00prelims.qxd
28/11/03
3:40 pm
Page i
CYCAD CLASSIFICATION CONCEPTS AND RECOMMENDATIONS
00prelims.qxd
28/11/03
3:40 pm
Page ii
00prelims.qxd
28/11/03
3:40 pm
Page iii
Cycad Classification Concepts and Recommendations Edited by
Terrence Walters and
Roy Osborne
CABI Publishing
00prelims.qxd
28/11/03
3:40 pm
Page iv
CABI Publishing is a division of CAB International CABI Publishing CAB International Wallingford Oxfordshire OX10 8DE UK
CABI Publishing 875 Massachusetts Avenue 7th Floor Cambridge, MA 02139 USA
Tel: +44 (0)1491 832111 Fax: +44 (0)1491 833508 E-mail:
[email protected] Web site: www.cabi-publishing.org
Tel: +1 617 395 4056 Fax: +1 617 354 6875 E-mail:
[email protected] ©CAB International 2004. All rights reserved. No part of this publication may be reproduced in any form or by any means, electronically, mechanically, by photocopying, recording or otherwise, without the prior permission of the copyright owners. A catalogue record for this book is available from the British Library, London, UK. Library of Congress Cataloging-in-Publication Data Cycad classification : concepts and recommendations / edited by Terrence Walters and Roy Osborne. p. cm Includes bibliographical references (p. ). ISBN 0-85199-741-4 (alk. paper) 1. Cycads--Classification. I. Walters, Terrence, 1955. II. Osborne, Roy. III. title QK494.C93 2004 585´.9´012--dc21 2003010044 ISBN 0 85199 741 4 Typeset by MRM Graphics Ltd, Winslow, Bucks Printed and bound in the UK by Cromwell Press, Trowbridge
00prelims.qxd
28/11/03
3:40 pm
Page v
Contents
Contributors About the Editors Preface Acknowledgements 1. ‘We Hold these Truths …’ Terrence Walters, Roy Osborne and Don Decker
vii ix xi xv 1
2. Saving Ghosts? The Implications of Taxonomic Uncertainty and Shifting Infrageneric Concepts in the Cycadales for Red Listing and Conservation Planning 13 John Donaldson 3. Character Evolution, Species Recognition and Classification Concepts in the Cycadaceae 23 Ken D. Hill 4. Morphological Characters Useful in Determining Species Boundaries in Cycas (Cycadaceae) Anders Lindström
45
5. Comments on Cycas, Dyerocycas and Epicycas (Cycadaceae) 57 Chia-Jui Chen, Ken D. Hill and Dennis Wm. Stevenson 6. Classification Concepts in Encephalartos (Zamiaceae) Piet Vorster
69
v
00prelims.qxd
28/11/03
vi
3:40 pm
Page vi
Contents
7. Classification Concepts in Macrozamia (Zamiaceae) from Eastern Australia Paul I. Forster 8. Classification Concepts in Ceratozamia (Zamiaceae) Loran M. Whitelock
85
95
9. Relationships and Phytogeography in Ceratozamia (Zamiaceae) 109 Andrew P. Vovides, Miguel A. Pérez-Farrera, Dolores González and Sergio Avendaño 10. A Morphometric Analysis of the Ceratozamia norstogii Complex (Zamiaceae) 127 Miguel A. Pérez-Farrera, Andrew P. Vovides, Luis Hernández-Sandoval, Dolores González and Mahinda Martínez 11. Hypotheses on the Relationship between Biogeography and Speciation in Dioon (Zamiaceae) 137 Timothy J. Gregory and Jeffrey Chemnick 12. Molecular Phylogeny of Zamia (Zamiaceae) Paolo Caputo, Salvatore Cozzolino, Paolo De Luca, Aldo Moretti and Dennis Wm. Stevenson
149
13. Systematics of Meso-American Zamia (Zamiaceae) Bart Schutzman
159
14. Zamiaceae of Bolivia, Ecuador and Peru Dennis Wm. Stevenson
173
15. In Search of the True Tree: Guidelines for Classification Roy Osborne and Terrence Walters
195
Appendix 1: The World List of Cycads Ken D. Hill, Dennis Wm. Stevenson and Roy Osborne Appendix 2: Glossary of Terms Encountered in Cycad Systematics Roy Osborne and Terrence Walters Index
219
237
259
00prelims.qxd
28/11/03
3:40 pm
Page vii
Contributors
Avendaño, S., Instituto de Ecologia A.C., Apartado Postal 63, Xalapa, Veracruz 91000, Mexico. Caputo, P., Dipartimento di Biologia Vegetale, Università degli Studi di Napoli Federico II, Via Foria 223, 80139 Napoli, Italy. Chemnick, J., Ganna Walska Lotusland, 695 Ashley Road, Santa Barbara, California 93108, USA. Chen, C.-J., Institute of Botany, Chinese Academy of Sciences, 20 Nanxincum, Xiangshan, Beijing 100093, China. Cozzolino, S., Dipartimento di Biologia Vegetale, Università degli Studi di Napoli Federico II, Via Foria 223, 80139 Napoli, Italy. Decker, D., Decker & Associates, Inc., PO Box 222153, Carmel, California 93923, USA. Donaldson, J., Kirstenbosch Research Center, National Botanical Institute, Private Bag X7, Claremont 7735, South Africa. Forster, P.I., Queensland Herbarium, Environmental Protection Agency, Brisbane Botanic Gardens, Mt Coot-tha Road, Toowong, Queensland 4066, Australia. González, D., Instituto de Ecologia A.C., Apartado Postal 63, Xalapa, Veracruz 91000, Mexico. Gregory, T.J., Montgomery Botanical Center, 11901 Old Cutler Road, Miami, Florida 33156-4242, USA. Hernández-Sandoval, L., Facultad de Biologia, Universidad Autónoma de Querétaro, Centro Universitario, Cerro Las Campanas S/N, Querétaro 76010, Mexico. Hill, K.D., Royal Botanic Gardens, Mrs Macquaries Road, Sydney 2000, Australia. Lindström, A., Nong Nooch Tropical Botanical Garden, 34/1 Sukhumvit Highway, Najomtien, Sattahip, Chonburi 20250, Thailand. De Luca, P., Dipartimento di Biologia Vegetale, Università degli Studi di Napoli Federico II, Via Foria 223, 80139 Napoli, Italy. vii
00prelims.qxd
viii
28/11/03
3:40 pm
Page viii
Contributors
Martínez, M., Facultad de Biología, Universidad Autónoma de Querétaro, Centro Universitario, Cerro Las Campanas S/N, Querétaro 76010, Mexico. Moretti, A., Dipartimento di Biologia Vegetale, Università degli Studi di Napoli Federico II, Via Foria 223, 80139 Napoli, Italy. Osborne, R., PO Box 244, Burpengary, Queensland 4505, Australia. Pérez-Farrera, M.A., Escuela de Biología, Universidad de Ciencias y Artes de Chiapas (UNICACH), Calzada Samuel León Brindis 151, Tuxtla Gutiérrez, Chiapas 29000, Mexico. Schutzman, B., Environmental Horticulture Department, University of Florida, 1525 Fifield Hall, Gainesville, Florida 32611-0670, USA. Stevenson, D.W., Institute of Systematic Botany, New York Botanical Garden, Bronx, New York 10458, USA. Vorster, P., Botany Department, University of Stellenbosch, Private Bag X1, Matieland 7602, South Africa. Vovides, A.P., Instituto de Ecología A.C., Apartado Postal 63, Xalapa, Veracruz 91000, Mexico. Walters, T., Montgomery Botanical Center, 11901 Old Cutler Road, Miami, Florida 33156-4242, USA. Whitelock, L.M., 4524 Toland Way, Los Angeles, California 90041, USA.
00prelims.qxd
28/11/03
3:40 pm
Page ix
About the Editors
Terrence Walters is the executive director of Montgomery Botanical Center, a botanical garden in Miami, Florida, that concentrates on scientifically documented population-based collections of cycads and palms. Since 1990, he has conducted numerous field expeditions to parts of Asia, Africa and the Americas to investigate and document the cycad flora of the world. He is on the board of directors of the Cycad Society, a member of IUCN’s Cycad Specialist Group and on the research faculty of Florida International University. (Photography by Mary Andrews.)
ix
00prelims.qxd
x
28/11/03
3:40 pm
Page x
About the Editors
Roy Osborne has been studying and growing cycads in Africa and Australia for more than 20 years. He is a member of the IUCN’s Cycad Specialist Group, and founder and first President of the Cycad Society of South Africa. Now living in Brisbane, Australia, he has published more than 100 scientific papers, books and book chapters and has participated in major international conferences on cycad biology. (Photography by Mary Andrews.)
00prelims.qxd
28/11/03
3:40 pm
Page xi
Preface
On 7 April 2002, the Cycad Classification Concepts (CCC) Workshop was convened at Montgomery Botanical Center in Miami, Florida, USA. Seventeen of the world’s leading authorities on cycad systematics were invited to participate in the workshop and to submit manuscripts for this volume. Fifteen of these systematists submitted manuscripts and 14 were able to attend the 3-day CCC Workshop. The purpose of the CCC Workshop was to develop a suite of classification guidelines in support of the advancement of an internationally accepted and stable evolutionary classification system for taxa in the Cycadales. Increased research activity in the field of cycad systematics has led in some cases to increased confusion. As researchers across the globe pursue the many new lines of inquiry provided by technological advances of the past two decades (e.g. DNA sequencing, random amplified polymorphic DNA analysis, etc.), focus on consensus for how the approximately 300 species of cycads should be classified has become clouded. There is an urgent need for guidelines that all cycad systematists can follow in the designation of species, species boundaries and species groupings. The CCC Workshop provided the venue for the development of these guidelines. Although workshops with a similar purpose have been held to examine critically the systematics of other plant groups, the CCC Workshop was uniquely designed using progressive business methodologies. Five arenas were identified as necessary for the planning and management of this event. The Personnel Arena dealt with the subject of who would be involved as CCC Participants, who would be on the CCC Support Team and who would be in leadership roles during the Workshop process. The Site Arena dealt with everything concerning the facilities required for the Workshop – such as rooms for the various events and work sessions, transportation, housing, furniture, catering and audio-visual equipment. The Operations Arena dealt with identifying and taking those actions required to xi
00prelims.qxd
xii
28/11/03
3:40 pm
Page xii
Preface
produce the major product of the Workshop – this volume. The Planning Arena dealt with determining all of the tasks required, their flow, their content and their sequence – from the overall purpose and concept of the Workshop to the minute details associated with the organization and objectives of the Workshop sessions themselves. Finally, the Management Arena dealt with how all of the above would be led and managed. The first step was to bring in a management consultant, Don Decker, to support the Management Arena objectives and to oversee development of the other four arenas. The next steps were to articulate the purpose, or reason, for having the CCC Workshop, and to determine the products, or results, required to meet the purpose successfully. The overall process of actions that would be required to obtain the products was outlined and then the functioning capabilities, or resources, required for the process were identified. These processes and the development of the above five arenas provided the overall planning and execution structure for the CCC Workshop. Bringing together a group of world-renowned cycad systematists representing several countries, cultures and languages for consensus building can be difficult. That this event was successful is a tribute to the considerable work that took place prior to, during and after the Workshop by the CCC Support Team and the CCC Participants. The CCC Participants were 14 of the world’s leading and most respected cycad systematists. Paolo Caputo from the Università degli Studi di Napoli Federico II in Italy represented one of the largest concentrations of cycad systematists at any one institution in the world. The Naples cycad group has worked extensively on New World taxa. Participants representing Asia included Chia-Jui Chen from the Institute of Botany in Beijing, China, an expert on the cycads of China, and Anders Lindström, the cycad curator at Nong Nooch Tropical Gardens in Thailand. Lindström is one of the leading experts on the cycads of Thailand. Cycas lindstromii was named in his honour. John Donaldson, from the National Botanical Institute, and Piet Vorster, from the University of Stellenbosch, were the workshop’s representatives from South Africa. Donaldson is Chairman of the IUCN (World Conservation Union) Cycad Specialist Group. Vorster is currently the President of the Cycad Society of South Africa and is an authority on the African genus, Encephalartos. Due to the large number of active cycad systematists in Australia, this country was well represented at the workshop. Attendees included Paul Forster of the Queensland Herbarium, Australia’s expert on Macrozamia. Ken Hill, from the Royal Botanic Gardens in Sydney, is the world’s expert on the taxonomically difficult genus Cycas. Roy Osborne, who currently resides in Queensland and formerly lived in South Africa, began the development many years ago of the world list of cycads. Hill and Osborne recently published the authoritative work Cycads of Australia. Andrew Vovides directs the National Cycad Collection of Mexico, and has developed the concept of local conservation of native cycads by initiating projects in which local villagers create nurseries to grow native cycads from sustain-
00prelims.qxd
28/11/03
3:40 pm
Page xiii
Preface
xiii
able seed harvests. Americans Tim Gregory and Jeff Chemnick continue to undertake extensive systematic fieldwork in Mexico. Their commitment to walking up every canyon in search of each and every population of a species is to be admired. Other participants from the United States included Bart Schutzman of the University of Florida, the editor of The Cycad Society Newsletter and expert on Meso-American Zamia. Dennis Stevenson of the New York Botanical Garden is the leading authority on Central and South American taxa and has published extensively on evolutionary concepts in the Cycadales. Loran Whitelock from California, after a decade of fieldwork, research and writing, has recently completed what will become the major reference work on the cycad flora of the world – The Cycads. Ceratozamia whitelockiana and Encephalartos whitelockii are named in recognition of Whitelock’s extensive research on the world’s cycad flora. The first session of the CCC Workshop, held on 7 April, created the opportunity for each CCC Participant to give a 20-minute oral presentation of their professional views on cycad classification concepts, systematics and taxonomy. This 1-day work session was organized as a symposium (CCC Symposium) that included invited guests. The second work session, conducted on day 2, focused on elucidating the beliefs and philosophies that the participants held to be true concerning cycad systematics. Also on day 2, during work session three, Katherine Kron of Wake Forest University presented a discussion on a relatively new and somewhat controversial approach to plant nomenclature called ‘phylocode’. On the third day of the Workshop, the fourth and fifth work sessions required that the CCC Participants come to alignment on a suite of classification concepts or guidelines that they, as a group, would support and encourage the use of presently and in the future. In this volume, Chapter 1 presents why the CCC Workshop was convened and the beliefs, or working hypotheses and assumptions, that the CCC Participants hold to be true for cycad classification. This chapter resulted from work sessions two and three. The final chapter, Chapter 15, based on the products obtained from work sessions four and five, summarizes the classification guidelines that the CCC Participants have agreed to follow, support and encourage the use of to produce a universally accepted stable classification system for the Cycadales. Prior to the Workshop, each CCC Participant submitted a manuscript to the editors. These manuscripts were detailed discussions of the oral presentations presented by the participants during the CCC Symposium (work session one). These manuscripts constitute Chapters 2–14 of this volume. In Chapter 2, John Donaldson discusses the practical need for a durable classification system in the Cycadales when dealing with cycad conservation issues and planning. In Chapters 3 and 4, Ken Hill and Anders Lindström critically examine the usefulness of various characters for defining species and species concepts within Cycas. Three of the CCC Participants, Chia-Jui Chen, Ken Hill and Dennis Stevenson, report on a study of a recently described genus in the Cycadaceae in Chapter 5. They present a methodology for how cycad systematists should critically evaluate proposed new taxa. Cycad experts Piet Vorster,
00prelims.qxd
xiv
28/11/03
3:40 pm
Page xiv
Preface
Paul Forster and Loran Whitelock present their individual thoughts on infrageneric classification concepts for African Encephalartos taxa, Australian Macrozamia taxa and New World Ceratozamia taxa in Chapters 6, 7 and 8, respectively. Chapters 9 and 10 were submitted by the participants from Mexico, Andrew Vovides and Miguel A. Pérez-Farrera. Unfortunately, Miguel was not able to attend the Workshop. These two researchers evaluate the usefulness of characters for defining species and species complexes within Ceratozamia. Tim Gregory and Jeff Chemnick develop in Chapter 11 an exciting hypothesis that extant species of Dioon are the result of rapid evolution in a dynamic group of plants. Two of the participants, Paolo Caputo and Dennis Stevenson, along with their colleagues, report on a molecular study in Chapter 12 that examines the usefulness of molecular and morphological data sets when trying to develop a phylogenetic tree for species of Zamia. Results from Bart Schutzman’s extensive and detailed morphological studies on Meso-American species of Zamia are given in Chapter 13. Dennis Stevenson, in Chapter 14, presents a monograph on the Zamiaceae from Bolivia, Ecuador and Peru. His chapter illustrates many of the guidelines the participants discussed during the last day of the workshop concerning content, style and format for the type of publication resulting from floristic cycad research. Two appendices are included in this volume. Firstly, the ongoing discovery of new species and the continuous refinements to the taxonomy of the alreadyknown taxa mean that the list of ‘officially recognized’ taxa needs to be timeously revised. Appendix 1 gives details of the ‘World List of Cycads’ at the time the manuscript for this volume was submitted to the publisher and is based on ‘The Cycad Pages’ website (http://plantnet.rbgsyd.nsw.gov.au/PlantNet/cycad). Secondly, the interdisciplinary nature of work on cycad systematics has led to a large and complex vocabulary of terms, the precise meanings of which are sometimes obscure and occasionally misused. Appendix 2 provides a glossary of these terms, drawn up after extensive consultations with specialists, and amplified where possible with cycad-specific examples. For consistency with author citations for taxa, we have followed the International Plant Names Index (IPNI Website: http://www.ipni.org/ index.html) for the chapters and appendices. In Chapters 1–15, authors’ names are unabbreviated. They are cited when the taxon first appears within a chapter and are also cited when appropriate in figure captions and tables. For Appendix 1, authors’ names are abbreviated. Taxa known to be distinct by the authors of each chapter, but as yet not ‘officially’ published, are indicated by double quotes in Chapters 1–15. The work presented in this volume is not only a report on the current state of affairs in cycad classification, but also highlights areas of difficulty and leads to guidelines for meaningful future advances. We hope it will become a widely used reference for the benefit of all cycad researchers, enthusiasts, conservationrelated public and private agencies and students of plant systematics. Terrence Walters and Roy Osborne
00prelims.qxd
28/11/03
3:40 pm
Page xv
Acknowledgements
The directors, members, staff and volunteers of Montgomery Botanical Center (MBC) deserve thanks and gratitude for their commitment to planning, hosting and sponsoring the Cycad Classification Concepts (CCC) Workshop. The Workshop was the first of its kind held at MBC, and therefore demanded an extremely fast learning curve by the MBC Team. We hope they will consider organizing and hosting CCC Workshop II in the future. The Workshop was 11 months in the planning stage. Numerous meetings were held during this period to define clearly the purpose, expected products and agenda for the Workshop. Tim Gregory and William Tang provided much needed guidance during the early planning stages. Don and Sonja Decker are thanked for hosting a pre-Workshop planning session at their home in December 2001. At this 1-day session, Don Decker, Tim Gregory, Deena Walters and Terrence Walters clearly identified the purpose and the products expected from the Workshop, as well as plotting the 3-day agenda. Jeff Chemnick, Tim Gregory and Loran Whitelock provided encouragement and support throughout the entire planning period for the Workshop. Their input was much appreciated. The CCC Participants’ commitment to the success of the Workshop was greatly valued, given their own responsibilities and time constraints. They were truly an amazing group of individuals with whom to collaborate. The CCC Support Team oversaw the planning and logistics of the Workshop beginning in April, 2001. Jean Stark of Stark Connections took care of all the details associated with the participants’ travel and housing needs before, during and after the Workshop. Don Decker of Decker & Associates was the management consultant and leader for the Workshop. Katherine Kron of Wake Forest University was kind enough to lecture and field numerous questions on ‘phylocode’ during one of the work sessions. Evelyn Young planned and coordixv
00prelims.qxd
xvi
28/11/03
3:40 pm
Page xvi
Acknowledgements
nated all of the on-site meals and events at MBC during the 3 days. Larry Noblick coordinated the audio-visual equipment and supplies required for the Workshop. Lee Anderson oversaw facility preparations for all of the events each day during the Workshop. Deena Walters and Mary Andrews documented the Workshop photographically. Mayna Hutchinson created magnificent cycad arrangements for all of the facilities used at MBC during the Workshop. Barbara Judd, Sue Katz and Eric Shroyer were scribes for the break-out groups. All of the above individuals, as well as their own support teams, ensured that every aspect of the Workshop ran smoothly and according to the agenda. The CCC Workshop was supported through grants from the Bressler Foundation, Ajax Foundation and General Mills Foundation. Libby and By Besse, Judith and Richard Bressler, Tim Gregory, Eileen and Loyd Kelly and Linda and Mark Smith were significant contributors to the Workshop. The commitment of these foundations and individuals to the Workshop was very much appreciated. The officers, directors and members of the Central Florida Palm & Cycad Society are acknowledged for providing support towards the publication of this volume. A number of individuals were involved in the preparation and publication of this volume. The Montgomery Botanical Center directors, members and staff gave the editors the time and resources to prepare the manuscript. The first editor gratefully acknowledges Deena Walters for graphic design and illustration production assistance and, most of all, for her support during the preparation of the manuscript. The second editor similarly acknowledges the support of the Osborne family and friends. Finally, we wish to express our thanks to Tim Hardwick and the many individuals at CAB International who have so professionally managed the publication aspects of this volume.
