Palaeobiology

PALAEOBIOLOGY A SYNTHESIS

An ichthyosaur embryo (skull length 6.5 cm) discovered in 1985 by collectors Robert and Peter

Langham; from the Lower Lias (Lower Jurassic) of the Somerset coast, U.K. On display at City of Bristol Museum & Art Gallery, U.K. (Photograph courtesy of Dept. of Geology, University of Bristol).

PALAEOBIOLOGY A SYNTHESIS

EDITED BY

DEREK E. G. BRIGGS Department of Geology University of Bristol Queen's Road Bristol BS8 lRJ

AND

PETER R. CROWTHER Department of Geology Bristol City Museums and Art Gallery Queen's Road Bristol BS8 lRL

ON BEHALF OF THE PALAEONTOLOGICAL ASSOCIA nON

b

Blackwell Science

© 1990 by Blackwell Science Ltd Editorial Offices: Osney Mead, Oxford 0)(2 OEL 25 John Street, London WClN 2BL

DISTRIBUTORS

Marston Book Services Ltd PO Box 269 Abingdon, Oxon OX14 4YN

(Orders:

23 Ainslie Place, Edinburgh EH3 6AJ 350 Main Street, Maiden MA 02148 5018, USA

54 University Street, Cariton Victoria 3053, Australia Other Editorial Offices: Blackwell Wissenschafts-Verlag GmbH

Tel:

01235 465500

Fax: 01235 465555) USA Blackwell Science, Ine. Commerce Place 350 Main Street Maiden, MA 02148 5018

(Orders:

Tel:

800 759 6102 617 388 8250

Kurrurstendamm 57

Fax: 617 388 8255)

10707 Berlin, Germany Canada Blackwell Science KK

Copp Clark Professional

MG Kodenmacho Building

200 Adelaide Street West, 3rd Floor

7-10 Kodenmacho Nihombashi

Toronto, Ontario M5H 1W 7

Chuo-ku, Tokyo 104, Japan

(Orders:

Tel:

416 597-1616 800 815-9417

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988,

Fax: 416 597-1617) Australia Blackwell Science Pty Ltd

54 University Street Cariton, Victoria 3053

(Orders:

copyright owner. First published 1990 Paperback reissue with corrections 1992 Reprinted with further corrections 1993, 1995, 1997

Tel:

3 9347 0300

Fax: 3 9347 5001)

without the prior permission of the

British Library Cataloguing in Publication Data Palaeobiology: A Synthesis 1. Palaeobiology

1. Briggs, D.E.G.

H. Crowther, P.R.

560 Set by Setrite Typesetters Ltd, Hong Kong Printed and bound in Great Britain at the University Press, Cambridge The Blackwell Science logo is a trade mark of Blackwell Science Ltd, registered at the United Kingdom Trade Marks Registry

ISBN 0-632-02525-5 (Hbk) ISBN 0-632-03311-8 (Pbk) Library of Congress Cataloging-in-Publication Data Palaeobiology: A Synthesis. Includes index. 1. Palaeobiology

1. Briggs, D.E.G. QE721.E53

1989

H. Crowther, P.R. 560'.3'21

ISBN 0-632-02525-5 (Hbk) ISBN 0-632-03311-8 (Pbk)

88-35060

Contents

1.12 Hominids, 88

List of Contributors, ix

R. L. SUSMAN

Foreword, xiii L. R. M. COCKS

1

2

Major Events in the History of Life

2.1

3

1.1

Origin of Life,

1.2

Precambrian Evolution of Prokaryotes and

2.2

9

2.3

Precambrian Metazoans,

17

2.4

Origin of Hard Parts - Early Skeletal Fossils,

24

2.5 2.6

30

Evolutionary Faunas,

I· I·

SEI'KOSKI.

2.7 37 2.8

If

Early Diversification of Major Marine Habitats w. I. AUSICH & D.

I.

2.9

BOTTlER

I.

119

BENTON

Hierarchy and Macroevolution,

124

Patterns of Diversification,

130

Coevolution,

136

139

Adaptation,

1'. W. SKELTON

2.10 Evolution of Large Size, 147

R. B. RICKARDS

M.

1.7.3 Reefs, 52

I.

BENTON

2.11 Rates of Evolution - Living Fossils, 152 D. C. FISHER

c. T. SCRUTTON

2.12 Mass Extinction: Processes 2.12.1 Earth-bound Causes, 160

Terrestrialization

1.8.1 Soils, 57 V. P. WRIGHT

A. HALLAM

2.12.2 Extra-terrestrial Causes, 164

1.8.2 Plants, 60

D. IABLONSKI

D. EDWARDS & N. D. BURGESS

2.12.3 Periodicity, 171

1.8.3 Invertebrates, 64

I. I

1'. A. SELDEN

1.8.4 Vertebrates, 68

SEPKOSKI,

Ir

2.13 Mass Extinction: Events 2.13.1 Vendian, 179

A. c. MILNER

Flight

M. A. S. McMENAMIN

2.13.2 End-Ordovician, 181

1.9.1 Arthropods, 72

P.

R. I. WOOTTON

I.

BRENCHLEY

2.13.3 Frasnian-Famennian, 184

1.9.2 Vertebrates, 75

G. R. McGHEE.

K. I'ADIAN

1.10 Angiosperms, 79

Ir

2.13.4 End-Permian, 187

M. E. COLLINSON

D. H. ERWIN

1.11 Grasslands and Grazers, 84 I.

Red Queen Hypothesis,

S. CONWAY MORRIS

1.7.2 Plankton, 49

1.9

111

1'. W. SIGNOR

1.7.1 Infauna and Epifauna, 41

1.8

106

N. ELDREDGE

s. CONWAY MORRIS

1.7

Heterochrony,

M.

Late Precambrian-Early Cambrian Metazoan Diversification,

1.6

Microevolution and the Fossil Record,

K. I. McNAMARA

B. RUNNEGAR & s. BENGTSON

1.5

100

Speciation,

P. R. SHELDON

M. A. FEDONKIN

1.4

95

B. CHARLESWORTH

A. H. KNOLL

1.3

Molecular Palaeontology, G. B. CURRY

c. R. WOESE & G. WACHTERSHAuSER

Protists,

The Evolutionary Process and the Fossil Record

2.13.5 End-Triassic, 194

R. THOMASSON & M. R. VOORHIES

M.

v

I

BENTON

2.13.6 Cretaceous-Tertiary (Marine), 198

3.11.6 Holzmaden, 282

F. SURLYK

R. WILD

2.13.7 Cretaceous-Tertiary (Terrestrial), 203

3.11.7 Solnhofen Lithographic Limestones, 285

L. B. HALSTEAD

G. VIOHL

2.13.8 Pleistocene, 207 E. L. LUNDELIUS,

3.11.8 Grube Messel, 289 j.

Jr

L. FRANZEN

3.11.9 Baltic Amber, 294 3

T. SCHLOTER

Taphonomy 3.1

3.12 Completeness of the Fossil Record, 298

Decay Processes,

213

c. R. c. PAUL

P. A. ALLISON

3.2

The Record of Organic Components and the Nature of Source Rocks,

217

P. FARRIMOND & G. EGLINTON

3.3

Destructive Taphonomic Processes and Skeletal Durability,

223

4

Palaeoecology 4.1 4.2

Transport - Hydrodynamics

4.3

3.4.1 Shells, 227 j.

4.4

R. L. ALLEN

R. A. SPICER

3.4.3 Bones, 232 Fossil Concentrations and Life and Death Assemblages,

4.5

235

4.6

Obrution Deposits,

239

4.7 4.8 4.9

244

Populations,

M. E. TUCKER

P. A. ALLISON

3.8.3 Pyrite, 253

Stromatolites,

336

Reefs and Carbonate Build-Ups, Encrusters,

341

346

3.8.4 Phosphate, 256 L. PREV6T &

Taphofacies,

A. c. SCOTT

4.11 Trace Fossils, 355 S. G. PEMBERTON, R. W. FREY & T. D. A. SAUNDERS

4.12 Evidence for Diet, 362 j.

P. A. ALLISON

j.

E. POLLARD

4.13 Predation 4.13.1 Marine, 368 C. E. BRETT

LUCAS

258

4.13.2 Terrestrial, 373 j.

c. E. BRETT & S. E. SPEYER

3.10 Anatomical Preservation of Fossil Plants, 263

A. MASSARE & c. E. BRETT

4.14 Parasitism, 376 s. CONWAY MORRIS

A. c. SCOTT

3.11 Taphonomy of Fossil-Lagerstatten 3.11.1 Overview, 266 A. SEILACHER

3.11.2 Burgess Shale, 270 s. CONWAY MORRIS

3.11.3 Upper Cambrian 'Orsten', 274 MOLLER

3.11.4 Hunsriick Slate, 277 j.

330

Coloniality,

4.10 Reconstructing Ancient Plant Communities, 351

3.8.2 Carbonate Nodules and Plattenkalks, 250

j.

326

P. D. TAYLOR

Diagenesis

K.

322

B. R. ROSEN

3.8.1 Skeletal Carbonates, 247

3.9

Hydrodynamics,

S. M. AWRAMIK

D. E. G. BRIGGS

3.8

318

B. R. ROSEN

c. E. BRETT

Flattening,

Biomechanics,

M. LaBARBERA

F. T. FORSICH

3.7

314

G. B. CURRY

A. K. BEHRENSMEYER

3.6

Composition and Growth of Skeleton,

P. A. SELDEN

3.4.2 Plant Material, 230

3.5

307

B. RUNNEGAR

c. E. BRETT

3.4

Morphology, L. LUGAR

BERGSTROM

3.11.5 Mazon Creek, 279 G. c. BAIRD

4.15 Palaeopathology, 381 L. B. HALSTEAD

4.16 Trophic Structure, 385 j.

A. CRAME

4.17 Evolution of Communities, 391 A.

j.

BOUCOT

4.18 Biofacies, 395 P.

j.

