanded sampling of species. Springer et al. (17), Stanhope et al. (15), and Douady et al. (23) demonstrated that two former members of Lipotyphla belong in Afrotheria. Interestingly, these two taxa, golden moles and tenrecs (Afrosoricida), had previously been associated in morphological phylogenies (24, 25, but see 8), but within an intact (earlier) concept of Lipotyphla. Based on these new results, remaining lipotyphlans (hedgehogs, moles, shrews, and solenodons) were placed in a new order coined Eulipotyphla (26). Phylogenetic inferences based on complete mitochondrial genomes even suggested that these remaining taxa

may not constitute a natural grouping. Indeed, while the European hedgehog (Erinaceus europaeus) was generally identified as the first offshoot of the placental tree (27, 28), other eulipotyphlans such as the European mole (Talpa europaea) seemed more closely related to Laurasiatheria (29) (bats, cetartiodactyls, perissodactyls, carnivores, and pangolins). However, this hypothetical diphyly of Eulipotyphla was soon rejected. First, nuclear studies based on up to 19 nuclear and three mitochondrial gene fragments (30, 31) and data from species representing all subfamilies (32) did not corroborate a basal hedgehog position (13, 30–33). Second, improved mitogenomic analyses favored eulipotyphlan monophyly (34–37). Denser taxonomic sampling and better suited models of sequence evolution indeed suggested that the basal position first reported for hedgehog was mostly due to an artifact potentially resulting from their peculiar base composition (35–39). Molecular evidences have also challenged intraordinal associations. Morphology had suggested a fundamental split of Lipotyphla into Erinaceidae and all other families (Erinaceomorpha vs. Soricomorpha sensu 25). Stanhope et al. (15) refuted the Soricomorpha concept by exclusion of tenrecs and golden moles from this group. Then, a definite rejection of shrews–moles affinities came from Murphy et al. (30, 31) and Douady et al. (23, 32). Considering hedgehogs, shrews and moles as representative, they both supported erinaceomorph hedgehogs as the closest relative of soricomorph shrews. Roca et al. (41) even further discredited the soricomorph assemblage in supporting solenodontids as the most basal lineage in Eulipotyphla. This evidence based on 19 nuclear and three mitochondrial gene fragments conclusively resolved the position of the sole living family that was absent from the already well-established eulipotyphlan phylogeny (Fig. 2). Now, ancient DNA studies are awaited to elucidate the phylogenetic affinities of the last family,

Eukaryota; Metazoa; Vertebrata; Mammalia; Eulipotyphla

497

Table 1. Divergence times (Ma) and their confidence/credibility (CI) intervals among hedgehogs, shrews, moles, and solenodons (Eulipotyphla). Timetree Node

Time


CI

Ref. (41)(b) Time

CI

Ref. (43)(a) Time

CI

Ref. (43)(b) Time

CI

Ref. (43)(c) Time

CI

1

80.5

85

95–75

76

81–72

–

–

–

–

–

–

2

74.3

73

86–61

73

78–68

69.8

81–59

72.9

82–64

82.8

93–72

3

66.2

65

80–51

65

71–60

62.8

74–52

65.2

75–56

72.8

83–62

Note: Node times in the timetree represent the mean of time estimates from different studies and methods. Results from ref. (41) are based on the analysis of (a) three mitochondrial rRNA and (b) 16 nuclear and three mitochondrial rRNA genes. Results from ref. (43) are from analysis of (a) 1st + 2nd codon positions of three nuclear genes, (b) amino acid sequences of three nuclear genes, and amino acid sequences of eight nuclear genes.

Nesophontidae. Indeed, West Indian shrews presumably became extinct during post-Colombian time. However, while waiting for molecular data, it has been proposed by Roca et al. (41) that this taxon could be closely related to shrews. This result was suggested by a reanalysis of Asher et al.’s (42) morphological data set constrained by a molecular scaffold corresponding to the most likely relationships among extant taxa. Most comprehensive molecular dating estimates for the eulipotyphlan family tree come from Douady and Douzery (43) and from Roca et al. (41). Both studies are strongly linked, as they rely on Bayesian relaxed molecular clocks as implemented in the Thorne–Kishino method (44, 45) and have partly overlapping genetic data. Douady and Douzery (43) used a denser taxonomic sampling, with lesser genetic coverage, but could not include the key taxa Solenodon, whose sequences were unavailable at the time. In regard to calibration times, both studies again show some degree of overlap. However, Roca et al. (41) followed Springer et al. (46) in choosing the most probable ages of fossils as calibrations, whereas Douady and Douzery (43) employed upper and lower bounds of the stratigraphic range of the geological epochs to which the fossils pertaining to the divergence under focus were assigned. The second approach is a more conservative one as it accounts for uncertainty in timing the fossil remains. Thus Douady and Douzery (43) assumed divergence between 24 and 5 Ma for the split between Mus/ Rattus and 72 and 49 Ma for Feliformia/Caniformia, Hippomorpha/Ceratomorpha, Hippopotamidae/Cetacea and Paenungulata. In contrast, Roca et al. (41) used >12 Ma, 63–50 Ma, 58–54 Ma, >52 Ma, and 65–54 Ma for these nodes, respectively. In addition, Roca et al. (41) constrained a basal divergence among extant xenarthrans

before 60 Ma, a maximum of 65 Ma for cetartiodactyl diversification, and a 60–43 Ma range for the divergence between pteropodid bats and the false vampire bat. While some differences in estimates exist, mainly for Roca et al.’s (41) three-gene mtRNA data and Douady and Douzery’s (43) eight-gene (nuclear) data set, overall results are similar (Table 1). Both analyses suggest that the interfamily diversification of extant eulipotyphlans took place in the late Cretaceous. The basal (early branching) position of Solenodon and its divergence estimate argue for a vicariant origin of this West Indian taxon. Indeed, it is well established that proto-Antilles separated from the North American mainland between 80 and 70 Ma (47) and the mean estimate of 80 Ma fits in this window of time. However, the complex history of the area cannot exclude other alternatives (41). One such alternative would disconnect speciation in the late Cretaceous and colonization of the West Indies later in the Paleogene (48). The split between hedgehogs and shrews is quite likely contemporaneous to the Cretaceous/Paleogene boundary. Thus, it seems plausible that events that triggered the mass extinction acted as a diversification agent for Eulipotyphla. This may have included subsequent adaptive radiation into newly available niches during the very early Paleogene. However, historical events leading to the origin of the mole lineage are much more elusive. One could argue that they were linked to some climatic or tectonic event that occurred at the boundary between the mid- and late Campanian (84–71 Ma). The time of divergence of the Nesophontes lineage is unclear. However, the phylogenetic position suggested by Roca et al.’s (41) reanalysis of Asher et al.’s (42) data would suggest that divergence occurred sometime after the hedgehog–shrew divergence but before separation

498


of the Subfamilies Crocidurinae and Soricinae. Douady and Douzery (43) estimated the time of divergence of these subfamilies 38 Ma (95% credibility interval: 47–29). Thus, if phylogenetic assumptions are correct, the Nesophontes–shrew split could have occurred between 65 and 38 Ma. Additional molecular studies are still required for a better understanding of the evolution of Eulipotyphla, one of the most recently delineated orders of placental mammals.

Acknowledgments Support was provided by Federative Research Institute 41 and Université Claude Bernard Lyon 1.

References 1.

2. 3.

4. 5. 6. 7.

8.

9. 10. 11.

12. 13. 14.

15. 16. 17. 18.

D. E. Wilson, D. M. Reeder, Eds., Mammals Species of the World. A Taxonomic and Geographic Reference, 3rd ed. (Johns Hopkins University Press, Baltimore, 2005). J. A. Wagner, Die saugethiere in abbildungen nach der natur (Weiger, Leipzig, 1855). E. Haeckel, Systematische Einleitung in die allgemeine Entwicklungsgeschichte. Generelle morphologie der organismem (Georg Reimer, Berlin, 1866). W. Leche, K. svenska Vetensk Akad. Handl. 21, 1 (1885). G. G. Simpson, Bull. Am. Mus. Nat. Hist. 85, 1 (1945). P. M. Butler, Proc. Zool. Soc. Lond. 126, 453 (1956). P. M. Butler, in Studies in Vertebrate Evolution, K. A. Joysey, T. S. Kemp, Eds. (Oliverand Boyd, Edingburgh, 1972), pp. 253–265. R. D. E. MacPhee, M. J. Novacek, in Mammal Phylogeny: Placentals, F. S. Szalay, M. J. Novacek, M. C. McKenna, Eds. (Springer-Verlag, New York, 1993), pp. 13–31. R. J. Asher, Cladistics 15, 231 (1999). T. H. Huxley, Proc. Zool. Soc. Lond. 43, 649 (1880). M. J. Stanhope et al., in Primates and their Relatives in Phylogenetic Perspective, R. D. E. MacPhee, Ed. (Plemum Press, New York, 1993), pp. 251–292. M. M. Miyamoto, Mol. Phylogenet. Evol. 6, 373 (1996). O. Madsen et al., Nature 409, 610 (2001). W. W. de Jong, J. A. M. Leunissen, G. J. Wistow, in Mammal Phylogeny, F. S. Szalay, M. J. Novacek, M. C. McKenna, Eds. (Springer-Verlag, New York, 1993), pp. 5–12. M. J. Stanhope et al., Proc. Natl. Acad. Sci. U.S.A. 95, 9967 (1998). M. J. Stanhope et al., J. Mol. Evol. 43, 83 (1996). M. S. Springer et al., Nature 388, 61 (1997). O. Madsen , P. M. T. Deen, G. Pesole, C. Saccone, W. W. de Jong, Mol. Biol. Evol. 14, 363 (1997).

19. M. J. Stanhope et al., Mol. Phylogenet. Evol. 9, 501 (1998). 20. C. J. Douady, F. Catzeflis, J. Raman, M. S. Springer, M. J. Stanhope, Proc. Natl. Acad. Sci. U.S.A. 100, 8325 (2003). 21. E. L. Simons, P. A. Holroyd, T. M. Bown, Proc. Natl. Acad. Sci. U.S.A. 88, 9734 (1991). 22. E. Douzery, F. Catzeflis, J. Mol. Evol. 41, 622 (1995). 23. C. J. Douady, F. Catzeflis, D. J. Kao, M. S. Springer, M. J. Stanhope, Mol. Phylogenet. Evol. 22, 357 (2002). 24. J. F. Eisenberg, The Mammalian Radiation (University of Chicago Press, Chicago, 1981). 25. P. M. Butler, in The Phylogeny and the Classification of the Tetrapods, vol 2. Mammals M. J. Benton, Ed. (Clarendon Press, Oxford, 1988), pp. 117–141. 26. P. J. Waddell, N. Okada, M. Hasegawa, Syst. Biol. 48, 1 (1999). 27. A. Krettek, A. Gullberg, U. Arnason, J. Mol. Evol. 41, 952 (1995). 28. U. Arnason et al., Proc. Natl. Acad. Sci. U.S.A. 99, 8151 (2002). 29. S. K. Mouchaty, A. Gullberg, A. Janke, U. Arnason, Mol. Biol. Evol. 17, 60 (2000). 30. W. J. Murphy et al., Nature 409, 614 (2001). 31. W. J. Murphy et al., Science 294, 2348 (2001). 32. C. J. Douady et al., Mol. Phylogenet. Evol. 25, 200 (2002). 33. C. J. Douady, M. Scally, M. S. Springer, M. J. Stanhope, Mol. Phylogenet. Evol. 30, 778 (2004). 34. M. Nikaido et al., J. Mol. Evol. 53, 508 (2001). 35. M. Nikaido, Y. Cao, M. Harada, N. Okada, M. Hasegawa, Mol. Phylogenet. Evol. 28, 276 (2003). 36. Y. H. Lin et al., Mol. Biol. Evol. 19, 2060 (2002). 37. M. T. Cabria, J. Rubines, B. Gomez-Moliner, R. Zardoya, Gene 375, 1 (2006). 38. S. K. Mouchaty, A. Gullberg, A. Janke, U. Arnason, Zool. Scripta 29, 307 (2000). 39. J. Sullivan, D. L. Swofford, J. Mammal. Evol. 4, 77 (1997). 40. P. J. Waddell, Y. Cao, J. Hauf, M. Hasegawa, Syst. Biol. 48, 31 (1999). 41. A. L. Roca et al., Nature 429, 649 (2004). 42. R. J. Asher, M. C. McKenna, R. J. Emry, A. R. Tabrum, D. G. Kron, Bull. Am. Mus. Nat. Hist. 217, 1 (2002). 43. C. J. Douady, E. J. Douzery, Mol. Phylogenet. Evol. 28, 285 (2003). 44. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 45. H. Kishino, J. L. Thorne, W. J. Bruno, Mol. Biol. Evol. 18, 3521 (2001). 46. M. S. Springer, W. J. Murphy, E. Eizirik, S. J. O’Brien, Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003). 47. S. B. Hedges, Ann. Rev. Ecol. Syst. 27, 163 (1996). 48. S. B. Hedges, Ann. Mo. Bot. Gard. 93, 231 (2006).

Bats (Chiroptera) Emma C. Teeling UCD School of Biology and Environmental Science, Science Center West, University College Dublin, Belfield, Dublin 4, Ireland (emma. [email protected])

Abstract Bats are grouped into 17–18 families (>1000 species) within the mammalian Order Chiroptera. Recent phylogenetic analyses of molecular data have reclassified Chiroptera at the interfamilial level. Traditionally, the non-echolocating megabats (Pteropodidae) have been considered to be the earliest diverging lineage of living bats; however, they are now found to be the closest relatives of the echolocating rhinolophoid microbats. Four major groups of echolocating microbats are supported: rhinolophoids, emballonuroids, vespertilionoids, and noctilionoids. The timetree suggests that the earliest divergences among bats occurred ~64 million years ago (Ma) and that the four major microbat lineages were established by 50 Ma.

Bats are nocturnal mammals that have achieved the ability of true self-powered flight and are members of the monophyletic Order Chiroptera (meaning “hand-wing”; Fig. 1). They are the second most species-rich mammalian order (>1000 species) and account for ~20% of all extant mammalian diversity (1). They are found throughout the globe and are only absent from the extreme polar regions, but some bat lineages show high levels of endemism (1). Bats exploit many environmental niches and can feed on insects, fish, fruit, pollen, nectar, mammals, birds, and blood. They are important pollinators and play an important role in the tropical ecosystems (2). There are two major types of bats: megabats and microbats. As the names suggests, the largest bats are megabats (40–220 cm wingspan) and the smallest bats are microbats (22–135 cm wingspan). Another major difference between these groups is their mode of sensory perception (3, 4). Microbats (17 families) are capable of using sophisticated laryngeal echolocation, whereby they acoustically perceive their environment by interpreting returning echoes of emitted sound (5). In contrast, megabats (one family) rely on large eyes specialized for nocturnal vision (4). Icaronycteris index is one of the

oldest bat fossils (~55 Ma) and is considered a microbat; however, the majority of the bat fossil record is fragmentary and missing key species (6, 7). Here I review the relationships and divergence times of the extant families of bats. Traditionally bats have been divided into two superordinal groups: Megachiroptera and Microchiroptera (see 8, 9 for reviews). Megachiroptera was considered basal and contained the Old World megabat family Pteropodidae, whereas Microchiroptera contained the 17 microbat families (8, 9). Although this division was based mainly on morphological and paleontological data, it highlighted the difference in mode of sensory perception between megabats and microbats. Because all microbats are capable of sophisticated laryngeal echolocation whereas megabats are not (5), it was believed that laryngeal echolocation had a single origin in the lineage leading to microbats (10). The 17 families of microbats have been subsequently divided into two infraorders, Yinochiroptera (rhinolophids, hipposiderids, megadermatids, craseonycterids, rhinopomatids, emballonurids, and nycterids) and Yangochiroptera

Fig. 1 An Old World leaf-nosed bat (Hipposideros larvatus), Family Rhinolophidae, in flight from the Kanchanaburi region in Thailand. Credit: S. Puechmaille.

E. C. Teeling. Bats (Chiroptera). Pp. 499–503 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

500

THE TIMETREE OF LIFE Phyllostomidae 16

Furipteridae

11 15

Thyropteridae

5

Yangochiroptera

Noctilionidae 17

Mystacinidae

Vespertilionidae 12

Miniopteridae

9 3

Molossidae

7

Natalidae Nycteridae 6

Rhinopomatidae 10 8

Craseonycteridae 13

Megadermatidae

2

Rhinolophidae Pteropodidae Paleogene

Rhinolophoidea

Emballonuridae

1

Emballonuroidea

4

Vespertilionoidea

Myzopodidae

Yinpterochiroptera

14

Noctilionoidea

Mormoopidae

Neogene CENOZOIC

50

25

0 Million years ago

Fig. 2 A timetree of bats (Chiroptera). Divergence times are shown in Table 1.

(vespertilionids, molossids, natalids, phyllostomids, noctilionids, furipterids, thyropterids, mormoopids, mystacinids, and myzopodids), based on whether their premaxillaries were moveable/absent or fused relative to their maxillaries (8, 9, 11). This arrangement was largely supported by recent morphological data sets (6) and supertree consensus studies (12). However, the number of superfamilial groupings varied in content and number among studies (6, 8, 9, 12). From the onset it became apparent that molecular data did not support the monophyly of Microchiroptera and, therefore, did not support a single origin of laryngeal echolocation. Rather, molecular data supported a basal division between Yinpterochiroptera (rhinolophoid microbats and pteropodids) and Yangochiroptera (all other bats; 7, 13–16). This topology suggested that laryngeal echolocation either originated in the ancestor of all bats and was subsequently lost in lineages leading to the megabats or originated more than once in the microbat lineages (10). Initially immunological distance data (17); single gene data sets (18, 19); whole genomic DNA–DNA

hybridization studies (20); repetitive genomic elements (21); and taxonomically limited consensus studies (22) all supported microbat paraphyly to different degrees (4). However, strong support and congruence for the association of the rhinolophoid microbats with the pteropodids was only derived from large concatenated nuclear data sets with representatives from nearly all putative bat families (7, 13, 14) and rare cytogenetic signature events (23). Molecular data in the form of large nuclear and mitochondrial concatenations provide strong support for the association of four major groups of echolocating microbat lineages (Fig. 2): (a) Rhinolophoidea, which includes the rhinolophids (which includes the hipposiderinids), rhinopomatids, craseonycterids, and megadermatids (7, 13–15); (b) Emballonuroidea, which includes the nycterids and emballonurids (7, 13–15); (c) Vespertilionoidea, which includes the vespertilionids, molossids, natalids, and miniopterids (7, 13–15); and (d) Noctilionoidea, which includes the noctilionids, phyllostomids, furipterids, thyropterids, mormoopids, mystacinids, and

Eukaryota; Metazoa; Vertebrata; Mammalia; Chiroptera

501

Table 1. Divergence times (Ma) among bats and their credibility/confidence intervals (CI) among bats (Chiroptera). Timetree Node

Estimates Ref. (7)

Time

Time

Ref. (13) CI

Ref. (14)

Ref. (30)

Time

CI

Time

CI

Time

CI

1

62.0

64

71–58

62

70–56

64

70–58

57.9

–

2

57.5

58

63–53

58

65–52

58

63–53

55.8

–

3

55.0

55

61–50

53

61–47

56

62–50

56

–

4

54.5

54

60–50

–

–

55

61–50

54.6

–

5

51.9

52

57–46

–

–

52

57–46

51.6

–

6

51.5

52

58–47

50

57–43

52

58–47

52.1

–

7

50.6

50

56–45

50

58–44

51

56–45

51.4

–

8

50.1

52

55–48

54

60–48

52

55–48

42.5

–

9

48.1

47

53–42

48

56–42

48

54–43

49.3

–

10

47.0

49

53–45

50

54–44

50

54–46

39

–

11

45.3

46

51–41

44

51–37

45

51–41

46.1

–

12

44.0

–

–

45

53–39

43

49–38

–

–

13

43.0

43

47–39

–

–

43

47–39

–

–

14

42.4

42

47–37

–

–

41

47–36

44.3

–

15

40.7

40

46–35

–

–

40

45–36

42.1

–

16

37.5

36

42–32

39

46–33

36

41–31

38.8

–

17

36.6

36

41–31

39

46–33

35

41–30

36.2

–

Note: Node times in the timetree represent the mean of time estimates from different studies. Number of genes analyzed are 17 (7), 16 (14), and 4 (13).

myzopopdids (7, 14). The position of Myzopodidae was not resolved by nuclear intronic or mitochondrial data (14, 15). Also, the relationships within Noctilionoidea differed when analyzed with nuclear introns (13) vs. either nuclear exons (7, 14) or mitochondrial data (15). Within these superfamilial groups, molecular data also has revealed novel interfamilial relationships with unique biogeographic and morphological implications. The grouping of Koopman’s yinochiropteran Families (8) Emballonuridae and Nycteridae with the other yangochiropteran families indicates that the unique mammalian condition of moveable premaxillaries must have arisen at least twice in bats (7, 11). Rhinolophoid microbats are united by the presence of pubic nipples, which are not found in any other bat lineage. All rhinolophoids uniquely posses an ossified first costal cartilage fused to the manubrium and first rib (9, 13). While it is believed that this structure may reduce the energetic costs of stationary echolocation emission and, therefore, may be evidence for a dual origin of laryngeal echolocation in bats

(13, 24), this has been questioned (4). Within the Superfamily Noctilionoidae, the monotypic New Zealand Family Mystacinidae is the closest relative of the Neotropical noctilionoid families and the monotypic Malagasy myzopodids are the earliest diverging lineage (7, 14). This topology suggests a previous Gondwanan distribution and perhaps vicariant origin; however, molecular dates indicate that the Noctilionoidea started to diversify long after the separation of the Gondwanan supercontinent (7, 13, 14). Molecular data support the elevation of miniopterids to familial status and suggests that they are closest relative of Vespertilionidae rather than a member of that family (13–15, 25). This change in rank is also supported by a unique suite of morphological characters (see 14, 25 for reviews) and deep divergence dates (45 Ma; 13, 14, 25). One of the first estimates of bat divergence dates was based on a concatenated data set of five nuclear and three mitochondrial genes and used a quartet dating approach (26). Due to limited taxonomic sampling, only the

502


divergence time of the earliest split among living bats was estimated at 54–52 Ma (26). Three recent, large molecular studies have estimated the divergence dates for crowngroup bat families (7, 13, 14). These dates are all based on a relaxed Bayesian clock method (27, 28) with similar constraints and priors available from the fossil record (7, 13, 14). Teeling et al. (7) used a concatenation of 13.7 kb from fragments of 17 nuclear genes (exons and untranslated regions) as representative of all bat putative bat families. Miniopteridae, which was considered a subfamily at the time, was not included. The data set included 30 bat genera and four laurasiatherian outgroups (29). Six fossil constraints were employed: (a) a maximum of 34 Ma for the base of the Family Phyllostomidae; (b) a minimum of 30 Ma for the Mormoopidae/Phyllostomidae split; (c) a minimum of 37 Ma for the split between Vespertilionidae/Molossidae; (d) a minimum of 37 Ma for the base of Emballonuridae; (e) a minimum of 37 Ma for the base of Rhinolophidae; and (f) a maximum of 55 Ma for the base of Rhinolophoidea. Bayesian dating analyses (27, 28) were used to estimate the branch lengths and divergence times for the entire concatenation and also for each gene considered as a unique partition within the data set. The earliest split among living bats was estimated to have occurred ~64 Ma at or following the Cretaceous–Paleogene boundary (Fig. 2). The four major echolocating microbat lineages all originated within a narrow time frame (~52–50 Ma) within the early Eocene. All extant bat families were estimated to have diversified by the end of the Eocene (~34 Ma). Crown-group pteropodids did not originate until the early Oligocene (28 Ma); however, they had diverged from the rhinolophoids by the late Paleocene (~56 Ma) (Fig. 2). The data set of Eick et al. (13) was based on a concatenation of 4 kb of DNA sequence from four nuclear introns for 17 of the 18 bat families (including Miniopteridae, but missing Craseonycteridae) for 55 bats and three laurasiatherian outgroups. Eick et al. (13) recovered a highly congruent topology with Teeling et al. (7); however, they reported a close relationship between Mystacinidae and Thyropteridae, a basal position for Myzopodidae within the Vespertilionoidea and a grouping of the Vespertilionoidea with the Noctilionoidea, although these alternate groupings received little bootstrap support (13). Like Teeling et al. (7) they incorporated Bayesian dating analyses with constraints from the fossil record, but only analyzed the concatenation as a single partition. Similar to Teeling et al. (7) they jackknifed fossil constraints, and found time estimates to be robust to use of different fossil constraints. The dating results from Eick et al.

(13) and Teeling et al. (7) were nearly identical (Table 1). Miniopteridae is estimated to have diverged from the Vespertilionidae at 45 Ma (Fig. 2, Table 1). The most recent estimate of interfamilial divergence times is an augmentation of the Teeling et al. (7) data set to include an additional basal representative of Vespertilionidae and two miniopterid species (14). The data set included 11 kb of DNA fragments from 16 genes (VWF is not included in this data set). The entire concatenation was analyzed as a single partition using Bayesian methods, with constraints from the fossil record as incorporated by Teeling et al. (7). The results were similar to previous divergence dates (Table 1). Miniopteridae is estimated to have diverged ~45 Ma (Fig 2; Table 1). These molecular results (Table 1) are corroborated by an independent dating analysis with larger taxonomic scope (30). The authors estimated the relative molecular dates for each node by fitting sequence data from six genes to a supertree consensus topology (12) and that of Teeling et al. (7). They incorporated local clocks, which were calibrated by nodal ages extracted from the fossil record and/or previously published absolute molecular dates (30). The timetree suggests that the four major lineages of echolocating microbats originated within a narrow time frame (~52–50 Ma). This was coincident with an ~7°C rise in the global temperature (Paleocene/Eocene thermal maximum), a significant increase in plant diversity and the peak of tertiary insect diversity (7). These dates imply that the major echolocating microbat lineages may have radiated in response to an increase in prey diversity and roost sites (7). Jones et al. (30) also reported an increase in bat diversification, particularly the phyllostomids, at 40–25 Ma, which correlates with an increase in flowering plant diversity. This suggests that phyllostomid bats may have radiated due to an increase in fruit and pollen food sources. Although the earliest divergence among living bats is highly supported (64 Ma), the geographic location is still contentious with biogeographic analyses of similar molecular topologies suggesting either Africa (13) or North America (7). This is partly attributed to a poor fossil record. Indeed, Teeling et al. (7) compared the oldest fossil dates with the molecular estimates for each branch on the tree and suggested that the fossil record underestimates first occurrences by on average 73%. They also suggested that at least 98% of fossil history is missing from the megabat lineages. Perhaps this explains the difficulty in assessing whether laryngeal echolocation was lost in the megabat lineages or never acquired in

Eukaryota; Metazoa; Vertebrata; Mammalia; Chiroptera

the first place. Stem megabat fossils, which may or may not show a gradual change in skull structure resulting from a switch in sensory perception (auditory–visual), are not found. Although the timetree has revolutionized our understanding of bat evolutionary history we are still not able to determine whether laryngeal echolocation evolved once or more in bats. Future comparative genomic studies are needed to establish the molecular mechanisms that underlie laryngeal echolocation. This would enable researchers to assess if all laryngeal echolocators are governed by the same molecular mechanisms (which would indicate a single origin of echolocation in bats) or not. Likewise, it is pertinent that we keep searching for key transition fossils that will shed light on the evolution of echolocation. The most basal bat fossil has only recently been found (Onychonycteris finneyi) and shows evidence of flight but not laryngeal echolocation capabilities (31). This fossil has enabled biologists to determine that flight most likely originated in bats before the ability to echolocate and answer the long standing question of which came first, flight or echolocation. More fossils of this nature are needed, indeed the timetree and inferred biogeographic hypotheses could suggest new areas and time transects to target for future fossil discoveries.