Cycad Classification Concepts Workshop support team seated from left to right: Terrence Walters; Jean Stark; Don Decker and Katherine Kron. (Reprinted by permission of Montgomery Botanical Center, Miami, Florida, USA.)
00prelims.qxd 28/11/03 3:40 pm Page xvii
Acknowledgements
xvii
Cycad Classification Concepts Workshop participants seated from left to right: Andrew P. Vovides; Ken D. Hill; Chia-Jui Chen; Roy Osborne; Paolo Caputo; John Donaldson; Loran Whitelock; Paul I. Forster; Dennis Wm. Stevenson; Jeffrey Chemnick and Piet Vorster. On the floor in front: Anders Lindström; Bart Schutzman and Timothy J. Gregory. (Reprinted by permission of Montgomery Botanical Center, Miami, Florida, USA.)
00prelims.qxd
28/11/03
3:40 pm
Page xviii
ch01.qxd
28/11/03
3:42 pm
Page 1
‘We Hold these Truths …’
1
Terrence Walters,1 Roy Osborne2 and Don Decker3 1Montgomery
Botanical Center, Miami, Florida, USA; 2PO Box 244, Burpengary, Queensland, Australia; 3Decker & Associates, Inc., Carmel, California, USA
Abstract In order to develop classification guidelines for the Cycadales, a workshop was held in April 2002, at Montgomery Botanical Center in Miami, Florida, USA. Fourteen internationally-renowned cycad systematists spent 3 days identifying and developing guidelines that would provide a stable, practical and informative classification scheme for cycads. The participants agreed that convening such a workshop was vital, timely and necessary to produce a universally accepted evolutionary classification for the Cycadales in the near future. Before developing the guidelines, the participants first needed to identify the assumptions, or beliefs, that they hold to be true about cycad classification. These beliefs are presented under three categories: (i) beliefs about biological relationships; (ii) beliefs about what systematists can and should do in order to understand biological relationships; and (iii) beliefs about what cycad systematists can and should do in order to understand relationships in the Cycadales.
Cycad Classification Concepts Workshop The field of cycad systematics, which focuses on all members of the plant order Cycadales, has seen a flurry of activity during the past 20 years. New species are being discovered and described on an annual basis. Existing species circumscriptions are being critically tested for their scientific soundness. Familial and generic circumscriptions and relationships are being re-evaluated by a number of laboratories worldwide. Certain key developments in recent years (e.g. advances in systematic technologies and tools; ease of international travel, including access to countries previously unavailable to systematists; horticultural demand for rare ©CAB International 2004. Cycad Classification: Concepts and Recommendations (eds T. Walters and R. Osborne)
1
ch01.qxd
28/11/03
2
3:42 pm
Page 2
T. Walters et al.
cycads; recognition of the rare and endangered status of cycads; and an urgency for cycad conservation in many countries) have collectively stimulated cycad systematists to try to better understand and manage the taxonomy of the world’s cycad flora. Today, field and laboratory equipment can quickly generate massive amounts of systematic data, which often far surpass the immediate needs of systematists. This is particularly true for molecular data, which are being analysed in a multitude of ways with a plethora of user-friendly software programs. Consequently, systematists sometimes find it difficult to decide which analyses are appropriate for their work and how to interpret the hundreds of statistical summaries produced from these computer programs. Also, with these wonderful new opportunities and ever-increasing knowledge of cycads, it is easy to lose sight of the ultimate mission, which is to provide a universally accepted, consistent and informative evolutionary classification scheme. Although scientists do not believe that any truth can be exactly known, it is the purpose of science to approach truths as closely as possible. Scientists are forced to perform this work in an unsteady grounding of assumptions. These assumptions are not self-evident, but arise from the observations and experimentations of previous researchers. So, for any particular scientific field, there is a collection of assumptions, or beliefs, based on previous work that forms the framework for further discovery. As the scientific method proceeds, these assumptions are subject to change, usually in the form of minor modifications but sometimes in the form of radical reassessment. But, whatever insights future inquiry may bring, current hypotheses and guidelines for future research must be rooted in presently held assumptions. The major objectives of this volume are to enumerate the currently held assumptions, or beliefs, in the field of cycad systematics, and to present guidelines for future systematic work within the Cycadales. These concepts were fleshed out during a Cycad Classification Concepts (CCC) Workshop held in April 2002, at Montgomery Botanical Center in Miami, Florida. The CCC Workshop provided a forum for cycad systematists to ‘regroup’ and clarify as a team what they believe to be true (the best working assumptions) and important in the realm of cycad systematics. The participants then went a step further, agreeing on a suite of guidelines that they would follow in support of actualizing the team’s beliefs when engaging in future research (see Osborne and Walters, Chapter 15 this volume). The participants agreed not only to follow these guidelines in their own systematic studies, but also to encourage the global use of these guidelines by all cycad systematists and students. This chapter attempts to record the beliefs raised by the participants during the CCC Workshop, whereas the final chapter enumerates the proposed guidelines for developing a useful, evolutionary-based classification system for cycads. Before presenting the beliefs, it is necessary to provide some background on cycads and systematics and to explain some terms that the reader will encounter either in the list of beliefs or in other chapters of this volume.
ch01.qxd
28/11/03
3:42 pm
Page 3
‘We Hold these Truths …’
3
What is a Cycad? Cycads are an ancient group of seed plants that evolved in the Carboniferous or early Permian, some 280 million years ago (Norstog and Nicholls, 1997). They reached their zenith of abundance and diversity in the Mesozoic era. Cycads are one of four groups (cycads, ginkgos, conifers and gnetophytes) that are collectively and commonly referred to as gymnosperms. The Cycadales (the order containing all cycad families) is considered to be monophyletic. A monophyletic group is composed of an ancestor and all of its descendants based on a suite of shared derived characters, called synapomorphies. Some synapomorphies within the Cycadales include girdling leaf traces, a specialized pattern of vascular bundles in the petiole, distinctive meristems, buffer cells surrounding the archegonium, and the presence of mucilage canals, methylazoxymethanol glycosides and the non-protein amino acid BMAA (β-n-methylamino-L-alanine). Coralloid roots (specialized roots that host cyanobacteria) are found in all cycad taxa. Cycads also bear cataphylls, which are scale-like leaves that serve to protect the apical meristem. Cycad reproductive structures typically occur in cones, with each strobilus consisting of an axis and a series of spirally arranged megasporophylls (‘leaves’ bearing ovules) or microsporophylls (‘leaves’ bearing pollen sacs). All cycads are dioecious, with male and female reproductive structures on separate plants. Insects appear to be the primary vectors for pollination, although wind may be a factor for some genera (see discussion by Grobbelaar, 2002). Although not fully substantiated yet, evidence is accumulating to suggest coevolutionary processes between cycads and their pollinators. Once these processes are uncovered, resulting data will probably have a significant impact on how cycad taxa are classified. All genera except Cycas Linnaeus form a determinate female cone. In Cycas, the female ‘cones’ are indeterminate. Ovules are borne on loosely arranged whorls of megasporophylls (for an interpretive discussion on female ‘cones’ in Cycas, see Norstog and Nicholls, 1997). Cycad seeds usually have a brightly coloured, fleshy outer layer called the sarcotesta that encourages dispersal by animals. Birds, rodents and probably many other animals disperse cycad seeds by digesting the sarcotesta and dropping the stony layer and its contents away from the mother plant (Hill and Osborne, 2001). Seeds of some species of Cycas have a thick layer of spongy tissue, instead of the usual fleshy layer. This spongy layer allows these seeds to remain buoyant and viable for long periods of time in salt water. This may explain the wide distribution of this genus compared with the narrower ranges of other cycad genera. With the exception of Cycas, all cycad genera are restricted to single landmasses (Jones, 2002). Encephalartos Lehmann and Stangeria T. Moore occur only on the continent of Africa. Bowenia Hooker ex Hooker filius, Lepidozamia Regel and Macrozamia Miquel are endemic to Australia. Microcycas (Miquel) A. de Candolle is restricted to the island of Cuba. Ceratozamia Brongniart, Chigua D.W. Stevenson, Dioon Lindley and Zamia Linnaeus are endemic New World genera. Cycas is found
ch01.qxd
28/11/03
4
3:42 pm
Page 4
T. Walters et al.
in subtropical and tropical countries of the Old World that have Pacific or Indian Ocean coastlines and in neighbouring countries. Although cycad taxa are widely distributed in subtropical and tropical regions worldwide, extant populations are often widely disjunct. A cycad population is frequently found as an isolated pocket of individuals quite far removed from other such pockets. A major dilemma that faces today’s cycad systematists is understanding the evolutionary histories and futures of these populations. Part of the problem is determining whether these populations have been artificially separated because of human fracturing of the habitat, or are naturally occurring entities that are either gradually going extinct, are restricted to a very specialized niche, or are continuing to evolve as separate entities.
Cycad Taxonomy and Systematics Taxonomy is the process of circumscribing and assigning scientific names to the diversity of taxa, and then ordering this diversity into an appropriate classification system. In the realm of biology, a ‘taxon’ (plural ‘taxa’) is a group of individuals given a proper name or a group that could be given a proper name. For example, the taxon Dioon includes all named and as yet unnamed groups of individuals within this genus. An important aim of the cycad systematist is to describe and name only ‘natural taxa’ and to place these in a classification system that represents the order of nature. A natural taxon is a taxon that exists in nature independent of human ability to perceive it. It can be discovered, but not invented (Wiley, 1981). The same assumption can be applied to the order of nature, i.e. that it can be discovered for what and how it is. The basic assumption for biological order is that it is based on reproductive ties (genealogy) as they are affected by the process(es) of evolution. Classification is the process of organizing knowledge so that it facilitates communication and comprehension. The objectives of classification are: (i) to define and distinguish among ‘kinds’; and (ii) to position these kinds in a system that reflects their natural relationships and imparts information about these kinds. A classification system is a human construct that attempts to make natural order comprehensible to the human mind. The classification of biological organisms has its own language and rules of language use. For assigning a taxonomic name and having the name recognized by the botanical community, cycad systematists must follow recommendations outlined in the most recent edition of the International Code of Botanical Nomenclature (ICBN; Greuter et al., 2000). These recommendations are built on a hierarchical system of classification wherein each level of the hierarchy is referred to as a distinct rank. Typical ranks of use in the field of cycad systematics start with the all-inclusive ‘order’ (Cycadales) and move down to increasingly less inclusive ranks such as family (e.g. Zamiaceae), genus (e.g. Macrozamia), section (e.g. Parazamia) and species (e.g. Macrozamia lucida). The basic rank of species holds a special place in terms of the usefulness and importance of bio-
ch01.qxd
28/11/03
3:42 pm
Page 5
‘We Hold these Truths …’
5
logical entities to humanity; therefore, most cycad systematists undertake studies at the species level. Cycad systematics is the study of the cycad diversity that exists on earth today and the evolutionary history of this diversity. One of the main objectives is to convey knowledge about the genealogical relationships among cycad taxa in a hierarchical system of classification. The field of cycad systematics often requires that systematists have knowledge from many other scientific disciplines, such as taxonomy, morphology, ecology, molecular biology, pollination biology, anatomy, embryology, genetics, physiology, phytochemistry and palaeontology, so we are better able to uncover the true genealogical relationships among taxa. The number of described cycad species has almost doubled since 1985. Today, over 300 species are known (see Hill et al., Appendix 1 this volume) and many researchers believe the number may reach as many as 400 species when all potential cycad habitats have been investigated and taxonomic studies have been completed. Exactly what constitutes a cycad species remains unclear. Defining what makes a species is not a problem limited to cycad taxonomists, but is a basic source of consternation throughout the biological world. Generally, delimitation of a population or suite of populations as a new species is based on the training, background, knowledge and the basic scientific philosophy of the describer. No unified concept is in place to guide cycad systematists in defining and circumscribing new species. A variety of species concepts are used throughout the biological world. One of these, the biological species concept, does not work particularly well with cycads. The major premise of the biological species concept is that individuals within a species, when tested, are interfertile, while interspecific individuals are not. However, clearly defined and widely accepted species within a number of cycad genera can produce viable offspring with one or more other species in the same genus (Norstog and Nicholls, 1997). Consequently, cycad systematists generally agree that interspecific fertility, when tested, is just another character for systematic studies, and that the character of interspecific sterility should not be unduly weighted in the determination of cycad species. Moreover, the determination of the production of fertile offspring from putative hybrids is not practical for those species, like cycads, with long life cycles. Another out-of-favour species concept for cycad taxonomists is the phenetic species concept. This concept defines a species based on the overall similarity of its individuals combined with a significant gap in variation when these individuals are compared with individuals of another species. In practice, this qualitative approach does not always define natural taxa (Judd et al., 1999). The CCC Workshop participants agreed that the most common (unstated, but de facto) species concept in use by cycad systematists is what they termed a ‘morphogeographic’ species concept. This concept recognizes the importance of both morphological characters and geographical isolation in circumscribing a species. The large geographical disjunctions among cycad populations have greatly influenced the cycad systematist’s species concept. These disjunct populations are viewed as maintaining separate identities and having their own evolu-
ch01.qxd
28/11/03
6
3:42 pm
Page 6
T. Walters et al.
tionary tendencies and fates. In this respect, the cycad systematist’s view is a form of the evolutionary species concept. By the end of the CCC Workshop, participants were still not able to agree on exactly what constitutes a cycad species. This was not surprising, since cycad lineages have a variety of unique and long histories. Species differ to varying degrees and, therefore, a single species concept does not work for all cycads. Today, data from a wide variety of sources, including molecular analyses, ecology, geography, pollination biology and life history strategies, are providing independent measures of the evolutionary reality of existing and proposed species in the Cycadales. In contrast with species, circumscriptions of cycad genera are clearly defined and stable, with the possible exception of Chigua, which may be congeneric within Zamia (Whitelock, 2002; and see Caputo et al., Chapter 12 this volume). Cycad genera can usually be identified using gross vegetative features and can always be identified with gross features of the female reproductive structures. Family circumscriptions within the Cycadales are still somewhat unclear, being confounded by the age of the group and the inability of cycad systematists to decide on the amount of character differentiation required for family recognition within the order. Three to four families are typically recognized, with the only uncertainties revolving around the placement of two genera, Stangeria and Bowenia. Given recent advances in molecular systematics and the number of laboratories actively studying generic and familial relationships, it is predicted that a stable familial classification will be available in the very near future. Historically, the characters chosen, the importance of specific characters for differentiating genera and species, and the analyses used for describing new cycad species have been left to the discretion of each investigator. Vegetative characters, especially those associated with the leaf, along with characters related to various aspects of the female reproductive structure, are commonly used for distinguishing taxa. Male cone characters are usually not used for differentiating taxa. Cycad systematists are well aware of the plasticity of various morphological features among plants within a taxon, especially when plants are brought under cultivation. However, the degree of plasticity and the taxonomic importance of this plasticity continue to remain unclear. Another ongoing problem for cycad systematists is the lack of a consistent terminology for describing morphological features that are unique within the Cycadales. This lack of standardized morphological terminology creates problems when trying to compare characters in one taxonomic description with those in another description, or when trying to identify an individual plant based on specific characters. For this reason, a glossary of terms commonly encountered in cycad systematics is included in this volume (see Osborne and Walters, Appendix 2 this volume).
ch01.qxd
28/11/03
3:42 pm
Page 7
‘We Hold these Truths …’
7
Cycad Phylogenetics – Uncovering Genealogical Relationships Although never really attainable, systematists work toward producing a natural classification system that arranges taxa in a way that reflects the natural evolutionary order of the taxa. Since the first basic assumption is that the natural order is created by the process(es) of evolution, systematists typically strive to produce an evolutionary classification scheme. More specifically, phylogeneticists aim to recover the broad genealogical lineages within a group of taxa and to produce a classification system that reflects these genealogical or phylogenetic relationships. The starting point in phylogenetic analysis is usually the divergence of a previously occurring lineage into two or more progeny lineages. The next step is to reconstruct the separation of these lineages by identifying changes, or modifications, in characters. A character is a feature having one or more states that can be described, figured, measured, weighed, counted, scored or otherwise communicated from one systematist to another. Certain characters are biologically connected to the concept of genealogy and these characters can provide cycad systematists with justification for group membership in a phylogenetic tree. These types of characters are called apomorphies. An apomorphic character (sometimes referred to as a specialized character or a derived character) has evolved directly from its pre-existing homologue (Wiley, 1981). The task of phylogeneticists is to attempt to discover those characters that reflect the phylogeny of natural taxa. Because species are considered to be naturally occurring entities, by inference, phylogenetic characters are inherent to species. A phylogenetic character is one in which its occurrence in two or more taxa is believed to be the product of descent from a shared ancestor. A phylogenetic character shared by two organisms implies a phylogenetic relationship. Of particular importance is the synapomorphy, which is a genealogically shared, derived character state that arose in an ancestor of a lineage and is present in all of that lineage’s descendants (Hennig, 1966). Synapomorphies are the strongest evidence for shared ancestry. They are distinguished from symplesiomorphies, which are earlier character states that are shared by members of a lineage and by a more ancient ancestor to the lineage. In practice, symplesiomorphic versus synapomorphic character states for a lineage are determined by comparison with an outgroup (i.e. a related taxon that is not part of the monophyletic lineage being examined). The outgroup of choice is the ‘sister group’ to the lineage, which is genealogically the closest non-ancestral relative of the lineage. In other words, two or more taxa are sister groups if they share an ancestor not shared by any other taxon. Phylogenetic studies depict results by a graphical representation of the genealogy of one or more descendants from a common ancestor. Phylogenetic trees are branching diagrams that portray the hypothesized genealogical relationships and sequence of historical events linking taxa. A clade within a phylogenetic tree incorporates the common ancestor of a group and all of its descendants.
ch01.qxd
28/11/03
3:42 pm
Page 8
8
T. Walters et al.
Cycad systematists use phylogenetic trees to try to produce a phylogenetic classification that reflects the best estimate of the evolutionary history of cycads. Construction of a classification based on a phylogenetic tree essentially involves two steps: (i) the delimitation and naming of groups that are monophyletic in the tree; and (ii) the ranking of these monophyletic groups and placement of them into a hierarchical classification system (Wiley, 1981). Phylogenetic studies do not always lead to a new classification. These studies can provide support to an existing classification. Also, naming every monophyletic group would become cumbersome and in some cases not provide any additional information to the end-user. Cycad systematists continue to put forth hypotheses about the genealogical relationships among taxa in the Cycadales. These hypotheses are tested with evidence derived from a wide variety of sources. Hypotheses and test results are published, usually peer-reviewed, and evaluated, and some phylogenetic trees are provisionally chosen over others. In other words, the evolutionary tree for the Cycadales continues to be tested as additional and new types of data become available. An important conclusion made by the CCC Workshop participants was that the knowledge of genealogical relationships among taxa should be placed in an unambiguous and stable natural classification system that is useful for a multitude of end-users and purposes. It is believed that such a system can orient human understanding of life and the world around us.