BRENCHLEY

4.19 Fossils as Environmental Indicators 4.19.1 Climate from Plants, 401 R. A. SPICER

vii

Contents 5.10 Global Boundary Stratotypes 5.10.1 Overview, 471

4.19.2 Temperature from Oxygen Isotope Ratios, 403

J. w. COWIE

T. F. ANDERSON

4.19.3 Salinity from Faunal Analysis and Geochemistry, 406 J

5.10.2 Precambrian-Cambrian, 475 J. W. COWIE

5.10.3 Ordovician-Silurian, 478

D. HUDSON

4.19.4 Oxygen Levels from Biofacies and Trace Fossils, 408

c. R. BARNES & S. H. WILLlAMS

5.10.4 Silurian-Devonian, 480 c. H. HOLLAND

D. J. BOTTJER & C. E. SAVRDA

5.11 Fossils and Tectonics, 482

4.19.5 Depth from Trace and Body Fossils, 411 G. E

5

R. A. FORTEY & L. R. M. COCKS

FARROW

Taxonomy, Phylogeny, and Biostratigraphy 5.1

6

Infrastructure of Palaeobiology 6.1

Computer Applications in

Rules of Nomenclature

Palaeontology,

5.1.1 International Codes of Zoological and Botanical Nomenclature, 417

J

6.2

Practical Techniques

6.2.1 Preparation of Macrofossils, 499

M. E. TOLLlTT

5.1.2 Disarticulated Animal Fossils, 419

P. J. WHYBROW & w. LlNDSAY

6.2.2 Extraction of Microfossils, 502

R. J. ALDRIDGE

5.1.3 Disarticulated Plant Fossils, 421

R. J. ALDRIDGE

6.2.3 Photography, 505

B. A. THOMAS

5.1.4 Trace Fossils, 423

D

Analysis of Taxonomy and Phylogeny

D. CLAUGHER & 1'. D. TAYLOR

6.2.5 Determination of Thermal Maturity, 511

5.2.1 Overview, 425

J. E

R. A. FORTEY

6.3

5.2.2 Cladistics, 430

6.3.1 Collection Care and Status Material, 515

P. L. FOREY

P. R. CROWTHER

6.3.2 Collection Management and Documentation Systems, 517

A. J. CHARIG

5.2.4 Stratophenetics, 437

P. R. CROWTHER

r. D. GINGERICH

5.2.5 Problematic Fossil Taxa, 442

6.3.3 Exhibit Strategies, 519

s. BENGTSON

R. S. MILES

5.3

Analysis of Taxonomic Diversity,

5.4

Vicariance Biogeography,

445

6.4

Societies, Organizations, Journals, and Collections,

A. B. SMITH

448 6.5

Palaeobiogeography,

History of Palaeontology

6.5.1 Before Darwin, 537

452

c. R. NEWTON

J. c. THACKRAY

6.5.2 Darwin to Plate Tectonics, 543

Biostratigraphic Units and the Stratotypei Golden Spike Concept,

461

P. J

Zone Fossils, M. G

5.8

466

J. W. VALENTINE

6.5.4 The Past Decade and the Future, 550

BASSETT

International Commission on Stratigraphy,

468

A. HOFFMAN

M. G. BASSETT

5.9

International Geological Correlation Programme, J. W. COWIE

BOWLER

6.5.3 Plate Tectonics to Paieobioiogy, 547

c. H. HOLLAND

5.7

522

J. NUDDS & D. PALMER

L. GRANDE

5.6

A. MARSHALL

Museology

5.2.3 Evolutionary Systematics, 434

5.5

J. SIVETER

6.2.4 Electron Microscopy, 508

S. R. A. KELLY

5.2

493

A. KITCHELL

469

Index,

557

List of Contributors

R. J. ALDRIDGE

A. J. B 0 UCOT

Department of Geology, University

J. R. L. ALLEN ,

P.

Postgraduate Research Institute for

P.

A. ALLISON

J. BRENCHLEY

C. E. BRETT Department of Geology, Univer­

D. E. G. B RIGGS Department of Geology,

University

of Bristol, Queen's Road, Bristol BS8 1RJ, U.K.

Department of Geological Sciences,

Ohio State University, Columbus, Ohio 43210, U.5.A.

S. M. A WRAMIK

Department of Geological Sciences, Uni­

versity of Rochester, Rochester, New York 14627, U.S.A.

sity of Illinois, Urbana, Illinois 61801, U.5.A.

W. I. A USICH

Department of Earth Sciences,

U.K.

U.K.

T. F. ANDERSON

Department of Social Anthropology,

University of Liverpool, PO Box 147, Liverpool L69 3BX,

Postgraduate Research Institute for

Sedimentology, University of Reading, Reading RG6 2AB,

N. D. BURGESS

Royal Society for the Protection of

Birds, Sandy, Bedfordshire SG19 2DL, U.K.

Department of Geological Sciences,

University of California, Santa Barbara, California 93106,

A. J. CH A RIG

U.S.A.

G. C. BAIRD

J. BOWLER

Queen's University, Belfast BT7 1NN, U.K.

Sedimentology, University of Reading, Reading RG6 2AB, U.K.

P.

Department of Zoology, Oregon State

University, Corvallis, Oregon 97331, U.5.A.

of Leicester, Leicester LE1 7RH, U.K.

clo Department of Palaeontology, The

Natural History Museum, Cromwell Road, London SW7 5BD, U.K.

Department of Geosciences, State Univer­

sity of New York College: Fredonia, Fredonia, New York

B. CHARLESWORTH

14063, U.5.A.

C. R. BARNES

Chicago, Illinois 60637, U.S.A.

School of Earth and Ocean Sciences,

University of Victoria, P.O. Box 3055, Victoria, British

D. CL A UGHER

Columbia V8W 3P6, Canada.

M. G. BASSETT

Department

L.

Institute of Palaeontology, University

Egham Hill, Egham, Surrey TW20 OEX, U.K.

S. CONWA Y MORRIS

Department of Geology, University of

Sciences,

J. BERGSTROM Swedish Museum of Natural

Department of Geology, Royal

Holloway & Bedford New College, University of London,

Bristol, Queen's Road, Bristol BS8 1RJ, U.K.

University

Department

of Cambridge,

of

Earth

Downing Street,

Cambridge CB2 3EQ, U.K.

History,

J. W. COWIE

S-104 05

Department of Geology, University of

Bristol, Queen's Road, Bristol BS8 1RJ, U.K.

Stockholm, Sweden.

D. J. BOTTJER

Department of Palaeontology, The

M. E. COLLINSON

of Uppsala, Box 558, S-751 22 Uppsala, Sweden.

PO Box 50007,

R. M. COCKS 5BD, U.K.

Institution, Washington DC 20560, U.5.A.

Section of Palaeozoology,

The

Natural History Museum, Cromwell Road, London SW7

of Paleo­

biology, National Museum of Natural History, Smithsonian

M. J. BENTON

of Mineralogy,

5BD, U.K.

Department of Geology, National

A. K. BEHRENSMEYER

Department

Natural History Museum, Cromwell Road, London SW7

Museum of Wales, Cathays Park, Cardiff CFl 3NP, U.K.

S. BENGTSON

Department of Ecology and

Evolution, University of Chicago, 1103 East 57th Street,

J. A. CRAME

Department of Geological Sciences,

British Antarctic Survey, High Cross,

Madingley Road, Cambridge CB3 OET, U.K.

University of Southern California, Los Angeles, California 90089, U.5.A.

P.

R. CROWTHER

Bristol City Museums & Art

Gallery, Queen's Road, Bristol BS8 1RL, U.K. ix

List of Contributors

x

G. B. CURRY Department of Geology

Applied

&

Geology, University of Glasgow, Glasgow G12 8QQ, U.K.

D. EDWARDS Department of Geology, University of Wales College of Cardiff Cathays Park, Cardiff CFl 3YE, U.K.

G. EGLINTON Organic Geochemistry Unit, School of Chemistry, University of Bristol, Bristol BS8 l TS, U.K.

N. ELDREDGE

Department

of

Invertebrates,

American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, U.S.A.

D. H. ERWIN Department of Palaeobiology, National Museum of Natural History, Smithsonian Institution, Washington DC 20560, U.S.A.

P. FARRIMOND The Organic Geochemistry Unit, The University, Newcastle upon Tyne NEl 7R U, U.K.

G. E. FARROW 19 Glenburn Road, Bearsden, Glasgow G61 4PT, U.K.

M. A. FEDONKIN

Palaeontological

Institute,

U.S.S.R. Academy of Sciences, Moscow 117321, U.S.S.R.

D. C. FISHER Museum of Paleontology, University of Michigan, Ann Arbor, Michigan 48109, U.5.A.

P. L. FOREY

Department

of

Palaeontology,

The

5BD, U.K. The

Natural History Museum, Cromwell Road, London SW7 5BD, U.K.

J. L. F RANZEN

Forschungsinstitut

Senckenberg,

Senckenberganlage 25, D-6000 Frankfurt am Main 1, Germany.

R. W. FREY

Institut

for

Paliiontologie

der

Universitiit, Pleicherwall 1, D-8700 Wiirzburg, Germany.

P. D. GINGERICH

C. H. HOLLAND Department of Geology, Trinity ' College, Dublin, Ireland.

J. D. HUDSON Department of Geology, University of Leicester, University Road, Leicester LEl 7RH, U.K. Department

D. JAB L 0 NSKI

of

the

Geophysical

Sciences, University of Chicago, 5734 S. Ellis Avenue, Chicago, Illinois 60637, U.5.A.

S. R. A. K ELLY British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 OET, U.K.

J. A. KITCHELL Museum of Paleontology, Univer­ sity of Michigan, Ann Arbor, Michigan 48109, U.5.A.

A. H. KN 0 LL Department of Organismic lutionary

Biology,

Harvard

University,

& Evo­

Cambridge,

Massachusetts 02138, U.S.A.

M. LaBARBERA Department of Anatomy, Univer­ sity of Chicago, 1025 East 57th Street, Chicago, Illinois 60637, U.S.A.

'vV. LINDSAY

Department

of

Palaeontology,

The

Natural History Museum, Cromwell Road, London SW7

J. LUCAS Institut de Geologie, Universite Louis Pasteur, 1 rue Blessig, Strasbourg 67084, France.

L. LUGAR Department of Geology, Franklin

& Marshall

College, Lancaster, Pennsylvania 17604, U.5.A.

E. L. LUNDELIUS, Jr Department of Geological Sciences, University of Texas, Austin, Texas 78713, U.S.A.

J. E. A. MARSHALL Department of Geology, The University, Highfield, Southampton S09 5NH, U.K.

J. A. MASSARE Department of Geological Sciences,

Deceased.

F. T. FURSICH

Deceased.

5BD, U.K.

Natural History Museum, Cromwell Road, London SW7

R. A. FORTEY Department of Palaeontology,

A. HOFFMAN

Museum

of

Paleontology,

University of Michigan, Ann Arbor, Michigan 48109, U.5.A.