Acknowledgment This work was supported by a PIYRA Science Foundation Ireland grant.

References 1. N. B. Simmons, in Mammalian Species of the World, A Taxonomic and Geographic Reference, D. E. Wilson, D. M. Reeder, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 312–529. 2. T. H. Kunz, M.B. Fenton, Eds. Bat Ecology (University of Chicago Press, Chicago, 2003). 3. R. M. Nowak, Walker’s Bats of the World (Johns Hopkins University Press, Baltimore, 1994). 4. G. Jones, E. C. Teeling, TREE 21,149 (2006). 5. G. Jones, Curr. Biol. 15, R484 (2005). 6. G. Gunnell, N. Simmons, J. Mamm. Evol. 12, 209 (2005). 7. E. C. Teeling et al., Science 307, 580 (2005).

503

8. K. F. Koopman, Chiroptera: Systematics, Handbook of Zoology, Vol. 8, Mammalia (Walter de Gruyter, New York, 1994). 9. N. B. Simmons, J.H. Geisler, Bull. Amer. Mus. Nat. Hist. 235, (1998). 10. E. C. Teeling et al., Nature 403, 188 (2000). 11. J. M. Hutcheon, J. A. Kirsch, Acta Chiropterol. 8, 1 (2006). 12. K. E. Jones, A. Purvis, A. MacLarnon, O. R. P. BinindaEmonds, N. Simmons, Biol. Rev. 77, 223 (2002). 13. G. N. Eick , D. S. Jacobs, C. A. Matthee, Mol. Biol. Evol. 22, 1869 (2005). 14. C. M. Miller-Butterworth et al., Mol. Biol. Evol. 24, 1553 (2007). 15. R. A. Van Den Bussche, S. R. Hoofer, J. Mammal. 85, 321 (2004). 16. J. M. Hutcheon, J. A. W. Kirsch, J. Mamm. Evol. 11, 17 (2004). 17. E. D. Pierson, Molecular Systematics of the Microchiroptera: Higher Taxon Relationships and Biogeography, Ph.D. dissertation (University of California, Berkeley, California, 1986). 18. C. A. Porter, M. Goodman, M. J. Stanhope, Mol. Phylogenet. Evol. 5, 89(1996). 19. M. J Stanhope, J. Czelusniak, J.-S. Si, J. Nickerson, M. Goodman, Mol. Phylogenet. Evol. 1, 148 (1992). 20. J. M. Hutcheon, J. A. W. Kirsh, J. D. Pettigrew, Phil. Trans. Roy. Soc. Lond. Ser. B 353, 607 (1998). 21. R. J. Baker, J. L. Longmire, M. Maltbie, M. J. Hamilton, R. A. Van Den Bussche, Syst. Biol. 46, 579 (1997). 22. F.-G. R. Liu, M. M. Miyamoto, Syst. Biol. 48, 54 (1999). 23. L. Ao et al., Chrom. Res. 15, 257 (2007). 24. J. R Speakman, W. C. Lancaster, S. Ward, G. Jones, K. C. Cole, in Echolocation in Bats and Dolphins, J. A. Thomas, C. F. Moss, M. Vater, Eds. (Chicago University Press, Chicago, 2004), pp. 361–365. 25. S. R. Hoofer, R. A. Van Den Bussche, Acta Chiropterol. 5, 1 (2003). 26. M. S. Springer, E. C. Teeling, O. Madsen, M. J. Stanhope, W. W. de Jong, Proc. Natl. Acad. Sci. U.S.A. 98, 6241 (2001). 27. J. L. Thorne, H. Kishino, I. S. Painter, Mol. Biol. Evol. 15, 1647 (1998). 28. H. Kishino, J. L. Thorne, W. J. Bruno, Mol. Biol. Evol. 18, 352 (2001). 29. W. J. Murphy et al., Science 294, 2348 (2001). 30. K. Jones, O. R. R Bininda-Emonds, J. L. Gittleman, Evolution 59, 2243 (2005). 31. N. Simmons, K. Seymour, J. Habersetzer, G. F. Gunnell, Nature 451, 818 (2008).

Carnivores (Carnivora) Eduardo Eizirika,b,* and William J. Murphyc a

Faculdade de Biociências, PUCRS, Av. Ipiranga, 6681, Porto Alegre, RS 90619-900, Brazil; bInstituto Pró-Carnívoros, Av. Horácio Neto, 1030, Atibaia SP 12945-010, Brazil; cDepartment of Veterinary Integrative Biosciences, Texas A&M University, College Station, TX 77843-4458, USA *To whom correspondence should be addressed (eduardo.eizirik@ pucrs.br)

Abstract Living members of the mammalian Order Carnivora have been traditionally placed in 11 families making up two suborders: Feliformia and Caniformia. Recent analyses based on morphological and molecular data have identified additional groups of species that warrant family-level recognition, leading to major changes in the current understanding of carnivoran evolutionary history and taxonomy. There are presently 16 recognized families, whose relationships are now well understood. The carnivoran timetree indicates that Feliformia and Caniformia diverged from each other ~55 million years ago (Ma). Within each suborder, suprafamilial nodes span a broad range of divergence times, from 53 to 22 Ma.

The Order Carnivora contains a diverse set of mammals, including well-known species such cats, dogs, lions, bears, and seals, as well as enigmatic animals such as the stink badgers (Mydaus spp.), the African Palm Civet (Nandinia binotata), and the Fossa (Cryptoprocta ferox). There are currently 286 recognized species of living carnivorans (e.g., Fig. 1), which vary widely in morphology, ecology, physiology, and behavior (1, 2). Size range among carnivoran species is broader than in any other mammalian order, with body weight varying 1000 times among its representatives. The Order Carnivora has a relatively rich paleontological record. The earliest fossils, dating from the Paleocene (66–56 Ma), are usually placed in the extinct families Viverravidae and Miacidae (3), both of which likely comprise early branching lineages relative to the living taxa (4). The carnivoran fossil record from the Paleocene to the Oligocene (34–23 Ma) is confined to Eurasia and North America (4), supporting the view that

this mammalian order has had its origin in Laurasia. Only in the Miocene (23–5 Ma) do carnivoran families appear in the fossil records of Africa and South America, indicating an initial period of intercontinental dispersal in this group. Here we review the current understanding of the phylogenetic relationships and divergence times among carnivoran families, focusing exclusively on living lineages, and emphasizing results from recent studies. Living members of the Order Carnivora are grouped into two monophyletic suborders: Feliformia and Caniformia. The former traditionally included Families Viverridae (e.g., civets and genets), Herpestidae (mongooses), Hyaenidae (hyenas), and Felidae (cats), while the latter comprised the Families Canidae (dogs, wolves, and foxes), Mustelidae (e.g., otters, weasels, and badgers), Ursidae (bears), Procyonidae (e.g., raccoons and coatis), Otariidae (sea lions and fur seals), Phocidae (true seals), and Odobenidae (Walrus). Otariidae, Phocidae, and Odobenidae are highly adapted for marine life and have been historically grouped in a taxon called Pinnipedia. The monophyly and phylogenetic placement of pinnipeds have been contentious for many years (2, 5, 6), but this issue seems to be mostly settled now (7–10). It is now clear that Pinnipedia is monophyletic and that

Fig. 1 An African Wild Dog (Lycaon pictus), representing the Family Canidae (Caniformia), from Kruger National Park, South Africa. Credit: E. Eizirik.

E. Eizirik and W. J. Murphy. Carnivores (Carnivora). Pp. 504–507 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Eukaryota; Metazoa; Vertebrata; Mammalia; Carnivora

505

Mephitidae Mustelidae 12

Procyonidae Ailuridae

7

Otariidae 14 5

Odobenidae

11

Caniformia

9

Phocidae 3

Ursidae Canidae Felidae 8

Prionodontidae Hyaenidae

4 10 2

6

Herpestidae 13

Eupleridae

Feliformia

1

Viverridae Nandiniidae Paleogene

Neogene CENOZOIC

50

25

0 Million years ago

Fig. 2 A timetree of carnivores (Carnivora). Divergence times are shown in Table 1.

it is contained in Arctoidea (pinnipeds + Ursidae + Mustelidae + Procyonidae + Red Panda + skunks (see later)), which is nested within the carnivoran Suborder Caniformia. In addition to the argument over the relationships of Pinnipedia, several other aspects of the carnivoran tree have been contentious over the last few decades, leading to the production of a large body of literature on the phylogeny of this mammalian order (e.g., 4–30). Most of the effort has been traditionally focused on the Caniformia, and particularly on the positions of the Giant Panda (Ailuropoda melanoleuca), Red Panda (Ailurus fulgens), and skunks (originally placed in the Mephitinae within Mustelidae, but now recognized as comprising a separate family, Mephitidae). On the feliform side, the monophyly of the Family Viverridae has been challenged multiple times, starting with the proposition that the African Palm Civet (N. binotata) was actually not a member of this family, but rather the only living representative of the most basal extant lineage of the Feliformia (e.g., 4, 24, 25). Another feliform whose phylogenetic affinities have historically been enigmatic is the Fossa (C. ferox), a Malagasy

carnivoran with unique morphological and behavioral characteristics often placed in Viverridae, Herpestidae, or in its own monotypic family (1, 2, 26). The last few years have seen a surge in studies on these and other topics of the carnivoran phylogeny (e.g., 7–10, 18–23) most of which used concatenations of multiple nuclear and/or mitochondrial genes. This has led to a consistent resolution of most suprafamilial nodes (Fig. 2), settling many of the disputes briefly outlined earlier. The Giant Panda was established as the most basal extant ursid, and the Red Panda is now placed in its own monotypic family (Ailuridae), nested in Arctoidea. Skunks and stink badgers (Mydaus spp.) are closely related, and together constitute the Family Mephitidae, which is not immediately connected to the Mustelidae (14, 15, 28). Another recent finding is that Mustelidae and Procyonidae are each other’s closest relatives, an observation which is supported by several studies (e.g., 8, 15, 24). There is a core group in Arctoidea containing Mephitidae, Ailuridae, and Mustelidae + Procyonidae, whose internal structure has still not been confidently resolved (Fig. 2). Pinnipedia is now seen as the closest relative of this core clade, with Ursidae being

506


Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among carnivores (Carnivora). Timetree Node

1

Estimates Ref. (19)

Time

52.9

Ref. (22)

Ref. (23)

Ref. (30)

Ref. (31)

Ref. (32)

Ref. (33)

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

–

–

–

–

–

–

–

–

55.0

60–51

57.5

62–52

46.2

57–35

2

49.0

–

–

54.6

–

–

–

43.3

54–33

–

–

–

–

–

–

3

44.0

–

–

–

–

44

–

–

–

–

–

–

–

–

–

4

41.8

–

–

47.0

–

–

–

36.5

47–29

–

–

–

–

–

–

5

40.0

–

–

–

–

40

–

–

–

–

–

–

–

–

–

6

39.7

–

–

44.2

–

–

–

35.2

45–28

–

–

–

–

–

–

7

38.0

–

–

–

–

38

–

–

–

–

–

–

–

–

–

8

37.8

33.3

35–32

42.3

–

–

–

–

–

–

–

–

–

–

–

9

35.0

–

–

–

–

35

–

–

–

–

–

–

–

–

–

10

33.8

–

–

38.3

–

–

–

29.2

38–23

–

–

–

–

–

–

11

30.0

–

–

–

–

30

–

–

–

–

–

–

–

–

–

12

28.0

–

–

–

–

28

–

–

–

–

–

–

–

–

–

13

28.0

–

–

31.6

–

–

–

24.4

32–18

–

–

–

–

–

–

14

22.0

–

–

–

–

22

–

–

–

–

–

–

–

–

–

Note: Node times in the timetree represent the mean of time estimates from different studies.

the most basal lineage in Arctoidea. Canidae is indeed the most basal family in the Caniformia, supporting the traditional view that it is the only extant lineage of the Superfamily Cynoidea. Among feliforms, recent studies have led to major changes in the prevailing views on phylogenetic structure and evolutionary history. Two separate studies published in 2003 have shown further evidence of viverrid paraphyly, and identified novel lineages that are now recognized as valid families (19, 20). Asian linsangs (Prionodon spp.), traditionally part of the Viverridae, have been shown to be the closest relative of the Felidae (19), and are now placed in their own family, Prionodontidae (21). Another remarkable finding was that all Malagasy carnivores (including the Fossa), traditionally placed in the Viverridae or Herpestidae, comprise a separate, endemic monophyletic lineage (20), which is now recognized as Family Eupleridae. In addition, all recent studies that included Nandinia confirmed that this taxon is indeed the most basal feliform, and now constitutes its own family, Nandiniidae (1). As a whole, these recent studies have challenged not only the monophyly of traditionally recognized Viverridae, but also the monophyly of Herpestidae, restructuring the feliform phylogeny to a large degree. Most suprafamilial nodes in this suborder have now been consistently resolved by independent

studies (Fig. 2). Felidae and Prionodontidae are each other’s closest relatives, as are Eupleridae and Herpestidae. Hyaenidae is the closest relative of the Eupleridae + Herpestidae clade. The relative position of Viverridae (now restricted to a monophyletic core group) has not been confidently established with high support, though most studies indicate that it is more closely related to the Hyaenidae + Eupleridae + Herpestidae clade (Fig. 2). Although many studies have addressed carnivoran relationships, few have assessed the age of the inferred clades using molecular data. The results reviewed here are drawn mostly from four recent studies, which have separately addressed each of the two carnivoran suborders (19, 22, 23, 30). The basal divergence between Feliformia and Caniformia seems to have occurred between the Paleocene and the middle Eocene (49–40 Ma), with the dates used here being derived from large studies involving all placental mammal clades (31, 32) or multiple vertebrate groups (33). Within Feliformia, the divergence between Nandiniidae and the other lineages was estimated by one study to be ~43 Ma (30), and by another to be ~55 Ma (22); the latter may be an overestimate given the branch length observed in multiple studies between the feliform–caniform split and this basal feliform node. The dates obtained in this study (22) are consistently older than equivalent divergence times

Eukaryota; Metazoa; Vertebrata; Mammalia; Carnivora

obtained by other papers (19, 30), in some cases lying outside of the estimated confidence intervals (Table 1). Further investigations are required to better understand this discrepancy and generate a more reliable and consensual view on feliform divergence times. Within Caniformia, the only published study describing molecular estimates of divergence times places the basal split between Canidae and Arctoidea at 44 Ma, consistent with the fossil record for this group (23). Overall, suprafamilial divergences in the Carnivora seem to occur almost exclusively in the Paleogene, mostly concentrating in the Eocene and early Oligocene (53–34 Ma). Additional divergence dating studies are needed for this group, especially using the same data set for Feliformia and Caniformia, and employing multiple fossil calibrations. The reliability of fossil calibrations may be an issue (e.g., addressed in 22), since the exact phylogenetic placement of some extinct carnivorans may be uncertain or incorrect, potentially leading to biased dating results. It is therefore important to evaluate multiple calibrations simultaneously, and to assess their consistency. The next few years will likely see the consolidation of the carnivoran phylogeny at family, genus, and possibly species level, with accompanying progress on the reliability and precision in divergence time estimates for all included nodes. This will allow a much better understanding of the evolutionary history of extant lineages, and an improved framework upon which to investigate phylogenetic, biogeographic, morphological, and ecological aspects of extinct carnivoran groups.

Acknowledgments E.E. is supported by CNPq, CAPES, FAPERGS, and FNMA/MMA, Brazil, and W.J.M. is supported by the U.S. National Science Foundation.

References 1.

W.C. Wozencraft, in Mammal Species of the World, A Taxonomic and Geographic Reference, 3rd ed., D. E. Wilson, D. M. Reeder, Eds. (Johns Hopkins University Press, Baltimore, 2005), pp. 532–628. 2. R. M. Nowak, Walker’s Mammals of the World, 6th ed. (Johns Hopkins University Press, Baltimore, 1999). 3. M. C. McKenna, S. K. Bell, Classification of Mammals above the Species Level (Columbia University Press, New York, 1997). 4. J. J. Flynn, G. D. Wesley-Hunt, in The Rise of Placental Mammals: Origins and Relationships of the Major Extant Clades, K. D. Rose, D. Archibald, Eds. (Johns Hopkins University Press, 2005), pp. 175–198.

507

5. C. Ledje, U. Arnason, J. Mol. Evol. 42, 135 (1996a). 6. C. Ledje, U. Arnason, J. Mol. Evol. 43, 641 (1996b). 7. I. Delisle, C. Strobeck, Mol. Phylogenet. Evol. 37, 192 (2005). 8. T. L. Fulton, C. Strobeck, Mol. Phylogenet. Evol. 41, 165 (2006). 9. J. J. Sato, M. Wolsan, H. Suzuki, T. Hosoda, Y. Yamaguchi, K. Hiyama, M. Kobayashi, S. Minami, Zool. Sci. 23, 125 (2006). 10. J. J. Flynn, J. A. Finarelli, S. Zehr, J. Hsu, M. A. Nedbal, Syst. Biol. 54, 317 (2005). 11. J. J. Flynn, H. Galiano. Am. Mus. Novit. 2725, 1 (1982). 12. S. J. O’Brien, W. G. Nash, D. E. Wildt, M. E. Bush, R. E. Benveniste, Nature 317, 140 (1985). 13. A. R. Wyss, J. J. Flynn, in Mammal Phylogeny 2, Placentals, F. S. Szalay, M. J. Novacek, M. C. McKenna, Eds. (Springer-Verlag, New York, 1993), pp. 32–52. 14. J. W. Dragoo, R. L. Honeycutt, J. Mammal. 78, 426 (1997). 15. J. J. Flynn, M. A. Nedbal, J. W. Dragoo, R. L. Honeycutt, Mol. Phylogenet. Evol. 17, 190 (2000). 16. O. R. P. Bininda-Emonds, J. L. Gittleman, A. Purvis, Biol. Rev. 74, 143 (1999). 17. J. J. Sato, T. Hosoda, M. Wolsan, H. Suzuki, Zool. Sci. 21, 111 (2004). 18. L. Yu, Y.P. Zhang, Genetica 127, 65 (2006). 19. P. Gaubert, G. Veron, Proc. Biol. Sci. 270, 2523 (2003). 20. A. D. Yoder, et al. Nature 421, 734 (2003). 21. P. Gaubert, W. C. Wozencraft, P. Cordeiro-Estrela, G. Veron, Syst. Biol. 54, 865 (2005). 22. P. Gaubert, P. Cordeiro-Estrela, Mol. Phylogenet. Evol. 41, 266 (2006). 23. U. Arnason, A. Gullberg, A. Janke, M. Kullberg, Mol. Phylogenet. Evol. 45, 863 (2007). 24. J. J. Flynn, M. A. Nedbal, Mol. Phylogenet. Evol. 9, 414 (1998). 25. R. M. Hunt, Jr., Am. Mus. Novit. 2886, 1 (1987). 26. G. Véron, F. M. Catzefl is, J. Mammal. Evol. 1, 169 (1993). 27. J. P. Slattery, S. J. O’Brien, J. Hered. 86, 413 (1995). 28. R. K. Wayne, R. E. Benveniste, D. N. Janczewski, S. J. O’Brien, in Carnivore Behavior, Ecology and Evolution, J. L. Gittleman, Ed. (Cornell University Press, Ithaca, 1989), pp. 465–494. 29. W. C. Wozencraft, in Carnivore Behavior, Ecology and Evolution, J. L. Gittleman, Ed. (Cornell University Press, Ithaca, 1989), pp. 495–535. 30. K. P. Koepfli, S. M. Jenks, E. Eizirik, T. Zahirpour, B. Van Valkenburgh, R. K. Wayne, Mol. Phylogenet. Evol. 38, 603 (2006). 31. M. S. Springer, W. J. Murphy, E. Eizirik, S. J. O’Brien, Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003). 32. W. J. Murphy, T. H. Pringle, T. A. Crider, M. S. Springer, W. Miller, Genome Res. 17, 413 (2007). 33. S. Kumar, S. B. Hedges, Nature 392, 917 (1998).

Rhinoceroses, tapirs, and horses (Perissodactyla) Oliver A. Ryder Conservation and Research for Endangered Species, Zoological Society of San Diego, 15600 San Pasqual Valley Road Escondido, CA 92027-7000 USA ([email protected])

Abstract Rhinoceroses, tapirs, and horses, comprising 16 species, constitute the three surviving families of the mammalian Order Perissodactyla. Based on recent DNA sequence data, the perissodactyl timetree supports previous views of diversification into two suborders: Hippomorpha (containing living horses, asses, and zebras) and Ceratomorpha (containing living rhinos and tapirs). Although once considered perissodactyls, recent DNA sequencing studies have provided new evidence that hyraxes are neither perissodactyls nor closely related to them. The earliest divergence among these three families occurred ~56 million years ago (Ma), and the extant species of Equidae diversified most recently.

Perissodactyls comprise the odd-toed ungulates, an order of mammals that was once more species-rich and that occupied a wide variety of terrestrial habitats. The surviving species of this order include the horses, asses, zebras of the Family Equidae (Fig. 1), four extant species of tapirs that are included in a monogeneric family, Tapiridae, and the surviving species of rhinoceros that constitute the four genera of the Rhinocerotidae (1). The Equidae contains seven extant species (2). Four species of tapir survive, and five extant species of rhinoceros survive in Africa and Asia (1). Early perissodactyls are thought to have diverged from condylarths. Combined mitochondrial DNA and nuclear DNA sequence data sets, as well as analysis of rare insertions and deletions, support carnivoriform and phioldotiform affinities for extant perissodactyls, these ordinal groups being part of the Laurasiatheria (3, 4). Hyracotherium, or a Hyracotherium-like hippomorph, is generally regarded as the ancestral hippomorph from which all living equids descend, while Hyrachyus is hypothesized to be the ancestral ceratomorph, the

ancestor of tapirs and rhinos. Although a remarkable radiation of hippomorphs took place and, eventually, all continents (except Australia and Antarctica) were occupied by hippomorph perissodactyls, the only extant genus of hippomorphs is Equus, regarded as monophyletic. Within the Ceratomorpha, cladogenesis of tapiriform and rhinocerosiform perissodactyls resulted in a remarkable radiation. Subsequently, extinctions have reduced the tapirs to a single genus and the rhinoceroses to four genera. Of the four species of tapirs, three occur in Central and South America and one inhabits Asia. The three extant families of Perissodactyls, Equidae, Tapiridae, and Rhinocerotidae, are regarded as monophyletic. Multiple studies of mammalian phylogeny and systematics have produced data estimating the divergence of Perissodactyla from other orders (3–9) and the divergence of the Ceratomorpha and Hippomorpha (3, 4). Nearly all of these estimates incorporate a fossil calibration for the horse–rhino divergence of 58–54 Ma and, accordingly, the consistency among the Perissodactyla divergence time estimates is influenced by this common

Fig. 1 Grévy’s Zebra (Equus grevyi) survive in Northern Kenya and Somalia. Credit: Zoological Society of San Diego.

O. A. Ryder. Rhinoceroses, tapirs, and horses (Perissodactyla). Pp. 508–510 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

Tapiridae

1

Equidae Paleogene

Neogene

509

Hippomorpha

Rhinocerotidae 2

Ceratomorpha

Eukaryota; Metazoa; Vertebrata; Mammalia; Perissodactyla

CENOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of rhinoceroses, tapirs, and horses (Perissodactyla). Divergence times are from Table 1.

calibration. Poux et al. (9) obtained similar results if the horse–rhino calibration point was omitted, incorporating five additional well-established fossil calibration points. By sequentially removing each calibration to exclude the possibility that individual calibration constraints produce bias in the dating analyses, the molecular clock dates remained highly congruent, lending credence to the estimates noted in Table 1. Time estimates for clade divergences based on molecular data have been developed for events within each perissodactyl family (Fig. 2). Tougard (10) evaluated the divergence of the four rhinoceros genera utilizing 12S and control region sequences. Additional control region sequence data have been provided by Fernando et al. (11). Mitochondrial 12S and control region sequences were developed for extant equids by Oakenfull et al. (12), and additional control region sequences have been generated by Weinstock et al. (13). For tapirs, mitochondrial cytochrome oxidase II and 12S sequences have been generated by Norman and Ashley (14). The evolution of monodactyly in horses became, for a time, a classic tale of orthogenesis, told as if morphological evolution was a straightforward pattern of cladogenesis that occurred with replacement of one species by another via a rather linear process (16). More recent considerations have emphasized patterns of species proliferation, migrations, extinctions, and survival of relatively few lineages to lead to new taxa that undergo a similar process (16). In this way, the evolution of Equus remains an instructive example of the scientific interpretation of fossil evidence and the changing patterns of scientific endeavor and interpretation itself (17). Another feature of note in equids is their rapid rate of chromosomal evolution, varying from a diploid number of 66 chromosomes in Przewalski’s Horses to 32 chromosomes in Mountain Zebras (18). Dramatic chromosomal

differences occur between the Old World and New World tapirs, while rhinoceros karyotypes appear to have changed little over the period of divergence of the four extant genera (15). Once dominant ungulates, perissodactyl species diversity declined as artiodactyls radiated (19). However, this trend has accelerated in the last centuries, largely as a result of human activities. Hunting and habitat loss have affected equids, tapirs and rhinos alike. Domestication of horses and asses has seemingly assured their continued survival, albeit under the selective influence of humans. The only extant perissodactyl not currently under some level of conservation concern is the Plains Zebra, which in some African regions, survives in numbers from hundreds of thousands to millions of individuals. That its extinct component subspecies, the Quagga, was once the most numerous zebroid in all of southern Africa, is not a fact that should lead to complacency. The Grévy’s Zebra has declined recently due to drought and human–wildlife conflicts in a region of the world undergoing human suffering and strife (20). But, perhaps the most threatened perissodactyls are the rhinoceroses. The Javan Rhinoceros numbers ~60 individuals in Indonesia and perhaps fewer than six in Vietnam (21). The Sumatran Rhinoceros has continued to decline as forest habitat is replaced by agricultural and agroforestry operations in concert with continued poaching impacts (22). The value of Rhinoceros horn and other body parts continues to place rhino species at great risk (23). The Northern White Rhinoceros may be extirpated from the Congo basin where, in its last stronghold in Garamba National Park, in spite of heroic efforts, numbers continue to decline with just five animals remaining (24). The Southern White Rhinoceros, once on the brink of extinction, has recovered dramatically (23)—with appropriate intervention, rhinos need not go extinct.

510


Table 1. Divergence times (Ma) and their credibility/confidence intervals (CI) among rhinoceroses, tapirs, and horses (Perissodactyla). Timetree Node

Time

Estimates Ref. (3)

Ref. (4)

Ref. (5)

Ref. (6)

Ref. (7)

Ref. (8)

Ref. (9)

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

1

55.1

56.5

58–54

56.3

58–54

55.2

59–52

56.3

59–54

53.4

56–51

53

58–50

54.8

59–50

2

48.6

48.8

52–45

48.4

52–44

–

–

–

–

–

–

–

–

–

–

Acknowledgments Support was provided by the U.S. National Institutes of Health and by the Caesar Kleberg Wildlife Foundation.