CCC Workshop Beliefs The CCC Workshop participants enumerated their beliefs concerning cycad classification during the second day of the Workshop. Clarity and consensus with regard to these beliefs were needed so that the participants could go on to produce a final set of guidelines for future research aimed at establishing a suitable classification scheme for cycads (see Osborne and Walters, Chapter 15 this volume). For purposes associated with the production of this chapter, the suite of beliefs generated during the work sessions at the CCC Workshop has been reworded and organized to provide consistency in wording, style and format. The authors have organized the beliefs under the following three categories: (i) beliefs about biological relationships; (ii) beliefs about what systematists can and should do in order to understand biological relationships; and (iii) beliefs about what cycad systematists can and should do in order to understand relationships in the Cycadales.
Beliefs about biological relationships ● ●
We believe there is value in the biological world. We believe there is a natural order to the biological world.
ch01.qxd
28/11/03
3:42 pm
Page 9
‘We Hold these Truths …’
● ●
●
9
We believe that the natural order is based on genealogical relationships. We believe that the pattern of genealogical relationships naturally produces a hierarchical structure of lineages. We believe that each species (as a natural group) is a monophyletic lineage that evolves independently of other such lineages.
Beliefs about what plant systematists can and should do in order to understand biological relationships ●
●
●
●
●
●
●
●
●
●
●
We believe we can construct hypotheses that are testable, and that the process of testing and refining hypotheses leads to a better understanding of the natural world. We believe that genealogical relationships can be recovered through hypothesis testing. We believe that as technology, resources and data increase and change, we will be better able to construct a classification scheme that approximates true genealogical relationships. We believe that we should construct hierarchical classification schemes that best reflect actual genealogical relationships. Such schemes have greater predictive power, have greater heuristic value, and improve our ability to understand and communicate about the biological world as it existed, as it exists, and as it may exist in the future. We believe that the process of refining classification schemes brings us closer to approximating true genealogical relationships and therefore converges towards stability of the classification. We believe that the most important evolutionary entity to define and circumscribe is the species. We believe that species are not evolutionarily static (i.e. they change through time). We believe that species can be difficult to recognize, and, therefore, the definitions and circumscriptions that we apply to particular species are hypotheses to be tested. We believe that the exploration of species and species concepts will provide the common language for understanding speciation, species interactions and plant systematics. We recognize the existence of the International Code of Botanical Nomenclature (ICBN) and support the beliefs, philosophies, and principles of the Code to provide one correct name for each taxonomic group within a stable classification system. We believe that systematists should share information through the publication of data and analyses.
ch01.qxd
28/11/03
3:42 pm
Page 10
10
T. Walters et al.
Beliefs about what cycad systematists can and should do in order to understand relationships in the Cycadales ●
●
●
●
●
●
●
●
●
●
We believe that the Cycadales forms a distinct monophyletic lineage, and genealogical relationships within this lineage can be inferred through the collection and analysis of data. We believe that higher ranks in the Cycadales (e.g. genera and above) are easily recognizable and definable. We believe that the greatest challenge in cycad systematics is recognizing appropriate units to call species. We believe that there are differing opinions concerning cycad species definitions and circumscriptions. Given the uncertainty of species definitions and the lack of infraspecific data on cycads, we believe that it is not yet appropriate to try to define and identify relationships of taxa below the species level. We believe that to better understand cycad species, we must concentrate our resources on variation and relationships at the population level. We believe that there is a wealth of available data on cycads that still must be captured and analysed. We believe that a classification system should be valuable for a variety of known and unknown end-users and purposes. We believe that the extinction of cycad species is accelerating and that access to native populations is decreasing rapidly. Actions must be undertaken immediately to describe, classify, conserve and preserve species for continuing scientific studies. We believe that the process of understanding cycad systematics should be a collaborative endeavour.
Acknowledgements The Cycad Classification Concepts Participants spent many hours developing and discussing the suite of beliefs presented in this chapter. Their commitment to and support of undertaking this long and arduous task is greatly appreciated. The authors are deeply indebted to Deena Walters for her critical comments on early drafts of this manuscript.
References Greuter, W., McNeill, J., Barrie, F.R., Burdet, H.M., Demoulin, V., Filgueiras, T.S., Nicolson, D.H., Silva, P.C., Skog, J.E., Trehane, P., Turland, N.J. and Hawksworth, D.L. (2000) International Code of Botanical Nomenclature (Saint Louis Code). Koeltz Scientific Books, Köningstein, Germany, 474 pp.
ch01.qxd
28/11/03
3:42 pm
Page 11
‘We Hold these Truths …’
11
Grobbelaar, N. (2002) Cycads – with Special Reference to the Southern African Species. Published by the author, Pretoria, South Africa, 331 pp. Hennig, W. (1966) Phylogenetic Systematics. University of Illinois Press, Champaign-Urbana, Illinois, 263 pp. Hill, K. and Osborne, R. (2001) Cycads of Australia. Kangaroo Press, New South Wales, Australia, 116 pp. Jones, D.L. (2002) Cycads of the World – Ancient Plants in Today’s Landscape, 2nd edn. Reed New Holland, Sydney, Australia, 456 pp. Judd, W.S., Campbell, C.S., Kellogg, E.A. and Stevens, P.F. (1999) Plant Systematics: a Phylogenetic Approach. Sinauer Associates, Inc., Sunderland, Massachusetts, 464 pp. Norstog, K.J. and Nicholls, T.J. (1997) The Biology of the Cycads. Cornell University Press, Ithaca, New York, 363 pp. Whitelock, L.M. (2002) The Cycads. Timber Press, Portland, Oregon, 374 pp. Wiley, E.O. (1981) Phylogenetics: the Theory and Practice of Phylogenetic Systematics. John Wiley & Sons, New York, 439 pp.
ch01.qxd
28/11/03
3:42 pm
Page 12
ch02.qxd
28/11/03
3:42 pm
Page 13
Saving Ghosts? The Implications of Taxonomic Uncertainty and Shifting Infrageneric Concepts in the Cycadales for Red Listing and Conservation Planning
2
John Donaldson Kirstenbosch Research Centre, National Botanical Institute, Claremont, South Africa
Abstract A comparison of the cycad Red Lists from 1978 to 2002 shows that taxonomy has had a profound influence on the outcomes of the Red List process. The descriptions of new species, as well as taxonomic revisions, have increased the number of recognized cycad taxa from 136 in 1978, to over 300 at the time of publication. During this time, the proportion of threatened taxa has fluctuated from a low of 46% of all cycad taxa in 1978, to a high of 82% in 1997, and is currently estimated to be 52%. At least one-third of the changes in the Red List are due to taxonomic changes, which reflects an increase in taxonomic activity between 1978 and 2002. In addition to new species, there were 48 changes in the Red List between 1978 and 2002 as a result of uncertainty about the infrageneric status of cycad taxa. Frequent changes in threatened status are not helpful for conservation planning and could even undermine the Red List process. The IUCN has introduced robust criteria for Red Listing to deal with inadequate population data and ecological uncertainty. The results of this analysis show that cycad taxonomists need to develop consistent and widely accepted concepts for infrageneric taxa to reduce the influence of taxonomic uncertainty on the Red Listing process.
Introduction Ongoing assessment of the threatened status of the world’s cycad flora (Red Listing) is an essential process for conservation planning and actions (Osborne, ©CAB International 2004. Cycad Classification: Concepts and Recommendations (eds T. Walters and R. Osborne)
13
ch02.qxd
28/11/03
14
3:42 pm
Page 14
J. Donaldson
1995; Donaldson, 2003). Having a clear picture of which cycads are most threatened with extinction, and where they are located, enables conservation agencies and funding bodies to allocate resources to the most threatened taxa or to the most effective conservation actions. Red Listing and conservation planning are dynamic processes. Typically, Red Listing assessments are based initially on inadequate information and they are modified as more information becomes available. As a result, changes in threatened status are unavoidable, and they can come about for several reasons: ●
●
● ● ●
Actual changes in plant distribution or abundance due to ongoing threats or to the success of conservation actions. The discovery of new populations or additional demographic data that provide more accurate estimates of population numbers and trends. Changes in the criteria used for assigning threatened status. The discovery of previously unknown species. Taxonomic revisions that result in the splitting of existing species into several taxa or the lumping of several described species into one species.
Frequent changes in Red List status are not helpful for conservation planning and they may even undermine the credibility of Red Listing as a conservation tool. The aim of Red Listing should be to create a reasonably stable and scientifically based system that can portray real changes in the distribution and abundance of threatened species as well as the relative status of different species. To achieve such a system, the IUCN introduced revised categories of threat and more rigorous criteria for assessing threatened status (IUCN, 1994; IUCN/SSC Criteria Review Working Group, 1999; Mace, 2000). The revised system provides a framework for dealing with the uncertainty that arises from using inadequate ecological data to assess threatened status. However, the IUCN system does not deal specifically with an important area of uncertainty in threat assessments, i.e. the taxonomic status of threatened species. Clearly, a credible Red List relies on a consistent and widely accepted system of classification. Within the Cycadales, there is still considerable uncertainty about what a taxon is, especially at the infrageneric level, and this problem has important implications for cycad Red Listing. This chapter examines changes in the threatened status of cycad taxa from 1978 to 2002 to show how taxonomy influences the outcome of Red Listing assessments and conservation planning. During this time there have been five Red Listing accounts of the world cycad flora, starting with the list drawn up by the IUCN Threatened Plant Unit and based on the criteria outlined in The IUCN Plant Red Data Book (Lucas and Synge, 1978). The initial list was followed by Gilbert (1984), Osborne (1995), the 1997 IUCN Red List of Threatened Plants (Walter and Gillett, 1998) and the Cycad Action Plan (Donaldson, 2003). These accounts provide a basis for evaluating changes in threatened status over a 24year period and for determining the relative contributions of taxonomic uncertainty and ecological uncertainty to these changes.
ch02.qxd
28/11/03
3:42 pm
Page 15
Red Listing and Conservation Planning
15
Taxonomy and the Threatened Status of Cycads In the first cycad Red List, based on The IUCN Plant Red Data Book (Lucas and Synge, 1978), there were 136 recognized cycad taxa (species and subspecies), comprising one Extinct species (Encephalartos woodii Sander), 63 Threatened taxa (46%), 45 taxa of Unknown threatened status (33%) and only 15 taxa (11%) that were Not Threatened (Fig. 2.1). In 1984, the number of cycad taxa had risen to 168, comprising one Extinct species, 65 Threatened taxa (38.6%), 86 taxa of Unknown status (51%) and 16 taxa (9.5%) that were classified as Not Threatened (Gilbert, 1984). Eleven years later, using the same criteria, Osborne (1995) recognized 197 taxa, comprising one Extinct species, 124 Threatened taxa (63%), 42 taxa that could not be assessed (21%) and 30 taxa that were Not Threatened (15%) (Fig. 2.1). In the last assessment using these criteria, Walter and Gillett (1997) recognized only 180 taxa, comprising three Extinct species, 147 Threatened taxa (82%), three taxa of Unknown status (1.6%) and 27 taxa that were Not Threatened (15%) (Fig. 2.1). Finally, the Cycad Action Plan
Not Threatened
Fig. 2.1. The number of cycad taxa (species, subspecies and undescribed species) classified as Extinct, Threatened, Not Threatened or Unknown, according to the IUCN Threatened Plant Unit database (1978), Gilbert (1984), Osborne (1995), the 1997 IUCN Red List of Threatened Plants (Walter and Gillett, 1998) and the 2002 assessments for the Cycad Action Plan (Donaldson, 2003). The total height of each bar represents the total number of recognized taxa. Taxa of unknown conservation status include those classified as ‘Indeterminate’, ‘Unknown’ and ‘Data Deficient’.
ch02.qxd
28/11/03
16
3:42 pm
Page 16
J. Donaldson
(Donaldson, 2003), based on the latest revised IUCN criteria (Mace, 2000), included 297 taxa, comprising two species that are ‘Extinct in the Wild’ (no ‘Extinct’ species), 154 Threatened taxa (52%), 18 taxa that are ‘Data Deficient’ (6%) and a further 123 taxa (41%) that are not classified as threatened (Fig. 2.1). Clearly there are substantial differences between these assessments and it is necessary to examine these differences in greater detail to determine how taxonomic uncertainty has influenced the outcome of the Red Listing process. The substantial increase in the number of recognized cycad taxa (at the species and subspecies level) between 1977 and 2002 (Fig. 2.1) has obviously influenced the overall threatened status of the world’s cycads. Even at this level, there is a degree of confusion about what constitutes a valid taxon for Red List evaluation. Osborne (1995) recognized 197 taxa whereas Walter and Gillett (1998) recognized only 180 taxa. The taxa omitted by Walter and Gillett (1998) comprised six species of Cycas Linnaeus, both species of Bowenia Hooker ex Hooker filius, and nine taxa in the Zamiaceae, including one undescribed species of Encephalartos Lehmann. Walter and Gillett (1998) also included Cycas celebica Miquel, which is now considered synonymous with C. rumphii Miquel (Hill et al., 2003). Species classified as Extinct (pre-1994 categories) or Extinct in the Wild (IUCN, 1994) provide a useful starting point for exploring the influence of taxonomic uncertainty on Red Listing. Until 1995, only Encephalartos woodii was classified as an Extinct species and this status has been consistent in later assessments. Walter and Gillett (1998) also included Cycas szechuanensis C.Y. Cheng, W.C. Cheng & L.K. Fu and Zamia monticola Chamberlain as Extinct taxa. Cycas szechuanensis was first classified as Extinct because it was known only in cultivation. It was later thought to be conspecific with the more widespread C. guizhouensis K.M. Lan & R.F. Zou (Hill et al., 2003), which would warrant a downlisting to ‘Near Threatened’. However, it is now regarded as a valid species (Chen, 2000), but it no longer warrants a status of Extinct in the Wild because two wild populations have been discovered. In this case, changes in both taxonomic interpretation and new discoveries of wild populations have influenced the Red List status. In contrast, the taxonomic status of Zamia monticola has not been in doubt and the species was downgraded to ‘Critically Endangered’ (Stevenson et al., 2003) after the discovery of new populations. The Cycad Action Plan (Donaldson, 2003) also lists Encephalartos relictus P.J.H. Hurter as Extinct in the Wild. This species was described from a cultivated specimen. Two undescribed taxa that were classified as Extinct (herbarium specimen) and Extinct in the Wild (cultivated specimen) in an assessment of African cycads by J. Golding and P.J.H. Hurter (unpublished results) were not included in the Cycad Action Plan, due to their uncertain taxonomic status. A clear and consistent approach to infrageneric delimitations would help to resolve the Red List status of these taxa. Some of the same problems recur with other categories of threat (Fig. 2.2). The overall change in the number of threatened taxa shows a gradual increase in threatened taxa from 1978 to 1997 using the pre-1994 criteria. Based on the
28/11/03
3:42 pm
Page 17
17
Red Listing and Conservation Planning
Pre-1994 Rare
Post-1994 Vulnerable
Pre-1994 Vulnerable
Post-1994 Endangered
Pre-1994 Endangered
Post-1994 Critical
180 160 140 120 Number of taxa
ch02.qxd
100 80 60 40 20 0 1978
1984
1995 1997 Year of assessment
1995
2002
Fig. 2.2. The number of cycad taxa (species, subspecies and undescribed species) assigned to IUCN threatened categories according to the IUCN Threatened Plant Unit database (1978), Gilbert (1984), Osborne (1995), the 1997 IUCN Red List of Threatened Plants (Walter and Gillett, 1998) and the 2002 assessments for the Cycad Action Plan (Donaldson, 2003). Assessments for 1978 and 1997 used the ‘old’ IUCN (pre-1994) criteria whereas the 2002 assessment used the criteria introduced after 1994. Osborne (1995) used both sets of criteria to evaluate cycad species.
post-1994 criteria, there is also an increase in the number of threatened taxa from Osborne (1995) to the Cycad Action Plan in 2002 (Donaldson, 2003) (Fig. 2.2). The low number of taxa of unknown status in the later assessments (Fig. 2.1) indicates that there is increasing ecological information on threatened taxa so that changes in the number of taxa assigned to different threatened categories may be due to better ecological information. However, an analysis of changes due to better ecological information (distribution and abundance data) compared with those due to taxonomic changes (new species, change in infrageneric status) (Fig. 2.3) shows that taxonomy accounts for a large number of changes in the number of threatened cycads. Clearly, many changes in the number of threatened species could arise from
28/11/03
3:42 pm
18
Page 18
Cy ca s Sta ng eri a Bo we nia Ch igu Ce a rat oz am ia
Dio En on ce ph ala r to s Le pid oz am ia Ma cro za mi a Mi cro cyc as Za mi a
ca s Sta ng eri Bo we nia Ch igu Ce a rat oz am ia
Dio En on ce ph ala r to Le s pid oz am ia Ma cro za mi a Mi cro cyc a Za mi a
J. Donaldson
Cy
ch02.qxd
Fig. 2.3. The number of cycad taxa where a change in threatened status was recorded between assessments. (A) Changes between the TPU Red List based on Lucas and Synge (1978) and Osborne (1995) using the pre-1994 IUCN criteria. (B) Changes between the 1997 IUCN Red List of Threatened Plants (Walter and Gillett, 1998) and the 2002 assessments for the Cycad Action Plan (Donaldson, 2003). Changes that resulted from taxonomic revisions (new species, split taxa and combined taxa) are represented by black bars, and those that resulted from more complete ecological data are represented by white bars.
ch02.qxd
28/11/03
3:42 pm
Page 19
Red Listing and Conservation Planning
19
the description of new species as a result of increased field work in areas such as Australia, Central Africa, Mexico and South-East Asia between 1978 and 2002. It is therefore important to distinguish between the influence of new taxa on Red List assessments, compared with the influence of other changes that reflect uncertainty about the infrageneric status of cycad taxa. The number of taxa included in one Red List assessment that were no longer considered to be valid taxa in the following assessment (Fig. 2.4A) can be compared with the number of new taxa that were included in each Red List assessment (Fig. 2.4B). The results show that a considerable number of changes could be attributed to the description of new species (Fig. 2.4B), especially in Cycas, but also in Encephalartos, Macrozamia Miquel and Zamia Linnaeus. However, a large number of changes also occurred because of changes in the infrageneric status of described species, especially in Zamia, Cycas and Macrozamia. The data indicate that there were fewer changes due to uncertain taxonomic status between 1997 and 2002 than between earlier assessments (Fig. 2.4A), suggesting that the species limits of cycad taxa are better understood now than they were in 1978. The data summarized in Figs 2.1–2.4 do not reveal the many subtle influences that taxonomic changes have on cycad Red Listing. The description of new species and changes in species concepts can have different outcomes on Red Listing depending on the genus under review. For example, the delimitation of new species and subspecies of Cycas resulted in an increase in recognized taxa from eight in 1978 to 98 in 2002, while the number of threatened taxa increased from four in 1978 (50%) to 38 in 2002 (39%). As a result, there has been a substantial decrease in the proportion of Cycas species that are listed as threatened with extinction. In Africa, the delimitation of new species and subspecies of Encephalartos over the same period resulted in an increase from 48 taxa in 1978 to 67 taxa in 2002, with an increase in threatened taxa from 32 (66%) in 1978 to 46 (68%) in 2002. In the case of Cycas, many new species were discovered and described in areas where cycads are still abundant. In contrast, several new species of Encephalartos were described as a result of the revision of already threatened taxa, such as the split of E. eugene-maraisii I. Verdoorn into several taxa, all classified as threatened.