L. GRANDE Department of Geology, Field Museum of Natural History, Chicago, Illinois 60605, U.S.A.

A. HALLAM School of Earth Sciences, University of Birmingham, PO Box 363, Birmingham B15 2TT, U.K.

L. B. HALSTEAD Deceased.

University of Rochester, Rochester, New York 14627, U.5.A.

G. R. MeGHEE, Jr Sciences,

Rutgers

Department

University,

New

of

Geological

Brunswick,

New

Jersey 08903, U.S.A.

M. A. S. MeMENAMIN Department of Geology Geography,

Mount

Ho/yoke

College,

South

&

Hadley,

Massachusetts 01075, U.5.A.

K. J. MeNAMARA Western Australian Museum, Francis Street, Perth, Western Australia 6000, Australia.

List of Contributors R.

S.

MILES Department

of

Public

Services,

The

Natural History Museum, Cromwell Road, London SW7 5BD, U.K.

A.

C.

MILNER Department

of

Palaeontology,

The

C.

A.

P.

A.

University

Chicago, Illinois 60637, U.S.A. P.

R.

of Wales, Cathays Park, Cardiff CFl 3NP, U.K. Department

of

Earth

Sciences,

U.K. G.

P.

PEMBERTON Department of Geology, Uni­

J. E. POLLARD Department of Geology, University of Manchester, Oxford Road, Manchester M13 9PL, U.K.

Strasbourg 67084, France. B.

P.

of

Earth

of

Palaeontology,

The

Natural History Museum, Cromwell Road, London SW7

A.

RUNNEGAR Institute of Geophysics & Planetary

California 90024, U.5.A. SAUNDERSDepartment of Geology, Uni­

SAVRDA Department

of

Geology,

Auburn

University, Auburn, Alabama 36849, U.S.A.

T. SCHLUTER Department of Geology, C.

B.

SMITH Department

of

Palaeontology,

The

E.

SPEYER Department of Geology, Arizona State

R.

A.

SPICERDepartment of Earth Sciences, Univer­

F. SURLYK Geological Institute, University of Copen­ hagen,

0ster

Voldgade 10,

DK-1350 Copenhagen K,

R.

Makerere

SCOTT Department of Geology, Royal Holloway

& Bedford New College,

University of London, Egham

Hill, Egham, Surrey TW20 OEX, U.K.

L.

SUSMAN Department of Anatomical Sciences,

State University of New York, Stony Brook, Long Island, New York 11794, U.S.A. P.

D.

TAYL0 R Department

of

Palaeontology,

The

Natural History Museum, Cromwell Road, London SW7 5BD, U.K.

J. C. THACKRAY Archivist, The Natural History Museum, Cromwell Road, London SW7 5BD, U.K.

University, PO Box 7062, Kampala, Uganda. A.

W. SKELTON Department of Earth Sciences,

University, Tempe, Arizona 85287-1404, U.S.A.

versity of Alberta, Edmonton, Alberta T6G 2E3, Canada. E.

University

5BD, U.K.

versity of California, 405 Hilgard Avenue, Los Angeles,

C.

Collections,

Denmark.

Physics and Department of Earth & Space Sciences, Uni­

A.

Geology

Natural History Museum, Cromwell Road, London SW7

5BD, U.K.

D.

J. SIVETER

sity of Oxford, Parks Road, Oxford OXl 3PR, U.K.

B. R. R0 SEN Department

T.

SIGNORDepartment of Geology, University of

U.K.

Sciences, Cambridge

CB2 3EQ, U.K.

B.

Sciences,

Open University, Walton Hall, Milton Keynes MK7 6AA,

S.

RICKARDS Department

Earth

Museum, Parks Road, Oxford OXl 3PW, U.K.

L. PREV6T Centre de Geochimie, CNRS, 1 rue Blessig,

University of Cambridge, Downing Street,

W.

D.

versity of Alberta, Edmonton, Alberta T6G 2E3, Canada.

R.

of

California, Davis, California 95616, U.S.A.

University of Liverpool, PO Box 147, Liverpool L69 3BX, S.

SHELDON Department

Open University, Walton Hall, Milton Keynes MK7 6AA, U.K.

D. PALMER Department of Geology, National Museum PAUL

SELDEN Department of Geology, University of

Sciences, University of Chicago, 5734 S. Ellis Avenue,

of California, Berkeley, California 94720, U.S.A.

C.

Connecticut

J. J. SEPK0 SKI, Jr Department of the Geophysical

PADIAN Department of Paleontology,

R.

Yale University, New Haven,

Manchester, Oxford Road, Manchester M13 9PL, U.K.

NEWTON Department of Geology, Syracuse

chester, Oxford Road, Manchester M13 9PL, U.K.

C.

Geological

06511, U.S.A.

J. NUDDS Manchester Museum, University of Man­

K.

of

10, D-7400 TUbingen 1, Germany, and Kline Geology

J. MULLER Rheinische F.-W. Universitiit, Institut

University, Syracuse, New York 13244, U.S.A.

Department

SEILACHER Institut und Museum fUr Geologie

Laboratory,

fUr Paliiontologie, Nusallee 8, D-5300 Bonn 1, Germany. R.

SCRUTTON

und Paliiontologie, Universitiit TUbingen, Sigwartstrasse

Natural History Museum, Cromwell Road, London SW7

C.

T.

Sciences, University of Durham, Durham DHl 3LE, U.K.

5BD, U.K. K.

xi

B.

A.

THOMAS Department

of

Botany,

National

Museum of Wales, Cathays Park, Cardiff CFl 3NP, U.K.

J. R. THOMASSON Department of Biology and Allied Health, Fort Hays State University, Hays, Kansas 67601, U.S.A.

List of Contributors

xii M.

E.

TOLLITT Department of Public Services, The

R.

5BD, U.K. M.

E.

S.

TUCKER Department of Geological Sciences,

Sciences,

University

of

VIOHL Jura

Department

of

Geological

California,

Santa

Barbara,

C.

R.

Museum,

University

Willibaldsburg,

0-8078

of

Nebraska,

W.A.CHTERSH.A.USER

of

Nebraska

R.

WILLIAMS Department

of

Earth

Sciences,

R.

WOESE Department of Microbiology, University

J. WOOTTON Department of Biological Sciences,

Lincoln,

4PS, U.K.

State

Nebraska

V.

P. WRIGHT Postgraduate Research Institute for Sedimentology, University of Reading, Reading RG6 2AB,

Tal

29,

0-8000

MUnchen 2, Germany.

P. J. WHYBROW Department of Palaeontology, The Natural History Museum, Cromwell Road, London SW7 5BD, U.K.

H.

University of Exeter, Prince of Wales Road, Exeter EX4

68588, U.S.A. G.

Naturkunde,

Urbana, Illinois 61801, U.5.A.

VOORHIES University

Museum,

fUr

of Illinois, 131 Burrill Hall, 407 South Goodwin Avenue,

Eichstatt, Germany. M.

Museum

Canada.

California 93106, U.S.A. G.

Staatliches

Memorial University, St John's, Newfoundland AIB 3X5,

University of Durham, Durham DHl 3LE, U.K.

J. W. VALENTINE

WILD

Rosenstein 1, 0-7000 Stuttgart 1, Germany.

Natural History Museum, Cromwell Road, London SW7

U.K.

Foreword L. R . M . C O C K S President of the Palaeontological Association 1986-1988

dition to its twin periodicals Palaeontology and Special Papers in Palaeontology, and we are particu­ larly pleased at the international response to our call for contributions, all of which have been received within a very tight timetable . However, the Association's most particular and special thanks must go to Derek Briggs and Peter Crowther, who, from the twin venues of the Univer­ sity and City Museum at Bristol, have cheerfully and enthusiastically master-minded the whole pro­ ject from its inception . Their contributions of time and effort, willingly given at the Association' s re­ quest, have culminated so effectively in the present volume . Blackwell Scientific Publications have also proved excellent partners, and have brought all their renowned publishing expertise into the pro­ duction of this book . I cannot close without reiterating what a chal­ lenging and exciting time this is for palaeontology . During the nineteenth century the dating of rocks by fossils was at the very leading edge of geological studies, but for the middle years of this century it was displaced from that central position as the new generation of machine-led scientists made quali­ tative comparisons of fossils seem old-fashioned and peripheral . However, this very volume demon­ strates how that latter position has changed, and that palaeontological and palaeobiological studies are now at the heart of a host of scientific themes ranging from evolutionary biology, through the dis­ position of continental plates in ancient oceans, to direct use in the search for oil . These changes have been accompanied by much quantitative reassess­ ment of biotas and much new machinery . Individual palaeontologists have responded vigorously to these challenges and our horizons are already expanding in all dimensions into the next century .

Scientists, both professional and amateur, have been describing fossils for over 200 years and the fruits of their labours make long library shelves groan with monographs and periodicals . These fruits have been distilled many times into the varied palaeontological textbooks and other encyclopaedic essays, which, in the case of the most common fossils, the inver­ tebrate animals, have culminated in the many vol­ umes of the Treatise on Invertebrate Paleontology. It is not our aim to compete with them . This is not an encyclopaedia of palaeontology . Why then another book? In fact the very virtues and comprehensiveness of the Treatise and other compilations have enabled many scientists to add extra dimensions to their studies over the past 20 years, and it is the fruits of this vintage crop which are assembled here . Palaeobiology has come to en­ compass the heady topics of evolution, ecology and the subsequent taphonomy of extinct animals and plants, and articles on these are gathered here in over 120 contributions by leading workers from a variety of countries . General descriptions of the morphology of fossils are omitted, but the book includes background sections on general taxonomy, biostratigraphy and techniques, and a tantalizing group of essays in which the historical background to our science is placed in perspective . Each of the contributions reflects the individuality of its authors, but we trust that each article is complete in itself (and many will no doubt directly refresh a continuing lecture course) . For over 30 years the Palaeontological Association has been the focal point in Britain for studies on fossils . This book is not merely sponsored by the Association, but was generated in outline round its Council table . It forms one of a line of continuing substantial publications by the Association in ad-

xiii

1 MAJOR EVENTS IN THE HISTORY OF LIFE

The Jurassic pterosaur Pterodactylus kochi from Solnhofen Limestone preserving impressions of the wing membranes, x 0.84. (Photograph courtesy of J.M.V. Rayner.)