References 1. D. E. Wilson, D. M. Reeder, Eds., Mammal Species of the World, 3rd ed. (Johns Hopkins University Press, Baltimore, 2005), pp. 634–637. 2. C. Groves, O. A. Ryder, in The Genetics of the Horse, A. T. Bowling, A. Ruvinsky, Eds. (CABI, Oxfordshire, 2000), pp. 1–24. 3. M. S. Springer et al., Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003). 4. W. J. Murphy et al., Genome Res. 17, 413 (2007). 5. U. Arnason et al., J. Mol. Evol. 50, 569 (2000). 6. Y. Cao et al., Gene 259, 149 (2000). 7. M. Hasegawa et al., Genes Genet. Syst. 78, 267 (2003). 8. A. D. Yoder et al., Mol. Ecol. 13, 757 (2004). 9. C. Poux et al., Mol. Syst. Biol. 54, 719 (2005). 10. C. Tougard et al., Mol. Phylogenet. Evol. 19, 34 (2001). 11. P. Fernando et al., Conserv. Genet. 7, 439 (2006). 12. E. A. Oakenfull, H. N. Lim, O. A. Ryder, Conserv. Genet. 1, 341 (2000). 13. J. Weinstock et al., PLoS. Biol. 3, e241 (2005).

14. J. E. Norman, M. V. Ashley, J. Mol. Evol. 50, 11 (2000). 15. M. L. Houck et al., Sonderdruck aus Verhandlumgsbericht des 37. Internationalen Symp. uber die Erkrankungen der Zootiere, Dresden, 25 (1995). 16. B. J. MacFadden, Science 307, 1728 (2005). 17. B. J. MacFadden, Fossil Horses: Systematics, Paleobiology, and Evolution of the Family Equidae (Cambridge University Press, New York, 1992), pp. 224–227, 323–331. 18. O. A. Ryder et al., Cytogenet. Cell. Genet. 20, 323 (1978). 19. R. L. Cifelli, Evolution 35, 433 (1981). 20. P. D. Moehlman, Ed., Equids: Zebras, Asses, and Horses: Status Survey and Conservation Action Plan (IUCN/ SCC Equid Specialist Group, IUCN, Gland, Switzerland, 2002), pp. 1–190. 21. International Rhino Foundation, Javan Rhino, http:// www.rhinos-irf.org/javan/ (2008). 22. T. J. Foose et al., Asian Rhinos: Status Survey and Conservation Action Plan (IUCN/SSC Asian Rhino Specialist Group. IUCN, Gland, Switzerland, 1997), pp. 1–122. 23. R. Emslie et al., African Rhino: Status Survey and Action Plan (IUCN/SCC African Rhino Specialist Group, IUCN, Gland, Switzerland, 1999), pp. 1–91. 24. International Rhino Foundation, Northern White Rhino, http://www.rhinos-irf.org/en/cms/?304 (2008).

Whales and even-toed ungulates (Cetartiodactyla) John Gatesy Department of Biology, University of California, Riverside, CA 92521, USA ([email protected])

Abstract Whales and even-toed ungulates are grouped into ~24 families within the mammalian Order Cetartiodactyla. Recent phylogenetic analyses of molecular and morphological data robustly support most interfamilial relationships, including a nested position of whales within the order. However, resolution among basal clades of toothed whales and groupings within Ruminantia remain elusive. The fossil record of Cetartiodactyla is rich and has inspired many molecular clock studies. The cetartiodactyl timetree suggests that the earliest divergences among living species may have occurred in the Cretaceous (146–66 million years ago, Ma) and that the majority of splits among families were in the Oligocene (34–23 Ma).

Whales (Cetacea) and even-toed ungulates (“Artiodactyla,” not a natural group) are represented by ~290 extant species (1), and have been grouped into 20–25 families within the placental Order Cetartiodactyla (Fig. 1). Extinct diversity is well represented in the clade, with about six extinct genera for every genus that contains extant species (2). This rich fossil record has facilitated molecular clock analyses (3–14), but phenotypic diversity in the group is extensive and has confounded systematic studies. Cetartiodactylans range in body size from the tiny ~4 kg mouse deer, Tragulus javanicus, to the enormous 190,000 kg Blue Whale, Balaenoptera musculus (1). It is often difficult to assess homology of structures in organisms that are as divergent as these in terms of anatomical organization, mass, and ecological specialization. Among taxa with living representatives, McKenna and Bell (2) recognized 11 families of Cetacea and 10 families of “Artiodactyla.” Three additional family-level groups (Eschrichtiidae, Neobalaenidae, and Kogiidae) are commonly accepted. Here I review the phylogenetic relationships and divergence times of the families of Cetartiodactyla (Fig. 2).

The systematic database for Cetartiodactyla is large. Nuclear genome sequences are completed or in progress for members of four cetartiodactyl families. Mitochondrial (mt) genomes have been sequenced from 18 families (11, 13, 15), insertions of transposons have been scored from most families (7, 16, 17), and cladistic analysis of paleontological data sets is at an advanced stage (18–25). Perhaps most importantly, several large matrices of fossil and molecular data have been compiled (26–29). Studies based on these combined data sets permit direct synthesis of DNA sequence data with temporal information from the fossil and geological record. Given the wealth of systematic evidence, there are some areas of strong congruence between analyses of morphological and molecular data sets. Suina, a grouping of Suidae (pigs) and Tayassuidae (peccaries), is robustly supported by both DNA sequences and phenotypic characters (3, 14, 16, 20, 23–25, 27, 28, 30–33). Likewise, a cluster of species that “chew the cud,” Ruminantia, has been consistently supported by diverse analyses, in combination with Pecora, a subclade of Ruminantia (3, 9, 14, 16, 18, 20, 23, 27, 28, 30–35). Pecora includes all extant cetartiodactylans with prominent cranial appendages (Bovidae = antelopes and cattle, Cervidae = antlered deer, Giraffidae = giraffes, and Antilocapridae = pronghorn

Fig. 1 An antilocaprid (Antilocapra americana; foreground) and a bovid (Bison bison; background). Credit: painting by C. Buell (J. Gatesy, copyright).

J. Gatesy. Whales and even-toed ungulates (Cetartiodactyla). Pp. 511–515 in The Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009).

512


Monodontidae 23 21

Phocoenidae

12

19

Pontoporiidae Lipotidae

11

Hyperoodontidae Platanistidae

9

Cetacea

Iniidae 20

Odontoceti

Delphinidae 16

Kogiidae 17

Physeteridae

7 18 4

Eschrichtiidae Neobalaenidae

14

Mysticeti

Balaenopteridae 22

Balaenidae Hippopotamidae

Moschidae

13

Cervidae

10

Giraffidae

8

2

Antilocapridae

5

Suidae

6

Tayassuidae

Suina

Tragulidae 1

"Artiodactyla"

15

Ruminantia

Bovidae

3

Camelidae Paleogene

Neogene

CENOZOIC 50

25

0 Million years ago

Fig. 2 A timetree of whales and even-toed ungulates (Cetartiodactyla). Divergence times are shown in Table 1. Hyperoodontidae = Ziphiidae.

antelopes), as well Moschidae (musk deer), which may be hornless primitively (28) or secondarily (9, 18). DNA data and the phenotypic evidence also generally agree in supporting Cetacea (whales), Odontoceti (toothed whales), Mysticeti (baleen whales), Eschrichtiidae (Gray Whale) + Balaenopteridae (rorqual baleen whales), Physeteroidea (Physeteridae = Giant Sperm Whale and Kogiidae = Dwarf and Pygmy Sperm Whales), Iniidae (Amazon River Dolphin) + Pontoporiidae (Franciscana Dolphin), and Delphinoidea (Delphinidae = oceanic dolphins, Phocoenidae = porpoises, and Monodontidae = Beluga and Narwhal) (4–8, 10, 11, 13–17, 19, 20, 22, 23, 26–33, 36, 37).

Despite this consensus, sharp conflicts between molecules and morphology have emerged over the past 20 years. Non-monophyly of “Artiodactyla,” even-toed ungulates, is perhaps the most striking molecular incongruence with traditional mammalian taxonomy. Multiple nuclear gene sequences (5, 6, 10, 30, 31), mt genomes (11), and insertions of transposons (16) support a close relationship between Cetacea and Hippopotamidae (hippos), which is closest to Ruminantia (Fig. 2). The clusters render “Artiodactyla” paraphyletic, in contrast to most cladistic analyses of phenotypic data that favor a monophyletic grouping of even-toed ungulates (20, 21, 25, 28; but see 23, 24). Recently discovered hindlimbs

Table 1. Divergence times (Ma) and their confidence/credibility intervals (CI) among whales and even-toed ungulates (Cetartiodactyla). Timetree Node

Time


Ref. (4)

Refs. (7, 9)

Ref. (8)(a)

Ref. (8)(b)

Ref. (10)(a)

Ref. (10)(b)

Refs. (13, 14)

Time

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

Time

CI

65–62

55.3

60–51

74.1

77–71

1

67.3

–

–

–

–

–

67.3

73–63

64.1

68–60

63.8

2

63.5

–

–

–

–

–

63.5

68–59

62.5

66–57

60.5

62–59

50.7

56–46

71.6

74–70

3

59.1

–

–

–

–

–

59.1

63–55

57.7

63–55

55.7

57–54

44.1

49–39

65.9

68–63

4

52.9

53.3

–

–

–

–

52.9

55–52

52.9

55–52

52.4

54–52

39.8

45–35

59.5

62–57

5

46.3

–

–

–

46.3

55–39

–

–

–

–

–

–

–

–

–

–

6

42.8

42.8

–

–

–

–

–

–

–

–

–

–

–

–

49.7

64–36

7

32.3

–

47.5

54–42

32.3

37–27

27.4

33–22

25.5

31–22

29.6

34–25

20.3

25–16

–

–

8

31.6

–

–

–

31.6

36–28

–

–

–

–

–

–

–

–

–

–

9

30.0

32.1

45.5

51–41

30.0

35–25

–

–

–

–

–

–

–

–

–

–

10

29.4

–

–

–

29.4

33–26

22.0

28–17

–

–

–

–

–

–

–

–

11

28.9

–

–

–

28.9

34–24

–

–

–

–

–

–

–

–

–

–

12

28.2

–

–

–

28.2

33–23

–

–

–

–

–

–

–

–

–

–

13

27.8

–

–

–

27.8

30–25

–

–

–

–

–

–

–

–

–

–

14

27.3

20.9

31.0

36–26

–

–

–

–

–

–

–

–

–

–

27.3

29–25

15

26.2

–

–

–

26.2

29–24

–

–

–

–

–

–

–

–

–

–

16

25.0

22.4

–

–

25.0

30–20

–

–

–

–

–

–

–

–

–

–

17

24.5

24.5

37.0

42–32

–

–

–

–

–

–

–

–

–

–

–

–

18

23.3

17.6

–

–

19

21.5

–

–

20

19.9

18.4

23.0

21

19.8

16

24.5

22

19.3

14.7

21.5

26–18

–

–

–

–

–

–

–

–

23

13.1

12.8

20.0

22–18

13.1

17–9

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

23.3

26–21

21.5

26–17

–

–

–

–

–

–

–

–

–

–

26–20

19.9

24–16

–

–

–

–

–

–

–

–

–

–

28–22

19.8

24–16

–

–

–

–

–

–

–

–

–

–

–

–

19.3

22–16

–

–

–

–

Note: Node times in the timetree are from refs. (3, 7, 8(a), 9, 11, 13). For ref. (8), times are included from (a) Bayesian analysis of mitochondrial and nuclear DNA from ref. (5), and (b) Bayesian analysis of mitochondrial genome data translated into amino acids. For ref. (10), times are included from (a) Bayesian analysis of mitochondrial and nuclear DNA from ref. (6) using all calibrations, and (b) Bayesian analysis of mitochondrial and nuclear DNA from ref. (6) using all calibrations except those within Cetartiodactyla.

514


from Eocene whales (21, 38) show that early cetaceans had a paraxonic tarsus with a typical “artiodactyl” ankle (39), but such characters provide support for a grouping of Cetacea with “Artiodactyla,” but not within “Artiodactyla” (21, 25, 28). Additional conflicts between morphology and molecules include monophyly vs. paraphyly of Balaenoidea (Balaenidae = right whales and Neobalaenidae = Pygmy Right Whale) (11, 13, 15, 17, 22, 29, 37), monophyly vs. non-monophyly of Phocoenidae + Monodontidae (4, 7, 11, 15, 22, 28, 32, 36), monophyly vs. polyphyly of river dolphins (Iniidae, Pontoporiidae, Lipotidae = Chinese River Dolphin, and Platanistidae = Indian River Dolphin) (4, 7, 11, 15, 22), monophyly vs. polyphyly of Selenodontia (Ruminantia and Camelidae = camels), and monophyly vs. non-monophyly of Suiformes (Suina and Hippopotamidae) (3, 5, 6, 10, 11, 16, 20, 21, 23, 25, 28, 30–32). Molecular resolutions of these conflicts are shown in Fig. 2. Neither separate nor combined systematic matrices have robustly resolved some interfamilial relationships within Cetartiodactyla. In particular, basal clades of Odontoceti and the five pecoran ruminant families (Bovidae, Cervidae, Giraffidae, Antilocapridae, and Moschidae) have been resistant to hierarchical grouping, and extensive disagreements among characters remain. The topology in Fig. 2 illustrates odontocete phylogeny consistent with the transposon analysis of Nikaido et al. (7) and ruminant phylogeny according to analysis of mt and nuclear genes by Hassanin and Douzery (9). However, these relationships remain highly controversial and are disputed by alternative, large data sets (e.g., 11, 22, 28). Furthermore, the basal positioning of Camelidae as the closest relative of all other extant cetartiodactylans has been debated, with various molecular and combined matrices favoring different local rearrangements. In general, the conflict here is between mt (11) and nuclear data (5, 6, 10, 16, 30), with the result favored by the nuclear evidence, the basal position of Camelidae (Fig. 2), prevailing in the most comprehensive compilation of characters to date (28; >600 phenotypic and >40,000 molecular characters). The timetree for Cetartiodactyla is based on six molecular clock analyses (Fig. 2; Table 1). These include four Bayesian analyses of multiple nuclear and/or mt gene sequences (7–9, 13), a distance analysis of mt genes (3), and a study that utilized maximum likelihood branch lengths for mt genomes (11). All six studies employed fossils to establish calibration points. Alternative estimates of time for various nodes are shown in Table 1. The temporal pattern of diversification in the timetree suggests

that much of the splitting among cetartiodactyl families (12 of 23 divergences) occurred in the Oligocene. These divergences include the difficult-to-resolve radiations at the base of Odontoceti and at the base of Pecora (Fig. 2). The 95% credibility intervals for estimates of divergence times among all five pecoran families overlap, and speciation events that separate odontocete families also are tightly spaced in time. The earliest branching point within Cetartiodactyla predates the Cretaceous/ Paleocene boundary in the timetree, but alternative analyses push multiple nodes to the Cretaceous (14) or restrict all nodes to the Cenozoic (66–0 Ma) (10), depending on choice of fossil constraints, database, and methodology. Molecular clock analyses within Cetartiodactyla have been common, in part because extinct taxa have been directly integrated into cladistic studies, and also because extensive genetic data have been compiled for members of this group. However, even with this effort, critical ambiguities remain. For example, the divergence of Cetacea from other cetartiodactylans is a very common calibration point for molecular clock analyses of mammals; van Tuinen (12) counted >30 studies that have utilized this divergence, yet the date remains poorly constrained because of the phylogenetic instability of some extinct taxa. The hippopotamid lineage apparently extends back to the Eocene through a paraphyletic series of extinct “anthracotheriids” (24, 28), but the identity of the earliest stem cetaceans remains controversial. Based on cranial and dental evidence, the extinct Pakicetidae (~52 Ma) has consistently been placed on the stem lineage of extant Cetacea (12, 19–21, 23, 26–28), but the position of Paleocene mesonychids (~62 Ma) remains undetermined. Systematic studies of basicranial characters (19), dental characters (40), and the largest cetartiodactyl matrix compiled to date (28) group Cetacea closer to Mesonychidae than to “artiodactylans,” but other data sets entirely exclude Mesonychidae from Cetartiodactyla (21, 23, 25, 27, 41). Alternative fossil calibrations have led to very different molecular estimates of time. In particular, divergences within Cetacea (4, 7, 11) and at the base of Cetartiodactyla (8, 10, 12, 14) differ by 10–15 million years (Table 1). These discrepancies illustrate the critical importance of robust fossil calibration points in molecular clock studies. Even a group with a well-documented fossil record, such as Cetartiodactyla presents challenges to researchers who wish to time particular divergences using molecular clocks. However, given the wealth of information for Cetartiodactyla, it is very likely that this group will continue to be an exemplar clade for testing

Eukaryota; Metazoa; Vertebrata; Mammalia; Cetartiodactyla

new and improved methods for dating evolutionary events using a combination of diverse systematic data.

Acknowledgment This work was supported by the U.S. National Science Foundation.

References 1. R. M. Novak, Walker’s Mammals of the World: Volume II, 5th ed. (Johns Hopkins University Press, Baltimore, Maryland, 1991). 2. M. C. McKenna, S. K. Bell, Classification of Mammals above the Species Level (Columbia University Press, New York, New York, 1997). 3. C. Montgelard, F. M. Catzeflis, E. Douzery, Mol. Biol. Evol. 14, 550 (1997). 4. I. Cassens et al., Proc. Natl. Acad. Sci. U.S.A. 97, 11343 (2000). 5. W. J. Murphy et al., Nature 409, 614 (2001a). 6. W. J. Murphy et al., Science 294, 2348 (2001b). 7. M. Nikaido et al., Proc. Natl. Acad. Sci. U.S.A. 98, 7384 (2001). 8. M. Hasegawa, J. L. Thorne, H. Kishino, Genes Genet. Syst. 78, 267 (2003). 9. A. Hassanin, E. J. P. Douzery, Syst. Biol. 52, 206 (2003). 10. M. S. Springer et al., Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003). 11. U. Árnason, A. Gullberg, A. Janke, Gene 333, 27 (2004). 12. M. van Tuinen, E. A. Hadly, J. Hered. 95, 200 (2004). 13. T. Sasaki et al., Syst. Biol. 54, 77 (2005). 14. O. R. P. Bininda-Emonds et al., Nature 446, 507 (2007). 15. J. Yan, K. Zhou, G. Yang, Mol. Phylogenet. Evol. 37, 743 (2005). 16. M. Nikaido, A. P. Rooney, N. Okada, Proc. Natl. Acad. Sci. U.S.A. 96, 10261 (1999). 17. M. Nikaido et al., Mol. Biol. Evol. 23, 866 (2006). 18. C. M. Janis, K. M. Scott, Am. Mus. Novit. 2893, 1 (1987).

515

19. Z. Luo, P. D. Gingerich, Univ. Mich. Pap. Paleontol. 31, 1 (1999). 20. M. A. O’Leary, J. H. Geisler, Syst. Biol. 48, 455 (1999). 21. J. G. M. Thewissen, E. M. Williams, L. J. Roe, S. T. Hussain, Nature 413, 277 (2001). 22. J. H. Geisler, A. E. Sanders, J. Mammal. Evol. 10, 23 (2003). 23. J. H. Geisler, M. D. Uhen, J. Vertebr. Paleontol. 23, 991 (2003). 24. J.-R. Boisserie, F. Lihoreau, M. Brunet, Proc. Natl. Acad. Sci. U.S.A. 102, 1537 (2005). 25. J. M. Theodor, S. E. Foss, J. Mamm. Evol. 12, 161 (2005). 26. M. A. O’Leary, Am. Zool. 41, 487 (2001). 27. J. H. Geisler, M. D. Uhen, J. Mamm. Evol. 12, 145 (2005). 28. M. A. O’Leary, J. Gatesy, Cladistics 24, 397 (2008). 29. T. A. Deméré, M. R. McGowen, A. Berta, J. Gatesy, Syst. Biol. 57, 15. 30. J. Gatesy, M. C. Milinkovitch, V. Waddell, M. Stanhope, Syst. Biol. 48, 6 (1999). 31. C. A. Matthee, J. D. Burzlaff, J. F. Taylor, S. K. Davis, Syst. Biol. 50, 367 (2001). 32. J. Gatesy, C. A. Matthee, R. DeSalle, C. Hayashi, Syst. Biol. 51, 652 (2002). 33. S. A. Price, O. R. P. Bininda-Emonds, J. L. Gittleman, Biol. Rev. 80, 445 (2005). 34. M. H. Fernandez, E. S. Vrba, Biol. Rev. 80, 269 (2005). 35. F. Kraus, M. M. Miyamoto, Syst. Zool. 40, 117 (1991). 36. S. L. Messenger, J. A. McGuire, Syst. Biol. 47, 90 (1998). 37. A. L. Rychel, T. W. Reeder, A. Berta, Mol. Phylogenet. Evol. 32, 892 (2004). 38. P. D. Gingerich, M. u. Haq, I. S. Zalmout, I. H. Khan, M. S. Malkani, Science 293, 2239 (2001). 39. B. Schaeffer, Evolution 2, 164 (1948). 40. M. A. O’Leary, in The Emergence of Whales, J. G. M. Thewissen, Ed. (Plenum Press, New York, 1998), pp. 133–161. 41. J. G. M. Thewissen, L. Cooper, M. Clementz, S. Bajpai, B. Tiwari, Nature 450, 1190 (2007).


Index 12S 509 12S rDNA 245 16S rDNA 245 16S rRNA 286–287, 303 18S rDNA 136, 164, 261–262, 264, 286, 288 18S rRNA 239–240, 252, 258, 266, 285, 287, 303 26S rDNA 180 28S 264 28S rDNA 262, 275 28S rRNA 252, 258, 266, 287, 303 40Ar/39Ar 31 40K/40Ar 30 A Aardvark 56, 471 AATS 261 abdominal marsupium 466 abdominal spinnerets 255 Abrocomidae 491 Abuta 169 Acalyptratae 274 Acanthaceae 179, 183, 185 Acanthamoebidae 117 Acanthisitta 424, 430 Acanthisittidae 423–424, 426 Acanthocephala 224 Acanthodii 64, 66, 310 Acanthopanax 185 Acanthopanax friedrichii 185 Acanthopanax gigantocarpus 185 Acanthopanax mansfeldensis 185 Acanthopanax obliquocostatus 185 Acanthopterygii 65 Acari 255 accentor 429 Accipitres 436 Accipitridae 439 Accipitriformes 437

ACCTRAN 164, 168 acellular mesoglea 233 aceols 76 Acer 200 Aceraceae 200 Acetobacteraceae 109 Achariaceae 189, 201 Achatocarpaceae 174 Achelata 294 Achoania 63 Acholeplasmataceae 107 Acidobacteria 109 Acipenser 333 Acipenser brevirostrium 332 Acipenser dabryanus 332 Acipenser oxyrinchus 333 Acipenser schrenckii 332 Acipenser sturio 333 Acipenser transmontanus 332 Acipenseridae 333 Acipenseriformes 328–329, 333 acipenseriforms 333 acoelomate flatworms 224 Acoelomorpha 224 Aconoidasida 117 Acoraceae 206 Acorales 203, 209 Acorus 203, 208 Acrantophis 393 Acraspeda 236 Acrobatidae 467, 469 Acrobeles complexus 247 Acrobolbaceae 147 acrocarpous haplolepidae 138 acrocarpous mosses 138 Acroceridae 272, 276 Acrochordidae 390–391, 394 acrochordoid 394 Acrochordus 394 Acrodonta 387

Acroechinoidea 303 Acrogymnospermae 157 Acropora 233 Actiniaria 235 actiniarian 233 Actinidiaceae 178, 182, 185 Actinistia 311 actinistian 310 Actinobacteria 94, 107 actinomycete 192 Actinopodidae 256, 258 Actinopteri 330 Actinopterygii 40, 62–64, 66, 227, 309–310, 325–326, 330 actinorhizal 192 Aculeata 7, 264, 267 adapoids 49 Adelanthaceae 147 Adelobasileus 58 Adephaga 278, 281, 283–284, 286, 288 Aderidae 279 adhesion protein 223 Adoxaceae 180, 184, 186 Adunator 54 advanced bird 420 adzebill 443 Aedes aegypti 274 Aegithalidae 424 Aegithinidae 425, 429 Aeglidae 294 Aegothelidae 454 Aegyptopithecus 49 Aenne liasina 74 Aenne triassica 74 Aesculus 200 Aethotaxis 341 Aextoxicaceae 170, 174 African barbet 445, 449 African lungfishes 349 African Palm Civet 505

518

INDEX

African phiomorph 493 African Savannah Elephant 471 African tenrec 480 African Wild Dog 504 Afrobatrachia 358 Afrophidia 391 afrophidian 390 Afrosoricida 472, 479, 480–481, 496 Afrotheria 39, 56–57, 471–472, 479–480, 496 Agamidae 383–384, 388 Agapostemon virescens 264 Agaricomycotina 216 Agathis 159 Agathosma 200 agave 206 Agdestidaceae 174 Ageratum 177 Agnatha 309–310, 312 Agoutidae 491 Agromyzidae 274 Agyrtidae 283 Ailuridae 505 Ailuropoda melanoleuca 505 Ailurus fulgens 505 aïstopod 62 Aizoaceae 171, 174 Akaniaceae 198 akinetes 113 Aksyiromys 52 Alangiaceae 180, 182, 185 Alaudidae 424 Albian 159 Albuliformes 336 Alcaligenaceae 108 Alcidae 433 alder 193 alderflies 260 Alethinophidia 390–391, 395 Alexiidae 280 algae 94, 114, 116, 119 Alismataceae 203, 206 Alismatales 203–204, 206, 211 alkaloids 177, 183 Allcock, A. Louise 239, 242 Alliaceae 206 Alligator 375, 405 Alligator mississippiensis 405 Alligator sinensis 405 Alligatoridae 405 Alligatoroidea 402 Allodapini 268 Allonautilus 242 Alloposidae 245

allozyme 4 Alnus 193 aloe 206 Alophosia azorica 143 Alopiidae 321 Alphaproteobacteria 89–91, 93, 95–96, 109, 113 Alseuosmiaceae 180, 184 Alsophis 393 Alstroemeriaceae 205 Altiatlasius 50 Altingiaceae 170 alveolates 117, 118 Alytidae 359 amaranth family 174 Amaranthaceae 171, 174 Amazon River Dolphin 512 Amblypygi 255 Amblysomus 496 Amblystegiaceae 142 Ambondro 58 Amborellales 163 ambulacral plate 303 Ambulacraria 224, 226, 228 Ambystomatidae 366 Ameridelphia 57, 467 Amerophidia 391 Amia calva 328 Amiiformes 328–329, 335 ammonoid 31, 246 Amniota 58–59, 62, 309–310, 312, 325, 349, 378 Amoebozoa 117 Amphibia 62 amphibian 6, 9, 59, 309, 325, 349 amphibian-amniote 62 Amphicoela 359 Amphidon 58 Amphiesmenoptera 262 Amphilestes 58 amphinomid 71 amphioxus 45 amphipithecids 49 Amphiprion ocellaris 309 Amphisbaena bakeri 383 Amphisbaenia 9, 309, 383–386, 388 Amphisbaenidae 383–384, 387 Amphisbaeniformata 387 Amphisbaenoidea 387 Amphitherium 58 Amphitretidae 245 Amphiumidae 366 Amphizoidae 284 Ampulicidae 267