The influence of higher-level classification on conservation planning The focus of this chapter is primarily on what happens at the infrageneric level, but it is worth noting that taxonomic changes at higher levels also affect conservation planning. Vane-Wright et al. (1991) argued that systematics should be used as a criterion for prioritizing conservation actions. Their reasoning was that, given the high number of species that are threatened with extinction, priority should be given to those species that represent more threatened lineages (represented by higher taxonomic groups). For cycads, the genus Chigua D.W. Stevenson comprises two Critically Endangered species (Stevenson et al., 2003) and represents the second most threatened genus within the Cycadales [after Microcycas
3:42 pm
Page 20
20
Za mi a Za mi a
Ma cro za mi a Ma cro za mi a
En ce ph ala r to s En ce ph ala r to s
Dio on
s ca Cy s ca Cy
Ce rat oz am ia
J. Donaldson
Dio on
28/11/03
Ce rat oz am ia
ch02.qxd
Fig. 2.4. Changes in recognized cycad taxa between different Red List assessments, from Gilbert (1984) to Osborne (1995), from 1995 to the 1997 IUCN Red List of Threatened Plants (Walter and Gillett, 1998), and from 1997 to 2002 (Donaldson, 2003). (A) The number of taxa listed on the first date but not recognized as valid taxa when the next list was published. (B) The number of taxa listed in the later assessment that were not recognized as distinct taxa in the earlier assessment. Only the major genera are included, as there were no differences in the genera Bowenia, Lepidozamia, Microcycas and Stangeria.
ch02.qxd
28/11/03
3:42 pm
Page 21
Red Listing and Conservation Planning
21
(Miquel) A. de Candolle]. Based on Vane-Wright et al. (1991), the two species of Chigua should receive higher conservation priority than equally threatened species in larger genera (Walters, 2003). However, if Chigua is viewed as part of the genus Zamia (sensu Schutzman and Dehgan, 1993), then the two species currently included in Chigua become part of a much larger genus and therefore have lower conservation priority.
Conclusions A comparison of the cycad Red Lists from 1978 to 2002 shows that taxonomy has had a profound influence on the outcomes of the Red List process. By far the greatest contribution has been from the description of new taxa, due to a resurgence of interest in cycad taxonomy within the past 25 years. Greater taxonomic activity has also resulted in revised species concepts and the sorting out of nomenclatural problems so that many early names are no longer recognized as valid species. In both cases, the changes reflect developments within the science of cycad taxonomy that are a positive contribution to our knowledge of the Cycadales. These changes do, however, influence the process of Red Listing and it is essential to ensure that taxonomic changes are both necessary and consistent. To achieve this, taxonomists need to agree on concepts used to delimit infrageneric taxa and then apply these concepts consistently.
References Chen, C.J. (2000) Cycadaceae. In: Fu, L.K. et al. (eds) Higher Plants of China, Vol. 3. Qingdao Press, Qingdao, China, pp. 1–11. Donaldson, J.S. (ed.) (2003) Cycads. Status Survey and Conservation Action Plan. IUCN/SSC Cycad Specialist Group, IUCN, Gland, Switzerland and Cambridge, UK, ix + 86 pp. Gilbert, S.G. (1984) Cycads: Status, Trade, Exploitation and Protection 1977–1982. TRAFFIC, Washington, DC, 74 pp. Hill, K.D., Chen, C.J. and Loc, P.K. (2003) Regional overview: Asia. In: Donaldson, J.S. (ed.) Cycads. Status Survey and Conservation Action Plan. SSC Cycad Specialist Group, IUCN, Gland, Switzerland and Cambridge, UK, 25–30 pp. IUCN (1994) IUCN Red List Categories. Prepared by the IUCN Species Survival Commission, IUCN, Gland, Switzerland, 22 pp. IUCN/SSC Criteria Review Working Group (1999) IUCN Red List Criteria review provisional report: draft of the proposed changes and recommendations. Species 31–32, 43–57. Lucas, G. and Synge, H. (1978) The IUCN Plant Red Data Book. IUCN Threatened Plants Committee, Kew, UK, 540 pp. Mace, G. (2000) Summary of the results of the review of IUCN Red List categories and criteria 1996–2000. In: Hilton-Taylor, C. (compiler) 2000 Red List of Threatened Species. IUCN, Gland, Switzerland and Cambridge, UK, pp. 57–61.
ch02.qxd
28/11/03
22
3:42 pm
Page 22
J. Donaldson
Osborne, R. (1995) The world cycad census and a proposed revision of the threatened species status for cycad taxa. Biological Conservation 71, 1–12. Schutzman, B. and Dehgan, B. (1993) Computer assisted systematics in the Cycadales. In: Stevenson, D.W. and Norstog, K.J. (eds) The Biology, Structure, and Systematics of the Cycadales. Proceedings of the Second International Conference on Cycad Biology. Palm & Cycad Societies of Australia Ltd, Milton, Queensland, Australia, pp. 281–289. Stevenson, D.W., Vovides, A. and Chemnick, J. (2003) Regional overview: New World. In: Donaldson, J.S. (ed.) Cycads. Status Survey and Conservation Action Plan. SSC Cycad Specialist Group, IUCN, Gland, Switzerland and Cambridge, UK, 31–38 pp. Vane-Wright, R.I., Humphries, C.J. and Williams, P.H. (1991) What to protect? Systematics and the agony of choice. Biological Conservation 55, 235–254. Walter, K.S. and Gillett, H.J. (eds) (1998) 1997 IUCN Red List of Threatened Plants. Compiled by the World Conservation Monitoring Centre, IUCN, Gland, Switzerland and Cambridge, UK, 862 pp. Walters, T. (2003) Off-site collections. In: Donaldson, J.S. (ed.) Cycads. Status Survey and Conservation Action Plan. SSC Cycad Specialist Group, IUCN, Gland, Switzerland and Cambridge, UK, 48–53 pp.
ch03.qxd
28/11/03
3:43 pm
Page 23
Character Evolution, Species Recognition and Classification Concepts in the Cycadaceae
3
Ken D. Hill Royal Botanic Gardens, Sydney, Australia
Abstract A number of systems of infrageneric classification of the family Cycadaceae have been presented by different authors. No two systems have been the same in structure or species recognition, and some have been strikingly discordant. Phylogenetic analysis of combined morphological and molecular datasets has yielded a cladogram with good resolution and support on many deeper branches. Plotting characters used by different authors in developing infrageneric systems of classification allows an independent assessment of the value of these characters and of the systems of classification derived from them. No classification system presented to date is wholly concordant with the results of the phylogenetic studies, and a number of characters previously heavily relied upon in defining groups are shown to be highly plastic. In particular, the recently described segregate genus Epicycas is shown to be polyphyletic, as are the four subgenera of Cycas erected by the same authors. The analysis supports the recognition of a single genus with five sections, although a number of taxa are insufficiently known to be clearly placed. Species recognition and species concepts are discussed.
Introduction The genus Cycas Linnaeus has long been accepted as the single constituent genus of the family Cycadaceae, itself the basal divergence within the extant Cycadales or Cycadophyta (Johnson, 1959; Stevenson, 1992). The genus (and family) has a present-day distribution concentrated in a zone between northern Australia and southern China, and extending westwards to Madagascar, the Comoros and the adjacent African mainland, and eastwards to Tonga. Many regional taxonomic ©CAB International 2004. Cycad Classification: Concepts and Recommendations (eds T. Walters and R. Osborne)
23
ch03.qxd
28/11/03
3:43 pm
Page 24
24
K.D. Hill
treatments have been published over the years, but all suffer to some extent from the lack of an up-to-date comprehensive monographic treatment (e.g. Backer and Bakhuizen van den Brink, 1963; Smitinand, 1971; Fu et al., 1978; Hiêp and Vidal, 1996; Chen and Stevenson, 1999). The comprehensive treatment by Schuster (1932) is generally acknowledged to have created more problems than it solved (Johnson, 1959), and the most recent overall treatment by de Laubenfels and Adema (1998), wherein a second genus is erected, has proved controversial and has not been widely accepted (see Chen et al., Chapter 5 this volume). The different regional and comprehensive treatments also differ markedly in species recognition and circumscription. Analysis of combined morphological and molecular data yields a resolved cladogram with good support on many deeper branches (Fig. 3.1). Plotting characters used by different authors in developing infrageneric systems of classification allows an independent assessment of the value of these characters and of the systems of classification derived from them.
Historical background A number of authors have attempted to divide the genus Cycas internally and present systems of infrageneric classification. These are most notable in their discordancy. This is partly an indication that the known species are closely related and form a coherent group, but also a reflection of the poor understanding of specific limits and relationships within the group. The first attempt to subdivide the genus was that of Miquel, who firstly separated C. revoluta Thunberg from all other species on the basis of the revolute leaflet margins (Miquel, 1843), and later recognized two informal groups separated by tomentose vs. glabrous ovules (Miquel, 1861). He was followed by de Candolle (1868). Other early authors recognized two informal groups on the basis of the degree of division of the lamina of the megasporophyll (Carruthers, 1893). Warburg (1900) divided the genus, again informally, on the basis of the number of ovules per megasporophyll. Warburg was followed by Pilger (1926). Schuster (1932) attempted the first formal infrageneric classification of Cycas (discussed below), recognizing three major groups as sections. A second formal system was presented by Smitinand (1971), with a very different major division into two sections. Dehgan and Dehgan (1988) published an informal infrageneric classification based on seed structure and pollen morphology that was completely different from either of the above, recognizing two subgenera. Hill (1994, 1996) recognized four sections, each with two or three subsections, in part merging the systems of Schuster and Smitinand. Wang (1996) erected two subgenera and within these a number of sections, subsections and series, many corresponding in circumscription, but not rank or placement, to taxa recognized by Hill. De Laubenfels and Adema (1998) presented an entirely different system, separating the new genus Epicycas de Laubenfels and dividing the remainder of Cycas
ch03.qxd
28/11/03
3:43 pm
Page 25
Classification Concepts in the Cycadaceae
25
C. wadei C. curranii C. revoluta C. panzihuaensis C. changjiangensis C. guizhouensis C. sexseminifera C. bifida C. diannanensis C. pectinata C. clivicola C. lindstromii C. pranburiensis C. condaoensis C. siamensis C. inermis C. scratchleyana C. beddomei C. circinalis C. spherica C. riuminiana C. thouarsii C. rumphii C. micronesica C. bougainvilleana C. seemannii C. apoa C. tuckeri C. yorkiana C. cairnsiana C. ophiolitica C. maconochiei C. calcicola C. furfuracea C. silvestris C. media C. papuana C. armstrongii
Fig. 3.1. Combined molecular and morphological analysis for species of Cycas (strict consensus tree). Bootstrap support values shown above branches. From Hill (1999 and in preparation).
into four subgenera. The classification systems of Schuster and subsequent authors are examined in the light of recent cladistic studies below. Species concepts Species definition in Cycadaceae is complicated by the variability of some of the characters that have been traditionally used to separate taxa and by the inade-
ch03.qxd
28/11/03
3:43 pm
26
Page 26
K.D. Hill
quacy of specimens and recorded observations in representing the differences and similarities occurring within and between populations and taxa. For example, development of spines on petioles is often variable within populations, and changes with age in most taxa. Microsporophylls vary considerably in size and shape from base to apex of the cone; in particular, the apical spine is reduced or absent on the lowermost sporophylls, gradually increasing in size towards the apex or towards the centre of the cone (Amoroso, 1986). Megasporophylls also vary, often widely, in size, shape of lamina and number of ovules. This can depend on their position within a growth flush, the first and last produced often being markedly smaller and less elaborate than those in the centre. Because herbarium collections are frequently fragmentary, incomplete and often sterile, many characters are not represented or recorded. Comprehensive field study is thus essential to understanding the ranges of variation of characters that may distinguish taxa. An example is in the recent segregation of seven taxa in Cape York Peninsula, Australia, from what had long been accepted as a single species (Hill, 1995). Many of the characters that discriminate these taxa were unrepresented in specimens or notes, and it was only by comprehensive and systematic field observation that the totality of morphological variation could be observed. This is achieved in a rigorous and reproducible manner by following a standard pro forma (Fig. 3.2) and making a standard set and number of observations on each population. Data recorded can then be statistically analysed if required, and groups can be defined. This synthesis can then lead to satisfactory species recognition, following which identification keys and classifications can be developed (Table 3.1). In the Cape York example, the cataphyll characters critical in discriminating taxa had not been previously recorded. It is also of note that these taxa occur in discrete, geographically separated populations. The genus Cycas in general shows a similar geographical replacement pattern throughout its range, often with many closely related entities. Although these taxa are often similar in many respects, the homoplasy evident in the defining characters does not always allow unequivocal aggregation into groups that could be treated as species with subordinate infraspecific taxa. In order to satisfactorily separate and recognize groups of populations that show real, albeit sometimes small differences, it is strongly recommended that these groups be treated as distinct species. This rather narrow view of species is considered preferable to the arbitrary submerging of these recognizably distinct and truebreeding groups of populations into broader and less meaningful ‘species’ and the consequent loss of information on the real diversity of these plants.
Phylogenetic studies Phylogenetic analyses of the family Cycadaceae sensu stricto have been conducted using sets of data obtained from DNA molecular sequences and from morphological and anatomical characters. Sequence data of the internal transcribed spacer region from the nuclear genome for 40 terminal taxa were analysed (Hill,
ch03.qxd
28/11/03
3:43 pm
Page 27
27
Classification Concepts in the Cycadaceae
–
– – –
Fig. 3.2. Pro forma used for rigorous systematic observations in populations of Cycas.
ch03.qxd
28/11/03
3:43 pm
Page 28
28
K.D. Hill
Table 3.1. Key to the species of Cycas in Cape York Peninsula, Australia. Characters observed in field studies are shown in bold. 1 Cataphylls hard, strongly spinescent 2 1 Cataphylls soft, not spinescent 4 2 Hypodermis absent; leaf margins flat; leaflet midrib equally raised above and below C. xipholepis K.D. Hill 2 Hypodermis present; leaf margins recurved; leaflet midrib almost flat above, strongly raised below 3 3 Pinnae > 9 mm wide; cataphylls deciduous C. silvestris K.D. Hill 3 Pinnae < 9 mm wide; cataphylls persistent C. media R. Brown 4 Crown and cataphylls densely orange-woolly or floccose 5 4 Crown and cataphylls very shortly grey to white-sericeus 6 5 Megasporophyll apex 60–100 mm long, with 24–32 lateral spines 3–6 mm long, apical spine 2–16 mm long C. yorkiana K.D. Hill 5 Megasporophyll apex 40–55 mm long, with 16–24 lateral spines 1–4 mm long, apical spine 21–25 mm long C. badensis K.D. Hill 6 Megasporophylls short (10–13 cm); apex dilated (35–50 mm wide, 50 –65 mm long) C. tuckeri K.D. Hill 6 Megasporophylls longer (15–22 cm); apex not dilated (18–35 mm wide, 60–75 mm long) C. semota K.D. Hill
in preparation), and morphological and anatomical data were taken for the same 40 taxa from a data set used in morphological studies (Hill, 1999 and in preparation). The combined data yielded a result that was not completely consistent with any published morphological studies or recent taxonomic classifications (Fig. 3.1). The resultant cladogram from these studies will be taken as a basis for character analysis below.
Character Analysis Previously published classifications have been examined and characters used to define groups recorded (Table 3.2). These were then plotted on to the cladogram previously obtained from analysis of combined morphological and molecular data (Fig. 3.1). Consistency indices (CI) are assigned to each character as a measure of the degree of homoplasy shown by that character, and also as a measure of the reliability and efficacy of that character in defining and recognizing natural groups. Most of the characters that have been widely used in infrageneric classification in the past are shown to be useful, although character polarities were not always as previously assumed. Each character state is discussed individually below.
ch03.qxd
28/11/03
3:43 pm
Page 29
29
Classification Concepts in the Cycadaceae
Table 3.2. Characters used by different workers in defining groups in Cycas.
Character
States
Outgroup conditions
Ovules
Glabrous/tomentose
Glabrous
Sclerotesta
Smooth/striate/ribbed/ Smooth/striate verrucose/crested Spongy Absent/present Absent megagametophytic tissue Fibrous sarcotesta Absent/present Absent Ovule number Seed cone Megasporophyll apex Megasporophyll length Pollen cones
1–14 Closed/open Pectinate/dentate/ entire Long/short
Pollen cones
Cylindrical/ovoid
Microsporophyll apex Base of plant
Attenuate/truncate
Soft/rigid
Swollen/not swollen
Habit
Caulescent/ acaulescent
Petiole
Long/short
Leaflet Leaflet margins
Entire/divided Flat/revolute
Leaflet apex Leaflet midrib
Soft/spinescent Raised/flat
Reference Miquel, 1861; Schuster, 1932 Hill, 1996; Wang, 1996 Deghan and Yuen, 1983; Hill, 1994
Hill, 1994; Wang, 1996 2 Warburg, 1900 Not comparable Wang, 1996 Entire Schuster, 1932 Not comparable Schuster, 1932 Soft/rigid
Smitinand, 1971; Hill, 1996, 1999 Cylindrical/ovoid de Laubenfels and Adema, 1998 Not comparable de Laubenfels and Adema, 1998 Swollen/not de Laubenfels and swollen Adema, 1998 Caulescent/ Smitinand, 1971; acaulescent de Laubenfels and Adema, 1998 Long/short de Laubenfels and Adema, 1998 Entire/divided Wang, 1996 Flat de Laubenfels and Adema, 1998 Soft/spinescent Schuster, 1932 Raised/flat de Laubenfels and Adema, 1998
Ovule tomentum Two species, Cycas revoluta and C. taitungensis C.F. Shen, K.D. Hill, C.H. Tsou & C.J. Chen, possess tomentose ovules. This is a synapomorphy uniting these taxa (CI 100%, Fig. 3.3A) but of no value in further grouping or taxonomically classifying this small group. Miquel (1861), Schuster (1932) and Hill (1996, 1999) had used this character state as a key defining character for section Asiorientales. This section proves to be a natural group, supported by several other synapomorphies,
ch03.qxd
28/11/03
3:43 pm
Page 30
30
K.D. Hill
and the character of tomentose ovules is clearly a useful character in identification of this group.