1.1 Origin of Life C. R . WOESE & G . WACHTERSHAuSER

Introduction

allusion has had a significant impact on later think­ ing, undoubtedly far more than its author intended or would have liked. Oparin (1924 in Bernal 1967) and Haldane (1929 in Bernal 1967) are generally credited with formulating the issue scientifically; not because they were the first to attempt it, but because their origin scenarios were more compre­ hensive than those of their predecessors. The details of these theories need not concern us; for they often reflected misconceptions, for example as to the nature of genes, viruses, and protoplasm, and how they replicate. However, in their general aspects the theories are of great interest, for, remarkably, these half century old proposals still remain the foundation of our understanding of life's beginnings. The Oparin ocean scenario has by now become almost catechismaL It begins with a primitive anoxic atmosphere, comprising gases such as carbon diox­ ide or methane, nitrogen or ammonia, hydrogen sulphide, water, and hydrogen. Current thinking invokes the less reduced forms of these elements, the fully reduced forms being postulated earlier, by Urey (1951 ) and others, on the mistaken as­ sumption that the nascent Earth possessed an at­ mosphere similar to those found on the large gaseous planets. Miller (1953 ) was the first to put such models to scientific tests when, as a student in the nineteen-fifties, he demonstrated that electrical discharge acting on Urey's atmosphere produced a conglomeration of organic compounds that included many of the familiar amino acids. The many experi­ ments that followed showed that not only amino acids, but also a variety of organic compounds of biological interest, can be produced by a variety of energy sources under a variety of conditions (pro­ viding that oxygen is absent). So today we believe that some anoxic, slightly reducing atmosphere, acted upon by ultraviolet light and/or electrical dis­ charge, served as a continual source of the simple reactive organic chemicals needed to begin and sustain the evolutionary process. The products of this atmospheric chemistry ended up in the primitive ocean, which over time became a vast repository of reactive organic chemicals. Oparin's and Haldane's primitive ocean was a 'hot

The origins of man and his fellow creatures are concerns perhaps as old as man himself. However, before the nineteenth century these could not be given scientific form. The prescientific notions of life's beginnings were an incongruous amalgam of biblical thought, philosophy, alchemy and folk wisdom. The Bible taught that all life arose through special acts of divine creation during the first days of the Earth's existence. Commonplace experience showed, however, that life can also arise spon­ taneously, as maggots seemed to, for example, from rotting meat. And vitalism saw life as an ever­ present non-material property of the universe. In the nineteenth century four great scientific achievements laid the groundwork for making the origin of life a scientific problem: (1 ) The realization that the cell was the fundamental unit of biology introduced an enormous gulf between the living and non-living worlds; (2 ) Darwin's theory of evo­ lution implied that all life came from some distant universal ancestor; (3 ) Pasteur painstakingly and convincingly refuted the claims of spontaneous generation (of microscopic life); so that, if life had arisen spontaneously on this planet, it must have done so under conditions no longer present and probably long gone; and (4 ) Mendel discovered genetics, whose origin is to this day one of biology's great mysteries. The picture we now have of life's origin, though scientific, is based upon very few facts. It derives mainly from metaphysical assumptions, cultural images we take for granted. Consequently, it is likely to share features with the prescientific ac­ counts of life's origin which go unrecognized and unchallenged. The present discussion of origins is framed along historical lines, a format that generally helps to reveal prejudices that impede or sidetrack the development of a scientific picture. The conventional primitive ocean scenario

Darwin apparently gave little thought to the origin of life; he understood the problem to be intractable in his time. However, his casual 'warm little pond' 3

4

1

Major Events in the History of Life

dilute soup' (Haldane's phrase); hot on the mistaken assumption (common in the nineteen-twenties) that the Earth had arisen as a fragment of the Sun. With the later acceptance of a cold accretion model for the Earth's formation (also incorrect), the oceanic soup was cool from the start. The primitive ocean, then, was a 'vast chemical laboratory', a cosmic retort in which the great alchemist Nature sought to concoct the first living cell. The most important but weakest element in the ocean scenario is the transition from prebiotic chemistry to actual living, self-replicating entities. The early models necessarily resorted to hand waving arguments - interactions occurring among the reactive chemicals in the primitive ocean led to ever more complex structures, to more and more complicated aggregates, that ultimately somehow became self-perpetuating. Later proposals, drawing upon the structure of nucleic acid, refined this no­ tion to that of macromolecular templating. As Haldane (1929 ) put it, in the ocean 'the first precursors of life found food available in consider­ able quantities'. Therefore, there was no need for them to evolve the capacity to produce these metab­ olites, and so they did not. The aboriginal organisms were total heterotrophs. This seldom questioned assertion determines the subsequent evolutionalY course. A heterotrophic life style will necessarily (and rather quickly in evolutionary terms) deplete the oceanic stores of nutrients. If organisms are then to survive, they must evolve an intermediary metabolism and eventually learn to transduce other forms of energy (light or chemical) into biochemical energy. The current explanation for how intermediary metabolism arose in the aboriginal heterotrophs was fashioned by Horowitz (1945 ). When the oceanic supply of a particular amino acid, for example, became exhausted, the supply of its im­ mediate chemical precursor, for example an hydroxy acid (for which organisms previously had had no need), still remained untapped. Were the organism to evolve an enzy me that converted this precursor to the needed amino acid at this point both the organism and its progeny would survive. When, sometime later, the supply of the precursor also became exhausted, the process would repeat - the organism evolving another enzyme to catalyse syn­ thesis of the precursor from its own (previously unutilized) precursor; and so on. In this manner all intermediary metabolism arose, the pathway s evolving 'backward', one step at a time. Neither Oparin nor Haldane initially postulated

cellular entities as a starting point for evolution, although Oparin did so later, with his coacervate model. There has been no subsequent consensus as to when cellularity arose. Criticisms of and refinements to the standard model

The essence of the primitive ocean scenario (i. e. its metaphysics) has never been seriously questioned. However, many weaknesses in its details have come to light over the years. In each case the tendency has been to correct the problem by adding some new feature to the model. As a result today's origin scenario is a Ptolemaic hodgepodge of ad hoc as­ sumptions. There is little point any longer in criti­ cizing the standard model simply in the standard way, adding another ad hoc feature to remedy each new difficulty. The basic (implicit) assumptions of the model must be questioned. Cultural roots of the primitive scenario. Although the Oparin ocean scenario was developed as a scientific alternative to the prescientific versions of the pro­ cess, its similarity to the Garden of Eden myth should be worrisome. An oceanic 'paradise' is pos­ tulated in which organisms can develop safely in the midst of plenty. In addition to having 'food available in considerable quantities', the first or­ ganisms 'had no competitors in the struggle for existence' (Haldane 1929 ). Scientists have tended to see the first organism as arising through a series of highly unlikely events; discussions centre about improbable happenings that, given long times and enormous numbers of trials, eventually come to pass. (Remember that before the discovery of micro­ fossils the Earth was generally thought to have been sterile for most of its existence, allowing billions of years for key events to happen. ) Yet none of this is regarded as miraculous! One can even detect something akin to the biblical banishment in the scientific account: because or­ ganisms ultimately destroyed their oceanic paradise (by consuming the store of biochemicals), they were thrust into a harsh world where they had to fend for themselves by developing intermediary metabo­ lism, autotrophy, and eventually phototrophy. This was the dog-eat-dog world of competitive existence: 'The further life progressed the less nutrient sub­ stances were available to the organisms and the more strongly and bitterly the struggle for existence was waged' (Oparin 1 924 ). Its truth aside, the Oparin ocean scenario seems a prime example of culturally

1.1

Origin of Life

determined imagery shaping a scientific concept. The sooner these cultural influences are recognized and understood, the sooner a proper scientific pic­ ture of the origin of life will emerge. With a need for major restructuring in mind, let us analyse the main elements in the standard scen­ ario in detail. Energy sources, the multi-theatre assumption, and the ocean repository. The ultraviolet light or electrical discharge invoked by the standard scenario to create the initial simple reactive compounds are so ener­ getic that they produce indiscriminate bond rup­ tures, ionizations and free radicals. These would be entirely destructive of any larger (organic) compounds, not to mention living systems. This dilemma forces the standard scenario into a 'two­ theatre' assumption: the initial simple reactive com­ pounds produced in one theatre (the atmosphere) are subsequently quenched and protected in a se­ cond theatre (the ocean), where they accumulate and further react to produce more complex struc­ tures, and ultimately living systems. The need for two widely separated theatres seems to underlie a pernicious paradox in the standard scenario: the notion of reactive chemicals is at odds with their transport over great distances, from the high atmosphere to the ocean. The ocean must accumulate reactive chemicals over long times be­ cause distance necessarily translates into dilution. A protracted accumulation (and storage) in turn is at odds with the reactivity of these chemicals and their rapid removal by hydrolysis or sedimentation. In retrospect it is strange that all attempts to correct the difficulties with this oceanic chemical repository-reaction pot have never questioned its underlying multi-theatre assumption. Rather, dif­ ficulties were overcome by invoking additional theatres. Mineral surfaces, particularly clays, were seen by Bernal (1 967) as a vehicle for concentrating and reacting organic chemicals. He pictured the dilute organic compounds in the ocean becoming concentrated in the froth that forms on its surface, the froth being driven shoreward, to end up in estuaries - where the already concentrated com­ pounds became even more so in the oozes that formed there. Organic compounds adsorbed in high concentrations on clay sediments (and perhaps oriented and/or activated in the process) might then undergo spontaneous condensations, to pro­ duce biopolymers, and so on. Considerable experi­ mental work has been done, e. g. by Katchalsky (1973 ) and co-workers, on the properties of clay