Anabantidae 346 Anabantinae 346 Anabantoidei 7, 344, 345 anabaritids 75 Anabas 346 Anacardiaceae 198, 200 Anacardium 200 Anadenobolus arboreus 251 anagenetic branch lengths 22 Anaplasmataceae 109 anapsid 376 Anarthriacae 209 Anarthriaceae 204 Anatidae 417 Anatinae 416 anatropous ovule 162 ancient fish 330 Ancistrocladaceae 170, 174 Anderson, Cajsa Lisa 203 Andreaea 142 Andreaeaceae 140 Andreaeales 141 Andreaeobryaceae 140 Andreaeobryales 140 Andrenidae 265, 267 Andreolepis 64 Andreolepis hedei 63 anemone 45 Anethum graveolens 177 Aneuraceae 148 Angarosphecidae 264 angel shark 322 angiosperm 23, 133–136, 156–159, 161–169, 174–175, 184, 188, 192, 203, 209, 267–268, 276, 288 Angiosperm Phylogeny Group 174, 188, 194, 201, 203 Anguidae 384, 386 Anguilliformes 336 Anguimorpha 383–384, 386–388, 395 Anguinae 386 Anguioidea 386, 388 Anhimidae 417 anhinga 436 Aniliidae 390–391, 393, 395 Anilioidea 391 Anilius 391, 394 animal pathogens 216 animals 93, 116–119, 223, 309 Anisakidae 249 Anisopodidae 271 Anisopsis calculus 70 Ankarapithecus 48 Annelida 68–70, 224, 237

Index

Anniellidae 384, 386 Annonaceae 167 Anobiidae 282 Anolis allisoni 309 Anolis baracoae 383 Anomala 294 Anomalepididae 390–391, 393 anomalous peristomes 143 Anomaluridae 493 Anomaluromorpha 493 Anomochilidae 391 Anomocoela 359 Anopheles 75 Anopheles gambiae 74–75, 248, 274 anoxia 235 Anseranatidae 417 Anseriformes 61, 409, 415–418, 434 Anserinae 415 ant 23 Antalis 240 Antalis torquatus 240 Antarctic Clade 342 antbird 424, 430 anteater 56, 459, 477 antelope 511 Antemus 53 Antemus chinjiensis 52 antennae 270 antheridia 133 Anthicidae 279 Anthoathecata 236 Anthocerophyta 133 Anthocerotophyta 134 anthocyanins 174 Anthomedusae 234 anthophyte 135, 158 Anthozoa 233, 234, 237 anthozoan 76 Anthracobunidae 57 anthracotheriids 514 Anthribidae 281 Anthropoidea 49, 482, 484 anthropoid 49 Antilocapra americana 511 Antilocapridae 511–512, 514 Antilopinae 56 Antipatharia 235 antlered deer 511 Antliophora 262 antpitta 424, 430 Antrodiaetidae 256, 258 ant 260, 268 Anura 354

Aotidae 482 ape 48, 483 Apectodinium augustum 66 APG II 182, 203, 206 apheliscine 57 aphid 45, 72 Aphloiaceae 188, 197–198, 201 Aphonopelma seemanni 255 Apiaceae 177, 180, 183, 185 Apiales 178, 180, 183, 185 apical meristems 153, 161 apicomplexans 117 Apidae 265, 267 Apioceridae 272, 274 Apis 74 Apis mellifera 74 Apium graveolens 177 Aplanulata 237 Aplodontidae 491 Aplopeltura 393 Aplysia 70 Apocrita 74, 262, 266 Apocynaceae 177, 179, 182, 185 Apodidae 455 apodiform 455 Apodiformes 420–422, 456 Apodora 393 Apoidea 265 Aponogetonaceae 206, 208 Apotreubia 148 Apsilocephalidae 274 Apsisaurus 60 Apterygidae 413 Apterygiformes 413 Aptornithidae 442 aquatic herbs 171 Aquificaceae 109 Aquificae 109 Aquifoliaceae 180, 183 Aquifoliales 178, 180, 183, 186 Aquilegia 169 Aquitanian 394 Arabidopsis thaliana 198 Araceae 203, 206, 209 Arachnida 252, 255 Arachnoididae 303 aragonite 236 Arales 211 Aralia antiqua 185 Araliaceae 180, 183 Aralidiaceae 180, 183, 185 Aramidae 442 Arandaspis prionotolepis 67 Araneae 255, 257

519

Araneidae 256, 258 Araneoidea 259 Araneomorphae 259 Araucariaceae 159 Arbacia punctulata 302 Arbaciidae 303 arboreal possum 469 arboreal songbird 419 arbuscular mycorrhizal symbionts 217 Arcanator 429 Archaebacteria 3, 6–8, 13, 89–92, 96 Archaeoglobi 102 archaeocyaths 76, 235 Archaeoglobaceae 102 Archaeonycteris 54 Archaeopneustes 303 Archaeopteryx 409 Archaeotetraodon winterbottomi 65 Archaeothyris 59 Archaeplastida 117 archaic bird 409 archegonia 133 Archescytinidae 72 Archexyelinae 73 Archiascomycetes 216, 217 Architeuthis dux 244 Archonta 50–51, 54, 56 Archontoglires 51 Archosauria 61, 376, 378 Archosauromorpha 60 archosaurs 378 Archostemata 278, 281, 284, 288 Arctoidea 507 Ardea herodias 419 Ardipithecus ramidus 46 Arecaceae 204, 211 Arecales 206, 208 Argentiniformes 335 argon 30 Argonautidae 243, 245 Argonautoidea 243, 245 Argophyllaceae 180, 184 Aristolochiaceae 167, 203 Arizonasaurus 61 Arkarua 68 armadillo 44, 56–58, 477 Armaniidae 264 armored catfish 335 army ants 267 Arnelliaceae 147 arrow poison 169 arrow worms 226 Arsinoitherium 57

520

INDEX

Artamidae 425 Artedidraconidae 340 arthrodontous mosses 140, 144 arthrodontous peristome 142, 143 Arthroleptidae 358 Arthropoda 7, 68–69, 71–72, 89, 119, 224, 226, 249, 251–253, 309 artichokes 177 articulated chitinous exoskeleton 251 articulated legs 251 artiodactyl ankle 514 Artiodactyla 45, 53–55, 471, 509, 511–512, 514 arum family 203 Ascalaphidae 291 Ascaphidae 359 Ascarididae 248, 249 Ascaridomorpha 249 Aschiza 273 Ascomycota 216 Asfaltomylos 58 Asher, Robert J. 35 Asian barbet 445, 449 Asian eosimiids 49 Asian linsang 506 Asiatherium 58 Asilidae 272, 274 Asiloidea 272 Asparagaceae 203, 206 Asparagales 205–206, 208, 210 asparagus 206 asphodels 206 Aspidodiadematidae 303 Aspidytidae 284 ass 509 Astacidae 294 Astacidea 294 Asteliaceae 205, 209 Aster 178 Asteraceae 177, 180, 184, 186 Asterales 178, 180, 186 Asterids 162–164, 169–170, 173–174, 177, 179, 181, 185 Asteropeiaceae 171, 174 Asterophryidae 358 Astriclypeidae 303 Astrochronologic recalibration 30 astrochronologic scale 30 Astylosternidae 358 Atdabanian 75 Atelestidae 272 Atelidae 480 Atelocerata 71, 251 Athecata 237

athecate hydroids 237 Athericidae 272 Atherinidarum 65 Atherinomorpha 65 Atherospermataceae 167 Atlantogenata 56, 472 atpB 136, 164 atractaspidid 390 atractaspidine 393 Atractaspis 393 Attelabidae 281 Attheyaceae 128 Atyidae 294 Atypidae 256, 258 Atypoidea 256, 258 Aucubaceae 182 Aulacomniaceae 139, 142 Aulacomniales 139, 142 Aulacomnium 142 Aurelia aurita 233 Ausktribosphenida 58 Australasian marsupial 466 Australian Lungfish 348 Australidelphia 57, 467 australidelphian 58 australopithecine 46 Australosphenida 58, 462 Austrobaileyales 162 authigenic sedimentary minerals 30 autocorrelated 41 autumn crocus 205 Avemetatarsalia 61 Aves 9, 375–376, 409–411, 423, 429–430, 437–439, 445, 449, 455 Avesuchia 61 avesuchian 61 Avicenniaceae 183 Avise, John C. 19 avocado 166 avocet 433 Axymyiidae 271, 275 Ayoub, Nadia A. 255 Azateaceae 195 B babbler 429 Bacillaceae 107 Bacillariophyceae 128 Bacillariophycidae 128 Bacillariophyta 117 bacteria 89–91, 93, 192 Bacteroidaceae 109 Bacteroidetes 109 badger 504

Baker, Allan J. 412, 415, 432 Balaenidae 512, 514 Balaenoidea 514 Balaenoptera musculus 511 Balaenopteridae 512 Balanerpeton woodi 62 Balanites 191 Balanoidea 299 Balanomorpha 299 Balanopaceae 189 Balantiopsidaceae 147 baleen whale 512 Balsaminaceae 178, 182, 185 balsaminoid 182 Baltodentialiidae 239 banana 206 bandicoot 58, 468 Bangiales 117 Bangiomorpha 116 baphetid 62 Baptornis 62 barbet 447 Barbeuiaceae 174 Barbeyaceae 190, 193 Barker, F. Keith 423 barley 208 Barn Owl 451 Bartonellaceae 109 Bartramiaceae 140, 142 Bartramiales 140 Barychelidae 256, 258 basal land plants 141, 143 basal meristem 203 Basellaceae 174 Basidiomycetes 215 Basidiomycota 215–216, 218 basil 177 basisphenoid bulla 479 Basommatophora 70 bat 44, 53–54, 56–58, 459, 471, 496, 499, 502 Bataceae 198 Bataguridae 399 Bathydraconidae 340 Bathyergidae 491 Bathyteuthidae 243 Bathyteuthoida 243 Batodon 53 Batoidea 320, 322, 326 Batrachia 354 Batrachomorpha 62 batrachomorphs 60 Battistuzzi, Fabia U. 101, 106

Index

bay 166 bay cedar 192 bay owl 451 Bayesian 150 Bayesian analysis 164 Bayesian approaches 42 Bayesian relaxed clock 186 Bdellovibrionaceae 109 bean 188, 192 bear 55, 504 beaver 492 bee fly 272 beech 193 bee 260, 264, 266 beetle 23, 45, 71–73, 260, 278, 286, 288 begonia 193 Begoniaceae 190, 193 Belidae 281 Belontia 346 Belontiidae 345 Beluga 512 Bennettitales 158 benthic octopus 242, 245 Benton, Michael J. 35 Berberidaceae 171 Berberidopsidaceae 170, 174 Berberidopsidales 163 Berberis 169 Bergenia 175 Berothidae 291 Bertone, Matthew A. 270 betalain pigments 174 Betaproteobacteria 108 Bethylidae 265 Bethylonymidae 264 Betta channoides 344 Betula 193 Betulaceae 190, 193 Bhattacharya, Debashish 116 Bibionidae 271 Bibionomorpha 271–272, 275 bichir 309, 328 Biebersteiniaceae 200 Bifidobacteriaceae 107 Bifurcata 386 bifurcatan 388 Big-headed Turtle 399 Bignoniaceae 179, 183 bilabiate flower 183 bilateral symmetry 223 Bilateria 67–69, 71, 75–76, 223–224, 226, 228, 234 bilby 467 bio-magneto-sequence 31

Biomphalaria 45, 70 biostratigraphic correlation 27 biostratigraphy 30 biota 249 biozonal units 30 bipartite perianth 162 Bipedidae 384, 387 Bipedoidea 387 Biphyllidae 280, 287 birch 193 bird 5–6, 13, 38–41, 58–62, 226, 309, 349, 375–378, 415 bird-mammal divergence 38, 40, 41, 58–60, bird of paradise 206, 429 Biretia 49 birth-death model of diversification 42 Bison bison 511 bivalve 57, 62 Bixaceae 198, 200 black coral 235 black pepper 166 blackbird 429, 431 Black-crested Coquette 454 Blair, Jaime E. 215, 223 Blandfordiaceae 205, 209 Blanidae 384, 387 Blanoidea 387 Blasiales 146, 148 Blastocladiomycota 217 Blatoptera 72 Blattodea 72 Blaxter, Mark 247 blechnoid fern 155 Blentosomys 52 Blephariceridae 271 Boa 393 boa 390 Bobolestes 53 bobtail squid 243 Boidae 391, 395 boine 394 Bolboceratidae 283 Bolitaenidae 243, 245 Bolyeriidae 391, 395 Bombacaceae 200 Bombinatoridae 359 Bombini 268 Bombycillidae 425, 429 Bombyliidae 272, 276 bonefish 335 bony endoskeleton 310 bony vertebrate 320 bonytongue 335

booby 436 Booidea 394 Boraginaceae 177, 179, 182 Boreoeutheria 53, 56, 472 Boreosphenida 462 Boridae 279 Boryaceae 205, 209 Bos 55 Boselaphini 56 Bossuyt, Franky 357 Bostrichidae 282 Bostrichiformia 287 Bostrichoidea 282 Bothrideridae 280 Bothriechis 393 bottlebrush 194 bottletail squid 242 Bovichtidae 341 Bovidae 55–56, 511–512, 514 Bovinae 56 Bovini 55 Bowfin 309, 328, 335 boxwood 173 Brachaeluridae 321 Brachinidae 284 Brachiopoda 224 Brachycephalidae 359 Brachycera 74–75, 275 Brachytheciaceae 139, 142 Brachyura 294 Brady, Seán G. 264 Bradynobaenidae 265 Bradypodidae 476 Bradypus 476 Bradyrhizobiaceae 109 Braithwaiteaceae 139 branchial canal 244 branchiopod 71 Branchiopoda 252 Branisella 49, 485 Brassica oleracea 198 Brassicaceae 188, 197–198, 200 Brassicales 201 Braunia squarrulosa 133 Bremer, Birgitta 177 Brentidae 281 Brevicipitidae 358 brine shrimp 252 Brissidae 303 broccoli 198 Brochu, Christopher A. 402 Bromeliaceae 203–204, 208, 211 Bromsgroveia 61 brontotheriid 55

521

522

INDEX

Brown, Joseph W. 436, 451, 454 Brucellaceae 109 Brunelliaceae 194 Bruniaceae 180, 183 brush-turkey 415 Bryaceae 140, 144 Bryales 140–141, 144 Bryidae 144 bryoflora 143 Bryophyta 133–134, 136–138, 141, 150 Bryozoa 224 bryozoan 237 buccal crown 244 Bucconidae 446, 448 buckthorn 193 Buddlejaceae 183 Bufonidae 359 bulldog ant 267 Bungarus 393 Bunolagus monticularis 487 bunting 429 Buprestidae 282 Buprestoidea 282, 287 Burhinidae 433 Burkholderiaceae 108 Burmanniaceae 206 Burramyidae 467, 469 Burseraceae 198, 200 bush shrike 429 bushbaby 44, 49–51, 58 bushfish 344 bustard 441 Buteo lagopus 436 Butomaceae 206 butterfly 260 butterfly ray 324 buttonquail 432, 441 Buxaceae 169, 171, 173 Buxbaumiaceae 140 Buxbaumiales 141 Byblidaceae 183 Byrrhidae 282, 287 Byrrhoidea 282, 287 Byturidae 280, 287 C cabbage 188, 198 cabbage family 197 Cactaceae 171, 174 cactus family 174 CAD 261, 274 caddisfly 260 Cadeidae 384, 387 Cadulus 240

Cadulus groenlandicus 240 caecilian 62, 309, 312 Caeciliidae 370 caecum 495 caenolestid 466 Caenolestidae 467 Caenophidia 390, 395 Caenorhabditis 236, 247, 249 Caenorhabditis briggsae 249 Caenorhabditis elegans 249 caffeine 177 caiman 405 Calabaria 393 Calamaria 393 Calanticidae 299 Calcarea 224 calcareous skeleton 223 calcisponge 76 calcitic endoskeleton 302 calcitic skeleton 236 calcium carbonate 236 calibration date 35, 38 calibration point 5 Calidris canutus 432 calla lily 203 Callaeatidae 426 Callianassidae 294 Calliphoridae 275 Callirhipidae 282 Callistemon 194 Callorhinchidae 322 Calomyscidae 491 Calorhamphus 447 caluromyid 467 Caluromyida 467 Caluromyidae 467 Calycanthaceae 167 Calyceraceae 180, 184 Calypogeiaceae 147 Calyptocephalellidae 359 Calyptratae 274 calyx 162 Camarodonta 303 Cambaridae 294 Cambaroidinae 294 Cambrian Explosion 6, 9, 13, 228 Camelidae 55, 512, 514 Camellia sinensis 177 camel 514 Campanula glomerata 177 Campanulaceae 177, 184 campanulid 162, 164, 177–178, 182–183, 186 Campephagidae 426

Campylobacteraceae 109 Campynemataceae 206 Cancridae 294 Candelabridae 237 Candida albicans 216 Candoia 393 Canellaceae 167 Canellales 167 Canidae 507 caniform 55, 506 Caniformia 14, 55, 507 Cannabaceae 193 Cannaceae 208 canna-lily 206 Cannatella, David C. 353 Cannnaceae 204 Cantharidae 282 Canthyloscelidae 27 Capitata 237 Capitella 45, 71 Capitellid polychaete 70 Capitonidae 446, 449 Caprifoliaceae 180, 184, 186 Caprimulgiformes 420–421, 436, 451–452, 455 Caprini 55 Capromyidae 491 Caprotragoides 56 Capsicum annuum 177 Captonidae 447 Captorhinidae 59 capybara 490 Carabidae 284 carbohydrate inulin 184 Carboniferous 36 Carboxydothermus 107 Carcharhinidae 321, 325 carcharhiniform 325 Carcharhiniformes 321, 322, 325 Carcharhinus 349 Carchariidae 321, 325 Carcharodon carcharias 320 cardamom 208 cardinal 429 Cardinalidae 424, 429 Cardiopteridaceae 180, 183, 185 Carettochelyidae 400 Carex digitata 203 Cariamidae 437 Caricaceae 199 Caridae 281 Caridea 294 Cariridris 267

Index

carnation family 174 Carnilestes 53 Carnivora 14, 53–55, 472, 507 Carnivoramorpha 54 carnivoran 56 carnivore 233, 471, 496, 506 carnivoriform 508 carpel 157, 162 Carphodactylidae 384, 387 Carpinus 193 carpolestid 51 carrot 177 cartilaginous fish 6, 309, 320, 323 Carya 193 Caryocaraceae 189 Caryophyllaceae 171, 174 Caryophyllales 162, 170–171, 174 Casarea 393 Caseidae 59 cashew 200 cassava 194 Cassidulidae 303 Cassiduloida 303, 304 cassowary 413 Castoridae 492 Castorimorpha 493 Casuariidae 413 Casuariiformes 413 Casuarinaceae 190, 193 cat 44, 53–58, 504 Catagoniaceae 139 Catagonium 142 catarrhine 49 Catarrhini 48–49, 480, 485 Catascopiaceae 142 catfish 336 Catharanthus roseus 177 Cathartidae 439 Cathartiformes 437 Cathaymyrus diadexus 67 Catopithecus 49, 484 Caudata 354 cauliflower 198 Caulobacteraceae 109 Cavia 52 Caviidae 491 caviomorph rodent 490 Caviomorpha 52, 491, 493 Caytonia 158 Cebidae 480, 482 Cecidomyiidae 271 Celastraceae 190, 194 Celastrales 188, 190–191, 193 celery 177

cell signaling 223 Celliforma favosites 268 cellular endosperm 182 cellular mesenchyme 233 Cellulomonadaceae 107 Celtidaceae 190 Cenomanian 394 centipede 251 Centrolenidae 359 Centrolepidaceae 204, 209 Centrophoridae 321, 325 Centrospermae 174 cephalic arterial pattern 476 cephalobomorph nematode 247 Cephalocarida 252 Cephalochordata 67, 224, 226, 309 cephalochordate 68 cephalon 251 Cephalopoda 7, 240, 244 cephalopod 241 Cephalotaceae 194 Cephalotaxaceae 159 Cephaloziaceae 147 Cephaloziellaceae 147 Cephaloziineae 150 Cerambycidae 281 Ceratobatrachidae 358 Ceratocanthidae 283 Ceratodontidae 349 Ceratodontoidei 349 ceratomorph 508 Ceratomorpha 509 Ceratophryidae 359 Ceratophyllales 161–163, 165 Ceratophyllum 208 Ceratopogonidae 271 Cercidiphyllaceae 170 cercomonad 117 Cercopithecidae 480, 483, 485 Cercopithecoidea 48–49, 485 cereal rust 216 ceres 415 Ceriantharia 235 Certhiidae 425, 429 Cervidae 511–512, 514 Cerylonidae 280 Cesium atom 26 cesium atomic clock 26 Cetacea 54, 471, 511–512, 514 cetacean 55 Cetartiodactyla 39, 54–55, 471–472, 496, 511, 515 cetartiodactyl 54 Cetorhinidae 321

523

chachalaca 415 Chadronian 55 Chaeropodidae 469 Chaeropus 468 Chaetognatha 226 chalaza 162 Chalcidoidea 74 Chamaeleonidae 383–384, 388 Chamaeleonoidea 388 Chambius 56 chameleon 388 chancelloriid 76 Channichthyidae 340 Channidae 344 Channoidei 344 Chaoboridae 271 Characiformes 336 characiform 337 Charadrii 433 Charadriidae 433 Charadriiformes 420, 432–434, 441 charadriiform 61 Charina 395 Charina prebottae 394 Charnia 75 charophycean outgroup 136 Chase, Mark W. 166, 169, 188, 197 checkered beetle 278, 287 cheliceral venom glands 255 Chelicerata 252 chelicerates 119 Chelidae 399 Chelonariidae 282 Cheloniidae 399 Chelydridae 399 Cheriogaleidae 480 chewing mouthparts 270 chicken 44, 60–62, 410, 415 chicken-like bird 423 chilli 177 Chilopoda 252 chimaera 90, 309, 320, 325 Chimaeridae 322 Chimaeriformes 322 Chimaeroidei 320 chimpanzee 44, 46, 48, 51, 56–58, 482 Chinchillidae 491, 493 Chinese River Dolphin 514 Chinese Swordfish 333 Chionidae 433 Chionodraco myersi 339 chironomid 74 Chironomidae 271 Chiroptera 54, 472, 499

524

INDEX

Chirostylidae 294 chitinozoan 31 Chlamydiaceae 109 Chlamydiae 109, 113 Chlamydoselachidae 321, 325 Chloranthales 161–163, 165 Chlorobi 109 Chlorobiaceae 109 Chloroflexi 107, 112 Chlorophyta 117 chlorophyte 116 Chloropidae 274 chloroplast gene 164, 182 choanoflagellate 117, 215 Choanoflagellida 117 chocolate 200 Choloepus 476 chondrichthyan 310 Chondrichthyes 7, 64, 66, 309–310, 320, 322–323, 326 Chondrophorina 234 Chordata 67–69, 224, 228, 249, 309 chordate 226 chorioallantoic placenta 466 choripetalous corolla 182 Chromadoria 248 chromadorid clades 249 Chromalveolata 117 Chromalveolate 118 Chromatiaceae 108, 112 chronogram 3 chronostratigraphic correlation 27 chronostratigraphy 31 Chrysididae 265 Chrysidoidea 264–265, 267 Chrysobalanaceae 189, 194 Chrysochloridae 56, 479–480, 496 Chrysomelidae 260, 278, 281, 287 Chrysomeloidea 281, 287 Chrysopidae 291 Chtenopterygidae 243 Chthamaloidea 299 chytrid 218 Chytridiomycota 217 Cicindelidae 278 Ciconiidae 437 ciconiiform 441 Ciconiiformes 419, 420, 421, 437 cidaroids 302, 304 Ciidae 279 ciliate 117 ciliated epithelia 247 Cinclidae 425, 429 Cingulata 476, 477

Cinnamomum 166 cinnamon 166 Circaeasteraceae 169, 171 circumscissile sporophyte dehiscence 143 Cirrata 243, 245 Cirripeda 252 Cirroctopodidae 245 Cirroteuthidae 245 Cissampelos 169 Cistaceae 198 Cistecephalus 60 Cisticolidae 424 Citrus 200 civet 504 cladistics 20 cladogenetic topology 22 cladogram 19–20, 22 Clambidae 283, 286 Claudiosaurus 60 Cleridae 280 Cleroidea 280, 287 Clethraceae 178, 182 click beetle 278, 286 Climacteridae 426 climber 188, 194 climbing gouramy 344 clitellate 71 cloaca 462 close-deye squid 242 Clostridiaceae 107 clove 194 clown beetle 286 clownfish 309 club moss 133 Clupeiformes 336 Clupeocephala 64 clupeocephalan 336 clupeomorph teleost 335 Clusiaceae 189, 194 Clypeasterina 305 Clypeasteroidea 303, 305 Cnemophilidae 426 cnidae 233 Cnidaria 7, 69, 75–76, 223–224, 226, 237 cnidocyte 233 coati 504 Cocatherium 58 Coccidia 117 Coccinellidae 280 Coccolithales 124 coccolithophorid 119 Cocculus 169 cocoa butter 200

cocoa plum family 194 codon 21 coelacanth 309–310, 312–313, 348 Coelacanthimorpha 309 Coelenterata 223 coelenteron 233 coelom 247 Coelomata 226 Coelopidae 274 Coelostegus 59 Coelurosauravus 60 Coffea arabica 177 coffee 177 COI 239–240, 244–245, 303 COII 303 COIII 303 cola 200 Colchiacaceae 205 Colchicaceae 205 colchicine 205 Coleoidea 242 Coleoptera 7, 73, 260–261, 278, 283–284, 287 Coleopteroidea 74 Coliiformes 420 collagenous cuticle 247 colleter 182 Colletidae 265, 267 Colluricinclidae 426 colonial anemone 235 colony 233 colony defense 268 colpi 169 Coluber 394 Colubridae 383, 390–391, 394 Columba livia 409 Columbiformes 420, 433 Columelliaceae 180, 183 Colwelliaceae 108 COM clade 188, 193 Comamonadaceae 108 Combretaceae 190, 195 Commelinaceae 204, 209 Commelinale 204–206, 210 commelinid 204, 206, 208 complete metamorphosis 260 complex stony coral 236 complex thalloid 146 condiments 189 conduplicate carpel 162 condylarth 57, 508 conifer 133, 135, 158, 159 Coniferophyta 157 Coniophis 394

Index

Coniopterygidae 291 Connaraceae 194 conodont 31, 67, 310, 312 Conopophagidae 426 continuous tectum 166 contraceptive 169 control region sequence 509 Convallaria majalis 203 convergence 255 Convolvulaceae 177, 179, 183 coot 440 Cophylidae 358 Coprinopsis 215 coprolite 248 coraciiform 446 Coraciiformes 420, 423, 430, 448 coral 24, 76, 233, 237 coral bleaching 236 Corallimorpharia 235 corallimorpharian 236 Corallus 394 corbiculate bee 264, 268 Corcoracidae 425, 429 Cordaitales 159 Cordylidae 384, 387 Cordyloidea 388 Cordylomorpha 384, 387 core angiosperm 161, 162 core campanulid 185 core Caryophyllales 174 core eudicot 169 core lamiid 186 core leptosporangiate 155 core monocot 208 Corethrellidae 271 C-org curve 34 coriander 177 Coriandrum sativum 177 Coriariaceae 190, 193 Corixidae 73 cormorant 436 corn 203 Corn Smut 216 Cornaceae 174, 177–178, 180, 182, 185 Cornales 162, 164, 177–179, 183, 186 cornea 244 Cornus 177 corolla 162, 183 Coronatae 236 Coronaves 420–421, 441 Coronuloidea 299 Corosaurus 60 Corsiaceae 206 Corvida 425