Ribbed sclerotesta Two species, Cycas wadei Merrill and C. curranii (J. Schuster) K.D. Hill, possess a strongly ribbed sclerotesta. This is a synapomorphy uniting these taxa (CI 100%, Fig. 3.3A) but of no value in further grouping or taxonomically classifying this small group. This section proves to be a natural group, supported by several other synapomorphies, and the character of ribbed sclerotesta is clearly a useful character in identification of this group. This clade was recognized as subsection Wadeanae by Hill (1996) and Wang (1996) on the basis of this character.
Verrucose sclerotesta A verrucose sclerotesta is shown to be a well-corroborated synapomorphy (CI 100%, Fig. 3.3A) supporting quite a large natural group of mainland Asian species occurring in southern China and northern Indochina. This group was formally named by Smitinand (1971) as section Stangerioides, although at that time with a single constituent species and not on the basis of sclerotesta characters. Section Stangerioides was adopted by Hill (1996) and Wang (1996), although with wider circumscriptions that differed from each other and that in both cases rendered the natural group non-monophyletic. The verrucose sclerotesta was recognized as a defining character for smaller groups within the broader groups by these authors.
Crested sclerotesta A small group of taxa from the Western Pacific and eastern Malesian regions possesses a crested sclerotesta. This is a synapomorphy uniting these taxa (CI 100% in the best-case scenario, Fig. 3.3B) but of no value in further grouping or taxonomically classifying this small group. While the crested sclerotesta is a useful character in recognition of the group, the relationships of this group with other more distantly allied species are unclear, and no formal nomenclatural recognition of this group has yet been attempted.
Spongy megagametophytic tissue The presence of spongy megagametophytic tissue is shown to be a synapomorphy (CI 50% in the strict consensus tree, 100% in the best-case scenario, Fig. 3.3A) supporting a small and widely distributed group of species that occurs in near-coastal situations from Tonga westwards to East Africa. When present, the
ch03.qxd
28/11/03
3:43 pm
Page 31
Classification Concepts in the Cycadaceae
31
Fig. 3.3. Character state distribution on the consensus cladogram. Species of Cycas indicated by their specific epithet. (A) Seed characters. (B) Seed characters. (C) Megasporophyll characters. (D) Female cone characters.
ch03.qxd
28/11/03
3:43 pm
32
Page 32
K.D. Hill
‘truncate’ microsporophyll
Fig. 3.3. (continued) Character state distribution on the consensus cladogram. Species of Cycas indicated by their specific epithet. (E) Megasporophyll length (cm). (F) Pollen cone. (G) Microsporophyll characters. (H) Stem characters.
ch03.qxd
28/11/03
3:43 pm
Page 33
Classification Concepts in the Cycadaceae
33
Fig. 3.3. (continued) Character state distribution on the consensus cladogram. Species of Cycas indicated by their specific epithet. (I) Petiole length (cm). (J) Leaflet characters.
spongy tissue may cause the seeds to float and facilitate aquatic dispersal on oceanic currents (Dehgan and Yuen, 1983), although not all seeds with spongy tissue float. Although the taxonomic implications of this character were recognized by these authors, misidentifications of study materials and a general lack of understanding of specific limits hindered application of these observations. The group of species, defined by presence of spongy megagametophytic tissue, was recognized as subsection Rumphiae by Hill (1994). The 50% CI arises from the uncertain placement of Cycas riuminiana Porte ex Regel, an inland forest species from the Philippines that is not aquatically dispersed and lacks other features of subsection Rumphiae. Placement of this species requires further study.
Fibrous sarcotesta A fibrous or corky layer within the sarcotesta that apparently grows outwards from the outer surface of the sclerotesta is present in a range of taxa, and has been recognized as a grouping character by Hill (1996) and Wang (1996). This character state appears to have been gained near the base of the tree, below Cycas pectinata Buchanan-Hamilton (Fig. 3.3B), and lost below C. riuminiana, although deltran optimization suggests two independent acquisitions on the clades below
ch03.qxd
28/11/03
3:43 pm
Page 34
34
K.D. Hill
C. pectinata and below C. inermis Loureiro. The moderate level of homoplasy shown (CI 33%) highlights the uncertain placement and need for further study of C. apoa K.D. Hill, an inland forest species from New Guinea. Nevertheless, this is shown to be a useful character for defining and recognizing several groups.
Ovule number The number of ovules per megasporophyll has been used both in defining species (Mueller, 1874; Wei, 1994; Chang et al., 1998) and in subdividing the genus (Warburg, 1900; Pilger, 1926). This character is highly variable in most if not all species, with only a slight tendency for mainland Asian species to carry fewer ovules (CI 9%, Fig. 3.3C). The degree of overlap, however, renders this character ineffective both in recognizing species and in defining groups.
Megasporophyll pectinate The pectinate state is shown to be ambiguous (Fig. 3.3D). In one scenario, the transition to the dentate or entire condition is a synapomorphy defining a large clade including Cycas inermis and all species above (corresponding to section Cycas), with the exception of two apparent reversals in C. scratchleyana F. Mueller and C. riuminiana (CI 33%). This section was first recognized informally by Carruthers (1893) and formally by Schuster (1932). This group has been recognized by most workers at some level, ranging from subsection (Smitinand, 1971) through section (above and Hill, 1996) to subgenus (Wang, 1996). In fact, only the systems presented by Deghan and Deghan (1988) and de Laubenfels and Adema (1998) have failed to discriminate this group.
Seed cone closed or open The closed seed cone is shown to be plesiomorphic (Fig. 3.3D). The transition to the open condition is a synapomorphy (CI 50%) defining a large clade including Cycas inermis and all species above (section Cycas) as well as a clade of two species – C. lindstromii S.L. Yang, K.D. Hill & Hiêp and C. pranburiensis S.L. Yang, K.D. Hill, W. Tang & Vatcharakorn. This character is correlated with the pectinate megasporophyll, but with less ambiguity.
Megasporophyll long or short A long megasporophyll was regarded by Schuster (1932) as a diagnostic character for subsection Pandemicae. The analysis indicates that this condition has arisen independently in at least four lineages (CI 17%, Fig. 3.3E). This character is con-
ch03.qxd
28/11/03
3:43 pm
Page 35
Classification Concepts in the Cycadaceae
35
sequently of no value in recognizing major groupings, although it is useful in discriminating related species and is a potential synapomorphy for Cycas thouarsii R. Brown ex Gaudichaud and C. micronesica K.D. Hill.
Pollen cone cylindrical or ovoid A cylindrical pollen cone was regarded by de Laubenfels and Adema (1998) as a diagnostic character for the new genus Epicycas, although the cylindrical state was acknowledged also to occur in Cycas. The analysis indicates that this condition is present in three separate lineages (CI 14% for cone shape, including the states ovoid and narrow-ovoid, Fig. 3.3F) and that the cylindrical condition may in fact be plesiomorphic for Cycas. Cone shape is also shown to be homoplastic within clades, and is consequently of little value in recognizing natural groups.
Microsporophyll soft, waxy or hard Microsporophyll texture shows no clear evolutionary progression, with the presence of all three states near the base of the tree and an ambiguous ancestral condition (Fig. 3.3G). The different states are, however, highly consistent (CI 100%). The soft condition defines the group discussed above as section Stangerioides; the waxy condition characterizes two small clades (the Cycas revoluta clade and the C. wadei clade) and C. panzhihuaensis L. Zhou & S.Y. Yang; and the hard condition prescribes a large clade that includes C. pectinata and all species above. The soft pollen cone character was used by Smitinand (1971) in defining section Stangerioides.
Microsporophyll ‘truncate’ Microsporophylls with a shortened apex (‘truncate’) were regarded by de Laubenfels and Adema (1998) as the diagnostic character for subgenus Truncata. The analysis indicates that such shortening has arisen independently in three separate clades (CI 22%, Fig. 3.3H) and is consequently of little value in recognizing major groupings, although it does suggest the possibility that Cycas rumphii Miquel, C. micronesica and C. seemannii A. Braun might form a clade.
Caudex with a bulbous base A caudex with a bulbous base was regarded by de Laubenfels and Adema (1998) as the key diagnostic character for the new genus Epicycas. The analysis indicates that this condition has arisen independently in at least two and possibly six separate lineages (CI 11%, Fig. 3.3H) and is consequently of little value in recognizing major groupings.
ch03.qxd
28/11/03
3:43 pm
Page 36
36
K.D. Hill
Caudex wholly subterranean Wholly hypogeous growth habits are shown by a few species. This character and the bulbous base (above) were concatenated by de Laubenfels and Adema (1998) in defining the genus Epicycas. The truly hypogeous habit is much less frequent than the bulbous base, and is shown to have arisen in two groups independently (CI 25%, Fig. 3.3H). This character was used by Smitinand (1971) in defining section Stangerioides.
Petiole long or short A short petiole (less than 30 cm long) was regarded by de Laubenfels and Adema (1998) as a key diagnostic character for Cycas subgenus Revoluta, and a long petiole for subgenus Pectinata. The analysis indicates that the long condition has arisen independently in at least seven different lineages (CI 20%, Fig. 3.3I), possibly with subsequent reversals, and is consequently of little value in recognizing major groups.
Leaflet dichotomously divided Dichotomously divided leaflets are a striking character present in four species occurring in southern China and Vietnam. This section was recognized as a series Multipinnata by Wang (1996), who regarded the divided state as primitive. This analysis shows the divided state to be advanced (Fig. 3.3J), but is not sufficiently detailed to examine the value of this character as a grouping character. Other morphological characters do not separate these four taxa and the divided leaflets are a useful grouping character on this basis.
Leaflet margins revolute Revolute leaflet margins were first used to subdivide the genus by Miquel (1843). In contrast, flat, often undulate leaflet margins were considered characteristic of the genus Epicycas by de Laubenfels and Adema (1998). Flat leaflets are here shown to be ancestral. The change to revolute margins is shown by the analysis to occur in several lineages (CI 22%, Fig. 3.3K), and is consequently ineffective in defining groups, although it is at times a useful character in discriminating species within groups.
Leaflet apex spinescent A spinescent leaflet apex was regarded by Schuster (1932) as a diagnostic character for subsection Endemicae. The analysis indicates that this condition has arisen independently in all major lineages (CI 10%, Fig. 3.3K), possibly more
ch03.qxd
28/11/03
3:43 pm
Page 37
Classification Concepts in the Cycadaceae
37
than once in some of these clades, and is consequently of little value in recognizing major groups.
Leaflet midrib raised or flat A flat midrib was regarded by de Laubenfels and Adema (1998) as a diagnostic character for Cycas subgenus Revoluta. The analysis indicates that this condition has arisen independently in three separate lineages (CI 12%, Fig. 3.3K), possibly more than once in some clades, and is consequently of little value in recognizing major groups. Alternatively, flat midribs can be almost as parsimoniously explained by only two origins and multiple subsequent losses.
Assessment of Previous Classifications Schuster (1932) attempted the first formal subgeneric classification of Cycas (Fig. 3.4A). He recognized three major groups (sections), corresponding basically to the pectinate megasporophyll (section Indosinenses) and non-pectinate megasporophyll (section Lemuricae) groups, with the further separation of C. revoluta (in which he included C. taiwaniana Carruthers) as a monotypic group defined by narrow revolute leaflets and tomentose ovules (section Asiorientales). He further subdivided section Lemuricae into two subsections, Pandemicae and Endemicae, on differences in tips of pinnae and length of megasporophylls. Although the major groups erected by Schuster take nomenclatural priority, two of his three sections are paraphyletic as he circumscribed them. He also allocated species and infraspecific taxa within these groups erratically and often apparently on the basis of guesswork, placing many previously recognized species as infraspecific taxa within complex hierarchies under C. circinalis Linnaeus and C. rumphii. Many of these are clearly distinct and distantly related species, rendering his concepts of C. circinalis and C. rumphii polyphyletic. In addition, Schuster’s work does not comply with the requirement for subgroups including the type of a genus to carry automatically the generic name (the autonym rule – International Code of Botanical Nomenclature (ICBN), Greuter et al., 2000). Hence section Lemuricae should correctly be section Cycas and subsection Pandemicae should be subsection Cycas. In summary, Schuster recognized five infrageneric taxa, three of which are shown to be polyphyletic (Fig. 3.4A). Only four infrageneric groups are marked on Fig. 3.4A. One of the marked groups (Endemicae) is monophyletic, while the other is monotypic (and thus monophyletic by definition). If the unmarked lineages collectively form another of his taxa (section Lemuricae?) then it might be said that it is poly- and not paraphyletic. Similarly, Pandemicae would be most appropriately described as polyphyletic. Smitinand (1971) proposed a very different major division into two groups (Fig. 3.4B), separating Cycas micholitzii Dyer in the monotypic section Stangerioides on the basis of the soft, shortly apiculate microsporophylls in very small, slender
ch03.qxd
28/11/03
38
3:43 pm
Page 38
K.D. Hill
Fig. 3.4. Previous systems of classification. Incorrect names under the ICBN are indicated by*. Monophyletic clades supported in this analysis shown in bold. Species of Cycas indicated by their specific epithet. Classifications by: (A) Schuster (1932); (B) Smitinand (1971); (C) Dehgan (1987); (D) Hill (1996, 1999).
ch03.qxd
28/11/03
3:43 pm
Page 39
Classification Concepts in the Cycadaceae
39
Fig. 3.4. (continued) Previous systems of classification. Incorrect names under the ICBN are indicated by*. Monophyletic clades supported in this analysis shown in bold. Species of Cycas indicated by their specific epithet. Classifications by: (E) Wang (1996); (F) de Laubenfels (1998).
cones, and the dwarf, mainly subterranean habit with very few leaves (‘Stangerioid’ habit). The remainder of the genus he placed in section Cycas, which he divided into two subsections on the basis of pectinate megasporophylls (subsection Pectinatae, also given as Pinnatidae in the key) and non-pectinate megasporophylls (subsection Circinnalidae). The former subsection included C. revoluta. Again, the requirement for autonyms was not strictly followed, and subsection Circinnalidae is correctly (and automatically) subsection Cycas. Subsection Pectinatae (or Pinnatidae) is also illegitimately described, no type species being designated. Of the four infrageneric taxa recognized by Smitinand, three are shown to be polyphyletic (Fig. 3.4B). Dehgan and Dehgan (1988) alluded to an infrageneric classification and published a series of names with a major division based on presence or absence of spongy tissue in seeds and incorporating differences in pollen structure (Fig. 3.4C). This classification was not formally published, and neither the requirement for autonyms nor the rule of priority was followed. Although the character basis for this subdivision was sound, incorrect identifications and a lack of understanding of specific limits made the arrangement untenable. The proposed classification recognized two subgenera with three sections and three subsections. No assignment of species to the lower groups was made, however, apart from the
ch03.qxd
28/11/03
3:43 pm
40
Page 40
K.D. Hill
listing of a single ‘typical’ species. On the basis of these inclusions, the two subgenera are shown to be polyphyletic, as is the one section that has more than one included species (Fig. 3.4C). Hill (1996, 1999) presented an arrangement incorporating four sections corresponding in name to the three erected by Schuster plus section Stangerioides as described above (Fig. 3.4D), in all cases with different circumscriptions, on the basis of morphological cladistic analyses. Nine subordinate groups corresponding to subsections were also proposed, most of them informal. Of the four sections, two are shown to be polyphyletic, and one of the nine subsections is shown to be polyphyletic, with two others possibly paraphyletic (Fig. 3.4D). Wang (1996) presented another formal infrageneric classification, recognizing two subgenera on the basis of pectinate vs. non-pectinate megasporophylls. Within these a number of sections, subsections and series were erected (Fig. 3.4E). These are not altogether internally consistent and several of the names are in contravention of the ICBN (Greuter et al., 2000). Of the two subgenera, one (subgenus Panzhihuaensis) is shown to be polyphyletic, and one of the sections is paraphyletic (Fig. 3.4E). Full enumeration of species included in a number group is not attempted, making assessment uncertain. De Laubenfels and Adema (1998) presented an entirely different system, recognizing a new genus Epicycas for taxa with a largely subterranean habit and a bulbous underground base. They divided the remainder of Cycas into four subgenera based on a combination of leaf, microsporophyll and megasporophyll morphology. Every generic and infrageneric group as circumscribed by these authors is shown to be polyphyletic (Fig. 3.4F).
Conclusions No previously published system of classification is wholly in accord with the results of molecular and morphological phylogenetic analyses. Many of the key characters on which previously published classifications were based are shown to be plesiomorphic, homoplastic or autapomorphic and of little value in defining infrageneric relationships and developing classifications based on these. In particular, separation of the genus Epicycas is clearly unwarranted (see Chen et al., Chapter 5 this volume). However, certain of the characters discussed above are shown to be synapomorphic characters that are useful in defining natural groups. Seed characters in particular are shown to be highly consistent, as are some aspects of sporophyll morphology. These characters are useful in developing a workable key to groups (Table 3.3) and to species within the genus Cycas. Data are still lacking on many recognizable species in the genus, and consequently placement of a number of species on the basis of the characters discussed above is uncertain. It is at present premature to develop a comprehensive system of infrageneric classification on this basis, but an interim arrangement recognizing demonstrably monophyletic groups and the nearest possible placements of species or groups incertae sedis can
ch03.qxd
28/11/03
3:43 pm
Page 41
41
Classification Concepts in the Cycadaceae
Section Asiorientales Section Wadeanae
Section Stangerioides
Section Indosinenses
Species incertae sedis
Subsection Cycas
Section Cycas
Subsection Rumphiae
Subsection Endemicae
Fig. 3.5. Comprehensive enumeration of species in Cycas. Consensus morphological and molecular tree shown in bold, other species interpolated from morphological cladistic analysis using available data.
ch03.qxd
28/11/03
3:43 pm
Page 42
42
K.D. Hill
Table 3.3. Key to major recognizable subdivisions of the genus Cycas. 1 1 2 2 3
Ovules tomentose Ovules glabrous Megasporophyll lamina pectinate Megasporophyll lamina not pectinate Male cones soft, microsporophyll apices not deflexed 3 Male cones rigid, waxy, microsporophyll apices deflexed 4 Sclerotesta not ribbed
Section Asiorientales 2 3 Section Cycas 4 5 Cycas panzhihuaensis L. Zhou & S.Y. Yang Section Wadeanae
4 Sclerotesta ribbed 5 Microsporophylls flexible, rounded; sarcotesta not fibrous; sclerotesta verrucose Section Stangerioides 5 Microsporophylls rigid, acuminate; sarcotesta fibrous; sclerotesta smooth Section Indosinenses
be informally presented here (Table 3.4). The primary division is made into sections rather than subgenera for the reason that basal branches are short and additional data may well collapse these or elucidate different relationships. The sections recognized are distinct and well-supported clades. A comprehensive enumeration of taxa in the genus Cycas that are recognizable as species or comparable terminal taxa is presented in Fig. 3.5. The tree is based on the strict consensus of a combined morphological and molecular analysis (shown in bold). Species for which molecular data are unavailable are interpolated on the basis of morphological cladistic analysis using available data (Hill, 1999 and in preparation). Not all species are adequately known, and data are incomplete for a number of taxa. This again represents an informal interim arrangement that shows known groups and relationships and highlights areas requiring further study.