5

minerals vis-a-vis the adsorption and reactivity of organic compounds. We shall encounter below a fundamentally different role for surfaces in the ori­ gin of life. Fox (1965 ) demonstrated experimentally that mix­ tures of amino acids (rich in the dicarboxylic amino acids) polymerized under hot (less than 200 0 C) non-aqueous conditions. Upon hydration these condensates produced 'proteinoid microspheres', which loosely resembled cells (in size, shape, and in having a few general catalytic properties). Because of this Fox argued that the high temperature con­ ditions associated with volcanic environments were those under which the organic compounds in the ocean repository bp came concentrated and re­ acted to give the prototypes of living systems. Invoking additional theatres is a Ptolemaic solu­ tion to the standard scenario's problems - which to a large extent are due to the multi-theatre as­ sumption. The true remedy may lie in single-theatre scenarios, in which the energy source can be in close proximity to or within the evolving system. These are conditions under which an energy flux can constantly generate a rich spectrum of organic biochemicals that are turned over rather than stored. The organism-environment dichotomy; heterotrophy and self-assembly. For a sufficiently primitive system, the organism-environment distinction does not exist. The dichotomy arises only when the evolving system has become sufficiently complex and physi­ cally separated from its surroundings that it can be viewed as an entity in its own right. However, the standard scenario (shaped by the properties of extant life) tends to see an organism-environment dichotomy early in the evolutionary process certainly too early. What occurs in the 'organism' is strongly distinguished from what occurs in its 'en­ vironment'. The dichotomy (together with the Garden of Eden image) then makes of the ocean repository a pre-existing store of 'food' for the abor­ iginal organisms. They, in turn, come into being as heterotrophs, and go on to deplete and ultimately exhaust their store of food. In other words, a prema­ turely forced distinction between organism and en­ vironment tends to place the replicative aspect of the primitive system in the former, its metabolic aspect in the latter. In such a dichotomous world the environment does not naturally, automatically, give rise to the organism. The latter has to 'strive' to bring itself into existence; it is the product of accidental self-assembly from simpler components in the

6

1

Major Events in the History of Life

environment - an improbable and, therefore, pro­ tracted process. (Subject to the vagaries of chance in this way the evolutionary process has to pass through a stage of instability, of uncertain outcome. ) A more extensive quote from Haldane shows these features rather clearly: 'When the whole sea was a vast chemical laboratory the conditions for the formation of such films [membranes, that is] must have been relatively favourable; but for all that life may have remained in the virus ... half­ living chemical ... stage for many millions of years before a suitable assemblage of elementary units was brought together in the first cell. There must have been many failures, but the first successful cell had plenty of food. ' Were life to have originated in an autotrophic rather than heterotrophic manner, the scenario would have been markedly different. Autotrophic evolution focuses on autocatalytic reaction net­ works, on metabolic pathways - whence all other evolutionary developments stem. It is a single theatre scenario, in which energy source, production of reactive compounds, and condensations to form complex organic structures, occupy the same locale. When life begins with autotrophic metabolic path­ ways, one avoids the kind of dichotomous sepa­ ration that exists between a heterotrophic organism and its host environment. An autotrophic system is a source of biochemical energy and complexity, not a sink for these (as are heterotrophs). With auto­ trophy the protracted, chancy trial and error period no longer seems required; the self-replicating entity (its genetics) can arise simply as a more refined and complex extension of the primary autotrophic and autocatalytic process. While some self-assembly might have to occur even in this process, that requirement too can be reduced by eliminating the constancy of the chemical conditions. In other words, major changes in the evolving system could be driven by, or be responses to, local or global changes in the state of the planet. The origin of genetics; templating and the genotype! phenotype dichotomy. Mendel's great discovery, that the cell has a phenotypic-functional aspect that is determined by a cryptic genotypic-reproductive aspect, has dominated our view of the origin of life. With the discovery in the nineteen-forties that each gene corresponds to a unique enzyme, the central question could then be phrased: 'Which came first, the gene or the enzyme?' Geneticists such as H. J. Muller felt that the gene had to have come first; only a few physiologists disagreed. (The gene at that

time was often thought of as proteinaceous and even as having its own primitive phenotype. ) Watson and Crick's discovery of the double stranded structure of nucleic acid rendered the question meaningless. Since all genes appeared to have the same basic structure, they could not have unique phenotypes, could not be functional in their own right; and, proteins (i. e. enzymes) could not evolve without genes - a chicken and egg paradox. At this point the central question should have become 'How did the genotype-phenotype re­ lationship (i. e. translation) arise?' However, the attractive and specific mechanism for the origin of gene replication inherent in the double stranded structure for nucleic acid (plus our near total ignor­ ance of the molecular mechanics of translation) took us in an opposite direction. The origin of the geno­ type (nucleic acid replication) separated completely from the origin of the phenotype (metabolism) the former question totally eclipsing the latter (Eigen et al. 1981 ; Orgel 1973 ). Recently it has be­ come popular to believe that (RNA based) 'nucleic acid life' must have preceded protein-based life; that initially nucleic acid was both the genotype and the phenotype. This point of view is supported by the facts: (1 ) that polypyrimidines can serve as templates that align complementary purine nucleo­ tides, which (when properly activated) then go on to condense into polypurine chains; and (2 ) that some RNAs possess certain limited enzymatic or catalytic properties. Eigen has also reported that in the presence of a particular protein (the replicase of the virus Q�) a certain type of RNA will spontaneously arise (in the absence of a pre-existing template). A fascinating variation on the templating theme is Cairns-Smith's (1985 ) proposal that life began with replicating patterns in clay layers (which could adsorb organic molecules and thereby influence the course of subsequent organic evolution). While daring in one way, this proposal is conventional in another; it takes for granted the need for templating as an initial step in the origin of life. A totally dichotomous view of the origins of the genotype (replication) and the phenotype (metab­ olism) is an extrapolation in the wrong direction. It has even led Dyson (1985 ) to propose that life arose twice; initially somehow as protein-based life, within which nucleic acid then separately arose as a 'disease'. The earliest life forms were almost cer­ tainly not incarnations of our dichotomies, of our attempts to define extant life. Rather, primitive living systems were undoubtedly less well de-

1.1

7

Origin of Life

fined, less compartmentalized, than their modern counterparts, and so in that unusual sense more 'integrated'. It is time to reassess the genotype­ phenotype dichotomy as a paradigm for the origin of life. A proper conceptualization of translation, the process that defines the genotype-phenotype re­ lationship, should have an integrating, unifying effect on our concept of the origin of life. Unfortu­ nately, the translation mechanism is large and com­ plex, and, therefore, its molecular workings and evolution are not understood. The fact that some of the proteins involved in translation are also com­ ponents of certain nucleic acid replication enzymes, however, suggests primitive connections between the two processes. The facts that cells today contain transfer RNA-like molecules as essential parts of non-translational ('non-programmed') polymeriz­ ations (e. g. polypeptide antibiotic and cell wall syn­ theses), and that nucleotides, other heterocyclic compounds, and even transfer RNAs play important roles in intermediary metabolism, hint at still deeper evolutionary connections. The suggestions are strong that the programmed polymerizations (translation and nucleic acid repli­ cation) have arisen out of more primitive metabolic interactions. Therefore, what seems called for at this juncture is a general view of polymerization pro­ cesses, one that attempts to relate polymer formation to the full spectrum of metabolic reactions in primi­ tive systems - e. g. the types of polymers arising under primitive conditions; the range of monomer units and chemical linkages involved; whether polymers were formed by monomer or oligomer condensations; chirality constraints; whether the sequences of the aboriginal polymers were random or (simply) ordered (e. g. homopolymers, poly­ mers of alternating sequence, etc. ); the extent to which templating is or is not involved; and oligonucleotide-amino acid interactions. Geological and phylogenetic constraints

time interval during which the evolutionary process could have started. The Earth's crust is now believed to have been initially quite hot, too hot to sustain liquid water. Any water present would have been partitioned between the primitive atmosphere and a semi­ molten crust. There is also geological evidence to suggest that the Archaean oceans were warm. The oldest sedimentary rocks (3800 Ma), although somewhat metamorphosed, give evidence of life at that time; and the better preserved 3500 Ma sedi­ ments give clear evidence of baderial life - showing both fossil stromatolite structures and microfossils (see also Section 1 .2 ). In that stromatolites today are produced by photosynthetic bacteria, principally cyanobacteria (or thermophiles of the Chloroflexus type), photosynthetic bacteria (probably) already existed 3500 Ma. The explosive developments in molecular phylo­ geny over the past decade have revealed a number of important facts: (1 ) the earliest phylogenetic branchings gave rise to three aboriginal lineages, the eubacteria, the archaebacteria and the eukaryo­ tes; (2 ) photosynthesis appears to have arisen (early) within the eubacteria. If so (given the stroma­ tolite evidence), eubacteria already existed at least 3500 Ma, so that the most recent ancestor common to all life lived at a still earlier time - probably far earlier, because of the enormous evolutionary dis­ tances that separate the three classes; (3 ) prokaryotic life (at least) arose in high temperature environ­ ments; (4 ) the ancestral environments were an­ aerobic; and (5 ) the ancestral forms of prokaryotic metabolism may have been autotrophic. Compari­ sons among the (sequences of the) genomes of diverse organisms will ultimately permit us to infer in some detail the nature of the most recent common ancestor of all extant life, and also certain things about still earlier evolutionary stages (see also Section 2 . 1 ) All the evidence to date, then, points to life having arisen quite early in the planet's history, and under thermophilic conditions. .

on the primitive ocean scenario

Knowing when the evolutionary process started is crucial to understanding how it occurred. Con­ ditions during the first few hundred million years of Earth's existence were certainly very different from those occurring 2000 million years later. The current understanding of the geological his­ tory of the Earth, Moon, and other planets, together with recent advances in the biologist's under­ standing of phylogeny, substantially restrict the

Alternatives to the primitive ocean scenario

An important methodological rule of K. Popper is that a new theory should have a greater explanatory power than its predecessors, i. e. it should explain a multitude of facts with a minimum of assumptions. Clearly, today's consensus theory of the origin of life is little more than a highly amended version of the original OparinlHaldane scenario it has replaced - which violates Popper's rule. Further