Corvidae 425 corvidan 425 Corvines 425 Corvini 429 Corvoidea 425–426, 431 Corylophidae 280 Corylus 193 Corymorphidae 237 Corynebacteriaceae 107 Corynocarpaceae 190, 193 Coscinodiscophyceae 128 Costaceae 204, 211 Costata 359 cotinga 424 Cotingidae 426 cotton 200 cotyledon 162, 203 Couloux, Arnaud 390 courser 432 cow 44, 53–58, 511 Coxiellaceae 108 CPSase 274 crab 251 Crab Plover 432 crabronid wasp 267 Crabronidae 265, 267 Cracidae 417 Craciformes 416 Cracraft, Joel 423 Cracticidae 425 crake 440 Crandall, Keith A. 293, 298 crane 420, 443 craniate 312 cranium 309 Craseonycteridae 499–500, 502 crassigyrinid 62 Crassula 175 Crassulaceae 170, 175 Craugastoridae 359 creeper 429 Crenarchaeota 91, 94–95, 102 crenarchaeotans 96 Crenotrichaceae 109 Creodonta 54 creosote bush 189 Cretaceous-Paleogene mass extinction 433 Cretotrigona prisca 268 Cricetidae 491 Crocidurinae 498 crocodile 44, 60–61, 375, 402 Crocodylia 7, 9, 60, 309, 375–378, 405 Crocodylidae 405

crocodyliform 403 crocodyline 403 Crocodyloidea 403 Crocodylus 402–403, 405 cross validation 38, 43 Crossosomataceae 198, 201 Crossosomatales 188, 197–198, 201 crow 429, 431 crown group 38 crown jellyfish 236 Crowsoniella 281 crurotarsan 61 Crurotarsi 61 Crustacea 251 crustacean 71 Crustaceomorpha 71 Cryphaeaceae 139 Cryptelytrops albolabris 390 Crypteroniaceae 195 Cryptobranchidae 366 Cryptobranchoidea 366 Cryptodira 398, 400 Cryptodires 398 Cryptophagidae 280 Cryptoprocta ferox 505 ctenglossan 246 Ctenizidae 256, 258 Ctenodactylidae 491 Ctenoglossa 243, 245 Ctenohystrica 52 Ctenomyidae 491 Ctenophora 223–224, 226 Ctenopilus elongatus 73 Ctenopominae 346 ctenosauriscid 61 Cubozoa 75, 234, 236 cuckoo 419 Cucujidae 280 Cucujiformia 288 Cucujoidea 280–281, 287 cuculiform 420 Cuculiformes 420 cucumber 193 Cucurbitaceae 190, 193 Cucurbitales 164, 188–190, 193 Culicidae 271 Culicomorpha 74–75, 271, 276 Cunoniaceae 189, 194 Cupedidae 284, 288 Cupressaceae 159 curassow 415 Curculionidae 281 Curculionoidea 281, 288

525

526

INDEX

currant 175 Curtisiaceae 180 custard-apple 166 cuticule 133 cuttlefish 244 Cyanidiales 117 Cyanobacteria 10, 91, 93–94, 107, 112–113, 116 Cycadaceae 159 Cycadophyta 157 cycadophyte 159 cycad 133, 159 Cyclanthaceae 206 cycle stratigraphy 30 Cyclocheilaceae 183 Cyclopedidae 477 Cyclopes didactylus 477 Cyclorrhapha 275 cyclorrhaphan fly 262 Cyclostomata 66, 309–312, 317–318, 325 Cyclura 393 Cygnodraco mawsoni 341 Cygnus olor 415 Cylindrophiidae 391 Cymodoceaceae 206 Cynara cardunculus 177 Cynocephalidae 495 Cynoidea 506 Cynosurus cristatus 203 Cyperaceae 203–204, 208, 211 Cypriniformes 336 cypriniform 337 Cypripedium calceolus 203 Cyrillaceae 178, 182 Cyrtaucheniidae 256, 258 Cysticyathus tunicatus 76 Cystodium 155 Cytinaceae 200 cytochrome b 332 cytochrome c 19, 249 cytochrome oxidase I 245, 286 D daddy long legs 255 Daedalornithes 455 Dahlia 178 Dalatiidae 321 dance fly 272 Danforth, Bryan N. 264 Danio 325, 349 Danio rerio 64 Daouitherium 56 Daphnia 45, 71 Daphniphyllaceae 170

Daphoenosittidae 426 Daphoenus 55 Dascillidae 282 Dascilloidea 282, 287 Dasyatidae 322, 325 Dasyatis 322 Dasypodaidae 265 Dasypodidae 476 Dasypogonaceae 204–205, 208, 210 Dasyproctidae 491 Dasyuridae 466, 467 Dasyuromorphia 466–467, 469 dasyuromorphian 467 Datiscaceae 190, 192 Daubentoniidae 480, 482, 484 Daucus carota 177 Decapodiformes 7, 245 decapodiform 244 deciduous tree 171 Decipomys 51 Decliniidae 283, 286 decomposers 218, 270 Decoredon 50 deep sea record 30 deep time 26, 28 deep-sea finned octopus 245 Degeneriaceae 167 Dehalococcoides 107 dehiscence mechanism 138 Deinococcaceae 107 Deinococcus 107 Deinopidae 256 Deinopoidea 256–257, 259 Delphinidae 512 Delphinoidea 512 Delsuc, Frédéric 475 Deltaproteobacteria 109 DELTRAN 164, 168 demosponge 76 Demospongiae 224 Dendroaspis 393 Dendrobatidae 359 Dendrobranchiata 294 Dendrocolaptidae 426 Dendrocygnidae 417 Dennstaedtiaceae 154, 155 dennstaedtioid 154, 156 Dentaliida 241 Dentaliidae 241 Dentalium 240 Dentalium acutoides 240 Dermatemydidae 399 Dermestidae 282 Dermochelyidae 399

Dermoptera 49–51, 472, 496 Dermotherium 50 Derodontidae 283, 286 Derodontoidea 283, 288 desclerotization 275 desert date 191 Desmatolagus 51 Desmostylia 56 Desulfovibrionaceae 109 Deuterophlebiidae 271, 276 Deuterostomia 6, 46, 68–69, 224, 226–228, 302 dewlap 415 Diabolepis 63 Diacodexis 55 Diadema 304 Diadematidae 303 diadematoid 304 Diadocidiidae 272 Diadophis 393 Dialipina markae 63 Diapensiaceae 182 Diapsida 37, 59–60, 376 Diarchineura 275 diatom 34, 119 Diatomyidae 491 Diatriata 370 Dibamidae 383–384, 386, 388 Dicaeidae 424 Dicamptodontidae 366 Dicentra 169 Dichapetalaceae 189 Dicranaceae 140 Dicranales 140 Dicranidae 138, 143 Dicroglossidae 358 Dicruridae 425 dictyopteran 72 Dictyostelium 95 Dicynodon 60 Didelphidae 57, 467 Didelphimorphia 466–467, 469 Didiereaceae 174 Didymelaceae 169, 171, 173 Diervilleaceae 184 digger wasp 264 Diguetidae 256 Dikarya 217 Dilaridae 291 Dilleniaceae 171, 174 Dilleniales 162 Dilleniidae 188 Dinilysia 394 dinoflagellate 34

Index

Dinomyidae 491 Dinomyidea 493 Dinophyceae 117 Dinornithidae 413 Dinornithiformes 413 dinosaur 51, 409, 473 dioecious shrub 182 dioecious tree 173 Diogo, Rui 332, 335 Dioncophyllaceae 170, 174 Diopsidae 274 Dioscoreaceae 206 Dioscoreales 205–206, 210 Dioscoreophyllum 169 Dipentodontaceae 197, 201 Diphysciaceae 140 Diphysciales 140 Diplodactylidae 384, 387 diplogasterid 248 Diplogasteromorpha 248 Diploglossidae 384, 386 Diploglossinae 386 diploid embryo 138 Diplolepidae 138 diplolepidous-alternate moss 138, 142, 144 Diplomonadida 117 diplomonad 93, 118 Diplopoda 252 Dipluridae 256, 258 Dipnoi 7, 309–311, 349 Dipodidae 491, 493 dipper 429 Diprotodontia 466, 467, 469 diprotodontian 468 Dipsacaceae 184 Dipsacales 178, 180, 183, 186 Dipsadidae 391, 394 Diptera 74–75, 260–262, 270, 275 dipteran 262 Dipteridaceae 155 Dipterocarpaceae 198, 200 Dirachmaceae 193 Disceliaceae 142 Discicristates 117 Discoglossidae 359 Discolomatidae 280 dispersal-vicariance analysis 211 dispersed pollen 184 Dissostichus 341 Disteniidae 281 distichous phyllotaxis 167 Ditomyiidae 272

Ditrichaceae 140 diurnal bird of prey 437 diving beetle 286 Dixidae 271 DNA-DNA hybridization methods 4 dobsonfly 260 Docodonta 58 docodont 58 dog 44, 53–58, 504 Dolichoderinae 266 Dolichopodidae 272 Doliodus problematicus 66 dolphin 44, 58 Donatiaceae 184 Donoghue, Philip C. J. 35 Doronicum 178 dorsal ocelli 287 Doryanthaceae 205 Dorylaimia 248 dorylomorph 267 Douady, Christophe J. 495 Douzery, Emmanuel J. P. 475, 495 Dracunculus muscivorus 203 dragonfish 340 Drilidae 282 Dromiciops 57, 466 Dromiciops gliroides 466 Droseraceae 170, 174 Drosophila 75, 226, 274 Drosophila melanogaster 74–75, 273 Drosophilidae 272, 275 Drosophyllaceae 174 Dryinidae 265 Dryolestoidea 58 Dryopidae 282 Dryopithecus 48 duck 61, 416 Duckbilled Platypus 462 duck-like bird 423 duckweed 203 Dwarf Sperm Whale 512 Dyscophidae 358 Dytiscidae 284 E E23r 74 eagle 436 eagle ray 324 Earth geological history 26 Ebenaceae 178, 182 ebenoid 182 Ecdeiocoleaceae 204, 209 ecdysis 224 Ecdysozoa 69, 224, 226, 228, 247

527

echidna 459, 463 Echimyidae 491 Echinacea 303 Echinerpeton 59 Echinidae 303 Echinocyamidae 303 Echinodermata 7, 67–69, 224, 226, 228, 249, 302, 325 Echinodiaceae 143 Echinoidea 303 echinoid 304 Echinolampadidae 303 Echinometridae 303 Echinoneidae 303 Echinoneus 304 Echinorhinidae 321, 325 Echinorhiniformes 322 Echinothuriidae 303 echinothurioid 304 echolocating microbat 499, 502 Ectocion 54 ectoderm 233 ectoparasite 274 Ecydsozoa 224 Ediacaran 69, 75 Edwards, Scott V. 375 eel 335 eelworm 247 EF-1y 259 Egyptian Plover 432 Ehretiaceae 182 Eizirik, Eduardo 471, 504 Elaeagnaceae 190, 193 Elaeocarpaceae 189 Elapidae 390–391, 393, 395 Elapsoidea 393 Elasmobranchii 320, 322, 326 Elateridae 282 Elateriformia 287 Elateroidea 282, 287 Eleaocarpaceae 194 electric ray 324 electroreception 462 electroreceptive bill 463 elephant 44, 56–58, 459, 471 elephant bird 412 Elephant Shark 44 elephant shrew 56, 471, 496 Eleutherodactylidae 359 Eleutherodactylus portoricensis 309 Ellerbeckiaceae 128 elm 193 Elmidae 282 elongation factor-1 gamma 258

528

INDEX

Elopiformes 330, 336 Elopocephala 336 elopocephalan 336 Elopomorpha 336 elopomorph 336 elytra 278, 281 Emballonuridae 502 Emballonuroidea 500 emberizid 429 Emberizidae 424 emberizine 431 Embrithopoda 57 embryo 161 embryo sac 162 Embryophyta 117, 133, 137, 150 embryophyte 136 Empididae 272 Empidoidea 272, 275 Emu 44, 60–61, 413 Emydidae 400 enantiornithine 409 Encalyptaceae 140, 142 Encalyptales 140 Encephalitozoon 93 endoderm 233 Endomychidae 280 Endopterygota 260 endosperm 162 endostome 141 endosymbiont 91 Enoplia 247 Entalimorpha 239 Entalina curvum 240 Entalinidae 241 Entelegynae 256, 258 entelegyne node 259 Enterobacteriaceae 108 Entodontaceae 139 Entomophthoromycotina 217 Entomoplasmataceae 107 Eoblattina complexa 72 Eodendrogale 50 Eogastropoda 70 Eolepadidae 299 Eomaia 50–51, 53, 58 Eomaia scansoria 57 Eopolytrichum 143 Eoredlichia wutingaspis 68 Eosimias 484 Eothyrididae 59 eotomarioid 70 Eotragus 56 Eotragus noyi 56 Ephedra 157, 159

Ephedraceae 159 ephemeris second 26 Ephydridae 274 epipetalous stamen 182 epiphyllous inflorescence 183 epiphyte 144, 146, 149, 206, 208 Episquamata 386 episquamatan 388 Epistylis 116 Epitheria hypothesis 472 Epsilonproteobacteria 109 Eptatretus burgeri 317 Equidae 509 Equisetaceae 154 Equisetum 153 Equus 509 Equus grevyi 508 Eremoneura 275 Eremosynaceae 180, 183 Erethizontidae 491 Erica 177 Erica vestita 177 Ericaceae 177, 182 Ericales 162, 164, 177–179, 182, 186 ericoid 182 Erinaceidae 479, 496 erinaceomorph 496 Erinaceomorpha 53, 496 Erinaceus europaeus 496 Erinacidae 496 Eriocaulaceae 204 Erotylidae 280 erycine 394 Erythrobacteraceae 109 Erythrosuchidae 61 Erythroxylaceae 189 Eryx 393 Escallonia 178 Escalloniaceae 180, 183 Eschrichtiidae 512 Esconites zelus 71 Esocidae 344 Esociformes 344 EST data 249 ethereal oil 203 Etmopteridae 321 Euarchonta 50 Euarchontoglires 50–51, 53, 471–472, 488 euasterid 177–179, 182, 186 Eubacteria 6–8, 13, 89–96, 114 Eublepharidae 384, 388 Eublepharoidea 387 Eucalyptus 194

eucalyptus family 194 Eucidaris 304 Eucinetidae 283 Eucnemidae 282 Eucommiaceae 182 Eucritta melanolimnetes 62 Eudicots 161–166, 169–170, 173, 175 Eudocimus albus 309 Euechinoidea 302 Euglenida 117 Eukaryota 7–8, 10, 13, 16, 89–96, 112, 114, 116–120, 215, 218 Eulichadidae 282 Eulipotyphla 471–473, 498 eulipotyphlan 53 eumagnoliid 208 Eumetabola 72, 74 Eumetazoa 75–76, 223 eumetazoan 226 eunicid 71 Eunotiophycidae 128 Euparkeriidae 61 Euphorbiaceae 189, 194 euphyllophyte 133–134, 137 euphyll 153 Eupleridae 506 Eupolypods 156 Eupomatiaceae 167 Euprimates 50 Eupteleaceae 171 Eureptilia 59 European Hedgehog 496 European Mole 496 European shrew 53, 58 Eurosid I 188–190, 192, 194, 197 Eurosid II 188, 194, 197, 201 Eurosids 170, 190 Eurotiomycetes 216 Eurya 185 Eurya crassitesta 185 Euryarchaeota 91, 102 euryarchaeotan 95 euryhaline shark 320 Eurylaimidae 426 Eurylaimides 424, 426 eurymylids 52 Eurypygae 442 Eurypygidae 441 eusporangiate 134 euteleost 65 Euteleostei 64–65, 336 Eutheria 45, 50, 53, 57, 459–460, 473 Euthyneura 70 evening primrose family 194

Index

even-toed ungulate 54, 513 evergreen tree 171, 173, 189, 194 Evocoidae 274 Excavata 117 excavates 117 Exiliboa 393 exostome 140 extrorse anther 203 F Fabaceae 188, 190, 192 Fabales 188–190, 192 fabid 162–164, 188 Fabroniaceae 139 Fagaceae 190, 193 Fagales 164, 188–190, 192 Fagus 193 falcon 436 Falconidae 439 falconiform 439 Falconiformes 420, 436–439, 451 Falconimorphae 437, 452 Falcunculidae 426 False Vampire Bat 497 Farrell, Brian D. 278 Felidae 506 feliform 55, 507 Feliformia 14, 55, 507 Fenhosuchus 61 fennel 177 fern 133–134, 153, 156 Ferungulata 54 fig 193 fighting fishe 344 Filifera 237 Filistatidae 256 fi lmy fern 155 finch 431 finfoot 443 firefly 278 Firmicutes 91, 107, 113 fish 23–24, 40, 66–68, 119, 237, 309–310, 313 Fissidentaceae 140 Fissidentalium 240 Fissidentalium pukaea 240 Flabelligeridae 71 Flacourtiaceae 201 Flagellariaceae 204, 209 flagellated spore 218 flamingo 421 flat bark beetle 286 flatworm 224 flea 260

flesh-eating placental mammal 54 fleshy fruit 183 fleshy receptacle 167 flightless bird 61 floral oil 267 floral primordia 164 flower beetle 286 flowering plant 161, 163, 177, 188 flowerpecker 429 flower 161, 179, 184 flufftail 440–441, 443 fly 274 flycatcher 424 flying lemur 50, 471, 495 Foeniculum vulgare 177 Folivora 477 Fontinalaceae 139 Foraminifera 117 foraminifer 34 Forest, Félix 166, 169, 188, 197 forget-me-not 177 Formicariidae 426 Formicidae 264–265, 267 Formicinae 266 formicoid clade 266 Fossa 504, 506 fossil events 31 fossil units 27 Fossombroniaceae 148 fossoriality 476 Fouquieriaceae 178, 182 Four-horned Antelope 56 fox 504 Fragillariophycidae 128 Franciscana Dolphin 512 Francisellaceae 108 Francoaceae 202 Frankeniaceae 170, 174 Frankia 193 Frankiaceae 107, 192 free petal 179, 183 freshwater snail 70 freshwater stingray 320 frigatebird 436 Fringillidae 424 frog 62, 309, 312 fruit 166, 170, 174–175, 184, 191, 194 fruitfly 45, 69, 71–74, 273 Frullaniaceae 147, 150 Fumariaceae 169 Funariaceae 140, 142 Funariales 140, 142 Fungi 6, 16, 116–119, 215–218, 223 fungivory 288

fur seal 504 Furipteridae 500 Furnarii 424 Furnariidae 426 Furnariides 424, 426, 430 Fusobacteria 109, 113 Fusobacteriaceae 109 Fustiaria 240 Fustiaria glabellum 240 Fustiariidae 241 fusulinids 31 Futaba Group 185 G Gadilida 241 Gadilidae 240 Gadilimorpha 240 Gadilinae 241 Gadilinidae 239 Galagonidae 480, 484 galago 50 Galatheidae 294 Galbulae 445–446, 450 Galbulidae 446, 448 Galbuliformes 446 Galea 322 Galeomorphii 321–322, 325 galeomorph 324 Gallia alsatica 74 Galliformes 61, 409, 418 gallinule 440 Galloanserae 61–62, 409–410, 415–417, 421, 423 Gallus 349 game bird 61, 415, 418 gametangia 138 gametophyte 133–134, 138, 141, 146 gametophytic leaves 134 gametophytic phase 137 Gammaproteobacteria 108, 113 gannet 436 garfish 335 Gargantuavis 61 Gargantuavis philoinos 61 garlic 206 Garryaceae 179, 182 Garryales 178, 185 gar 309, 328 Gasterorhamphosus zuppichini 65 Gasterosteiformes 65 Gastrophrynidae 358 Gastropoda 70 Gastrotricha 224 Gatesy, John 511

529

530

INDEX

Gavialidae 405 Gavialis 405 Gavialis gangeticus 402 gavialoid 405 Gaviiformes 61, 421 Geadephaga 286 gecko 386, 388 Geissolomataceae 188, 201 Gekkomorpha 384, 387 gekkomorph 388 Gekkonidae 384, 387 Gekkonoidea 387 Gekkota 384, 386, 388 Gelsemiaceae 179, 182, 185 genet 504 Genocidaridae 303 Gentianaceae 179, 182, 185 Gentianales 178–179, 182–183, 185 Gentrytragus 56 Geobacteraceae 109 geochemical curve 31 geochronologic correlation 27 Geoemydidae 399, 400 geologic clock 4 Geologic Time Scale 26, 29 Geologic Time Scale 2004 34 Geologic Time Scale 2010 34 Geomyidae 492 Geraniaceae 198, 202 Geraniales 188, 194, 197–198, 202 Geranium 202 Gerbera 178 Gerrardina 201 Gerrardinaceae 197, 201 Gerrhonotinae 386 Gerrhosauridae 384, 387 Gesneriaceae 179, 183, 185 Gettysburg 27 gharial 402 Giant Anteater 476 giant kelp 117 Giant Panda 505 Giant Sperm Whale 512 giant squid 244 Giardia 93 Giardia lamblia 92 gibbon 44, 48–49, 51, 58 Gigantopithecus 48 Gigaspermaceae 142 gill slits 226 ginger 208 Ginglymostomidae 321, 326 Ginkgo 158 Ginkgo biloba 157

Ginkgoaceae 158 Ginkgophyta 157 ginkgo 133 giraffe 511 Giraffidae 511–512, 514 Gisekiaceae 174 gladius 244 glandular hair 203 Glaphyridae 283 Glareolidae 433 Glaresidae 283 glauconite 30 Glaucophyta 117 glaucophyte algae 117 Gleicheniaceae 155 gleichenioids 155 Glires 45, 50–52, 488 Gliridae 491 Global Stratotype Sections and Points 27 Globigerina tapuriensis 66 globin 249 Globulariaceae 183 Glochiceras lithographicum 64 Gloeobacteraceae 107 Glomeromycota 149, 217 Glossinidae 274 glucosinolate 197 gnatcatcher 429 Gnathostomata 66–67, 309–312, 318 Gnathostomulida 224 gnepine 135 Gnetaceae 159 gnetifer 135 Gnetophyta 133–135, 159 gnetophyte 157 Gnetum 158 Gobiolagus 51 Gobionotothen acuta 341 Gobionotothen gibberifrons 341 Goebeliellaceae 147, 150 golden mole 56, 471, 479–481, 496 golden plover 432 golden spike 28 Gomortegaceae 167 Gongylophis 393 Gonorhynchiformes 336 Goodeniaceae 180, 184 goose 410, 416 gooseberry 175 gooseberry family 175 gorilla 48 Gossypium 200 Goupiaceae 189

gouramy 344 Gower, David J. 369 Gracilitarsidae 450 Gradstein, Felix M. 26 Grallinidae 425, 429 graminid 155, 211 grapefruit 200 grapevine 188 Grapsidae 294 graptolite 28, 31 graptolite fossil 27 graptolite zone 31 grass 203, 209 grass-tree 206 Grauvogelia 275 Grauvogelia arzvilleriana 75 grauvogeliid 74 Gravesia 64 Gray Plover 432 Gray Whale 512 Grayia 393 great ape 48 Great Blue Heron 419 Great Grey Owl 451 Great Oxidation Event 13 Great White Shark 320 Greater Antilles 22 grebe 409, 421 green algae 117, 119, 141 green pepper 177 Grès à Meules 74 Grès-a-Voltzia Formation 74 Grévys Zebra 509 Greyiaceae 202 Grimmiaceae 140, 143 Grimmiales 140 Grimpoteuthidae 245 Griseliniaceae 180, 183, 185 Grossulariaceae 170, 175 ground beetle 278 ground sloth 476 grouse 417 Grubbiaceae 178, 180 Grues 443 Gruidae 442 Gruiformes 420, 432, 437, 442 Gruoidea 440, 441, 443 GSSP 28 GTS2004 31 Guaiacum 191 Guamatela 201 Guamatelaceae 201 guan 415 guava 194

Index

guinea pig 44, 51–52, 56–58, 490, 493 guineafowl 415 guitarfish 320, 324 gull 432 Gunneraceae 171, 173 Gunnerales 162–163, 173 Gymnobucco 447 Gymnodraco acuticeps 341 Gymnomitriaceae 147 Gymnophiona 354 Gymnophthalmidae 384, 386, 388 gymnosperm 7, 133–136, 158 Gymnotiformes 336 Gymnuridae 322, 324 gynoecium 167 Gyrinidae 284 H Hackett, Jeremiah D. 116 Haematopodidae 433 Haemodoraceae 204, 209 hagfish 66, 309, 318 Hahellaceae 108 Haikouella 68 Haikouichthys 68 hair 459 Halictidae 264–265, 268 Haliplidae 284 hallux 416 Halobacteria 102 Halobacteriaceae 102 Halophytaceae 174 Haloragaceae 170, 175 halteres 262, 270 Halteria 262 Hamamelidaceae 170, 175 Hamamelidae 188 hammerhead shark 324 Hanguanaceae 204 Haplogynae 256, 259 Haplolepidae 138, 142, 143 Haplomitriales 148 Haplomitriopsida 146–148, 150 Haplomitrium 148 haplorrhine 485 Haplorrhini 480, 482 Hapsidophrys 393 Haptophyta 117 haptophyte 118 hard minimum constraint 40 hare 51, 488 Harpagiferidae 340 Harpoceras falciferum 74 harvestmen 255

hawk 436 Hayashi, Cheryl Y. 255 hazel 193 He, Shunping 332, 335 head capsule 281 head sac 273 heath 177 hectocotylus 245 hedgehog 44, 53, 56–58, 479, 496 Hedges, S. Blair 3, 89, 101, 106, 116, 309, 320, 348, 383, 390 Hedwigiaceae 140, 142 Hedwigiales 140, 142 Heinicke, Matthew P. 320, 348 Helagras 394 Heleophrynidae 359 Helianthus annuus 177 Helichrysum 178 Helicobacteraceae 109 Heliconiaceae 204 Heligmonellidae 248, 249 Heliophila juncea 197 Heliornithidae 442 helmet shrike 429 Helobdella 71 Heloderma 393, 395 Helodermatidae 384, 386, 393 Helodermatoidea 386 Helostoma temminkii 345 Helostomatidae 346 Helotidae 280 Helwingiaceae 180, 183 hematophagy 276 Hemerobiidae 291 Hemerobius 287 Hemiascomycetes 216 Hemicentetes semispinosus 479 Hemichordata 224, 226, 228 hemichordate 68 Hemigaleidae 321, 325 hemiparasite 173 Hemiprocnidae 455 Hemiptera 73 Hemiscylliidae 321 Hemisotidae 358 hemocyanin 259 Henophidia 391 henophidian 395 He-Nygrén, Xiaolan 146 Heomys 52 herb 166, 173–174, 191, 194, 199–200, 211 Herbertaceae 147 herbivorous insect 23 Hercynian Question 27

Hernandiaceae 167 Herpestidae 506 Herpetotherium 57 Hesperis 198 Hesperocyon 55 Hesperornis 62 Hesperornithiformes 62 Heteralepadidae 299 Heteralepadomorpha 299 Heteroceridae 282 Heterodontidae 321, 325 Heterodontiformes 321–322, 324 Heteromyidae 492 Heteroptera 73 heteropteran 73 Heteropyxidaceae 194 heterosporous fern 155 heterospory 137 Hevea brasiliensis 194 Hexacorallia 235 hexactinellid 76 Hexactinellida 224 Hexanchidae 321, 325 Hexanchiformes 322 Hexapoda 252 Hexathelidae 256 hexathelid 258 Hexatrygonidae 322 Hibiscus 200 hickory 193 higher fungi 215 higher Nematocera 271 Hilu, Khidir W. 133 Himalayacetus 55 Himantandraceae 167 Himantura 322 Hippidae 294 Hippoboscidae 274 Hippoboscoidea 274 Hippocastanaceae 200 Hippolytidae 294 Hippomorpha 509 Hippopotamidae 55, 512, 514 hippopotamus 471 hipposiderid 499 hipposiderinid 500 Hipposideros larvatus 499 Hironoia fusiformis 185 Hirundinidae 424, 454 Hispinae 287 Histeridae 283 Histeroidea 286 HKY85 model 164 Hoatzin 420