Acknowledgements The Hermon Slade Foundation is warmly thanked for the financial support that allowed this and related studies to proceed. The Vietnamese Institute of Ecology and Biological Resources and the Chinese Academy of Science are thanked for assistance with laboratory and field studies in Vietnam and China. Kampon Tansacha and the Nong Nooch Tropical Garden are gratefully acknowledged for hospitality and logistical assistance. Anders Lindström assisted in the field and in valuable discussions of the taxonomic and distributional limits of the cycads of Asia. The keepers of the herbaria at A, B, BKF, BM, G, K, L, NY and P are acknowledged for access to their collections. Peter Weston is thanked for constructive comment on earlier drafts of the manuscript.
ch03.qxd
28/11/03
3:43 pm
Page 43
Classification Concepts in the Cycadaceae
43
Table 3.4. A provisional classification of Cycas. Section Asiorientales Synapomorphies: encrypted stomata 3 species. No subgroups Section Wadeanae Synapomorphies: ribbed seeds 2 species. No subgroups Section Stangerioides Synapomorphies: verrucose seeds, soft pollen cones About 25 species. No clear subgroups. Section Indosinenses Synapomorphies: fibrous sclerotesta, hard pollen cones About 15 species. No clear subgroups. Section Cycas Synapomorphies: open seed cones, non-pectinate megasporophylls About 53 species. Three distinct monophyletic subgroups and a number of unplaced species Subsection Cycas: about 4 species Synapomorphies: fibrous sclerotesta Subsection Rumphiae: about 10 species Synapomorphies: spongy megagametophyte Subsection Endemicae: about 32 species Synapomorphies: palisade tissue in lower mesophyll species incertae sedis; about 8 species
References Amoroso, V.B. (1986) Morphological study of the sporophylls of Philippine Cycas. Philippine Journal of Science 115(3), 177–198. Backer, C.A. and Bakhuizen van den Brink, R.C. (1963) Cycadaceae. In: Flora of Java, Vol. 1. Noordhoff, Groningen, The Netherlands, p. 87. Carruthers, W. (1893) On Cycas taiwaniana sp. nov. and C. seemannii A.Br. Journal of Botany 31, 1–3; t. 330–331. Chang, H.T., Huang, Y.Y. and Zheng, H.X. (1998) Acta Sci. Nat. Univ. Sunyatseni 37, 8. [Cycas septemsperma.] Chen, C.J. and Stevenson, D.W. (1999) Cycadaceae. In: Wu, Z.Y. and Raven, P.H. (eds) Flora of China, Vol. 4, Cycadaceae through Fagaceae. Science Press, Beijing and Missouri Botanical Garden Press, St Louis, Missouri, pp. 1–7. De Candolle, A.P. (1868) Cycadeae. In: Prodromus Systema Natura and Regnum Vegetabile 16(2). Victor Massen, Paris, pp. 361–521. De Laubenfels, D.J. and Adema, F. (1998) A taxonomic revision of the genera Cycas and Epicycas gen. nov. (Cycadaceae). Blumea 43, 351–400. Dehgan, B. and Dehgan, N.B. (1988) Comparative pollen morphology and taxonomic affinities in Cycadales. American Journal of Botany 75, 1501–1516.
ch03.qxd
28/11/03
44
3:43 pm
Page 44
K.D. Hill
Dehgan, B. and Yuen, C.K.K.H. (1983) Seed morphology in relation to dispersal, evolution and propagation of Cycas L. Botanical Gazette 144, 412–418. Fu, S.H., Cheng, W.C., Fu, L.K. and Chen, C.J. (1978) Cycadaceae. In: Cheng, W.C. and Fu, L.K. (eds) Flora Reipublicae Popularis Sinicae 7. Science Press, Beijing, China, pp. 4–17. Greuter, W., McNeill, J., Barrie, F.R., Burdet, H.M., Demoulin, V., Filgueiras, T.S., Nicolson, D.H., Silva, P.C., Skog, J.E., Trehane, P., Turland, N.J. and Hawksworth, D.L. (2000) International Code of Botanical Nomenclature (Saint Louis Code). Koeltz Scientific Books, Köningstein, Germany, 474 pp. Hiêp, N.T. and Vidal, J.E. (1996) Cycadaceae. In: Morat, Ph. (ed.) Flore du Cambodge, du Laos et du Viêtnam, Vol. 28, Gymnospermae. pp. 6–23. Hill, K.D. (1994) The Cycas rumphii complex (Cycadaceae) in New Guinea and the Western Pacific. Australian Systematic Botany 7, 543–567. Hill, K.D. (1995) Infrageneric relationships, phylogeny and biogeography of the genus Cycas (Cycadaceae). In: Vorster, P. (ed.) Proceedings of the Third International Conference on Cycad Biology. Cycad Society of South Africa, Stellenbosch, South Africa, pp. 139–162. Hill, K.D. (1996) A taxonomic revision of the genus Cycas (Cycadaceae) in Australia. Telopea 7, 1–64. Hill, K.D. (1999) Cycas – an evolutionary perspective. In: Chen, C.J. (ed.) Biology and Conservation of Cycads, Proceedings of the Fourth International Conference on Cycad Biology. International Academic Publishers, Beijing, China, pp. 98–115. Johnson, L.A.S. (1959) The families of cycads and the Zamiaceae of Australia. Proceedings of the Linnaean Society of New South Wales 84, 64–117. Miquel, F.A.W. (1843) Genera et species Cycadearum viventium. Linnaea 17, 675–744. Miquel, F.A.W. (1861) Prodromus Systematis Cycadearum. Van der Post, Utrecht, Holland, 35 pp. Mueller, F.A.W. (1874) Fragmenta Phytographie Australiae, Vol. 8. Government Printer, Melbourne, Australia, 304 pp. Pilger, R. (1926) Cycadaceae. In: Engler, A. (ed.) Die Naturlichen Pflanzenfamilien, 2nd edn 2, 13, pp. 44–82. Schuster, J. (1932) Cycadaceae. In: Engler, A. (ed.) Das Pflanzenreich, Fascicle 99, Vol. 4, Part 1, pp. 1–168. Smitinand, T. (1971) The genus Cycas Linn. (Cycadaceae) in Thailand. Natural History Bulletin of the Siam Society 24, 163–175. Stevenson, D.W. (1992) A formal classification of the extant cycads. Brittonia 44, 220–223. Wang, D.Y. (1996) Systematic classification and brief introduction to Cycadales (Chapter 2) and Taxonomy of Cycas in China (Chapter 3). In: Wang, F.X. and Liang, H.B. (eds) Cycads in China. Guangdong Science and Technology Press, Guangdong, China, pp. 9–142. Warburg, O. (1900) Cycadaceae. In: Warburg, O. (ed.) Monsunia. Engelmann, Leipzig, Germany, pp. 178–181. Wei, F.N. (1994) A new cycad from Guangxi. Guihaia 14, 300. [Cycas ferruginea.]
ch04.qxd
28/11/03
3:44 pm
Page 45
Morphological Characters Useful in Determining Species Boundaries in Cycas (Cycadaceae)
4
Anders Lindström Nong Nooch Tropical Botanical Garden, Najomtien, Sattahip, Chonburi, Thailand
Abstract Morphological characters within the genus Cycas, based on previously published descriptions and on recent field research, are evaluated for their usefulness in distinguishing among distinct, yet morphologically similar, pairs of taxa within the genus. A standardized suite of taxonomically useful morphological characters in support of future designation of species within Cycas is recommended.
Introduction Taxonomy has traditionally used certain morphological characters or suites of morphological characters to define taxa. These characters are often chosen in a way that permits convenient measurement from herbarium specimens. For cycads, this usually restricts the choice to those characters derived from limited aspects of leaf and sometimes cone morphology, since these are usually the only parts represented in herbarium accessions. Annotations of herbarium vouchers often lack data significant to taxonomic studies or even identification. Leaf length, petiole length, stem height, stem profile and branching pattern are examples of these characters that cycad researchers require in their evaluation of taxa. Cycads, especially representatives of the genus Cycas Linnaeus, typically have long leaves that are difficult to process as herbarium specimens. Therefore, the ©CAB International 2004. Cycad Classification: Concepts and Recommendations (eds T. Walters and R. Osborne)
45
ch04.qxd
28/11/03
46
3:44 pm
Page 46
A. Lindström
majority of Cycas vouchers in herbaria are represented only by portions of leaves – usually a group of about four to six pinnae. Specimens often lack the petiole, a feature of the plant that can be of particular taxonomic significance. Labels accompanying Cycas specimens most often lack the necessary data to describe the whole leaf and seldom indicate from which portion of the leaf the pinnae were obtained. Within a leaf, terminal, central and basal pinnae each have distinctive features. Herbarium accessions only infrequently include reproductive material from Cycas plants and representation of both male and female carpological material for any given taxon is rare. A major problem associated with past herbarium-based Cycas studies has been that different suites of characters had to be used in the analysis of different specimens. The taxonomist is often faced with having to study specimens that have different portions of a leaf represented, or a portion of a leaf on one specimen and a complete leaf on another, different reproductive material present or reproductive material absent, and few descriptive or quantitative data in the annotations. A further limitation arises in that herbarium specimens often do not necessarily represent the ‘average’ plant for any given population. It is known that plants within a population of cycads express a wide variation for numerous morphological characters. Identifying a plant that represents a population can be time consuming, difficult and is often not a priority when herbarium specimens are collected. Very few specimen labels include details of the morphological variation within the population. The assignment of sectional and subsectional ranks in Cycas has been based primarily on the sharing of characters associated with the seed, and secondarily with other reproductive and pinna features (see Hill, Chapter 3 this volume). For any given section or subsection, seed structure is in several aspects linked to animate and inanimate dispersal processes and thus implicit in part for infrageneric genetic exchange. However, these seed characters are usually not available from herbarium specimens or accompanying data, making it difficult to assign the sectional or subsectional ranks. A suite of useful characters, that can be conveniently documented in the field and later included in the specimen annotation, is needed to assist future taxonomic studies at the section and subsection ranks. To facilitate future taxonomic studies at the sectional, subsectional and specific ranks within Cycas, it is critical to identify a suite of morphological characters that can be accessed, whether through herbarium specimens with their associated annotations and/or through publications. Hill (see Chapter 3 this volume) proposes a suite of such characters. Understanding the significance of a specific suite of characters is essential in the classification of new taxa into the appropriate section or subsection within the genus. The aim of this chapter is to test a suite of characters that may reflect measurable and taxonomically useful features in a consistently reliable classification within Cycas.
ch04.qxd
28/11/03
3:44 pm
Page 47
Determining Species Boundaries in Cycas
47
Materials and Methods The author has visited wild populations of almost every non-Australian species in the genus Cycas and herbarium specimens have been made from each population. Prior to collecting specimens, measurements were taken from a number of mature plants within each population using a pre-defined suite of characters. The characters chosen for this study were based on their use in past taxonomic studies within the genus, as well as those that showed potential for supporting future taxonomic studies. The 30 quantitative and qualitative characters (Table 4.1) are associated with seven aspects of a plant: plant habit, leaf, pinna, cataphyll, megastrobilus, microstrobilus and seed. Most of these characters and their measurement aspects are presented in Fig. 4.1. To understand the potential value of the chosen morphological characters for future studies within Cycas, three pairs of distinct yet morphologically similar species were identified, and measurements were obtained from selected wild populations. One population was sampled for four of the six species (C. edentata de Laubenfels, C. elephantipes A. Lindström & K.D. Hill, C. micronesica K.D. Hill and C. siamensis Miquel) and two populations were sampled for the remaining species (C. seemannii A. Braun and C. rumphii Miquel). The first pair of taxa studied was Cycas siamensis and C. elephantipes. Both species are arborescent and native to Thailand. The second pair, C. edentata and C. rumphii, are tall, erect, arborescent cycads from islands in the Pacific Ocean. Cycas edentata and C. rumphii are believed to hybridize naturally where their distributions overlap. The third pair examined was C. seemannii and C. micronesica. Cycas seemannii is a tall, erect, arborescent and usually unbranched species found in Fiji, Vanuatu, the Tonga Islands and New Caledonia. Cycas micronesica (often misidentified in collections as either C. circinalis Linnaeus or C. rumphii) is an erect, arborescent species up to 8 m tall from the Mariana Islands. Populations of these two species are widely separated geographically and no spontaneous hybridization is likely to occur. Nevertheless, molecular evidence indicates that, in addition to being morphologically similar, they are evolutionarily closely related (Ken Hill, Australia, 2002, personal communication). Measurements were recorded for multiple plants within each population for the six species. In some cases, not all characters were measured – male cones and seeds were often not available for all species during the time of the fieldwork. For each population sampled, the number of plants measured, the mean for each character and the standard deviation for each character are listed in Tables 4.2–4.4. Data derived from each species pair were compared using a t-test with an alpha level of 0.05. The t-test assesses whether the means of two groups are statistically different from each other and yields the same results as a one-way analysis of variance (ANOVA). Results from the t-tests are included in Tables 4.2–4.4.
ch04.qxd
28/11/03
3:44 pm
Page 48
48
A. Lindström
Table 4.1. Vegetative and reproductive characters for taxonomic studies in Cycas. Characters of questionable use for taxonomic studies are indicated by *. Characters determined to be unstable and variable among species at the onset of the investigation are indicated by **. Unstable and variable characters were removed from the analyses. Plant part
Character/character state
Plant habit
Stem height (cm) Stem minimum diameter (cm) Leaf number Length (cm)* Petiole length (cm) Petiole thickness (mm) Number of spines on petiole** Petiole length percentage spines Terminus (spine, single leaflet, leaflet pair)** Number Length (cm) Width (mm) Basal width (mm) Arrangement of pinnae (opposing, alternate)** Spacing between pinnae (mm) Pinna length (cm) Length (cm) Megasporophyll length (cm) Megasporophyll lamina length (mm) Megasporophyll lamina width (mm)* Megasporophyll apical spine length (mm)* Megasporophyll lateral spine number Megasporophyll number of ovules Microstrobilus length, excluding peduncle (cm) Microstrobilus diameter (cm) Microsporophyll length (mm) Microsporophyll width (mm) Microsporophyll apical spine length (mm) Length (mm) Width (mm)
Leaf
Pinna
Cataphyll Megastrobilus
Microstrobilus
Seed
Results Characters of questionable use Three characters preliminarily included in the study as potentially useful taxonomic characters were proven to be unstable and variable within and among
ch04.qxd
28/11/03
3:44 pm
Page 49
Determining Species Boundaries in Cycas
49
Fig. 4.1. Vegetative and reproductive characters measured from Cycas specimens: (1) leaf length; (2) petiole length; (3) pinna number; (4, 8) pinna length; (5) pinna width; (6) pinna basal width; (7) spacing between pinnae; (9) megasporophyll length; (10) megasporophyll lamina length; (11) megasporophyll lamina width; (12) megasporophyll apical spine length; (13) megasporophyll lateral spine number; (14) megasporophyll number of ovules; (15) microstrobilus length (excluding peduncle); (16) microstrobilus diameter; (17) microsporophyll length; (18) microsporophyll width; (19) microsporophyll apical spine length. Characters not shown (see Table 4.1), but included in the study, are those associated with the stem, cataphyll, seed, petiole thickness and petiole length percentage spines.
ch04.qxd
SD
n
ELEb Mean
P
SD
8.43 11.43 18.86
4.72 1.72 3.24
10 10 10
87.6 22.7 37.4
48.20 3.23 13.60
< 0.001* < 0.001* < 0.001*
12 12 12 12
75.58 13.79 7.4 69.75
10.35 3.22 1.3 26.37
10 10 10 10
167.3 25.7 15.0 54.0
12.08 6.52 2.7 19.6
< < < =
0.001* 0.001* 0.001* 0.124
12 12 12 12 12 12
73.54 10.04 5.83 3.25 4.33 2.82
8.67 1.6 0.58 0.45 0.78 1.05
10 10 10 10 10 9
117.3 18.76 9.1 3.9 8.2 39.3
20.0 2.8 0.9 0.7 1.3 23.4
< < < = <
2 cm wide, the widest being found in C. euryphyllidia Vázquez Torres, Sabato & D.W. Stevenson. Exceptions to the generally wide leaflet morphology are seen in C. norstogii D.W. Stevenson, C. mirandae Vovides, Pérez-Farrera & Iglesias and C. alvarezii, all three of which have leaflets < 1.5 cm wide. Moretti et al. (1980) were the first to recognize species complexes within the genus highlighting separate Ceratozamia mexicana and C. matudae complexes (Fig. 9.1). Stevenson et al. (1986) proposed two main groups in Ceratozamia, based on leaflet morphology: (i) a C. mexicana group with long narrow leaflets; and (ii) a C. euryphyllidia/miqueliana group with shorter and wider leaflets. However, throughout the distribution of Ceratozamia, various other related groups or complexes can also be found. We have confirmed the complexes identified by the earlier workers, as well as other new complexes – principally in southern and south-eastern
ch09.qxd
28/11/03
3:48 pm
Page 111
Relationships and Phytogeography in Ceratozamia
111
Fig. 9.1. Distribution in Mexico and Guatemala of four Ceratozamia species complexes: A = C. kuesteriana species complex; B = C. latifolia species complex; C = C. mexicana species complex; D = C. matudae species complex.
Mexico. Our main two additional groups are a C. norstogii complex (Fig. 9.2) (Pérez-Farrera et al., 1999; Vovides et al., 2001) and a C. miqueliana complex (Fig. 9.3) (Pérez-Farrera et al., 2001). The distribution of these complexes is apparently related to their habitats. For instance, C. miqueliana H. Wendland and related species occur in evergreen tropical rainforest and are not generally found at elevations > 1000 m; these plants are characterized by having both decumbent and erect female cones, and leaves up to 2.5 m long with oblanceolate leaflets. By contrast, species of the C. norstogii group are largely found in oak and pine/oak forests at elevations between 800 and 1400 m; characteristic features of these species are branching or cylindrical and non-branching trunks, erect male and female cones, and linear, flat or channelled leaflets with a straight or spirally twisted rachis. The C. robusta species complex (sensu Stevenson et al., 1986) (Fig. 9.4) appears to have a wide range of habitat type ranging over evergreen and deciduous tropical forests, oak forests and cloud forests; plants within this group are large and robust with leaves up to 3 m long and bear large erect or decumbent male and female cones. On the basis of gross morphology we now believe that there are at least seven species complexes in the genus: 1. The Ceratozamia mexicana complex is mainly distributed on the transMexican Neovolcanic mountain range, covering central Veracruz, parts of Puebla and Hidalgo states and possibly north-eastern Oaxaca (Fig. 9.1). Included
ch09.qxd
28/11/03
112
3:48 pm
Page 112
A.P. Vovides et al.