8

1 Major Events in the History of Life

amendments to the standard scenario are not what is needed; true alternatives to it are. Wachtershauser (1988 ) has proposed one such alternative, which dispenses with the multi-theatre assumption, the ocean repository, heterotrophic origin, and modular self-assembly. This theory, moreover, is sufficiently detailed to make testable assertions regarding the nature and evolution of primitive biochemical pathways. The first organ­ isms are assumed to be truly autotrophic (not hetero­ trophic) - the result of de novo biosynthesis of organic constituents by the uptake of inorganic material (e. g. CO2), and subsequent rearrangement reactions. They are not the products of accidental modular assembly. The theory's central idea is that life began with autocatalytic, metabolic pro­ cesses occurring in an essentially two-dimensional fashion, within organic monolayers anionically bonded to positively charged surfaces of minerals, such as pyrite, and in contact with water at high temperature. Adherence to the mineral surface is not the result of adsorption but of an in situ auto­ trophic growth of organic constituents that acquire their anionic surface bonding in statu nascendi. The concentration of dissolved organic constituents in the water phase is negligible. Hence the process by which a constituent loses its surface bonding is irreversible; detachment is tantamount to dis­ appearance. (In this respect the theory is the op­ posite of Bernal's clay theory, which is based upon adsorption). On these pyrite surfaces large poly­ anionic constituents, with ever stronger surface bonding, are automatically selected - to begin with polyanionic coenzymes, eventually nucleic acids and polypeptides. The primitive system grows by spreading onto vacant surfaces, reproduces by producing its autocatalytic coenzymes, and its evolution is driven by environmentally induced ignitions of new autocatalytic cy cles. The system evolves toward higher complexity, since the ther­ modynamic equilibrium in a surface metabolism would favour synthesis, not degradation (as would occur with solution reactions). High energy phospho-anhydride groups are not required for the formation of covalent bonds. Phosphate groups (whose source is taken to be the mineral substrate) have the sole function of surface bonding. The energy for carbon fixation is provided by the redox process of converting ferrous ions and hydrogen sulphide into pyrite, which is not only a waste product but provides the all-important binding sur­ face for the organic constituents. This initial laminar organism is succeeded by two

further stages. The second stage organisms are semicellular entities still supported by a pyrite sur­ face but having an (autotrophically grown) lipid over-layer, with an internal broth of detached con­ stituents. In this 'bleb' stage a membrane metabo­ lism and a cytosol metabolism appear, first as a supplement to, and later as a substitute for, the aboriginal surface metabolism. Membrane-bound electron transport chains allow the tapping of other redox energy sources and ultimately of light energy. The cytosol metabolism allows the salvaging of detached constituents by catabolic processes and the development of modular modes of synthesis that rely upon energy coupling. Eve�tually hetero­ trophy appears, as a by-product of the catabolic salvage pathways. The cell's genetic machinery develops from surface-metabolic precursors. It produces self-folding enzymes which compete with the mineral surface for bonding the metabolic con­ stituents. In this stage evolution becomes double tracked, an evolution of metabolic pathways and one of the bonding surfaces for their constituents. In the third stage the pyrite support is abandoned and true cellular organisms arise. Since the ocean cannot reasonably function as a reaction pot in which life originated, and its role as a repository is suspect, the question is whether it played any significant role at all in the origin of life. Two types of scenarios exist that make minimal use of the ocean. One is the idea that hydrothermal vents served as the aboriginal environment. Since hydrothermal vents create chemical gradients, a single-theatre vent scenario can be developed that has no need for the ocean repository assumption. How the model would cope with the fact that vents, and so their products, are ephemeral (especially so on an evolutionary time-scale) is unspecified. It was suggested by Woese (1979 ) that evolution began in the primitive atmosphere, at a time when the planet's surface was too hot to sustain liquid water. The early Earth can be pictured as surrounded by vast cloud banks, as Venus is today. The severe weather conditions that must then have existed would have caused large quantities of minerals (dust), from the dry surface, to be swept into the atmosphere. Atmospheric water vapour then con­ densed on the dust, dissolving it (in part). As a consequence, the primitive Earth was enshrouded in clouds of salt water. In addition to containing (possibly high concentrations of) minerals, the droplets in these clouds would accumulate organic compounds, produced by interactions among at­ mospheric gases and other constituents (or with

1.2 Precambrian Evolution compounds produced by thermal reactions on the Earth's surface and swept into the atmosphere). These droplets are natural precursors of cells their surfaces coated with mixtures of the larger organic compounds, their interiors solutions of re­ active (organic and inorganic) compounds. The dif­ ferent layers of the atmosphere would each have characteristic chemistries, the whole being in effect a connected series of chemostats. Droplets (and hydrated dust) offer enormous amounts of surface, and so surface chemistry becomes all important in life's beginnings. As the primitive Earth cooled, its surface would pass from a dry condition, through cycling damp/dry stages, to one where large bodies of (hot) water could accumulate. These major global transitions would bring about major changes in the evolution­ ary course (see above). The cloud setting suggests a single theatre scenario, requiring no repository as­ sumption; it also suggests that major stages in evol­ ution were driven by (were responses to) major changes in the state of the planet. In one sense the origin of life problem today remains what it was in the time of Darwin - one of the great unsolved riddles of science. Yet we have made progress. Through theoretical scrutiny and experimental effort since the nineteen-twenties many of the early naive assumptions have fallen or are falling aside - and there now exist alternative theories. In short, while we do not have a solution, we now have an inkling of the magnitude of the problem.

9

References Bernal, J.D. 1967. The origin of life. World Publishing Co., Cleveland, Ohio. Cairns-Smith, A.G. 1985. Seven clues to the origin of life. Cambridge University Press, Cambridge. Dyson, F.J. 1985. Origins of life. Cambridge University Press, Cambridge. Eigen, M., Gardiner, W., Schuster, P. & Winkler-Oswatitsch, R. 1981. The origin of genetic information. Scientific

American 244, 88-1 18. Fox, S.W. (ed.) 1965. The origins of prebiological systems: and of their molecular structure. Academic Press, New York. Horowitz, N.H. 1945. On the evolution of biochemical syn­ theses. Proceedings of the National Academy of Sciences,

USA 31, 153-157. Katchalsky, A. 1973. Prebiotic synthesis of biopolymers on inorganic templates. Naturwissenschaften 60, 215-220. Miller, S.L. 1953. A production of amino acids under possible primitive earth conditions. Science 117, 528-529. Oparin, A.l. 1938. The origin of life. (Translation of 1936 Russian Edition.) Macmillan, London. Orgel, L.E. 1973. The origins of life: molecules and natural selection. John Wiley & Sons, New York. Urey, H.C. 1951. The origin and development of the earth and other terrestrial planets. Geochimica et Cosmochimica

Acta 1, 209-277. Wachtershauser, G. 1988. Before enzymes and templates: theory of surface metabolism. Microbiological reviews. 52,

452-484. Wald, G. 1964. The origins of life. Proceedings of the National

Academy of Sciences, USA 52, 595-611. Woese, C. R. 1979. A proposal concerning the origin of life on the planet Earth. Journal of Molecular Evolution 13, 95-101.

1.2 Precamb rian Evolution of Prokaryotes and Protists A . H . KNOLL

Introduction

The Phanerozoic Eon, the interval under discussion in most of this volume, encompasses the most recent 13% of our planet's history. A sedimentary record documenting more than 3000 Ma of Archaean and Proterozoic time extends below the base of the Cambrian System, and research conducted over the

past three decades has demonstrated that this entire sweep of history is the proper domain of palaeon­ tology. Stromatolites, microfossils, and geochemical markers provide fragmentary, sometimes frustrat­ ing, but critically important evidence for early evo­ lution. Like younger invertebrate fossils, fossil

1 Major Events in the History of Life

10

prokaryotes and protists must be studied as popu­ lations characterized by a measurable range of morphological variation, reproductive pattern, behavioural orientation, taphonomic features, and distribution within and among sedimentary en­ vironments. Unlike invertebrate fossils, significant questions of metabolism may remain after popu­ lations have been otherwise characterized. The interpretation of early metabolic diversity requires that morphological investigations be supplemented by trace fossil studies (stromatolites and oncolites being the preserved traces of microbial communi­ ties) and geochemical analyses of ancient metabolic and environmental indicators. Geological data must be integrated with information from molecular phylogeny and the comparative physiology of living organisms, and interpreted with a clear appreciation of our incomplete understanding of both living micro-organisms and their geological record.

The Archaean E on: the early diversification of micro-organisms

The age of the earliest palaeobiological record has not changed appreciably in more than 20 years, but the quality of interpretable evidence has improved significantly at decadal intervals. Palaeobiological investigations of Early Archaean rocks have con­ centrated on two successions, the Onverwacht Group of South Africa and the Warrawoona Group, Western Australia. Both sequences are dated at c . 3500 Ma. Both are little-metamorphosed greenstone belt successions characterized by thick mafic and ultramafic lavas, subordinate felsic volcanics, and intercalated sedimentary rocks. Sediments origi-

Fig 1

nated largely as volcaniclastics and chemical pre­ cipitates, including carbonates, but most have been extensively silicified. Stratiform, domal, and colum­ nar to pseudocolumnar stromatolites occur locally in both areas (Byerly et al. 1986 ; Walter in Schopf 1983 ). These structures have generally been inter­ preted as the trace fossils of microbial communities. Although this interpretation is reasonable, no Early Archaean stromatolites are known to contain micro­ fossils. Thus, abiological alternatives must be con­ sidered, and biogenicity defended on the basis of gross morphology and microstructure (Buick et al. 1981 ). Microfossils have also been reported from both groups. Simple carbonaceous spheroids of varying size were reported from several horizons in the Onverwacht and overlying Fig Tree groups during the nineteen-sixties but the biogenicity of many of these structures is open to question. During the nineteen-seventies, several authors reported popu­ lations of spheroidal carbonaceous microstructures that show a number of features more consistent with a biological interpretation. These include a narrow, nearly normal size frequency distribution about a mean diameter of 2 . 5 f! m, clear evidence for binary division, a sedimentary context com­ parable to that of younger, undisputed microfossils, and taphonomic features comparable to younger fossils such as flattened and wrinkled vesicles and the occasional preservation of internal carbon­ aceous contents (Fig. 1K). Rod-like and filamentous microstructures have also been reported from the Swaziland succession, but their antiquity and mode of origin remain subjects for debate. Undoubted filamentous microfossils have re­ cently been described from cherts of the Warra-

Representative Archaean and Proterozoic fossils. A, Gunflilltia (filaments) and Huroniospora (spheroids) in stromatolitic

chert from the Lower Proterozoic Gunflint Iron Formation, Ontario. E, Stromatolites from the Upper Proterozoic Backlundtoppen Formation, Spitsbergen. C, D, Low and high magnification views of a surface-encrusting cyanobacterial population from the Upper Proterozoic Limestone- Dolomite 'Series', central East Greenland - the nested cups are successive extracellular envelopes produced by coccoidal cyanobacteria that jetted upward from the sediment surface, much as morphologically similar populations in peritidal environments of the Bahama Banks do today. E, F, Chroococcalean cyanobacteria from silicified playa lake carbonates of the Upper Proterozoic Bitter Springs Formation, Australia. G, Vase-shaped protist from the Upper Proterozoic Elbobreen Formation, Spitsbergen. H, Low magnification view of oscillatorian cyanobacteria from the Upper Proterozoic Backlundtoppen Formation, Spitsbergen, showing the alternation of vertical and horizontal orientations characteristic of many mat-building populations. I, Acritarch isolated from shales of the Upper Proterozoic Chuar Group, Arizona. J, Endolithic hyellacean cyanobacterium in silicified ooids from the Upper Proterozoic Limestone - Dolomite 'Series', central East Greenland - ooid surface is toward the top of the photograph. K, Spheroidal microstructure from a population showing various stages of binary division, Early Archaean Onverwacht Group, South Africa . L, Large, process-bearing acritarch preserved in chert nodules within a moderately metamorphosed succession of latest Proterozoic age, Prins Karls Forland, Svalbard . Bar