531

532

INDEX

Høeg, Jens T. 298 holasteroid 302, 304 holly 178 holocephalan 325 Holocephali 320, 322, 324, 326 Holometabola 72–73, 260, 262 holometabolous insect 270 Holostei 329 Homalopsidae 391, 394 Homalopsis 393 Hominidae 480, 483, 485 hominin 46, 48 hominoid slowdown 483 Hominoidea 46, 48–49, 480, 485 Homo 325, 349 Homo erectus 46 Homo sapiens 46 Homo sapiens neanderthalensis 46 homocostate pleurocarp 144 homopteran insect secretion 276 homoscleromorph 76 Honey Possum 469 Honeybee 45, 74 Honeycutt, Rodney L. 490 honeyguide 445 Hookeriaceae 139 Hookeriales 139, 142, 144 hop family 193 Hoplophrynidae 358 horizontal linear axis 29 horn shark 324 hornebeam 193 hornets 264 hornwort 133–137, 142, 150 horse 19, 44, 53–54, 56–58, 510 horse chestnut 200 horseshoe crab 251–252, 258 horsetail 133–134, 153–155 host phylogeny 249 Houde, Peter 440 Housefly 273 Hox genes 224, 252 Huaceae 192 Huertea 201 Huerteales 197, 201 human 44, 46, 48–49, 51, 53, 56–58, 60, 62, 69, 459, 482 Humiriaceae 189 hummingbird 419–421, 442, 448, 456 Huso 333 Hyaenidae 506 Hybonoticeras hybonotum 64 Hybosoridae 283 Hydatellaceae 208

Hydra 45, 75 Hydradephaga 288 Hydraenidae 283 Hydrangeaceae 180, 185 Hydridae 237 Hydrobacteria 108–109, 113 Hydrochaeris hydrochaeris 490 Hydrocharitaceae 204, 206 hydrogen sulfide 235 Hydrogenophilaceae 108 hydrogenosome 92 Hydroidolina 234, 237 hydroid 233, 237 Hydroleaceae 183 hydrophiine 394 Hydrophilidae 283 Hydrophiloidea 283, 286 Hydrophyllaceae 182 Hydroscaphidae 278, 284 Hydrostachyaceae 180, 182, 185 Hydrostachys 182 hydrothermal vent 236 hydroxyapetite 310 Hydrozoa 75, 237 hyena 55, 504 Hygrobiidae 284 Hylidae 359 Hylobatidae 48, 480, 483, 485 Hylocichla mustelina 423 Hylocomiaceae 139 Hylonomus 59 Hymenophyllaceae 154 Hymenophytaceae 148, 150 Hymenoptera 73–74, 260–262, 264, 266 Hymenostomatida 117 Hynobiidae 366 Hypericaceae 189, 194 Hyperoliidae 358 Hyperoodontidae 512 hyperthermophile 96, 113 hyphal clamp 215 Hypnaceae 139 Hypnales 139, 142, 144 Hypnidae 142 Hypnodendraceae 139 Hypnodendrales 139, 142 Hypnosqualea hypothesis 322, 324 Hypochilidae 256, 258 Hypodematium 155 Hypopterygiaceae 139, 144 hypostase 162 Hypoxidaceae 205, 209 Hypsiprimnodon moschatus 468 Hypsiprymnodontidae 467

Hyrachyus 508 hyracoid 471 Hyracoidea 472 Hyracotherium 508 hyrax 56, 496, 508 Hystricidae 491, 493 Hystricognathi 52, 490, 493 hystricognathous 490 Hystricomorpha 491, 493 I Ibidorhynchidae 433 Ibisbill 433 Iblidae 299 Iblomorpha 299 Icacinaceae 179, 182–183, 185 Icacinicarya budvarensis 185 Icaronycteris 54 Icaronycteris index 499 Ichthyophiidae 370 Ichthyornis 62 Ichthyornithiformes 62 Ichthyosauria 60 Icteridae 424, 429 icthyosporean 215, 223 Idiomarinaceae 108 Idiopidae 256, 259 Idiosepiidae 244 Iguania 383–384, 386–388, 395 Iguanidae 383–384, 386–387, 393 Iguanoidea 388 Ilex 178, 185 Ilex antique 185 Incirrata 243, 245 India-Madagascar plate 413 Indian River Dolphin 514 Indicatoridae 445–446, 448 Indonotothen 341 Indridae 480 infraorbital foramen 490 Iniidae 512, 514 Insectivora 53, 471, 479, 495 insect-pollinated flower 178 insect 23–24, 72, 75, 174, 251, 260, 262 Inshore Hagfish 317 integument 161, 182 intercalary meristem 134 interlaboratory recalibration 30 International Commission on Stratigraphy 27 International Stratigraphic Guide 27 intracellular algae 233 Introverta 226 invertebrate 6, 8, 16, 44, 312

Index

Ipomoea batatas 177 Irenidae 424 Iridaceae 205–206, 208 iridoid 178 iris 206 Irregularia 304 Irvingiaceae 189 Ischyromidae 52 Isochrysidales 124 isotopic decay 4 Iteaceae 170 Ithonidae 291 Ithyceridae 281 Ixerbaceae 188, 197–198, 201 Ixiolirionaceae 205 J jacamar 445–446, 448 jacana 432 Jacanidae 433 jacobid excavate 116 Jakoba ibera 116 Janßen, Thomas 203 Japanese Lamprey 317 Javan Rhinoceros 509 jawless fish 9, 309, 318 jawless vertebrate 46 jaws 309 jay 431 jellyfish 223, 233–234, 236 Joinvilleaceae 204, 209 Joubiniteuthidae 243 Jubulaceae 147, 150 Juglandaceae 190, 193 Juglans 193 Juncaceae 204, 208, 211 Juncaginaceae 206 Jungermanniaceae 147 Jungermanniales 151 Jungermanniidae 148 Jungermanniineae 150 Jungermanniopsida 150 K Kagu 441, 443 Kalanchoe 175 Kalophrynidae 358 Kaltanidae 73 Kambara 403 Kamoyapithecus 49, 485 kangaroo 44, 57–58, 459, 466, 468 Karatoserphus 74 karnal bunt of wheat 216 Karnimata 52

Kentisuchus 403 Kentuckia 330 Keroplatidae 272 Khasia 57 Khaya 200 Kickxellomycotina 217 Kim, Jungwook 260 Kimberella 69, 71 kinetoplastid 93, 95 kinglet 429 Kinorhynca 226 Kinosternidae 399 Kirkiaceae 200 kite 436 kiwi 413 Klausmuelleria salopiensis 71 Klikov sequence 185 Koala 466, 468 Koeberliniaceae 198 Kogiidae 512 Kokopellia 57 Kokopellia juddi 57 Kollikodon 58 Kotatherium 58 Krajewski, Carey W. 462, 466 Krameriaceae 190 Krebsophis thobanus 394 Kumar, Sudhir 3 Kuraku, Shigehiro 317 Kuratani, Shigeru 317 L labrum 281 labyrinth fish 346 Lacertibaenia 386 lacertibaenian 386 lacertid 386, 388 Lacertidae 384 lacertiform 386 lacewing 260 Lacistemataceae 189 Lactobacillaceae 107 Lactuca sativa 177 ladybird beetle 278, 286 Laemophloeidae 280 Laevidentaliidae 239 Laganidae 303 Lagerpeton 61 Lagomorpha 39, 45, 50–51, 471–472, 488 Lagonimico 49 Lambdotherium 55 lamellae 143 Lamiaceae 177, 179, 183

Lamiales 178–179, 182–183, 185 lamiid 162, 164, 177–178, 182, 186 lamina 162 Lamium maculatum 177 Lamnidae 321, 325 Lamniformes 321–322, 324 lamprey 44, 66–68, 309, 318 Lampropeltis triangulum 383 Lamprophiidae 391, 394 Lamprophis 393 Lampyridae 282 Lanariaceae 205, 209 land plant 117, 133, 135–136, 143, 146 landbird 419 landfowl 409, 419 Langiomedusae 237 Laniidae 425, 429 Lanthanotidae 383–384, 386 Lapparentophis defrennei 394 Lardizabalaceae 169, 171 Lari 433 Laridae 433 Larkin, Leah 264 larval exuvium 281 larval head capsule 270 larval mandible 271 laryngeal echolocation 501 lasmobranchs 325 lateral canal 243 Laterata 384, 386, 388 Laticauda 394 Latouchella 69–71, 75 Latridiidae 280 Lauraceae 167 Laurales 166–167, 208 Laurasiatheria 53–54, 471–472, 480, 496, 502, 508 laurasiatherian 53 laurel 166 Laurus 166 Lauxaniidae 274 lavender 177 Lavendula 177 layrngeal echolocation 503 leaf beetle 278, 287 leaf margin 162 leaf morphology 144 leaf venation 167 leafbird 429 leaf-nosed bat 499 leafy liverworts 146, 151 Lecythidaceae 178, 182 Ledocarpaceae 202 leech 45, 70

533

534

INDEX

leek 206 Legionellaceae 108 Leguminosae 192 Leiodidae 283 Leioheterodon 393 Leiopelmatidae 359 Lejeuneaceae 147, 151 Lembophyllaceae 143 Lemna 203 Lemnaceae 203 lemon 200 lemur 44, 49, 51, 56–58, 482, 484 Lemuridae 480 Lemuriformes 480 lemuriforms 49 Lemuroidea 480 Lentibulariaceae 183 Lepadidae 299 Lepadomorpha 299 Lepiceridae 288 Lepicoleaceae 147 Lepidobotryaceae 194 Lepidobotrys 193 Lepidolaenaceae 147, 150 Lepidonotothen 341 Lepidoptera 262 lepidopteran 262 Lepidosauria 376, 378, 383 Lepidosauromorpha 60 lepidosauromorph 60 Lepidosirenidae 350 Lepidosirenoidei 349 Lepidoziaceae 147 Lepisosteiformes 328–329, 335 Leporacanthicus triactis 335 Leporidae 9, 51, 488 Leptanillinae 266 Leptochariidae 321 Leptodactylidae 359 Leptodeira 393 Leptodontaceae 143 Leptolepides 64 Leptomedusae 234 Leptosiropsis torulosa 116 Leptospiraceae 109 leptosporangiate 133–134, 155 Leptothecata 236 Leptotyphlopidae 390–391, 393 Leptotyphlops 393 Lepuropetalon 193 Lepyrodontaceae 139 Lethenteron japonicum 317 Lethiscus 62 Lethiscus stocki 62

lettuce 177 Leucodontaceae 139 Leucodontales 142 Leucomiaceae 139 leukemia 177 liana 166, 173–174, 195, 200 Lias 36 Liasis 393 Lichanura 394 lichen 216, 218 lignin 134, 215 lignitized cell 137 lignum-vitae 191 lilac 177 Liliaceae 205 Liliales 205, 208, 211 lily 203, 205 Lily-of-the-Valley 206 limb 309 Limnanthaceae 198, 200 Limnichidae 282 Limnocharitaceae 206 Limnodynastidae 359 Limnomedusae 237 limpet 45, 69 Limpkin 443 Linaceae 189 Lindgren, Annie 242 Lindsaeoids 154, 156 linear interpolation 31 linear time unit 27 Linnaeaceae 184 Lion 504 Liotyphlops 393 Lipotidae 512, 514 Lipotyphla 53–54, 479, 496 lipotyphlan 53 Liriodendron tulipifera 161 Lissamphibia 62, 309–310, 312 Listeriaceae 107 litchi 200 Lithodidae 294 lithornid 413 lithornithid 61 lithostratigraphic correlation 26 Lithotryidae 299 liverwort 133–137, 142–143, 146–148, 151 lizard 9, 44, 60, 309, 375, 377, 383, 385–388, 394 Loasaceae 178, 180, 185 lobe-finned fish 5, 309–310, 349 lobose amoebae 117 lobule 149 local molecular clock 41, 186

Loganiaceae 179, 182, 185 Loliginidae 243 Lomidae 294 Lonchitis 155 longhorn beetle 278, 287 long-legged fly 272 long-tongued bee 267 loon 61, 409, 419, 421 Lophocoleaceae 147 Lophocoleineae 150 lophophorate 224 Lophornis helenae 454 Lophosteus superbus 63 Lophotrochozoa 69, 224, 228 Loranthaceae 171, 173 Loricifera 226 Loridae 480, 484 Loriformes 480, 484 loris 50, 482 lorisiform 50 lotus 171 louse 45, 72 Loveniidae 303 Lowiaceae 204 Loxocemidae 391, 393, 395 Loxocemus 393 Loxodonta africana 471 LSU 235–236 Lucanidae 283 Luciocephalidae 345 Luciocephalus 345 Luciocephalus pulcher 344 Lucy 46 Ludwig, Arne 332 Lufengnacta 73 Lufengopithecus 48 lungfish 309–310, 312–313, 350 lung 312 lunularic acid 133 Lushilagus 51 Luzuriagaceae 205 Lybiidae 449 Lycaon pictus 504 Lycidae 282 lycophyte 133–134, 137, 153 Lycopsida 134 Lygistorrhinidae 272 Lymexylidae 279 Lymexyloidea 287 Lythraceae 190, 195 M macaque 44, 48–49, 51, 58 Mackinlayaceae 183, 185

Index

Macropis 267 Macropodidae 468 macropodiform 469 Macropodiformes 468 Macroscelidae 495 Macroscelidea 56, 472, 480 macrostomatan 390 Macrotis 468 madder 177 Madsen, Ole 459 Maesaceae 182 Magallón, Susana 133, 161 Magellanic Plover 432 maggot 275 magnetic reversal 31 magnetochronology 31 magnetochron 31 magnetostratigraphy 30 magnetozone 74 Magnolia 166 Magnolia sprengeri 166 Magnoliaceae 167 Magnoliales 167 Magnoliidae 188 Magnoliids 161–168, 203 Magnoliophyta 161 Magpie Goose 415 mahogany 200 Majidae 294 Makinoaceae 148 malaconotid 429 Malaconotidae 425 malaconotine 431 Malaconotines 425 Malaconotini 429 Malacostraca 252 Malagasy primate 50 Malagasy tenrec 480 Mallee Fowl 415 Malphigian tubule 287 Malpighiaceae 189 Malpighiales 188–189, 191, 193–194, 197, 201 Maluridae 426 Malvaceae 198, 200 Malvales 197–198, 201 malvid 162, 164, 188, 197 mammal 5–6, 13, 23–24, 38–41, 58–60, 62, 119, 227, 309, 349, 375–376, 460 Mammalia 46, 58, 375–376, 459–460, 490 mammalian clades 46 mammary gland 459 mammary nipple 466

manakin 424 Mandasuchus 61 mandibular canal 463 Mandibulata 71, 252 Mangifera 200 mango 200 mangosteen family 194 Manihot esculenta 194 Manotidae 272 manta ray 324 Mantelliceras mantelli 65 Mantellidae 358 Mantispa 287 Mantispidae 291 maple 200 Marantaceae 204 Marasuchus 61 Marattiaceae 154 Marattidae 134 Marattioids 153–155 Marcgraviaceae 178, 182 Marchantiales 148, 150 Marchantiophyta 133–134, 146 Marchantiopsida 146–148, 150 margo 161 marine bird 419 marine turtle 398 marjoram 177 marmoset 44, 49, 51, 58 Marsileaceae 154 marsupial 57, 459–460, 462, 466–469, 471 Marsupial Mole 466 Marsupialia 57–58, 466, 469 Marsupionta hypothesis 459 Martyniaceae 179, 183 masseter muscle 490 Mastixiaceae 180 matK 136, 159, 180 Matonia pectinata 153 Matoniaceae 155 Matthee, Conrad A. 487 Mauritiinae 209 maxillary 500 maximum contsraint 35 maximum likelihood 31 Mayoa portugallica 211 Mayulestes 58 McKenna, Duane D. 278 Mecicobothriidae 256, 258 Mecistops 402 Mecistops cataphractus 402 Mecoptera 261 Mecopterida 262

mecopteroids 73 Medaka 44, 62, 65 medicine 189, 191, 194 Mediophyceae 128 Medlin, Linda K. 123, 127 medusa 233, 235 Medusagynaceae 189 Medusozoa 75–76, 237 Meemania 64 Meemania eos 63 Meesiaceae 140 megabat 499, 503 Megachasmidae 321 Megachilidae 265, 267 Megachiroptera 499 Megadermatidae 500 Megalaimidae 448 Megalonychidae 476 Megalopidae 480 Megalopodidae 281 Megaloptera 261 megapode 415, 417 Megapodiidae 417 megaspore 31 Megatheriidae 476 Megophryidae 359 Mehelya 393 Melampittidae 425, 429 Melandryidae 279 Melanerpes erythrocephalus 445 Melanobatrachidae 358 Melanocharitidae 426 Melanophyllaceae 180, 183, 185 Melanthiaceae 205 Melastomataceae 190, 195 Meleagrididae 416 Meliaceae 198, 200 Melianthaceae 198, 202 Meliphagidae 426 Meliphagoidea 426 Melittidae 267 Melittidae s.s. 265 Mellitidae 303 Meloidae 279 Melyridae 280 Menispermaceae 171 Menotyphla 495 Mentha 177 Menuridae 426 Menyanthaceae 180, 184 Mephitidae 505 Mephitiniae 505 Meredith, Robert W. 466 Merismopediaceae 107

535

536

INDEX

meristematic ring 164 Mesangiospermae 161 mesite 441 mesocoxal cavity 287 mesodermal skeleton 302 mesodermal tissue 223 Mesonychidae 54, 514 Mesoserphidae 74 Mesoserphus 74 Mesothelae 255 metacoxae 281 metallic wood-boring beetle 278 metameric segmentation 247 metasternal suture 287 Metatheria 58, 459–460, 469 metathoracic wing 270 Metaves 420–421, 441 Metazoa 7, 69, 76, 117, 215, 223, 226, 228, 234, 247 Meteoriaceae 139, 142 metepisterna 287 Metgeriidae 148 Methanobacteria 102 Methanobacteriaceae 102 Methanocaldococcaceae 102 Methanococcaceae 102 Methanococci 102 Methanomicrobia 102 Methanopyraceae 102 Methanopyri 102 Methanosarcinaceae 102 Methanospirillaceae 102 Methylococcaceae 108 Metoldobotes 56 Metopaulias depressus 251 Metridium senile 233 Metzgeriaceae 148 Metzgeriidae 146, 148, 150 Miacidae 55, 504 Micodon 49 Micrixalidae 358 Microbacteriaceae 107 microbat 499 microbe 90 Microbiotheria 57–58, 467 microbiotherian 466 Microbiotheridae 467 microbivore 248 Microchiroptera 500 Microgale 479 Microhylidae 358 Microhyloidea 358 Micromalthidae 278, 288 Micromalthus 281

Microsporidia 92, 215, 218 Microstigmatidae 256 Micrurus 393 Migidae 256, 259 Miglia, Kathleen J. 445 milkwort 192 millipedes 251 Mimia 330 Mimidae 425, 429, 431 Mimotona 52 Mindell, David P. 436, 451, 454 minimum constraint 6, 35, 40, 42 Miniopteridae 502 mint 177 minute bog beetle 278 Miocene ape 48 Miomoptera 72 Misodendraceae 173 Mississippi Paddlefish 333 mistletoe 173 mites 251–252, 255 mitochondria 89, 92–93, 96 mitochondrial cytochrome oxidase I gene 239 mitochondrial cytochrome oxidase II 509 mitochondrial DNA 5 mitosome 93 Mitsukurinidae 321 Miya, Masaki 328 Mniaceae 142 moa 413 mockingbird 429 Moclaybalistes danekrus 66 modern bird 423, 430 Modulatrix 429 mole 53, 479, 496 molecular clock 3, 4, 41 Mollicutes 112 Molluginaceae 171, 174 Mollusca 7, 69–71, 224, 239, 242 mollusk 69 Molossidae 500, 502 Momotus 446 monad 138 monarch flycatcher 429 Monarchidae 425, 429 mongoose 14, 55, 504 Monilophyta 133, 153 Monimiaceae 167 monitor lizard 383 monkey 50 monocot 7, 161–165, 203, 205–206, 211 Monodontidae 512, 514

Monograptus uniformis 28 Monommatidae 279 Monotomidae 280 Monotremata 58, 459, 462–464, 471 Monotrematum 463 Montanalestes keeblerorum 57 Montiniaceae 179, 183, 185 mooneye 335 moonwort 134 Moore, William S. 445 Moorella 107 moorhen 440 Moraceae 190, 193 Moraxellaceae 108 Mordellidae 279 Morelia riversleighensis 394 Morganucodontidae 58 Morinaceae 184 Moringaceae 198 Mormoopidae 500, 502 Morotopithecus 49, 485 Moschidae 512, 514 mosquito 45, 74 moss 133–135, 138, 143 moss beetle 287 Motacillidae 424 moth 260 moulting 224, 247 mound-builder 415 Mountain Zebra 509 mouse 44, 51–52, 56–58, 488, 490 mouse lemur 50 Mouse-like 491 mouthparts 261 Moythomasia 330 MPL 211 mtDNA 234 mucilage cell 134 Mucor 217 Mucoromycotina 217 mud turtle 400 multicellular algae 76 multicellular gametophytic rhizoids 134 multiplicative testa 166 Multipolar Centrics Group 128 Multituberculata 58 Muntingiaceae 198, 200 Muridae 52, 491 Murinae 52 Muroidea 52, 491, 493 Murphy, William J. 471, 504 Murtoilestes abramovi 57 Mus 52, 349, 497 Mus musculus 52

Index

Musaceae 204, 206 Musca domestica 273 muscicapid flycatcher 429 Muscicapidae 425 Muscicapoidea 425, 430 Muscidae 272, 274 muscle insertion 261 Muscomorpha 272 musk deer 512 Musky Rat-kangaroo 468 musophagiform 420 Musophagiformes 420 mustard oil glucoside 197 Mustelidae 505 Mustelus 322 Mute Swan 415 Mutillidae 265 Mycetophagidae 279 Mycetophilidae 272 Mycetozoa 117 Mycobacteriaceae 107 Mycoplasmataceae 107, 112 mycorrhizae 215, 218 mycorrhizal symbionts 216 Mydaus 505 Mydidae 272, 274 mygalomorph 259 Mygalomorphae 258 Mylia taylorii 150 Myliaceae 147, 150 Myliobatidae 322, 325 Myliobatiformes 322, 324 Myllokunmingia 68 Mylodon darwinii 476 Mylodontidae 476 Myobatrachidae 359 Myobatrachoidea 359 Myocastoridae 491 Myodocarpaceae 183, 185 Myomorpha 52, 493 Myoporaceae 183 Myopsida 244 Myosotis 177 Myriapoda 252 Myricaceae 190, 193 Myriochelata 253 Myriophyllum 175 Myristica 166 Myristicaceae 167 Myrmeciinae 267 myrmecobiid 467 Myrmecobiidae 467 Myrmecobius fasciatus 466 Myrmecophaga 476

Myrmecophaga tridactyla 476 Myrmecophagidae 477 myrmecophagy 476 Myrmeleon 287 Myrmeleontidae 291 Myrmeleontiformia 291 Myrmicinae 266 Myrothamnaceae 171, 173 Myrsinaceae 178, 182, 185 Myrtaceae 190, 195 Myrtales 188, 190, 195 myrtle 194 Myrtus 194 Mystacinidae 501 Mysticeti 512 Mytonolagus 51 Myxinidae 318 Myxiniformes 317 Myxococcaceae 109 Myxophaga 278, 281, 283–284, 286, 288 Myxozoa 224, 234 myxozoan 237 Myzopodidae 502 N Nageiaceae 157 naked coral 235 naked ladies 205 naked polyp 236 Nakundon 58 NALMA 55 Namacalathus 76 Namacalathus hermanastes 75 Namapoikia 76 Namapoikia rietoogensis 75 Nandinia 506 Nandinia binotata 505 Nandiniidae 506 nannofossil 31 nannoplankton 66 Nanoarchaeota 102 Nanoarchaeum 102 Narcinidae 322 Narcininae 322 Narcomedusae 236 Narkinae 322 Nartheciaceae 206 Narwhal 512 Nasikabatrachidae 358 Nasonia 74 nasturtium family 199 Natalidae 500 Natatanura 358

537

Natricidae 391, 393, 395 Natrix mlynarskii 394 natural rubber 194 Nautiloidea 242 Nautilus 242 nautilus 242 Navajosuchus mooki 403 Naxilepis 63 Naylor, Gavin J. P. 320 ndhF 178 neanderthal 44, 46, 48, 51, 58 Near, Thomas J. 328, 339 Neckeraceae 139, 143 nectar 267 Nectariniidae 424, 454 nectivory 276 Neisseriaceae 108 Nelumbo 169, 171 Nelumbonaceae 171 Nematocera 270 Nematoda 7, 45, 68–69, 71–72, 226, 249 nematode 247 nematodontous moss 143 Nematomorpha 226 Nematostella 75, 223 Nemertea 224 Nemesiidae 256, 258 Nemestrinidae 272, 276 Nemestrinoidea 272 Nemonychidae 281 Nemopteridae 291 Neoanguimorpha 386, 388 Neoaves 61, 409–410, 415, 419–421, 423, 430, 436, 438, 441–442, 446, 448, 451, 455 Neobalaenidae 511–512, 514 Neobatrachia 359 Neocaecilia 370 Neocallimastigomycota 217 Neoceratodus 348 neodiapsids 60 Neodiptera 272, 274 neognath waterbird 420 Neognathae 61, 409–410, 412, 421, 423 neognath 61 Neoiguania 387 Neomys fodiens 495 Neoptera 72 Neopterygii 335 Neornithes 61–62, 409, 419, 423, 437 neornithine 410 neoselachian shark 326 Neoselachii 320 Neoteleostei 335

538

INDEX

Neotrichocoleaceae 150 Neoverrucidae 299 Nepenthaceae 170, 174 Nephropidae 294 Nephrozoa 69 Nepomorpha 73 nerve cord 309 Nesomyidae 491 Nesophontes 498 Nesophontidae 495, 497 nest founding 268 nest parasite 445 nettle 193 Neuradaceae 198 Neuroptera 261 Neuropterida 261 Neuropteroidea 73–74, 262 Nevrorthidae 291 New World barbet 445, 448 New World monkey 482 New World opossum 466 New World passeroid 430 New World suboscine 424 New World tapir 509 New World vulture 437 New Zealand wren 423 Newton, Angela E. 138 Nicotiana tabacum 177 nicotine 177 nidamental gland 244 Nigerophiidae 394 nightbird 436 nightjar 419 nimravids 55 Nitidulidae 280, 286 Nitrariaceae 200 Nitrosomonadaceae 108 Nobleobatrachia 359 Nocardiaceae 107 Nocardiopsaceae 107 Noctilionidae 501 Noctilionoidea 500, 502 nocturnal owl 419 nodal structure 167 noddy 432 noncoelomate invertebrates 247 non-echolocating megabat 499 nonparametric rate smoothing 186 non-passerine 441 Northern White Rhinoceros 509 Nosodendridae 282 Nostocaceae 107 Noteridae 284 Nothofagaceae 193