Fig. 9.2. Distribution in Mexico (encircled area) of the Ceratozamia norstogii species complex.
here is C. mexicana Brongniart var. robusta Dyer (sensu Thiselton-Dyer, 1882–1886; Vovides et al., 1983) and C. mexicana var. mexicana. 2. The Ceratozamia latifolia complex, a species group to the north and north-east of the trans-Mexican Neovolcanic mountain range, is a somewhat
Fig. 9.3. Distribution in Mexico (encircled area) of the Ceratozamia miqueliana species complex.
ch09.qxd
28/11/03
3:48 pm
Page 113
Relationships and Phytogeography in Ceratozamia
113
Fig. 9.4. Distribution in Mexico, Guatemala and Belize (encircled area) of the Ceratozamia robusta species complex.
loose assemblage and may subsequently prove to comprise more than one complex. As presently circumscribed, the group consists of C. latifolia Miquel, C. hildae G.P. Landry & M.C. Wilson, C. microstrobila Vovides & J.D. Rees, C. huastecorum S. Avendaño, Vovides & Castillo-Campos, C. brevifrons Miquel (= C. mexicana Brongniart; see Hill et al., Appendix 1 this volume) and C. morettii Vázquez Torres and Vovides (the latter two species are on the Neovolcanic transverse mountain range). 3. The Ceratozamia kuesteriana complex occurs in north-eastern Mexico and includes C. kuesteriana Regel, C. sabatoi Vovides, Vázquez Torres, Schutzman & Iglesias, and C. zaragozae Medellin-Leal. These are three taxa with narrow lanceolate to linear leaflets. Other taxa in this region are at present under investigation. 4. The wide leaflet Ceratozamia miqueliana complex which occurs to the south and south-east of the trans-Mexican Neovolcanic belt, comprises about five taxa. These are C. miqueliana, C. euryphyllidia, C. zoquorum Pérez-Farrera, Vovides & Iglesias, a new taxon that will be described in the near future (Ceratozamia sp. “becerrae”) and another possibly new species from Chiapas. 5. The Ceratozamia norstogii complex is largely distributed to the south in Chiapas and eastern Oaxaca, with three apparently well-defined species: C. norstogii, C. alvarezii and C. mirandae Vovides, Pérez-Farrera & Iglesias (see Pérez-Farrera et al., Chapter 10, this volume). We are also tentatively placing a new taxon, yet to be described, in this complex (Ceratozamia sp. “chimalpensis”).
ch09.qxd
28/11/03
3:48 pm
Page 114
114
A.P. Vovides et al.
6. The Ceratozamia robusta (sensu Stevenson) group is found in Chiapas, Oaxaca, southern Veracruz, Guatemala and Belize. This is another loose assemblage in which we place C. robusta Miquel, C. whitelockiana Chemnick & T.J. Gregory, C. mixeorum Chemnick, T.J. Gregory & Salas-Morales, and a further new species of Ceratozamia, yet to be described, from Oaxaca. 7. The Ceratozamia matudae complex, which is distributed from the extreme south of Chiapas into Guatemala consists of C. matudae Lundell and a newly discovered species yet to be described (Ceratozamia sp.). The taxonomy of Ceratozamia is far from resolved and there is still much confusion at the species level within these complexes. It is hoped that the molecular, anatomical and phytogeographical studies at present under way on the genus will eventually contribute to the clarification of the status of species within Ceratozamia.
Habitats Most species of Ceratozamia can be found within 15–17.5° north and 92.5–98.5° west. Typical habitats for Ceratozamia in north-eastern Mexico are broadleaf deciduous cloud forest and pine/oak forest at elevations between 500 and 2000 m. In southern and south-eastern Mexico, Ceratozamia can be found in lowland evergreen tropical rainforest at elevations < 100 m, to cloud forests and pine/oak forests from about 600 to 1400 m. Populations can be locally abundant, and in some cases dominant in the herbaceous undergrowth of cloud forests, such as in the cases of C. mexicana in Veracruz and C. mirandae in Chiapas. Using multivariate analysis on taxa and their habitat types (McCune and Mefford, 1997) species with wide oblanceolate to oblong leaflets (C. euryphyllidia, C. mexicana var. robusta, C. miqueliana, C. zoquorum and C. sp. “becerrae”) are confined to evergreen tropical rainforest, whilst a number of species with narrow, linear or falcate leaflets are frequently associated with cloud forests, oak and pine/oak forests (Fig. 9.5).
Molecular Studies Molecular phylogenetic analyses on New World cycads are scarce, and those so far published have examined restriction fragment length polymorphism variation data at generic level (Caputo et al., 1991, 1993; Moretti et al., 1993; De Luca et al., 1995). In order to gain insight into phylogenetic relationships within the genus Ceratozamia, we undertook a study of variation in sequences of non-coding regions from the chloroplast and nuclear genomes. We selected two regions that have been used in phylogenetic studies at infrageneric level for a variety of plants (Schilling et al., 1998; Potter et al., 2000). One is the nuclear internal transcribed spacer (ITS) region, which has proved to be an excellent source of information
28/11/03
3:48 pm
Page 115
Relationships and Phytogeography in Ceratozamia
115
Number of species
Types of vegetation
Principal component 2
ch09.qxd
Principal component 1 Fig. 9.5. Relationships between taxa of Ceratozamia and habitat (BTSC, tropical sub-deciduous rainforest; BTP, evergreen tropical rainforest; BMM, cloud forest; BQ, oak forest; BC, coniferous forest; BTC, tropical deciduous forest). (A) Number of taxa of Ceratozamia present in six vegetation types. (B) Summary of a principal components analysis of habitat and leaflet width for taxa of Ceratozamia.
from the nuclear genome (Baldwin et al., 1995). However, only 29 characters out of 1083 of the ITS region were informative among 24 exemplars. The second region we examined is the chloroplast genome trnL-F non-coding region and trnL-trnF spacer. Here, the alignment for a subsample of 16 Ceratozamia sequences from the trnL-F non-coding region generated 998 characters, from which only two were informative. The selection of Zamia as an outgroup was based on the
ch09.qxd
28/11/03
3:48 pm
116
Page 116
A.P. Vovides et al.
results obtained with the analyses of the trnL-F non-coding region, where Zamia appeared as the sister group to Ceratozamia. The sequence data generated in this study were used to evaluate: (i) the hypothesis that Ceratozamia is monophyletic; and (ii) the species relationships within the genus. The results of these analyses yielded 112 trees of 51 steps; the strict consensus of these trees is shown in Fig. 9.6A (with bootstrap values and decay index) and one randomly chosen minimal length tree is presented in Fig. 9.6B to show average branch and distribution of the taxa. The low level of variation detected among species limits the conclusions that could be drawn from the study; however, the cladograms did enable us to infer a tentative but interesting scenario for the genus (González and Vovides, 2002). Other molecular techniques, such as random amplified polymorphic DNA (now in progress), might help to identify the more variable regions of the genome in order to enable us to design primers to sequence those regions in the near future. Microsatellite techniques also appear to be promising.
Leaflet Anatomy Leaflets from five adult plants from each of five taxa were taken from the midportion of a mature leaf. Leaflets were fixed in formalin acetic alcohol then hand microtome sections were made. These were stained in phloroglucinol HCl to show lignified tissues and in alcoholic Sudan III and IV mixture for cutin and other fatty acid tissues. Nineteen measurements were obtained from each individual plant using a calibrated eyepiece scale (Table 9.1). Analysis was done using analysis of variance (ANOVA) and multiple range tests from Statistica (StatSoft, Inc. Tulsa, Oklahoma, USA). Results so far from leaflet anatomy appear to be promising in our attempts to isolate useful characters for delimiting species. Nevertheless, a homogeneity of states has been observed in about 50% of the characters studied, reflecting a similar situation to that found in DNA sequencing of the ITS region mentioned previously. From a suite of 19 characters analysed using morphometric and phenetic techniques, nine were found to be taxonomically useful and enabled us to separate a mean of nine groups from a total of 18 operational taxonomic units.
Discussion Ceratozamia is largely found along the Sierra Madre Oriental, in southern and south-eastern Mexico and through Guatemala and Belize to Honduras. The genus presents disjunct populations in Mexico (Sabato and De Luca, 1985; Moretti et al., 1993) with the exception of C. robusta (sensu Stevenson) where the majority of populations are found in south-east Veracruz, north-west Chiapas and the southern sierra of Oaxaca, Belize and Guatemala. Future work will help
ch09.qxd 28/11/03
C. morettii
C. matudae
117
Fig. 9.6. Phylogenetic relationships within Ceratozamia based on ITS sequences (length 51, consistency index = 0.8235, retention index = 0.9151, rescaled consistency index = 0.7536) (from González and Vovides, 2002). (A) Strict consensus tree. (B) Minimal length tree chosen at random.
Page 117
C. matudae
C. mirandae
Relationships and Phytogeography in Ceratozamia
C. mirandae
3:48 pm
C. morettii
ch09.qxd
28/11/03
3:48 pm
118
Page 118
A.P. Vovides et al.
Table 9.1. Leaflet anatomical characters used to distinguish Ceratozamia taxa. Five plants from each of five taxa were sampled for the total 19 characters. Characters
Dimensions
Abaxial epiderm cells Cuticles abaxial and adaxial Epiderm cells abaxial and adaxial Palisade parenchyma cells Sclereids associated with vascular bundles Sclereids not associated with vascular bundles Bulliform-like cells Stomatal pores
Length and width Thickness Length and width Length and width Number, length and width
2 2 4 2 3
Number, length and width
3
Length and width Depth
2 1
Total number of characters examined
Number
19
to resolve whether these represent a single disjunct species or a complex of related taxa. Stevenson et al. (1986) proposed that leaflet morphology and habitat are correlated, with those species of Ceratozamia having wide and thin leaflets being confined to the more humid environments and taxa with narrow and more coriaceous leaflets to the more xeric habitats. We have confirmed this with our observations. So far, leaflet anatomy appears to group the species into nine groups. Figures 9.7 and 9.8 show examples of two useful anatomical characters: (i) there appear to be lower leaflet cuticle thickness differences between C. miqueliana and C. zaragozae but no differences in stomatal pore depth; and (ii) there are differences in the number of associated and non-associated fibres with the vascular bundles of the leaflets of C. mexicana var. robusta and C. kuesteriana (S. Avendaño, Veracruz, 2002, personal communication). Although there appears to be a certain correlation of the anatomical features with distribution, it is too early to draw conclusions. Stevenson et al. (1986) mentioned vicariance as a possible explanation for the present distribution of Ceratozamia and that the genus is monophyletic. PérezFarrera (1999) suggested that at least some species could be dispersed by mammals such as the peccari (Tayassu tajacu). Some of the present unusual distribution patterns could be the result of now-extinct dispersal agents, as suggested by Jansen and Martin (1982) for tropical trees bearing large fruits. It appears that Ceratozamia (Cycadopodites) was present in Miocene floras in Pichucalco, Chiapas, southern Mexico (Palacios and Rzedowski, 1993) and in Cenozoic Engelhardtia forests in Oaxaca, where the fossil pollen spectrum is remarkably similar to that of the modern; it is noted that Engelhardtia used to be widely distributed over the northern hemisphere in the Tertiary (Rzedowski and Palacios, 1977). However, owing to climatic changes during the Pleistocene
28/11/03
3:48 pm
Page 119
Relationships and Phytogeography in Ceratozamia
119
11 10 Cuticle thickness (micron)
ch09.qxd
9 8
± Std. Dev. ± Std. Error Mean
7 6 5 4 3 2 C. miqueliana #1 C. zaragozae #1 C. miqueliana #2 C. zaragozae #2
Fig. 9.7. Significant differences in cuticle thickness of lower leaflet epidermis between two samples of Ceratozamia miqueliana H. Wendland and two samples of C. zaragozae Medellin-Leal.
(Burnham and Graham, 1999) some members of Ceratozamia survived in the south of Mexico, where the greatest diversity appears to be concentrated, and very probably an ancestral form became isolated in Honduras, a territory with a history of geographical isolation (Coney, 1982). Isolation and subsequent speciation in the genus appear to be relatively recent, since DNA sequencing data show a low level of variation amongst species (Gonzáles and Vovides, 2002). However, some interesting insights into the phytogeography of this genus can be seen where a correlation with geographical distribution patterns appears to be evident. The three main clades in Fig. 9.9A are consistent with distributional ranges of the species. The two basal clades contain the species that are distributed in southern and south-eastern Mexico, both at and south of the Neovolcanic mountain range of Pliocene–Quaternary (Pleistocene) age (Fig. 9.9B). Here the Ceratozamia norstogii clade integrates well with what we consider the C. norstogii complex. Though we consider C. zoquorum to be part of the C. miqueliana complex, it appears to be paraphyletic to the C. miqueliana clade, but, owing to the low number of molecular characters that generated this tree, this is a tentative hypothesis and further research is still in progress. By and large, there is consistency with the findings of Marshall and Liebherr (2000), who identified two biogeographic assemblages, one north of the Neovolcanic range and another to the south. The large unresolved clade contains a group of morphologically distinct species to the north and north-east of the Neovolcanic range which appears to be of more recent speciation (Fig. 9.9). The vegetational history of southern Mexico and the Maya Mountains
ch09.qxd
28/11/03
120
3:48 pm
Page 120
A.P. Vovides et al.
af
af
Fig. 9.8. Differences in number of associated (af) and non-associated (nf) fibres with the vascular bundles of leaflets (ue, upper epidermis; pa, palisade mesophyll; sm, spongy mesophyll; s, bundle sclerenchyma; x, xylem; ph, phloem; le, lower epidermis). (A) Ceratozamia mexicana Brongniart var robusta Dyer. (B) Ceratozamia kuesteriana Regel.
(Belize) has been studied by Lundell (1939, 1945) and Miranda (1957, 1959), who agreed that the region contains relict floral elements of great age. During the past 40,000 years, tropical forests in Mexico have been disrupted and displaced to lower latitudes due to the onset of Pleistocene climatic changes with cycles of cold-dry, cold-wet and warm-dry climates, thus leaving relict pockets or refuges that protected the biota from lowering of temperature, rainfall and precipitation (Toledo, 1982). Graham (1976), using palynological analysis, maintained that modern tropical rainforest in Mexico is recent. However, there is a general consensus on the existence of floristic and faunistic refuges of great age in southern Mexico (Brown, 1976; Toledo, 1982, 1988), but these are apparently absent in the areas north of the Neovolcanic mountain range. From the data presented in this study we put forward the hypothesis that Ceratozamia has its origin in southern/south-eastern Mexico. The unresolved clade of Ceratozamia species at and north of the Neovolcanic range is very likely the result of adaptive radiation or
ch09.qxd
28/11/03
3:48 pm
Page 121
Relationships and Phytogeography in Ceratozamia
121
C. matudae
Fig. 9.9. Relationships among species of Ceratozamia (from González and Vovides, 2002). (A) Phylogenetic relationships using ITS sequences. (B) Geographical distribution in Mexico of the three major clades shown in Fig. 9.9A.
recent speciation processes. Those populations north of the range are morphologically distinct, with gross morphological characters that do not change when cultivated under similar conditions over several years; these are, however, homogeneous in DNA sequencing. It may be argued that there are very few mutations in those species that are resolved, but the relatively long generation times for Ceratozamia must be considered: species have a generation time (germination to maturity to seed set) of at least 15 years under optimal cultivation conditions and this period can safely be doubled for conditions in the wild. Therefore, only about 300 generations of a putative Ceratozamia
ch09.qxd
28/11/03
3:48 pm
122
Page 122
A.P. Vovides et al.
would have occurred since the end of the Pleistocene (10,000 years ago); thus any speciation within the genus would be understandably slow. It is hoped that anatomical, morphometric and other molecular studies will throw more light on this matter. Examining the cladogram (Fig. 9.9A), better-resolved clades (all south of the Neovolcanic belt) can easily be placed into the known refuges (Fig. 9.9B). The Ceratozamia matudae clade is found in the Soconusco refuge of Chiapas on the Sierra Madre del Sur while C. mixeorum is from the adjacent northern mountains of Oaxaca near the Sierra de Juarez refuge. The C. miqueliana, C. euryphyllidia, C. sp. “becerrae”, C. whitelockiana and C. zoquorum clades are situated on and adjacent to the Los Tuxtla refuge of Toledo and the ‘arc refuge area’ of Wendt (1987), covering southern Veracruz, northern Chiapas, northern Oaxaca and south-western Tabasco. These species comprise what might be called the C. miqueliana complex. On the Neovolcanic range we find the C. mexicana complex with a separate species on the Cordoba refuge of Toledo (1982, 1988). The rest of the unresolved cladogram consists of Ceratozamia species that are north of the Neovolcanic belt. These species appear to be much younger, suggestive of adaptive radiation with a northwards migratory pattern following a general warming of climate. The C. mexicana clade (on the Neovolcanic range) appears to be the most recently derived.
Conclusions We put forward the hypothesis that Ceratozamia, a relict palaeoendemic, is currently in the process of active speciation. This process was probably initiated during the early Cenozoic and was stimulated by climate changes of the late Tertiary and Quaternary periods giving rise to the present taxa, which are apparently well adapted to modern ecosystems. From a conservation point of view, we reiterate that cycads, which are phylogenetically basal to the living seed plants, should be given top priority in the agendas of countries that are rich in their cycad diversity. With 303 species (see Hill et al., Appendix 1 this volume), cycads are relatively few worldwide; efforts for their conservation should not be too expensive and the returns very great, considering the vast information contained in their genomes. As Norstog and Nicholls (1997) commented, ‘they are the Rosetta Stone of spermatophyte evolution’.