=

30 Ilm for A, 10 cm for B, 400 Ilm for C, 100 Ilm for D, 20 Ilm for E, F and I, 50 Ilm for G, H and J, and 75 Ilm for L.

1 .2 Precambrian Evolution

11

12

1

Major Events in the History of Life

woo na Gro up (Scho pf & Packer 1987), where they o ccur in asso ciatio n with clusters o f sphero idal unicells encased in multiple extracellular envelo pes. These micro fo ssils are mo rpho lo gically similar to extant cyano bacteria, and may be early represen­ tatives o f this gro up; ho wever, that interpretatio n is by no means assured. Even if the fo ssils do represent early cyano bacterial ancesto rs, there is no assurance that they were o xygenic pho to auto tro phs using two pho to systems. In the presence o f H2S, many living blue-greens pho to synthesize ano xygenically using o nly pho to system I, i. e. H2S, H2, o r o rganic mo l­ ecules do nate electro ns, and no O2 is pro duced. Co mparative bio chemistry indicates that this pho to synthetic system evo lved earlier than the cyano bacterial (and higher plant) pathway in which water do nates electro ns. The apparent lo w mo rpho lo gical diversity o f described Early A rchaean micro fo ssils canno t be taken too literally. Studies o f Early Pro tero zo ic assemblages fro m Western A ustralia have demo n­ strated that, as mo rpho lo gically varied assemblages o f fo ssils undergo increasing diagenetic and in­ cipient metamo rphic alteratio n, they beco me ' archaeanized' - i. e. they appear to co nverge mo rpho lo gically o n the simple micro structures fo und in weakly metamo rpho sed Early A rchaean cherts (Kno ll et al. 1988 ). The bio lo gical fixatio n o f CO2 is acco mpanied by a marked fractio natio n o f the stable iso to pes o f 2 carbo n, 1 C and 1 3c . Carbo n iso to pic ratio s in Onverwacht and Warrawoo na carbo nates and kero gens indicate significant fractio natio n between o xidized and reduced species, suggesting an Early A rchaean carbo n cycle fuelled by pho to synthesis, po ssibly under co nditio ns o f elevated Pe0 2. Sulphur isoto pes are likewise fractio nated during dissimilato ry sulphate reductio n, but in co ntrast to carbo n, Early A rchaean sulphur-bearing samples sho w little fractio natio n between sulphides and sulphates. A t t he same time, sedimento lo gical evi­ dence indicates that sulphate was an impo rtant anio n in the water bo dies beneath which bo th the Onverwacht and Warrawoo na beds accumulated. Th is apparent parado x has several po ssible expla­ natio ns: (1 ) it is po ssible that Early A rchaean o ceans co nt ained negligible sulphate co ncentratio ns, and that ro cks co ntaining evidence fo r sulphates in bo th the Onverwacht and the Warrawoo na gro ups accumulated under no n-marine co nditio ns - an explanat io n that is unsatisfacto ry to many geo lo gists familiar with the ro cks; (2 ) it is po ssible that signi­ ficant co ncentratio ns o f sulphate existed in o cea ns

fo r several hundred millio n years befo re pro karyo tes learned to use it - an explanatio n unsatisfacto ry to micro bio lo gists, who no te that bacteria evo lve rap­ idly to explo it no vel substrates; o r (3 ) perhaps almo st all sulphate in po re fluids was reduced bio ­ lo gically to sulphide in an essentially clo sed system with little fractio natio n because o f high ambient temperatures (70°C o r mo re) - a theo ry fo r which the geo lo gical reco rd pro vides little suppo rting evi­ dence. A generally acceptable so lutio n to this pro b­ lem has no t yet been pro po sed. Despit e o utstanding pro blems o f palaeo bio lo gical interpretatio n, it seems clear that 3500 Ma the Earth suppo rted co mplex pro karyo tic eco systems driven by pho to synthesis. Oxygen may have been gener­ ated by Early A rchaean cyano bacteria, but geo ­ chemical evidence indicates that any O2 pro duced was largely co nsumed by the o xidatio n o f o rganic matter, ferro us iro n, and sulphides. A mbient P0 2 appears to have been lo w and physio lo gical path­ ways, co nsequently, anaero bic. Oxide facies iro n fo rmatio n is fo und in Early A rchaean basinal facies, but no t in shallo w vo lcanic platfo rm sequences, pro mpting speculatio n that o xygenic pho to syn­ thesis may have o riginated in ' mid-gyre' enviro n­ ments far fro m sites o f vo lcanic o r sedimentary H2S generatio n. Co mpariso ns o f info rmatio nal macro ­ mo lecules in extant micro -o rganisms independently suggest rapid metabo lic diversificatio n early in evo lutio nary histo ry. Early branching gro ups in bo th the eubacteria and archaebacteria are pre­ do minantly anaero bic, thermo philic, and sulphur­ dependent; several are auto tro phic (Wo ese 1984 ). The search fo r o lder bio lo gical reco rds is limi ted by the paucity o f pre-3500 Ma sedimentary se­ quences. 3800 Ma ro cks fro m Isua, so uthwestern Greenland, co ntain reduced carbo n that is iso ­ to pically fractio nated relative to carbo nates in the same successio n, but the metamo rphism o f these ro cks to amphibo lite grade has o bliterated any un­ ambiguo us indicatio ns o f bio lo gical activity. L ater A rchaean successio ns in A ustralia, A frica, and No rth A merica co ntain diverse stro mato lites, rare micro fo ssils o f cyano bacterial aspect, and lo cal evi­ dence o f unusually stro ng carbo n iso to pe fractio n­ atio n. Mo st o f the iso to pically light kero gens co me fro m no n-marine depo sits, so their interpretatio n in terms o f glo bal co nditio ns is no t straightfo rward; ho wever, it has been suggested \that iso to pically light kero gens fix a minimum age fo r the evo lutio n o f a ero bic methylo tro phy (the metabo lic o xidatio n o f methane o r o ther o ne-carbo n co mpo unds; Hayes in Scho pf 1983 ).

1 .2 Precambrian Evolution The Early Proterozoic Eon: the diversification of aerobes

The modem era of Precambrian palaeontology began in 1954 with the brief description by S. Tyler & E. S. Barghoorn of microfossils preserved in cherts from the 2000 Ma Gunfl int Iron Formation, Canada. Sub­ sequent research has demonstrated that several discrete microfossil assemblages occur in Gunflint rocks. Stromatolitic cherts near the base of the for­ mation contain abundant microfossils preserved as organic, haematitic, or pyritic structures. Although more than a dozen valid species have been described from this facies, two taxa together comprise more than 99% of all individuals (Fig. lA). Gunflintia minuta is a thin (usually 1 -2 !-! m) fila­ mentous sheath that has been compared to both nostocalean cyanobacteria and iron bacteria. Its affinities remain uncertain; locally inflated areas along filaments interpreted as akinetes and hetero­ cysts (distinct cell types produced by nostocalean blue-greens) are probably diagenetic in origin. Small (2 - 15 !-! m) spheroidal fossils assigned to the genus Huroniospora occur in the same beds. The phylogenetic relationships of these populations are also unclear, but their recent interpretation as bac­ terial spores merits serious consideration. Other microfossils in the Gunfl int stromatolitic assem­ blage are uncommon; they include probable iron­ oxidizing bacteria, possible cyanobacteria, and problematica, but no strong candidates for eukaryotic assignment. Although these fossils occur within laminated stromatolitic structures, Gunflintia and Huroniospora populations do not display the orientations charac­ teristic of mat-building micro-organisms in younger rocks. Thus, like their phylogenetic relationships, their ecological interpretation as mat-builders is open to question. Non-stromatolitic Gunfl int assemblages include microbenthos preserved in silicified muds and probable planktic populations. The mud micro­ benthos is dominated by stellate microfossils inter­ preted as iron and manganese oxidizing bacteria, while the apparent planktic forms are 6 - 31 !-! m dia­ meter spheroids of uncertain systematic position. Whatever the taxonomic affinities of Gunfl int microfossils, it is clear that generally similar assem­ blages were widely distributed 2000 Ma. Assem­ blages comparable to Gunflint mud, mat, and plankton fl orules occur in Labrador, the Canadian N orthwest Territories, and two areas in Western Australia (references in Knoll et al. 1988 ). Not all of

13

these occur in iron formations, and several contain microfossils not found in the Gunflint Formation itself. For example, silicified carbonate muds of the Duck Creek Dolom ite, Western Australia, contain septate filaments as much as 63 !-! m in diameter among the largest such fossils known from any Proterozoic formation. Although Gunflint-like as­ semblages are widely distributed in Lower Protero­ zoic formations, they are not the only fossils in rocks of this age. Assemblages from hypersaline peritidal roc ks of the Belcher Supergroup, Hudson Bay, Canada, contain populations that are indis­ tinguishable from cyanobacteria found today in comparable environments (Hofmann 1976 ). Stromatolites are abundant and morphologically diverse in Lower Proterozoic platform carbonates (Walter 1976 ). It is not certain whether the observed increase in stromatolite diversity between the Late Archaean and Early Proterozoic eras refl ects a radi­ ation in mat-building prokaryotes, a preservational consequence of Late Archaean continental crustal growth and stabilization, or both. What may have been the most profound evo­ lutionary changes of the Early Proterozoic Era are events that must be inferred from sedimentological and geochemical data. During the Early Proterozoic, the degree of isotopic fract ionation recorded in sulphur-bearing minerals increased substantially. Detrital uraninite ceased to be a significant con­ stituent of f luviatile and deltaic sediments , while red beds became widespread. Limited data suggest that iron retention in palaeosols developed on mafic parent materials decreased by the end of this interval. Beginning with Preston Cloud, numerous com­ mentators have suggested that these phenomena refl ect a significant increase in the partial pressure of oxygen in the Earth's atmosphere. This has sometimes been interpreted as meaning that the Early Proterozoic atmosphere shifted from reducing to a composition comparable to the present; how­ ever, such a black-and-white view no longer seems tenable. The Archaean (especially the late Archaean) atmosphere undoubtedly contained some molecular oxygen, albeit in low concentrations. At the end of the E arly Proterozoic Era, the atmosphere probably contained only one to a few per cent of present day O2 levels. The difference, however, is metabolically significant; aerobic respiration is possible in the latter atmosphere, but not in the former. Some palaeontological evidence supports the idea of Early Proterozoic aerobic prokaryotes, but clearer insights come from molecular phylogeny and comparative