Nothofagus 192, 469 Nothrotheriops shastensis 476 notochord 226 Notoryctemorphia 467 notoryctemorphian 467 Notoryctes 467 Notoryctidae 467 Notorynchidae 321 Notothenia 341 Nototheniidae 341 Notothenioidei 342 Notothenoidei 7 NP23 66 NPRS 209–211, 258 Nubianophis afaahus 394 nucellar cap 162 nuclear rRNA 237 nucleariid 215, 223 Numbat 466 Numididae 418 Nummulites atacicus 55 nurse shark 326 nuthatch 429 nutmeg 166 Nyctaginaceae 171, 174 Nycteribiidae 274 Nycteridae 501 Nyctibatrachidae 358 Nymphaeaceae 166, 203 Nymphaeales 161–162, 165, 208 Nymphidae 291 Nymphomyiidae 271, 275 Nyssaceae 180, 185 O oak 193 Obdurodon insignis 463 Obudurodon 463 ocean acidification 236 oceanic anoxic event 246 oceanic dolphin 512 Ochnaceae 189, 194 Ochodaeidae 283 Ochotona 51 Ochotonidae 9, 51, 488 Ocimum 177 Ocoidae 274 Octapodiformes 7 Octocorallia 233, 235, 237 Octodontidae 491 octopine dehydrogenase 245 Octopoda 242–243, 245 Octopodidae 243 Octopodiformes 242–243, 246

octopod 242, 245 Ocythoidae 245 odd-toed ungulate 508 Odobenidae 505 Odontaspididae 321, 324 odontocete 514 Odontoceti 512, 514 Odontophoridae 418 Oedemeridae 279 Oedipodiaceae 140 Oedipodiales 140 Oedipodium 142 Oegopsida 245 Oepikodus communis 70 Oestridae 270, 274 Ogg, James G. 26 oil cell 167 oil collecting 267 oil-rich fruit 177 Olacaceae 171, 174 Old Red Sandstone 36 Old World finch 429 Old World megabat 499 Old World monkey 48, 483 Old World muscicapoid 430 Old World suboscine 424 Old World tapir 509 Old World vulture 437 Olea europaea 177 Oleaceae 177, 179, 183 oleaster 193 oleoresin 200 Olfactores 67, 69 Oliarces 287 oligochaete 71 Oligopithecidae 49 Oligopithecus 49 Oliniaceae 195 olive 177 Olivooides 76 Omalisidae 282 Omethidae 282 Ommastrephidae 243 Ommatidae 288 Ommatius gemma 270 omomyoid 49 Onagraceae 190, 195 Oncothecaceae 179, 182 one-whorled androecium 182 onion 206 ontogeny 302 Onychonycteris finneyi 503 Onychophora 223, 226 onycophoran 309

Index

Oomycetes 117, 215 open-eye squid 242 Opheodrys aestivus 375 ophiacodontid 59 Ophioglossaceae 154 Ophioglossidae 134 ophioglossoid 155 Opiliaceae 171, 173 Opiliones 255 Opisthobrachia 70 opisthobranch 70 Opisthokonta 117–118, 215, 218, 223 opisthokont 117, 118 Opisthoteuthidae 243, 245 Opisthothelae 255 opium 169 opossum 44, 57–58, 459, 466 orange 200 orangutan 44, 48, 51, 56, 58 Orbiculariae 258 Orbitulina concava 65 orb-web weaver 256 Orchidaceae 203, 205–206, 208, 210 orchid 203, 206 Orectolobidae 321 Orectolobiformes 321–322, 325 oregano 177 Oreopithecus 48 organelle 92 organismal phenotype 24 Origanum 177 Origin of Species 3 Oriolidae 426 ornamental 166, 169, 173–175, 194 ornamentation 144 Ornithodira 61 Ornithorhynchidae 463 Ornithorhynchus anatinus 459, 462 Orobanchaceae 179, 183, 185 Orrorin 48 Orrorin tugenenis 46 Orsodacnidae 281 Orthodontiaceae 139, 142 Orthodontiales 139, 142 Orthodontium 142 Orthogastropoda 70 orthogastropod 70 Orthogonikleithrus 64 Orthonychidae 426 Orthopalpae 256, 259 orthopalp 258 Orthorrhapha 272 Orthotrichaceae 140, 142 Orthotrichales 140, 144

Oryctolagus 349 oscine 423–425, 431 Osmeriformes 335 Osmundaceae 154, 155 Osmylidae 291 Osphronemidae 346 Osphronemus 346 Osprey 437 Ostarioclupeomorpha 335 ostariophysean 65, 335 Ostariophysi 330 Osteichthyes 40, 62–64, 66, 309–310, 320 osteichthyian 311 Osteoglossidae 336 Osteoglossomorpha 330, 336 Osteolaemus 402, 405 osteolepiform 312 Ostracoda 252 Ostrich 410, 413 Ota, Kinya G. 317 Otariidae 505 Otlestes 53 Otocephala 336 otocephalan 336 otophysan 64 otter 504 Oulodus elegans detorta 63 Ouranopithecus 48 ovenbird 424, 430 ovipositor 264 Ovis 55 ovules 157, 161 ovuliferous 157 owl 420, 436, 453 owlet-nightjar 454 Oxalidaceae 188–189, 194 Oxalidales 188–189, 194 Oxalis namaquana 188 oxygen isotope stratigraphy 30 Oxynaspididae 299 Oxynotidae 321 oystercatcher 433 Ozarkodina crispa 63 P Pachycephalidae 426 Pachyneuridae 271 paddlefish 309, 328, 333 Paenungulata 39, 56, 471 Paeoniaceae 170, 175 painted snipe 432 Pakicetidae 514 Palace of Minos 27 Palaechthon 51

Palaemonidae 294 Palaeochiropteryx 54 Palaeodictyopterida 72 palaeognath 61 Palaeohypsodontus zinensis 56 Palaeomanteida 73 Paleoanguimorpha 386, 388 paleobiology 37 Paleognathae 61, 409–410, 412–414, 417, 419, 423 paleognathous plate 412 Paleoheterodon arcuatus 394 Paleomacropis eocinicus 267 Paleomelittidae 267 paleontology 35 Paleophragmodictya 76 Paleopneustidae 303 Paleosuchus 402 Paleothyris 41 Palinuridae 294 palisade exotesta 167 Pallaviciniaceae 148, 150 Pallaviciniites devonicus 150 Pallid Sturgeon 332 palm 206, 209 palynoflora 73 palynology 60 palynomorph 57, 62 Pan troglodytes 482 Panarthrpoda 226 Pancrustacea 253 Pandanaceae 206 Pandanales 205–206, 210 Pandion 437 Pandion haliaetus 437 Pandionidae 439 Pangea 305 pangolin 53–54, 471, 496 Panheteroptera 73 Panorpida 73, 260 Panorpoidea 74 panray 325 Pantophthalmidae 272 Papaver 169 Papaveraceae 169, 171 papaya 200 papaya family 199 papillae 141 Parabasalidea 117 parabasalid 118 Parachaenichthys charcoti 341 Paracryphia 183 Paracryphiaceae 180, 183 paracytic stomata 161

539

540

INDEX

Paradisaeidae 425, 429 paradise fish 344 Paradoxopoda 252 parallelism 255 Paramys 52 Paramythiidae 426 Paraneoptera 72–73, 261 Paranyctoides 53 Parascylliidae 321, 325 parasite phylogeny 249 parasite 217, 224, 233, 248 parasitic fly 276 parasitic hyphal stage 216 parasitic myxozoan 237 parasitoid 270 Parastacidae 294 Paratropididae 258 paraxonic tarsus 514 Pardalotidae 426 Pareas 393 Pareatidae 391, 394 Pareledone charcoti 242 Paridae 424 Parnassia 193 Parnassiaceae 190, 194 paromomyid 50 Parosphromenus ornaticauda 344 parrot 419 Parsley 177 partridge 415 Parulidae 424, 429 Parviraptor 387 Passalidae 283 Passandridae 280 Passeri 426 Passerida 424–425, 429 Passeridae 424 passeridan 431 Passeriformes 410, 419–420, 423–424, 426–427, 429–431, 454 Passeroidea 424, 429, 430 Passifloraceae 189, 194 passion flower family 194 Pasteurellaceae 108 Patagonotothen 341 Patagopteryx 61 Patasola 49 PATHd8 211 pathogens 218 Paucituberculata 467 paucituberculate 466 Paulowniaceae 179, 183 Pavlovales 124 Pavlovophyceae 124

pax-6 245 PCM 22 peccary 511 Pecora 511, 514 Pedaliaceae 179, 183 Pederpes 62 Pedetidae 492 Pedinidae 303 pedinoid 304 Pedionomidae 432, 433 pedipalpi 255 Peditidae 491, 493 Pedunculata 299 Pelagibacter 89 pelagic octopus 245 Pelargonium 202 Pelecaniformes 420, 436 pelecaniform 61 Pelecorhynchidae 272 pelican 436 Pelliaceae 148 Pelliidae 146, 148, 150 Pelobacteraceae 109 Pelobatidae 359 Pelodryadidae 359 Pelodytidae 359 Pelomedusidae 399 Peltanthera 183 Penaeaceae 195 Penaeidae 294 penalized likelihood 186, 209 pendulum 26 Peng, Zuogang 332, 335 penguin 409, 419, 421, 434 Peniculida 117 Pennantiaceae 183 Pennatulacea 234 pennatulid 234 Pennipollis 211 Pennistemon 211 pentameral symmetry 302 pentamerous eudicot 163 Pentanchinae 322 Pentaphragmataceae 180, 184 Pentaphylacaceae 178, 182, 185 Penthoraceae 170, 175 Pentoxylon 158 Peperomia 166 Peptococcaceae 107 Peramelemorphia 466–467, 469 Peramelidae 468 Perciformes 336, 344 percomorph 65 Percomorpha 65

Pereira, Sérgio L. 412, 415, 432 peremelemorphian 467 Pérez-Losada, Marcos 293, 298 perigynous flower 167 Perimylopidae 279 Peripatus juliformis 223 Perissodactyla 45, 53–55, 471–472, 496, 510 Perissommatidae 271, 275 peristomate moss 142 peristome 138–141, 143 peristome reduction 144 peristome ring 140 peristome structure 143 periwinkles 177 Permian extinction 235 Permopsocidae 72 Persea 166 Perssoniellineae 150 Petauridae 467, 469 Petauroidea 467, 469 Peteroselinum crispum 177 petiole 162, 183 Petroicidae 425 Petromuridae 491 Petromyzon 325 Petromyzonidae 318 Petromyzoniformes 317 Petropedetidae 358 Petrosaviaceae 205–206, 208, 210 Pezizomycetes 216 Pezizomycotina 218 Pezosiren 56 PGD 261, 274 Phaeocystales 124 Phaeostigma 287 Phalacridae 280 phalanger 469 Phalangerida 468 Phalangeridae 467, 469 phalangeridan 466 phalangeriform 469 Phalangeroidea 467 Phascolarctidae 467, 468 Phascolotherium 58 Phasianidae 418 pheasant 416 Phellinaceae 180, 184 Phenacodus 54 phenetics 20 Phengodidae 282 phenotypes 23 Philepittidae 426 Philesiaceae 205

Index

Philydraceae 204 phioldotiform 508 Phiomorpha 491, 493 phiomorph rodent 490 phloem 182, 203 Phloeostichidae 281, 287 Phloiophilidae 280 Phocidae 505 Phocoenidae 512, 514 Phodilus 451 Phoenicopteridae 421 Phoenicopteriformes 420 Pholidophorus 330 Pholidota 54, 472 Phoridae 273 Phoronida 224 Phosphatherium 39, 56 photosynthetic lamellae 143 phragmocone 244 Phrymaceae 179, 183 Phrynobatrachidae 358 Phrynomeridae 358 Phyllanthaceae 194 Phyllobacteriaceae 109 Phyllocladaceae 157 Phyllodactylidae 384, 387 Phyllodocida 71 Phyllomedusidae 359 Phyllonomaceae 180, 183 Phyllorhynchus 393 Phyllostomidae 500, 502 phylochronology 3 phylogenetic tree 3 phylogeny 3, 20 phylogram 21 Physeteridae 512 Physeteroidea 512 Phytolaccaceae 171, 174 phytophagy 288 Picathartidae 425 Pici 445–446, 450 Picidae 445–446, 448 Piciformes 7, 420, 423, 430, 445–447, 450 Picnogonida 252 Picramniaceae 188 Picrodendraceae 189 Picrophilaceae 102 Pictodentalium 239 Pictodentalium vernedei 239 piculet 445, 448 pig 19, 44, 53–58, 511 pigeon 410, 419, 433 Pig-footed Bandicoot 467, 469 pigmy squid 243

Pig-nosed Turtle 400 pika 44, 51, 56–58, 488 Pikaia graciliens 67 pike-head 344 Pilosa 477 Pilotrichaceae 139 Pinaceae 135, 159 pincer-like anterior extension of premaxillae 462 pineapple 208 pink peppercorn 200 pinniped 505 Pinnipedia 505 Pinus sylvestris 157 Piperaceae 166–167, 203 Piperales 166–167, 208 Pipidae 359 Pipridae 426 Pisani, Davide 251 Pisauridae 256 Piscirickettsiaceae 108 pistachio 200 Pistacia 200 pitcher plant family 174 Pitheciidae 480, 482 Pitilidiales 148 Pittidae 426 Pittosporaceae 180, 183, 185 PL 211 placental mammal 57, 459, 471–473, 498 Placentalia 56–57, 466, 471–473, 488 placoderm 64 Placodus 60 Placozoa 223, 226 Plagiochilaceae 147 Plagiotheciaceae 139 Plains Zebra 509 Plains-wanderer 432, 441 Planctomycetaceae 109 Planctomycetes 109, 113 plane tree 171 Planorbidae 70 Planorboidea 70 Plantae 117 Plantaginaceae 179, 183, 185 plant 6, 46, 93, 116 plastid 119 plastid genome 153 plastron 281 Platanaceae 171, 175 Platanistidae 512, 514 Platanus 171 Platcanthomyidae 491 plate tectonics 37

Platyhelmintha 224 Platypezidae 272 platypus 44, 58, 459, 463 Platyrhinidae 322, 324 Platyrhiniformes 322 platyrrhine 49 Platyrrhini 49, 480, 482, 485 Platysteiridae 425 Platysternidae 399 Platyzoa 224 plectid 248 Plectocretacicus clarae 65 plectomycete 216 Plectreuridae 256 plesiadapid 50 plesiadapiform 51 Plesiobatidae 322 Plethodontidae 366 Plethodontoidea 366 Pleuragramma antarcticum 341 pleurocarp 138, 142, 144 pleurocarpid 138, 144 Pleurodira 398, 400 pleurodire 398 Pleuroziaceae 148 Pleuroziales 148 Plexechinidae 303 Ploceidae 424 Plocospermataceae 179, 183 plover 433 Plumbaginaceae 170, 174 Pluvialidae 434 Pluvialis 432 Pluvianellidae 433 Pluvianidae 434 Pluvianus 432 Poaceae 203–204, 211 Poached Egg Plant 200 Poales 204–206, 211 pocket gopher 492 pocket mouse 492 Podicipedidae 421 Podicipediformes 421 podocarp 469 Podocarpaceae 159 Podocnemidae 399 Podostemaceae 189, 194 Poecilasmatidae 299 polarity 270 polarity zone 31 Polarornis 62 Polarornis gregorii 61 Polemoniaceae 178, 182 polemonoid 182

541

542

INDEX

POLII 261 Polioptilidae 425, 429 pollen 164, 167, 175, 186, 267, 278 Pollicipedidae 299 pollinator 267, 270 pollinivory 276 polychaete 71 polydolopimorph 469 Polygalaceae 190, 192 Polygonaceae 170, 174 Polyodon 332 Polyodontidae 333 Polyosmaceae 180, 183 polyp 233 Polyphaga 278–283, 288 polypod fern 136, 156 Polypodiidae 134 Polypodiozoa 234 Polypods 154 Polypteriformes 329 polypteriform 330 Polyschides 240 Polyschides arnoensis 240 Polystoechotidae 291 Polytrichaceae 140 Polytrichales 138, 140–141, 143 Pomatostomidae 426 pomegranate family 195 Pompilidae 265 poneroid 266 Pongidae 48 Pontederiaceae 204, 209 Ponto-Caspian 332 Pontoporiidae 512, 514 poposaurid 61 porcupine 493 Porellaceae 147, 150 Porellales 151 Porellineae 150 Porifera 76, 223–224, 226 Porphyromonadaceae 109 Porpitidae 234 porpoise 512 Porter, Megan L. 293 Portulacaceae 171, 174 Portunidae 294 Posidoniaceae 206, 208 possum 468 Potamidae 294 Potamogetonaceae 204, 206, 208 Potamotrygonidae 322 potato 177 Potoroidae 468 potoroo 468

Pottiaceae 140 Pottiales 140 Poux, Céline 479 Praeanthropus afarensis 46 Praeanthropus anamensis 46 pratincole 432 predation 276, 288 predator 233, 270 premaxillary 501 presternal cervical sclerite 281 Priapulida 7, 224, 226 Primates 49–51, 349, 471–473, 482–485, 496 primitively hornless 512 Primulaceae 178, 182, 185 primuloid 182 Prionoceridae 280 Prionodon spp. 506 Prionodontidae 506 prionopid 429 Prionopidae 425 Pristidae 322, 324 Pristiformes 322, 324 Pristionchus pacificus 248 Pristiophoridae 321 Pristiophoriformes 322 Proboscidea 472 proboscidean 56 Procaprolagus 51 Procavia 496 Procellariiformes 61, 421 Procerites 58 Proconsul 49, 485 Proctotrupoidea 74 Procyonidae 505 Prodentaliidae 239 Prodentalium fredericae 240 Proeleginops grandeastmanorum 339, 342 Progonomys 52 Progonomys hussaini 53 Prohylobates 48, 485 Prohylobates simonsi 49 Prohylobates sp. 49 Prohylobates tandyi 49 prokaryote 6, 13, 16, 76, 89–96, 112–113, 119, 215 Prokennalestes minor 57 Prokennalestes trofimovi 57 prolacertiform 60 Prolixocupes 287 Promeropidae 424, 429 Pronaidites carbonarius 70 pronghorn antelope 512

Propalticidae 280 prophyll 167 Propionibacteriaceae 107 propleuron 281 Propliopithecidae 49 Propliopithecus 49 Prorastomus 56 Prosarcodon 53 Proscylliidae 321 prostome 141 Proteaceae 170–171, 175 Proteales 171 Proteidae 366 Proteobacteria 113 Proteoidea 366 Proterochampsidae 61 Proterosuchidae 61 Protictis 54 protist 6, 16, 116–117, 237 Protoclepsydrops 41, 59 Protoclepsydrops haplous 59 Protocucujidae 280 protonema 138 Protopteridae 350 Protorosaurus 60 Protorosaurus speneri 60 Protorothyrididae 59 Protostomia 6, 69, 224, 228 Prototheria 459–460, 464 Prunellidae 424 Pryer, Kathleen M. 153 Prymnesiales 124 Prymnesiophyceae 124 Przewalski's Horse 509 psaA 136, 164 Psammophylax 393 Psarolepis 64 Psarolepis romeri 63 Psaronius 153 psbA 150 psbB 136, 164 Psephenidae 282 Psephurus 332 Pseudaphritidae 341 Pseudoalteromonadaceae 108 Pseudocarchariidae 321 Pseudocheiridae 467, 469 pseudocoelomate 226 Pseudocryphaea 142 Pseudolagosuchus 61 Pseudolepicoleaceae 147 Pseudomerope gallei 73 Pseudomonadaceae 108 Pseudoscaphirhynchus 333

Index

Pseudoscaphirhynchus gladius 333 Pseudoscaphirhynchus spathula 333 Pseudotriakidae 321 Pseudoxenodon 393 Pseudoxenodontidae 391, 393, 394 Psidium 194 Psilotaceae 154 Psilotidae 134 Psilotum 159 Psiloxylaceae 194 Psittaciformes 420 psittaciform 61 Psophiidae 441, 442 Psychodidae 271, 276 Psychodomorpha 74, 271–272, 275 Psychopsidae 291 Pteridaceae 154, 155 pteridophyte 153 Pterobryaceae 139 Pterobryellaceae 139 Pterocles 433 Pteroids 154, 156 pteropodid 497, 500, 502 Pteropodidae 500 Pterostemon 170 Ptilidae 283 Ptilidiaceae 150 Ptilidiineae 150 ptilinal fissure 273 ptilinum 273 Ptilodactylidae 282 Ptilogonatidae 425 Ptilonorhynchidae 426 Ptilotodon 387 Ptychadenidae 358 Ptychomitriaceae 140, 143 Ptychomniaceae 139 Ptychomniales 139, 142, 144 Ptychopteridae 271, 276 Ptychopteromorpha 271 pubic nipple 501 Pucadelphys 58 Pucapampella 66 Puccinia graminis 216 Pucciniomycotina 216 puffbird 445–446, 448 Pulmonata 70 pulmonates 70 Pulsellidae 241 Pulsellum 240 Pulsellum infundibulum 240 pupal stage 260 puparium 273, 275 Pupation 281

Purella 71, 75 Purgatorius 49 Putranjivaceae 189, 197 Pycnogonida 252 Pycnonotidae 424 Pygmy Anteater 477 Pygmy Right Whale 514 Pygmy Sperm Whale 512 pygmy-possum 469 Pygodus anserinus 67 Pygopodidae 384, 387 Pygopodoidea 387 Pygopodomorpha 384, 388 pygopodomorph 388 pyrenoid 134 pyrenomycete 216 Pyrochroidae 279 Pyroteuthidae 244 Pythidae 279 Python 393 Pythonidae 391, 393, 395 python 390 Pyxicephalidae 358 Q Quagga 509 quail 415 quality of the fossil record 39 queen/worker dimorphism 268 Quercus 193 Quiinaceae 189 Quillajaceae 192 Quintinia 183, 186 R r8s 143, 240 rabbit 44, 50–52, 56–58, 488 rabbitfish 320 raccoon 14, 55, 504 Racopilaceae 139 radial canal system 236 Radial Centrics Group 128 Radiata 224 radiation of placental mammal orders 51 radiolarian 34 radiometric dating 28–31 radiometric decay clock 30 radish 198 radula 245 Radulaceae 147, 150 Rage, Jean-Claude 390 rail 420, 432, 443 Rajidae 322, 325 Rajiformes 322, 324

Rallidae 442 Ralloidea 443 Rams Horn Squid 242 Ramphastidae 449 Ramphotyphlops 393 Ranidae 358 Ranixalidae 358 Ranunculaceae 171 Ranunculales 162–163, 171 Rapateaceae 204, 211 raphal bundle 162 Raphanus 198 Raphidioptera 260 raphistomatid 70 raptor 419, 452 Rastelloid 258 Rastelloidina 258 rat 44, 51–52, 56–58, 488, 490 rate constancy 5 rate extrapolation 5 ratfish 320 ratis 412 ratite 61, 409–410, 412–414, 419, 423 Rattus 52, 497 Rattus norvegicus 52 rauisuchian 61 Ravenictis 54 ray-finned fish 6, 8–9, 16, 309, 328–330, 349 ray 309, 320, 322 rbcL 136, 150, 159, 164, 170, 178, 180, 186 Recurvirostridae 433 red algae 94, 117, 119 Red Knot 432 Red Panda 505 red pepper 177 Red-chested Flufftail 440 Red-headed Woodpecker 445 reduced eyes 495 reef-building coral 233 Regulidae 425, 429 relative time units 27 relaxed clock 4, 5, 42, 227 Remipedia 252 Renieria 76 Renner, Susanne 157 Reptilia 6, 9, 59, 375–377, 383, 390, 398, 402 reptiliomorph 62 Reptiliomorpha 60, 62 Resedaceae 200 Restionaceae 204, 208–209, 211 reticulate venation 203

543

544

INDEX

reticulated beetle 278 retrolateral tibial apophysis 256 Reutterodus andinus 70 Rhabdidae 241 Rhabditida 249 Rhabditidae 248 Rhabditoidea 248 Rhabditomorpha 249 Rhabdodendraceae 171, 174 Rhabdus 240 Rhabdus paralelum 240 Rhachiberothidae 291 Rhacophoridae 358 Rhagionidae 272 Rhagophthalmidae 282 Rhamnaceae 190, 193 Rhaphidiophorus hystrix 71 Rhaphidioptera 261 rhea 413 Rheidae 413 Rheiformes 413 Rheobatrachidae 359 Rhinatrematidae 370 Rhincodontidae 325 Rhineuridae 384, 388 Rhineuriformata 387 Rhineuroidea 387 Rhinidae 322, 324 Rhiniformes 322 Rhinobatidae 322 Rhinobatiformes 322, 324 rhinoceros 510 rhinocerosiform 508 Rhinocerotidae 509 Rhinocheilus lecontei 390 Rhinochimaeridae 322, 324 Rhinoclemmys 399 Rhinocryptidae 426 Rhinodermatidae 359 Rhinolophidae 499–500, 502 Rhinolophoidea 502 Rhinophis 393 Rhinophrynidae 359 Rhinopomatidae 500 Rhipiceridae 282 Rhipiduridae 425 rhipidurine flycatcher 429 Rhipiphoridae 279 Rhipogonaceae 205 Rhizaria 118 Rhizobiaceae 109, 192 rhizobial symbiosis 192 Rhizogoniaceae 139 Rhizogoniales 139, 142

Rhizophoraceae 189 Rhizopus 217 Rhizostomeae 236 Rhodnius 45, 72 Rhodobacteraceae 109 Rhodocyclaceae 108 Rhodophyta 117 rhodophyte 116 rhodopsin 245 Rhodosorus 116 Rhodospirillaceae 109 Rhopalosomatidae 265 Rhyacotritonidae 366 Rhychocalycaceae 195 Rhynchobatidae 322, 324 Rhynchobatiformes 322 Rhynchocephalia 383 Rhynchopidae 433 Rhyniognatha hirsti 73 Rhynochetidae 442 Rhysodidae 284 Rhytiodentalium kentuckyensis 240 Ribes 175 ribosomal DNA 317 rice 203, 208 Rickettsiaceae 109 right whale 514 Rigodiaceae 143 Rissikia 159 river dolphin 514 Riverine Rabbit 487 robber fly 270, 272 robust stony coral 236 Rock Pigeon 409 rock record 26 Rock units 26 rocket 198 Rodentia 39, 45, 50–52, 349, 471–473, 494 Roelants, Kim 357 Rogers, Alex D. 233 Romeria primus 59 root node 158 root nodule 192 root parasite 200 Roridulaceae 178, 182 rorqual baleen whale 512 Rosaceae 188, 190, 192–193, 201 Rosales 188–190, 193 Rosanae 188 rose 188, 193 rosemary 177 Rosidae 188

Rosids 162–164, 169, 173, 188, 192, 194, 197, 201 Rosmarinus officinalis 177 Rostratulidae 433 Rotifera 224 Rough Greensnake 375 Rough-legged Hawk 436 roundworm 247 Rousseaceae 180, 184 rove beetle 278 RPB2 178 rps16 intron 178 rps4 136, 150 rRNA 226, 234, 236, 239 RTA clade 256 Rüber, Lukas 344 Rubiaceae 177, 179, 182, 185 Rudbeckia 178 rudist bivalve 236 Rugosa 235 rugosan coral 235 Ruminantia 55, 511–512, 514 Ruppiaceae 206, 208 Ruptiliocarpon 193 rush 208 Rusophycus 69, 71 russellophiid 394 rust fungi 216 Rutaceae 198, 200 Rutenbergia 142 Rutenbergiaceae 139 Ruzophycuslike 72 Ryder, Oliver A. 508 S Sabiaceae 169, 171 Sacabambaspis janvieri 67 Saccharomyces cerevisiae 216 Saccharomycetes 216 Saccharomycotina 216, 218 Saccoloma 155 saff ron 206 Sagittariidae 438 Sagittariiformes 437 Sagittarius 437 Sahelanthropus 48 Sahelanthropus tchadensis 46 sailors by the wind 234 salamander 62, 309, 312 Salamandridae 366 Salamandroidea 366 Salicaceae 189, 201 salmon 340 Salmoniformes 64, 336