References Balduzzi, A., De Luca, P. and Sabato, S. (1982) A phytogeographical approach to the New World Cycads. Delpinoa 23–24, 185–202. Baldwin, B.G., Sanderson, M.J., Porter, J.M., Wojciechowski, M.F., Campbell, C.S. and Donoghue, M.J. (1995) The ITS region of nuclear ribosomal DNA: a valuable source
ch09.qxd
28/11/03
3:48 pm
Page 123
Relationships and Phytogeography in Ceratozamia
123
of evidence of angiosperm phylogeny. Annals of the Missouri Botanical Garden 82, 247–277. Brown, K.S. Jr (1976) Centros de evolução, refugios Quaternários, e conservação de patrimônios genéticos na região neotropical: Padrões de diferenciação em Ithomiinae (Lepidoptera: Nymphlidae). Acta Amazonica 7, 75–137. Burnham, R.J. and Graham, A. (1999) The history of neotropical vegetation: new developments and status. Annals of the Missouri Botanical Garden 86, 546–589. Caputo, P., Stevenson, D.W. and Wurtzel, E.T. (1991) A phylogenetic analysis of American Zamiaceae (Cycadales) using chloroplast DNA restriction fragment length polymorphisms. Brittonia 43, 135–145. Caputo, P., Marquis, C., Wurtzel, T., Stevenson, D.W. and Wurtzel, E.T. (1993) Molecular biology in cycad systematics. In: Stevenson, D.W. and Norstog, K.J. (eds) The Biology, Structure, and Systematics of the Cycadales. Proceedings of the Second International Conference on Cycad Biology. Palm & Cycad Societies of Australia Limited, Milton, Queensland, Australia, pp. 213–219. Coney, P.J. (1982) Plate tectonic constraints on the biogeography of middle America and the Caribbean region. Annals of the Missouri Botanical Garden 69, 432–443. De Luca, P., Moretti, A., Siniscalco Gigliano, G., Caputo, P., Cozzolino, S., Gaudio, L., Stevenson, D.W., Wurtzel, E.T. and Osborne, R. (1995) Molecular systematics of cycads. In: Vorster, P. (ed.) Proceedings of the Third International Conference on Cycad Biology. Cycad Society of South Africa, Stellenbosch, South Africa, pp. 131–137. González, D. and Vovides, A.P. (2002) Low intralineage divergence in the genus Ceratozamia Brongn. (Zamiaceae) detected with nuclear ribosomal DNA ITS and chloroplast DNA trnL-F non-coding region. Systematic Botany 27, 654–661. Graham, A. (1976) Late Cenozoic evolution of tropical lowland vegetation in Veracruz, Mexico. Evolution 29, 723–735. Jansen, D.H. and Martin, P.S. (1982) Neotropical anachronisms: the fruits the Gomphotheres ate. Science 215, 19–27. Lundell, C.L. (1939) Studies of Mexican and Central American Plants,VII. Lloydia 2, 73–76. [Ceratozamia matudae.] Lundell, C. (1945) Vegetation and natural resources of British Honduras. In: Lundell, C.L. (ed.) Plant and Plant Science in Latin America. Chronica Botanica, Waltham, Massachusetts, pp. 270–273. Marshall, C.J. and Liebherr, J.K. (2000) Cladistic biogeography of the Mexican transition zone. Journal of Biogeography 27, 203–216. McCune, B. and Mefford, J. (1997) Multivariate Analysis of Ecology Data. Version 3.17. MJM Software, Gleneden Beach, Oregon, 126 pp. Miquel, F.A.W. (1870) À la connaissance des cycadées. Adansonia 9, 154–180. Miranda, F. (1957) Vegetación de la vertiente del Pacífico de la Sierra Madre de Chiapas (México) y sus relaciones florísticas. Proceedings of the Eighth Pacific Congress 4, 438–452. Miranda, F. (1959) Estudios acerca de la vegetación. In: Beltran, E. (ed.) Los recursos naturales del sureste y su aprovechamiento. D.F. Instituto Mexicano de Recursos Naturales Renovables, Mexico, pp. 215–217. Moretti, A. (1990) Karyotypic data on North and Central American Zamiaceae (Cycadales) and their phylogenetic implications. American Journal of Botany 77, 1016–1029. Moretti, A. and Sabato, S. (1984) Karyotype evolution by centromeric fission in Zamia (Cycadales). Plant Systematics and Evolution 146, 215–223.
ch09.qxd
28/11/03
124
3:48 pm
Page 124
A.P. Vovides et al.
Moretti, A., Sabato, S. and Vázquez Torres, M. (1980) The distribution of Ceratozamia Brongn (Zamiaceae). Delpinoa 20, 13–21. Moretti, A., Caputo, P., Cozzolino, S., De Luca, P., Gaudio, L., Siniscalco Gigliano, G. and Stevenson, D.W. (1993) A phylogenetic analysis of Dioon (Zamiaceae). American Journal of Botany 80, 204–214. Norstog, K.J. and Nicholls, T.J. (1997) The Biology of the Cycads. Cornell University Press, Ithaca, New York, 363 pp. Palacios, C.R. and Rzedowski, J. (1993) Estudio palinológico de las floras fósiles del Mioceno inferior y principios del Mioceno Medio de la región de Pichucalco, Chiapas, México. Acta Botánica de México 24, 1–96. Pérez-Farrera, M.A. (1999) Dinámica poblacional de Ceratozamia matudae en la Reserva de la Biosfera el Triunfo, Chiapas, México. Tesis de Maestria, El Colegio de la Frontera Sur. San Cristóbal de Las Casas, Chiapas, Mexico. Pérez-Farrera, M.A., Vovides, A.P. and Iglesias, C.G. (1999) A new species of Ceratozamia (Zamiaceae, Cycadales) from Chiapas, Mexico. Novon 9, 410–413. [Ceratozamia alvarezii.] Pérez-Farrera, M.A., Vovides, A.P. and Iglesias, C.G. (2001) A new species of Ceratozamia (Zamiaceae) from Chiapas, Mexico. Botanical Journal of the Linnean Society 137, 77–80. [Ceratozamia zoquorum.] Potter, D., Luby, J.J. and Harrison, R.E. (2000) Phylogenetic relationships among species of Fragaria (Rosaceae) inferred from non-coding nuclear and chloroplast DNA sequences. Systematic Botany 25, 337–348. Rzedowski, J. and Palacios, C. (1977) El bosque de Engelhardtia (Oreomunnea) mexicana en la región de la Chinantla (Oaxaca, México). Una reliquia del Cenozoico. Boletín de la Sociedad Botánica de México 36, 93–123. Sabato, S. and De Luca, P. (1985) Evolutionary trends in Dion [sic] (Zamiaceae). American Journal of Botany 72, 1353–1363. Schilling, E.E., Linder, C.R., Noyes, R.D. and Rieseberg, L.H. (1998) Phylogenetic relationships in Helianthus (Asteraceae) based on nuclear ribosomal DNA transcribed spacer region sequence data. Systematic Botany 23, 177–187. Stevenson, D.W., Sabato, S. and Vázquez Torres, M. (1986) A new species of Ceratozamia (Zamiaceae) from Veracruz, Mexico with comments on species relationships, habitats, and vegetative morphology in Ceratozamia. Brittonia 38, 17–26. [Ceratozamia euryphyllidia.] Thiselton-Dyer, W.T. (1882–1886) Cycadaceae. Biologia Centrali-Americana 3, 190–195. Toledo, V.M. (1982) Pleistocene changes of vegetation in tropical Mexico. In: Prance, G.T. (ed.) Biological Diversification in the Tropics. Proceedings of the Fifth International Symposium of the Association for Tropical Biology. Columbia University Press, New York, pp. 93–111. Toledo, V.M. (1988) La diversidad biológica de México. Ciencia y Desarrollo 14, 17–30. Vovides, A.P. (1983) Systematic studies on the Mexican Zamiaceae. I. Chromosome numbers and karyotypes. American Journal of Botany 70, 1002–1006. Vovides, A.P. (1985) Systematic studies on the Mexican Zamiaceae II. Additional notes on Ceratozamia kuesteriana from Tamaulipas, Mexico. Brittonia 37, 226–231. Vovides, A.P. and Olivares, M. (1996) Karyotype polymorphism in the cycad Zamia loddigesii (Zamiaceae) of the Yucatan Peninsula, Mexico. Botanical Journal of the Linnean Society 120, 77–83. Vovides, A.P., Rees, J.D. and Vázquez Torres, M. (1983) Familia Zamiaceae. In: Sosa, V. and Gomez-Pompa, A. (eds) Flora de Veracruz. INIREB, Xalapa, Veracruz, Mexico, Fascículo 26.
ch09.qxd
28/11/03
3:48 pm
Page 125
Relationships and Phytogeography in Ceratozamia
125
Vovides, A.P., Pérez-Farrera, M.A. and Iglesias, C.G. (2001) Another new species of Ceratozamia (Zamiaceae) from Chiapas, Mexico. Botanical Journal of the Linnean Society 137, 81–85. [Ceratozamia mirandae.] Wendt, T. (1987) Las selvas de Uxpanapa, Veracruz-Oaxaca, México: evidencia de refugios florísticos Cenozoicos. Anales del Instituto de Biología UNAM – Serie botánica 58, 29–54.
ch09.qxd
28/11/03
3:48 pm
Page 126
ch10.qxd
28/11/03
3:49 pm
Page 127
A Morphometric Analysis of the Ceratozamia norstogii Complex (Zamiaceae)
10
Miguel A. Pérez-Farrera,1 Andrew P. Vovides,2 Luis Hernández-Sandoval,3 Dolores González2 and Mahinda Martínez3 1Escuela
de Biología, Universidad de Ciencias y Artes de Chiapas (UNICACH), Tuxtla Gutiérrez, Chiapas, Mexico; 2Instituto de Ecología A.C., Xalapa, Veracruz, Mexico; 3Facultad de Biología, Universidad Autónoma de Querétaro, Centro Universitario, Querétaro, Mexico
Abstract The three species forming the Ceratozamia norstogii complex (C. alvarezii, C. mirandae and C. norstogii) are found in adjacent areas of the Sierra Madre de Chiapas in Mexico. Taxonomic limits within this complex have not yet been fully defined, but are investigated in this project. Twenty morphological variables from 90 individuals from three populations have been analysed using ANOVA and discriminant analysis. The results reveal a clear number of differences for these variables among the three species.
Introduction The taxonomy and distribution of the genus Ceratozamia Brongniart (Zamiaceae) is little known due to various problems: (i) there is no clear delimitation of some of the species, principally in the C. mexicana Brongniart, C. latifolia Miquel and C. miqueliana H. Wendland complexes or species groups; (ii) many of the type specimens on which the original descriptions were based have been lost or destroyed and the descriptions themselves are often based on sterile or juvenile material and ©CAB International 2004. Cycad Classification: Concepts and Recommendations (eds T. Walters and R. Osborne)
127
ch10.qxd
28/11/03
3:49 pm
128
Page 128
M.A. Peréz-Farrera et al.
on plants cultivated in botanic gardens (Vovides et al., 1983); (iii) species complexes are found throughout the distribution of the genus (Moretti et al., 1980); and (iv) some neotypifications appear to be doubtful (e.g. C. robusta Miquel) (Stevenson and Sabato, 1986). Ceratozamia norstogii D.W. Stevenson was described in 1982 as having the distinctive characteristics of channelled leaflets and a straight rachis, but the type specimen assigned (C.A. Purpus 6) has a twisted rachis (Stevenson, 1982). On the basis of herbarium vouchers, Stevenson et al. (1986) assigned some populations and forms with channelled leaflets and a straight rachis to C. norstogii under the assumption that this species was polymorphic. However, it now appears that C. norstogii is not polymorphic and we believe that this species belongs in a group with two recently described related species (Pérez-Farrera et al., 1999, 2001; Vovides et al., 2001). Unfortunately the range of morphological variation within this complex and others of the genus is not yet fully understood. This lack of information, coupled with the difficulty in obtaining morphological characters from fertile specimens, has contributed to taxonomic confusion in the genus. In this study we examined the morphological variation within the Ceratozamia norstogii complex using multivariate statistical techniques to test the morphological differentiation between C. alvarezii Pérez-Farrera, Vovides & Iglesias, C. mirandae Vovides, Pérez-Farrera & Iglesias and C. norstogii.
Study Area The study was done in the La Sepultura Biosphere Reserve on the western side of the Sierra Madre de Chiapas in Mexico. This physiographical region runs parallel to the Pacific coastal plain, from the extreme south of the Isthmus of Tehuantepec, across Chiapas and reaches as far as Guatemala. It ranges in altitude from 1000 m in the north to 4000 m (Mount Tacaná) on the Guatemalan border with Chiapas (Müllerried, 1957). The predominant vegetational types are deciduous tropical forest, oak forest, conifer forest, cloud forest and evergreen tropical rainforest (Rzedowski, 1978). According to De La Rosa et al. (1989) the major part of the area is formed by Palaeozoic igneous rocks (granites and granidiorites) and Precambrian rocks (pink granitic gneiss and granidiorite gneiss). Specific metamorphic substrate is present only on slopes in some areas (Hernández-Yañez, 1995). Most populations within the Ceratozamia norstogii complex are found in oak and cloud forests or intermediate zones of these two ecosystems. The cycads grow in poor clay soils on steep slopes within an altitude range of 800–1200 m. The majority of populations are subjected to periodic fires at least once a year. All species within this complex are characterized by erect male and female cones, and plain or channelled leaflets.
ch10.qxd
28/11/03
3:49 pm
Page 129
Ceratozamia norstogii Complex: Morphometric Analysis
129
Materials and Methods Three natural populations were analysed (Fig. 10.1) using 30 individuals each of Ceratozamia norstogii, C. alvarezii and C. mirandae (Figs 10.2–10.4) and recording data from 14 vegetative and six reproductive morphological variables (Table 10.1). Measurements were done using a flexible tape measure (3 m) and a precision digital vernier (0.01 mm resolution). Data were captured and electronically stored. Analysis of variance (ANOVA) tests were done using JMP version 3.2 statistical software (SAS Institute, Cary, North Carolina, USA) and discriminant analyses (McCune and Mefford, 1997) with Statgraphics version 2.0 software (Manugistics, Rockville, Maryland, USA).
Results Univariate analysis Tables 10.2 and 10.3 summarize the ANOVA results. Species characters do not overlap, with the exception of intervein distance, leaf length, microsporophyll
Fig. 10.1. Location in Mexico of populations of the Ceratozamia norstogii complex sampled in this study.
ch10.qxd
28/11/03
130
3:49 pm
Page 130
M.A. Peréz-Farrera et al.
Fig. 10.2. Species in the Ceratozamia norstogii complex: C. alvarezii Pérez-Farrera, Vovides & Iglesias.
Fig. 10.3. Species in the Ceratozamia norstogii complex: C. norstogii D.W. Stevenson.
ch10.qxd
28/11/03
3:49 pm
Page 131
Ceratozamia norstogii Complex: Morphometric Analysis
131
Fig. 10.4. Species in the Ceratozamia norstogii complex: C. mirandae Vovides, Pérez-Farrera & Iglesias. Table 10.1. List of morphometric variables used in the analysis of populations in the Ceratozamia norstogii complex. No.
Abbreviation
Character and measurement units (parentheses)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
LARGTR PERMTR NHOJA LARHOJ ANCHHOJ LARESPPEC LARGOPEC PERIMPEC LARGORAQ NFOLIOL LARGOFOL ANCHOFOL NVENAS DISTVEN LARGOMICRO ANCHOMICRO ANCHOMEGAS LARGOMEGAS DIAMSEMI LARGSEMI
Trunk length (cm) Trunk perimeter (cm) Number of leaves per trunk Leaf length (cm) Leaf width (cm) Petiole prickle length (mm) Petiole length (cm) Petiole perimeter (cm) Rachis length (cm) Number of leaflets per leaf Leaflet length (cm) Leaflet width (cm) Number of veins per leaflet Intervein distance (mm) Microsporophyll length (mm) Microsporophyll width (mm) Megasporophyll distal face width (mm) Megasporophyll distal face length (mm) Seed diameter (mm) Seed length (mm)
ch10.qxd
28/11/03
3:49 pm
Page 132
132
M.A. Peréz-Farrera et al.
Table 10.2. Mean values and standard deviations (SD) for the ratios used in the analysis of populations in the Ceratozamia norstogii complex (C. alvarezii PérezFarrera, Vovides & Iglesias, C. mirandae Vovides, Pérez-Farrera & Iglesias and C. norstogii D.W. Stevenson). Species/character LARGTR PERMTR NHOJA LARHOJ ANCHHOJ LARESPPEC LARGOPEC PERIMPEC LARGORAQ NFOLIOL LARGOFOL ANCHOFOL NVENAS DISTVEN LARGOMICRO ANCHOMICRO ANCHOMEGAS LARGOMEGAS DIAMSEMI LARGSEMI
C. alvarezii
SD
C. mirandae
SD
C. norstogii
SD
20.55 42.68 8.91 79.64 45.36 0.34 20.59 2.45 47.00 45.00 22.98 0.65 7.55 0.08 14.90 6.99 35.54 15.54 16.0 22.36
± 7.67 ± 6.64 ± 3.99 ± 15.5 ± 6.61 ± 0.09 ± 8.52 ± 0.57 ± 10.2 ± 9.97 ± 3.28 ± 0.11 ± 0.93 ± 0.03 ± 0.83 ± 0.76 ± 5.32 ± 2.26 ± 1.34 ± 2.73
58.23 71.30 13.33 151.87 69.15 0.46 39.92 3.42 104.55 65.67 36.07 0.98 8.67 0.10 14.55 7.17 52.41 12.14 16.58 27.45
± 18.5 ± 7.28 ± 4.74 ± 17.9 ± 8.69 ± 0.07 ± 9.97 ± 0.63 ± 61.7 ± 9.05 ± 4.80 ± 0.12 ± 1.56 ± 0.02 ± 0.99 ± 0.70 ± 3.82 ± 1.97 ± 0.97 ± 1.12
32.07 58.31 4.83 98.60 94.24 0.40 18.18 2.61 62.11 51.17 35.55 0.41 5.31 0.09 14.32 7.74 46.82 9.88 17.55 27.13
± 21.76 ± 6.79 ± 0.97 ± 17.35 ± 135.3 ± 0.51 ± 6.93 ± 0.56 ± 9.90 ± 9.32 ± 7.65 ± 0.07 ± 0.71 ± 0.01 ± 0.99 ± 0.80 ± 1.64 ± 1.61 ± 1.37 ± 1.30
length, petiole length and petiole prickle length. The rest of the variables analysed were all highly significant (P < 0.0001).
Discriminant analyses Figure 10.5 shows the scatter diagram of data derived from discriminant function analysis. The three species separate ordinately in bidimensional space and do not present any overlapping between groups. Of the 20 variables included in the standardized discrete canonical function, the six variables with the highest values in factor 1 were trunk perimeter, petiole length, leaflet width, vein number, microsporophyll length and megasporophyll width. In factor 2, the highest variables were trunk perimeter, leaf length, petiole length, leaflet length and width. The first canonic variable showed that 70% of the variation is largely due to vegetative morphology. The positive correlations (Table 10.4) of all the variables show differences between species. The Wilks lambda test was highly significant (P < 0.0001) for the two factors (Table 10.5) thus showing that all the species were classified correctly.
28/11/03
3:49 pm
Page 133
133
Ceratozamia norstogii Complex: Morphometric Analysis
Table 10.3. Summary of analysis of variance of the 20 morphometric characters used in the analysis of populations in the Ceratozamia norstogii complex (R2 = correlation coefficient, F = F value, P = probability). Character
R2
F
LARGTR PERMTR NHOJA LARHOJ ANCHHOJ LARESPPEC LARGOPEC PERIMPEC LARGORAQ NFOLIOL LARGOFOL ANCHOFOL NVENAS DISTVEN LARGOMICRO ANCHOMICRO ANCHOMEGAS LARGOMEGAS DIAMSEMI LARGSEMI
0.47 0.71 0.52 0.76 0.06 0.02 0.06 0.37 0.30 0.45 0.49 0.87 0.56 0.01 0.61 0.09 0.76 0.62 0.25 0.61
40.04 108.18 43.41 138.79 2.81 1.02 3.02 25.63 19.04 36.61 42.92 241.20 55.90 0.77 1.36 4.83 145.52 71.80 15.27 68.8
■
C. alvarezii × C. mirandae
●
P < < <