Index

Salpingidae 279 Salviniaceae 154 Sanango 183 Sander, Jennifer M. 348 sandgrouse 420, 433 sandpiper 432 Sanguinaria 169 Santalaceae 171, 173 Santalales 162, 171, 174 Sapindaceae 198, 200 Sapindales 197–198, 201 Sapindus 200 Sapotaceae 178, 182 saprobic yeast 216 Sapygidae 265 SAR11 109 Sarcobataceae 174 Sarcolaenaceae 198 Sarcophagidae 274 Sarcopterygii 62–64, 309–310, 325, 328–329, 348 Sarothrura 442 Sarothrura rufa 440 Sarraceniaceae 182 Saurauia 185 Saurauia alenae 185 Saurauia antique 185 Sauropsida 59, 312 Sauropterygia 60 Saurosternidae 60 Saurosternon 60 Saurosternon bainii 60 Saururaceae 167 sawfish 320, 324 sawshark 322 Saxifragaceae 170, 175 Saxifragales 162, 170, 173–175, 192 saxifrage family 174 Scalpellidae 299 scaly-tailed squirrel 492 Scandentia 49–51, 472, 496 Scapaniaceae 147 Scaphiophrynidae 358 Scaphiopodidae 359 Scaphirhynchus 333 Scaphirhynchus albus 332 Scaphopoda 7, 240 scaphopod 241 Scarabaeidae 283, 288 Scarabaeiformia 287 Scarabaeoidea 283, 286, 288 scarab 278 Scatopsidae 271 Scenopinidae 272, 274

Scheuchzeriaceae 206, 208 Schinus 200 Schistochila aligera 146 Schistochilaceae 147 Schistostegaceae 140 schizaeoid fern 155 Schizaeoids 154 Schizasteridae 303 Schizomida 255 Schizopea typical 70 Schizophora 273 Schizosaccharomyces pombe 216 Schizosaccharomycetes 216 Schlegeliaceae 179, 183 Schuettpelz, Eric 153 Sciadopityaceae 159 Sciaridae 271 Scincidae 384, 388 Scinciformata 384, 386, 388 Scincoidea 387 Scincomorpha 387 Scirtidae 283, 286 Scirtoidea 283, 288 Sciuridae 491 Sciurognatha 52 Sciurognathi 490 Sciuromorpha 52, 493 Scleractinia 233–234, 237 Scleractinia (Complex forms) 234 Scleractinia (Robust forms) 234 scleractinian 236 Scleractiniomorpha 235 sclerite 261 Scleroglossa 383, 387 scleroglossan 383 Scleromochlus 61 Scolecomorphidae 370 Scolecophidia 390–391, 395 Scoliidae 265 Scolopaci 433 Scolopacidae 433 Scorpiones 255 scorpionfly 260 scorpion 255 Scouleria 143 Scouleriaceae 140 Scouleriales 140 Scraptiidae 279 screamer 415 Scrophulariaceae 179, 183, 185 scrub fowl 415 Scutellina 305 Scydmaenidae 283 Scyliorhinidae 321–322, 325

Scyliorhininae 322 Scyllaridae 294 Scyphozoa 75, 237 scyphozoan 233 sea anemone 75, 223, 233, 235 sea fan 233 sea hare 45, 70 sea lion 504 sea pen 234 sea spider 252 sea urchin 45, 68, 303 seagrass 204 seal 55, 504 secondarily hornless 512 Secretary Bird 437 sedge 208 sediment 26, 164, 185 sediment core 29 sedimentary cycle 28 Sedum 175 seed 166, 184 seed oil 194 seed plant topology 158 seed plants 133, 135–136, 159 seed-bearing plant 157 seedsnipe 432 selabacteria 113 Selenodontia 514 Seligeriaceae 140 Seligeriales 140 Semaeostomeae 236 Semiformiceras. darwini 64 semiparametric penalized likelihood method 164 semiparametric rate smoothing 184 Semnornis 448 Semnornis ramphastinus 447 Sepiadariidae 243 Sepiidae 245 Sepioidea 244 Sepiolida 243 Sepiolidae 244 sequence stratigraphy 30 seriema 437, 441 Serpentes 7, 383–384, 390, 392, 394 Serravallian 394 sessile polyp 233 Sessilia 299 Setchellanthaceae 198 Shaffer, H. Bradley 398 shag 436 shank 433 shark 40, 66–68, 309, 320, 322, 326 Shasta Ground Sloth 476

545

546

INDEX

Shaw, A. Jonathan 138, 146 sheath 183 sheathbill 432 Shedlock, Andrew M. 375 sheep 44, 58 Shewanellaceae 108 Shinisauridae 384, 386 Shinisauroidea 386 Shinisaurus 386 ship-timber beetle 287 shorebird 419–420, 432–434, 442 Short-beaked Echidna 462 short-horn fly 270 short-tailed whip-scorpion 255 short-tongued bee 267 shrew 44, 53, 479, 495–496, 498 shrike 429 shrimp 251 shrub 166, 171, 173–174, 188, 191–194, 199 Sialis 287 Sierolomorphidae 264 sieve cell plastid 203 Sikhotealinia 281 Silicea 223 siliceous sponge 223 silk glands 255 silky flycatcher 429 Silphidae 283 Siluriformes 337 siluriform 337 Silvanidae 281, 287 Silvianthemum 186 Simaroubaceae 198, 200 Simmondsiaceae 171, 174 simple thalloid gametophyte 147 Simplicidentata 52 simplified hindgut 495 Simuliidae 271 Sinoconodon 58 Sinocylcocyclicus guizhouensis 76 Sinodelphys 50–51, 53, 58 Sinodelphys szalayi 57–58, 469 Sinuopea sweeti 70 Siphonaptera 261 Siphonodentaliinae 241 Siphonodentalum 240 Siphonophora 234, 236 siphonophores 233, 237 Sirenia 472 sirenian 56, 471 Sirenidae 366 Sirenoidea 366 Sisyridae 291

Sittidae 425, 429 Sivapithecus 48 skate 320, 324 skeletal dissolution 236 skeletogenesis 228 skeletonized coral 236 skiff beetle 278 skimmer 432 skunk 505 slime mold 117, 215 sloth 56, 477 Smilacaceae 206 Smith, Andrew B. 302 smut 216 snake 9, 60, 309, 375, 383, 385–388, 390–392, 395 snakehead 344 snapping tutle 399 SNF 261 soap bark 192 soft cryptic plumage 452 soft maximum constraint 5, 9, 42–43, 46 softshell turtle 400 soil-dwelling fungi 216 Solanaceae 177, 179, 183 Solanales 177–179, 182–183, 186 Solanum lycopersicum 177 Solanum tuberosum 177 soldier beetle 278 Solenodon 53, 479, 497 Solenodontidae 53, 496 Solibacteraceae 109 Somniosidae 321 songbird 410, 423 Sooglossidae 358 Sooglossoidea 358 Sordariomycetes 216, 218 Soricidae 479, 496 Soricinae 498 Soricomorpha 53, 479, 496 soricomorph 53 South American barbet 449 South American caviomorph 493 South American Lungfish 349 southern beech 192 Southern Three-banded Armadillo 475 Southern White Rhinoceros 509 spadix 203 Spalacidae 491 Spanish Moss 208 Sparganiaceae 204 sparrow 429 Spatangidae 303

Spatangoida 303 spatangoid 302, 304 sperm ultrastructure 153 Spermatophyta 134 Spermatophytes 133–137, 157 Sphaeritidae 283 Sphaeriusidae 284 Sphaerocarpales 148, 150 Sphaerodactylidae 383–384, 387 Sphaerodactylus richardsoni 383 Sphagnaceae 140 Sphagnales 138, 140–141, 143 Sphagnum 143 sphearomorphic acritarch 76 Sphecidae 265 spheciform wasp 267 Sphecomyrminae 264, 267 Sphenisciformes 421 Sphenocleaceae 179, 183 Sphenodon 377 Sphenodontia 376 sphenodontid 378 Sphenolithus predistentus 66 Sphenostemon 183 Sphenostemonaceae 183 Sphindidae 280, 287 spice 166, 177 spider 251–252, 255, 257, 259 spider wasp 264 spiderwort 208 spindle tree family 193 Spirochaetaceae 109 Spirochaetes 109 spirochete 92 Spirula spirula 244 Spirulida 244 Spirulidae 243, 245 Spiruromorpha 248 Splachnaceae 140 Splachnales 140, 144 sponge 45, 76, 223 sponging mouthparts 270 spookfish 320 sporangium 133–134, 137, 155 spore tetrad 143, 150 spores 136 sporophyll 157 sporophyte 133–134, 137–138, 142 sporophyte anatomy 153 sporopollenin 133 Springer, Mark S. 462, 466 Springhaas 492 spurge family 194 Squalidae 321

Index

Squaliformes 321–322, 325 Squalimorphii 321–322, 325 squalimorph 324 Squamata 60, 309, 375–378, 388 Squatinidae 321, 325 Squatiniformes 321, 322 squid 242 squirrel 44, 51–52, 56–58, 490 Squirrel-like 491 SSU rRNA 237 stable isotope sequence 26, 29 Stachyuraceae 198, 201 stage-boundary outcrop 34 Stagonosuchus 61 stamen 178–179, 183 Stangeriaceae 157 Staphyleaceae 198, 201 Staphylinidae 262, 283, 288 Staphyliniformia 287 Staphylinoidea 283, 286, 288 Staphylococcaceae 107 starfish 302 starfruit 194 starling 429 stationary echolocation emission 501 Stauromedusae 234 Stauroteuthidae 243 Staurozoa 75, 236 Stegostomatidae 321, 326 Steiper, Michael E. 482 stem group 38 stem teleost 64 stem-Metatheria 50 Stemonaceae 206 Stemonuraceae 183, 185 stem-orang 48 Stenopodidae 294 Stenopodidea 294 Stenostiridae 424 Stenotrachelidae 279 Stenotritidae 265, 267 Stercorariidae 433 Sterculiaceae 200 stereom 302 Stereophyllaceae 139 Sternidae 433 Sternotherus carinatus 398 Steropodon 58, 463 stickleback 44, 62, 65 Stilbaceae 179, 183 stilt 433 stinging wasp 264, 266 stingray 324 stink badger 505

stipule 183 stochastic clock 4 Stoliczkaia 393 stomata 162 Stomopneustidae 303 stone plant family 174 stonecrop family 174 stork 434, 437 Strabomantidae 359 Stramenopiles 117, 118 Strasburgeriaceae 188, 197, 201 strata 26 stratified phloem 166 stratified sedimentary succession 30 stratigraphic interval 34 Stratiomyidae 272, 275 Stratiomyomorpha 272, 274 stratotype 28 stratotype section 27 Streaked Tenrec 479 Streblidae 274 Strelitziaceae 204, 206 Strengulidae 51 Strenulagus 51 Strepsiptera 262 strepsipteran 262 Strepsirrhini 50, 480 strepsirrine 484 Streptococcaceae 107 Streptomycetaceae 107 Strigidae 452 Strigiformes 420, 436–437, 453 Strix nebulosa 451 strobili 157 stromatoporoid 76 strongylid parasite 249 Strongylida 249 Strongylocentrotidae 303, 304 Strongylocentrotus 325 Strongyloidea 247 strontium isotope 34 Strugnell, Jan M. 239, 242 Struthio camelus 412 Struthionidae 413 Struthioniformes 413 sturgeon 309, 328, 333 Sturnidae 425, 429 Stylidiaceae 180, 184 Styracaceae 178, 182, 185 suboscine 423–424, 430 succulent herb 208 sugarbird 429 Suidae 512 Suiformes 55, 514

Suina 511, 514 Sulfolobaceae 102 sumac 200 Sumatran Rhinoceros 509 sunbird 429, 454 Sunbittern 441, 443 sundew family 174 sunflower 177 Sungrebe 440, 442 Surianaceae 190, 192 suspension feeders 233 SVG 34 swan 416 sweet flag 203 sweet potato 177 sweetener 169 Swietenia 200 swift 419–421, 442, 456 Sylphornithidae 450 Sylviidae 424 Sylvioidea 424, 430 Symbiobacterium 107 Symbiobacterium thermophilum 112 symbiont 90, 92 Symmetrodonta 58 sympetalous corolla 178, 182 sympetaly 164 Symphyla 252 Symphyta 266 Symplocaceae 178, 182 sympodial growth 162 Synapsida 37, 59, 312, 376 syndactyl hind feet 467 Synechococcaceae 107, 112 Synneuridae 271 Synteliidae 283 Syringa 177 Syrphidae 274 Syrphoidea 274 Systema Naturae 251 Syzygium 194 T Tabanidae 272 Tabanomorpha 272, 274 Tabulata 235 tabulate coral 235 Tachinidae 274 Tachyglossidae 463 Tachyglossus 462 Tachyglossus aculeatus 462 tailless whip-scorpion 255 Takakia 143 Takakiaceae 140

547

548

INDEX

Takakiales 140 Takifugu 44, 62, 65 Takifugu rubripes 64 Talpa europaea 496 Talpidae 53, 479, 496 talus 134 Tamandua 476 Tamaricaceae 170, 174 tanager 429, 431 Tanyderidae 271, 275 tapaculo 424 taphonomic fi ltering 211 Taphrinomycetes 216 Taphrinomycotina 216 tapioca 194 tapir 510 Tapiridae 509 tapiriform 508 Tapiscia 201 Tapisciaceae 197–198, 201 Tapocyon 55 tarantula 255 Tardigrada 226 tardigrade 309 taro 203 tarpon 335 tarsier 482 Tarsiidae 480, 482, 484 Tarsipedidae 467, 469 Tarsipes rostratus 469 Tarsius eocaenus 484 Tasmanian Wolf 469 Taxaceae 159 taxic approach 37 Taxodiaceae 157 Tayassuidae 512 tea 177 Tecophilaeaceae 205 Teeling, Emma C. 499 Teiformata 386 teiformatan 386 Teiidae 384, 388 Teiioidea 384, 386, 388 Teinolophos 58, 463 Telegeusidae 282 Teleostei 309, 328–330, 335–337, 339, 344 telephone-pole beetle 278 Telmatobiidae 359 telome 159 Temnopleuridae 303 temnopleuroid 304 temnospondyl 62 temporal banding 24 Tenebrionidae 279, 288

Tenebrionoidea 279, 288 tenrec 44, 56–58, 471, 479–481, 496 Tenrecidae 479–480, 495 tentacle pocket 244 tentacular club 244 tentaculitid 75 Tephritidae 270, 274 Tephritoidea 274 Teresomata 370 termite 268 Termitidae 268 tern 432 Terrabacteria 107, 113 terrapin 398 Testudines 375–376, 400 Testudinidae 400 Testudinoidea 400 Tethytragus 56 Tetracarpaeaceae 170, 175 Tetrachondraceae 179 Tetraclitoidea 299 Tetracondraceae 183, 185 Tetramelaceae 190 Tetrameristaceae 178, 182, 185 Tetraodon 44, 62, 65 Tetraodon nigris 64 Tetraodontidae 65 Tetraodontiformes 66 Tetraonidae 416 Tetraphidaceae 140 Tetraphidales 141 Tetraphis 142 Tetrapoda 5, 37, 46, 59, 62, 66–68, 309–313, 348–349, 387 tetrasporangiate anther 161 Tetratomidae 279 Teuthoidea 244 Teviornis 62 Teviornis gobiensis 61 Thalassinidea 294 Thamnobryaceae 143 Thamnophilida 424 Thamnophilidae 426 Thaumaleidae 271 Theaceae 177–178, 182, 185 thecate hydroids 237 Theobroma 200 Theophrastaceae 178, 182 Theraphosidae 256, 258 Theraphosoidina 258 therapsid 462 Therevidae 272, 274 Theria 57–58, 460 Theridiidae 256, 258

Theriimorpha 58 Thermoanaerobacteriaceae 107, 112 Thermococcaceae 102 Thermococci 102 Thermoplasmata 102 Thermoplasmataceae 102 Thermoprotei 102 Thermotogaceae 109 Thermotogae 109, 112 Thermus 107 theropod 409 Thinocoridae 433 thoracic sclerite 272 thornback ray 324 thrasher 429 Thraupidae 424, 429 thread-horn fly 270 three-toed sloth 476 Throscidae 282 thrush 429 Thryonomyidae 491 thryopterid 500 Thuidiaceae 139 Thurniaceae 204 Thylacinidae 469 Thylacomyidae 467 thyme 177 Thymelaeaceae 198, 200 Thymus 177 Thyropteridae 500, 502 tick 255 Ticodendraceae 193 tiger beetle 278 Tilia 200 Tiliaceae 200 Tillandsia usneoides 203 Tilletia indica 216 Timaliidae 424 timber 173–174, 191, 194, 200 Timmia 143 Timmiaceae 140, 143 Timmiales 140 Tinamidae 413 Tinamiformes 413 tinamou 409, 412–414, 419, 423 Tiphiidae 265 Tipulidae 272 Tipulomorpha 271, 276 Tischlingerichthys viohli 64 Tityridae 426 Tiupampa 58 toad 44, 60, 62, 309 tobacco 177 Tofieldiaceae 206

Index

Tolypeutes matacus 475 tomato 177 Tomistoma 405 Tomistoma schlegelii 402 tomistomine 403, 405 Tommotian 75 toothcombed prosimian 50 toothed whale 512 Torpedinidae 322 Torpediniformes 322, 324 Torricelliaceae 180, 183, 185 Torridincolidae 284 tortoise 399 Tortula muralis 138 torus 161 toucan 448 toucan-barbet 447 Toxicofera 384, 386, 388 Toxopneustidae 303 TPI 261, 274 tracheal gill 281 Tracheophyta 133–137, 150 Trachyboa 391, 393, 395 Trachylina 234, 236 Trachyloma 142 Trachylomataceae 139 Trachymedusae 237 Trachypachidae 284 Trachyphonus 447 Tragulidae 512 Tragulus javanicus 511 transformational approach 37 Trautwein, Michelle D. 260 tree fern 155 Tree of Life 40, 46 tree shrew 44, 50–51, 56–58, 471, 495 trees 166, 173–174, 182, 191, 194 treeswift 456 Trematomus 341 Trematomus scotti 340 Tremoctopodidae 243, 245 Treubia 148 Treubiales 148 Triakidae 321–322, 325 Triakis 322, 349 Tribelaceae 180, 183 Tribolium casteneum 275 tribosphenic molar 462 Trichoceridae 271 Trichocoleaceae 147 Trichoptera 262 Trichostrongylidae 249 tricolpate pollen 162–163, 169, 175, 178 Triconodonta 58

triconodont 58 Trictenotomidae 279 Trigona 267 Trigoniaceae 189 trilobed aedeagus 287 Trilobita 251 trilobite 28, 31, 252 trimerous flower 203 Trionychidae 399 triploblast bilaterian 76 triploid endosperm 162 Triuridaceae 209 trnL 178 trnT-trnF 178 trnV-atpE 178 trochanter-femur depressor muscle 255 Trochilidae 455 Trochiliformes 420–421, 454 Trochodendraceae 169, 171, 173 Trochozoa 224 Trogidae 283 Troglodytidae 425, 429 Trogoniformes 420 Trogonophidae 384, 387 Trogossitidae 280 Tropaeolaceae 199 tropicbird 420, 436 Tropidophiidae 390–391, 393, 395 Tropidophis 391, 393, 395 Tropidophis feicki 390 trout 340 true fly 260, 270, 272 true leaf 137 true seal 504 trumpeter 440–441, 443 Trypanosomatidae 117 Tsaganomys 52 tuatara 309, 375–378, 383 tube-dwelling anemone 235 tubenosed bird 421 Tubulariidae 237 Tubulidentata 56, 472 tuffaceous bed 30 tulip poplar 161 tulip 205 Tullbergs classification 490 tuna 19 tunicate 45, 68 Tupaiidae 495 Turdidae 423, 425, 429 Turnicidae 433 Turniciformes 432 Turrilina alsatica 27 Turritoma acrea 70

turtle 9, 309, 375–378, 400 tusk shell 239, 241 twisted-wing insect 260 two-toed sloth 476 Tylenchina 248 Tylopoda 55 Typhaceae 204 Typhlonectidae 370 Typhlopidae 390–391, 393 Typhlops 393 Typhlops arator 390 typical owl 451 Tyranni 423–424, 426 Tyrannidae 426 Tyrannides 424, 426, 430 Tyto 451 Tytonidae 452 U U/Pb 31 Ulkenulastomys 52 Ulmaceae 190, 193 Uloboridae 256 ungaliophiine 395 Ungaliophis 393 ungulate 509 ungulatomorph 45 uniaperturate pollen 203 unicellular choanoflagellate 223 Unidentata 386 unidentatan 388 unisexual flower 162, 183 UPGMA 20 Uraeotyphlidae 370 Urediniomycetes 216 Urochordata 224, 226, 228 urochordate 309 urogenital aperture 466 Urolophidae 322, 325 Uropeltidae 390–391, 393, 395 Uropeltis 393 Uropeltoidea 391, 395 Uropygi 255 Urotrygonidae 322 Ursidae 505 Urticaceae 190, 193 urticalean rosid 193 Ustilaginomycotina 216 Ustilago maydis 216 V Vahliaceae 179, 182 Valerianaceae 184 valvate corolla aestivation 184

549

550

INDEX

vampire squid 245 Vampyromorpha 242, 245 Vampyroteuthidae 243 van Tuinen, Marcel 409, 419 vanga shrike 429 Vangidae 425 Varanidae 387 Varanoidea 386, 388 Varanopseidae 59 variable rate 41 vascular plant fossil 164 vascular plant 133, 142–143, 150, 153, 164 Vegavis 62 Vegavis iaai 61, 416 vegetable 177 Velloziaceae 206 vena cava 244 venom 386 ventral coiling 244 Verbenaceae 179, 183 Vermileonidae 272 Vermilingua 477 Verrucidae 299 Verrucomicrobia 113 Verrucomicrobium 113 Verrucomorpha 299 Vertebrata 8, 45, 66–69, 72, 224, 226, 228, 249, 309–310, 312–313, 433 vertebrate taxa 23 Vesperidae 281 Vespertilionidae 502 Vespertilionoidea 499–500, 502 Vespidae 264–265, 268 Vespoidea 265 vetigastropod 70 vetulicolian 68 vetulicystid 68 Vibrionaceae 108 vicariance hypothesis 23 Victoriapithecus 49, 485 Victoriapithecus macinnesi 48 Vidal, Nicolas 383, 390 Vieites, David R. 353, 365 Vincelestes 58 vinegar fly 273 Violaceae 189 Viperidae 390–391, 393, 394 Vireonidae 426 vireo 431 Viridiplantae 117, 119 Viscaceae 173 Vitaceae 170, 173–175, 188, 190, 192, 194

Vitales 162 Vitis vinifera 188 Vitreledonellidae 243, 245 Viverravidae 55 Viverridae 506 Vivianiaceae 198, 202 viviparous mammal 466 Vjushkovisaurus 61 Vochysiaceae 190, 195 volant landbird 419 volcanic ash 31 volcanic tuff 29 Vombatidae 468 vombatiform 466 Vombatiformes 468 W wading bird 419 wagtail 429 Wake, David B. 353, 365 Wake, Marvalee H. 353 Wallaces Line 449 walnut 193 Walrus 504 Wanganui Basin 30 Wangisuchus 61 warbler 429, 431 Warburgella rugulosa 28 wasp 45, 71–74, 260 water milfoil family 175 water plant 182 Water Shrew 495 waterbird 410, 421 waterfowl 409, 419 waxwing 429 weasel 504 weed 169 weevils 278, 286 Weigeltisauridae 60 Welwitschia 157, 159 Welwitschiaceae 159 Wemersoniellidae 239 West Indian shrew 495, 497 Westlothiana 62 Westlothiana lizziae 62 whale 53–54, 514 Whale Shark 325 Whatcheeria 62 whatcheeriid 62 whippomorph 55 Whippomorpha 55 whip-scorpion 255 whirligig 278, 286 whisk fern 133–134, 153–155

whistling duck 415 White Ibis 309 white-blooded fish 339 Widanelfarasia 56 Wiegmann, Brian M. 260, 270, 290 Wikström, Niklas 138, 146 Wilkinson, Mark 369 Winteraceae 167 Winterton, Shaun L. 290 witch hazel 175 wolf 504 Wollemia 159 wombat 468 wood thursh 423 wood-boring beetle 287 woodcreeper 424 woodpecker 419, 449 woody plant 182 wren 429, 431 wrinkled bark beetle 278 Wryneck 445, 448 X Xanthomonadaceae 108 Xanthorhysis 484 Xanthorrhoeaceae 206 Xantusiidae 384, 388 Xantusioidea 387 Xenarthra 56, 471–473, 475–477, 497 xenarthrale 475 xenarthran 56 Xenoanura 359 Xenochrophis 393 Xenodermatidae 391, 393, 394 Xenopeltidae 391, 395 Xenopeltis 393 Xenophidiidae 394 Xenopus 325, 349 Xenopus laevis 62 Xenosauridae 384, 386 Xenosauroidea 386 Xenoturbella 226 Xeronemataceae 205 Xiphosura 252, 258 Xyelidae 73 xylem 203 Xylocopini 268 Xylomyidae 272 Xylophagidae 272 Xylophagomorpha 272, 274 xyrid 211 Xyridaceae 204

Index

Y yams 205 Yangochiroptera 500, 501 yeast 19, 216 Yinochiroptera 499 yinochiropteran 501 Yinpterochiroptera 500 Yoon, Hwan Su 116 Young, Nathan M. 482 Younginiformes 60 Youngolepis 63 Ypresian 56 Yunnanozoon 68 Z Zaglossus 462 Zaglossus robustus 463

zalambdalestid 44–45, 50 zalambdodont molar 479 Zamiaceae 159 Zanobatidae 325 zebra 508 Zebra Shark 326 Zebrafinch 44, 62 Zebrafish 44, 62, 64 Zebrawood 194 Zhang, Peng 353, 365 zhelestid 44–45, 53 Zhongjianichthys 68 Zingiberaceae 204, 208 Zingiberales 204–206, 208, 211 Zoanthidea 235

zoanthid 235 zonal composite 31 Zooamata 54, 56 Zoopagomycotina 217 Zopheridae 279 Zorocratidae 256 Zosteraceae 204, 206, 208 Zosteropidae 424 zygodactyl feet 451 Zygogramma 260 zygomasseteric system 490, 492 zygomorphic flower 183 zygomycetes 216 Zygomycota 215, 217 Zygophyllaceae 192

551

Timetree of life

The Timetree of Life

The Abode of Life

equations of life

Equations of Life

Life of Pi

Life of Pi

The Abode of Life

The Secret of Life

Time of Your Life

The Race of Life

Love Of Life

Battle of Life

The Life of Objects

The Abode of Life

Fact of Life #31

Book of Life

The Colour Of Life

The Abode of Life

Love of Life

Knowledge of Life

The Life of Death

Facts of Life

Circle of Life

Tree of Life

Time of Your Life

Diversities Of American Life

Music of a Life

The Book Of Life

Timetree of life

The Timetree of Life

The Abode of Life

equations of life

Equations of Life

Life of Pi

Life of Pi

The Abode of Life

The Secret of Life

Time of Your Life

The Race of Life

Love Of Life

Battle of Life

The Life of Objects

The Abode of Life

Fact of Life #31

Book of Life

The Colour Of Life

The Abode of Life

Love of Life

Knowledge of Life

The Life of Death

Facts of Life

Circle of Life

Tree of Life

Time of Your Life

Diversities Of American Life

Music of a Life

The Book Of Life

Recommend Documents