1 Major Events in the History of Life

14

physiology. In many aer obic physiological path­ ways, oxygen-r equir ing steps ar e appended to an other wise anaer obic ser ies of r eactions (C hapman & 5 chopf in 5 chopf 1983 ). Molecular data, specifically compar isons of nucleotide sequence in 1 65 r ibo­ somal RNA molecules among differ ent living micr o­ or ganisms, suggest that aer obic r espir ation evolved independently in a number of gr oups, most of which ar e fundamentally photoautotr ophic (Woese 1984 ). If one accepts that br oad constr aints on the timing of evolutionary events can be gleaned fr om molecular data, then it can be inferr ed fur ther that the polyphyletic evolution of aer obic pr okar yotes occurr ed dur ing a r elatively br ief per iod following a long per iod of anaer obic evolution (Fig. 2 ). The later Proterozoic Eon: the emergence of protists

Although tr eated last in this chr onological account, the later Pr oter ozoic Eon might have justifiably been discussed fir st, because its palaeobiological r ecor d, especially for the per iod 900 -600 Ma, is far mor e extensive and better pr eser ved than that of ear lier epochs. Near ly 200 Late Pr oter ozoic

o ( Ma)

Anaerobes

A n i m a llPlant associ ated ( Polyphyl e t i c)

Aerobes ( Polyphyletic)

570±20

P

I

,

, I ?

C H

I 4000

I , ,

I I I

Y

3000

A E A

?

I I ?

Endosymbiotic o r i g i n s of m itoc h o n d r i a ( Polyphyletic)

?

I

I I I , ,

Eu karyotic cytosol ancestor I

?

P R O KARYOTES 4600 Fig. 2

Endosymb of p l a s ti ds ( polyphylet i c )

L----- : � :I

2000

2500

N

?

v

,iotic � lI origins

C

A R

Seaweeds

1 000

T E R

o Z o

1 -c e l l ed p rot ists seen in fos s i l reco rd

f;;:J

P

H

�

micr ofossil biotas ar e known fr om sev. en continents (Knoll 1985 ). Envir onmental sampling is far better than for ear lier er as. Thus, it is in later Pr oter ozoic sequences - wher e the r ecor d is cl ear est - that pr inciples of palaeoecological, palaeogeogr aphical, taphonomic, systematic and, hence, evolutionar y inter pr etation can best be established. Late Pr oter ozoic micr ofossil assemblages have been r epor ted fr om silicified car bonates r ep­ resenting a var iety of per itidal depositional en­ vir onments. In situ micr obenthic populations occur in str atifor m str omatolites an d, much less fr e­ quently, in conoidal, domal, or columnar for ms (Fig. 1 C -F, H). Micr obenthos can also be found in silicified micr ites, oncoids, and ooids, as well as in shales and, r ar ely, in unsilicified car bonates. Ther e is a str ong corr elation between facies and assem­ blage composition. Many populations ar e con­ vincingly inter pr eted as cyanobacter ia, although under exceptional cir cumstances bacter ial heter o­ tr ophs can be r ecognized. Less amenable to inter ­ pr etation ar e populations of unor namented 1 0 -20 [tm spher oids that ar e distr ibuted spor adically thr oughout most fossilifer ous r ocks. Although their simple mor phology pr ecludes confident systematic

Summary chart illustrating generalized patterns of prokaryotic and protistan evolution.

PROTI STS

?

1.2 Precambrian Evolu tion classification, some o f these fossils r esemble the cells and cysts of gr een algae and protozoans that occur in moder n micr obial communities of per it idal and hyper saline lake envir onments. J udging fr om their spatial distr ibution within and among facies, other spher oid populations appear to be allo­ chthonous, pr obably planktic, elements. Many Late Pr oter ozoic pr okar yotes differ little in mor phology, development, or behaviour fr om living c yanobacter ial populations found in physical en­ vir onments like those inferr ed for the fossils. For example, endolithic microfossil assemblages found in silicified ooids fr om the 700 -800 Ma Eleonor e Bay Gr oup, East Gr eenland, contain half a dozen discr ete populations which have close moder n counter par ts in present day Bahamian ooid shoals (Fig. In . Late Pr oter ozoic cyanobacter ia appear to be essentially moder n in their diver sity and en­ vir onmental distr ibution. One can hypothesize that the appar ent incr ease in cyanobacter ial diver sity r ecor ded in the Pr oter ozoic as a whole is mainly a function of mor e complete sampling in younger successions; that is, the major featur es of cyano­ bacter ial diversity wer e established dur ing the Ear ly Pr oter ozoic Er a or ear lier . This hypothesis cannot be r ejected on the basis of curr ently available data. The r ecord of other pr okar yotes is less clear , although the pr esence of Late Pr oter ozoic sulphate r educer s, methanogens, methylotr ophs and other bacter ia can be established or inferr ed on the basis of geochemical evidence. Str omatolites pr ovide sedimentar y evidence for the continued wide distr ibution of micr obial mat communities in later Pr oter ozoic envir onments (Fig. lB). It has been suggested that Pr oter ozoic str omatolites changed systematically as a function of age, and that this provides indir ect evidence for Pr oter ozoic cyanobacter ial evolution. Sever al objec­ tions can be r aised against this view: (1 ) the debate over the str atigr aphic distr ibution of str omatolite for ms continues unr esolved - hinder ed by the failur e of many r epor ts to place stromatolites in their pr oper sedimentological per spective and by the absence of a r ational, internationally accepted system of nom enclature; and (2 ) it may well be tr ue that cer tain str omatolites char acter ize par ticular time inter vals, but this does not necessar ily say anything about cyanobacter ial evolution. Differ ­ ences between Ear ly and Late Pr oter ozoic str omato­ lites may as easily r eflect the addition of eukar yotic algae to mat-building communities, tempor al changes in featur es of the physical envir onment (such as C aC03 super satur ation), the evolution of

15

uncalcified metaphytes that outcompeted micr o­ or ganisms for space in cer tain envir onments, or the evolution of meiofaunal gr ade metazoans. Undisputed pr otistan fossils ar e abundant in Upper Pr oterozoic r ocks. Lar ge (up to 2 mm) acr i­ tar chs occur in both silicified car bonates and shales (Fig. 1 1 ); some of these may r epr esent the phycomata of planktic pr asinophyte algae, but the systematic r elationships of most ar e uncer tain. Latest Pr oter o­ zoic cher ts a nd finely laminated shales fr om C hina, A ustr alia, and Svalbar d contain par ticular ly complex for ms, including spiny and pr ocess-bear ing populations (Fig. lL). In their gener al level of mor phological complexity, these r esemble younger Palaeozoic acr itar chs, but the Pr oter ozoic for ms ar e invar iably much lar ger and ar e cer tainly distinct at the specific and, usually, the gener ic level. Recent discover ies in Spitsber gen and Ar ctic C anada demonstr ate that the r ecor d of spinose and pr ocess­ bear ing acr itar chs goes back at least to 800 Ma. Vase-shaped micr ofossils of uncer tain systematic position also occur in Upper Pr oter ozoic shales and car bonates (Fig. I G); in some successions, they ar e among the most abundant fossils pr eserve d. Like fossil pr okar yotes, pr otistan micr ofossils r eflect palaeoenvir onments in their distr ibution, but unlike pr okar yotes, they change systematically thr ough time. Ther efor e, acr itar chs have pr oved useful in at least Late Pr oter ozoic biostr atigr aphy (Vidal & Knoll 1983 ; Hofmann 1987 ). The r ecor d of eukar yotes can be tr aced though time at least back to 1 700 Ma, when both the mor ­ phological and molecular geochemical r ecor ds of pr otists begin (J ackson et al. 1986 ). The r ecor d of metaphytes may be almost as long. Diver se multi­ cellular algae occur in Upper Pr oter ozoic r ocks (Hofmann 1 985 ); with somewhat less confidence, both car bonaceous and tr ace fossil r emains in 1300 - 1400 Ma r ocks can be inter pr eted as sea­ weeds. No unequivocal r emains of metazoans have been descr ibed fr om pr e-Ediacar an deposits. Thus, either seaweeds and animals or iginated at str ikingly differ ent times or , for the fir st half of their histor y, animals must have been tiny, meiofaunal gr ade or ganisms unlikely to sur vive as fossils or pr oduce r ecognizable tr aces. While the palaeobiological tr ail of ear ly eukar ­ yotes curr ently turns cold at about 1 700 Ma, it must be admitted that nucleated cells th at wer e incapable of fossilization or , at least, unlikely to be r ecognized as eukar yotic, almost cer tainly existed ear lier . How much ear lier is unclear . The ancestor s of the eukar­ yotic cytosol (nucleus and cytoplasm) appear to

16

1 Major Events in the History of Life

have aris en early in Earth his tory, either directly from the progenote or later from archaebacterial ances tors . The Early Proterozoic P02 increas e prob­ ably fos tered endos ymbiotic couplings between an­ ces tral cytosols and purple nons ulphur bacteria, leading to the polyphyletic evolution of hetero­ trophic, mitochondria-bearing protis ts . The later acquis ition of endos ymbiotic cyanobacteria res ulted in the origin of eukaryotic algae, again indepen­ dently in s everal lineages . I ndeed, it appears that the plas tids of s ome algal groups are des cended from endos ymbiotic eukaryotic algae, giving s uch organis ms a truly complicated phylogeny.

References Barghoorn, E . 5 . & Tyler, S . M . 1965 . Microorganisms from the GunfIint Chert. Science 147, 563-577. Buick, R . , Dunlop, J . 5 . R . & Groves, D J . 1981 . Stromatolite recognition in ancient rocks: an appraisal of irregularly laminated structures in an Early Archaean chert- barite unit from North Pole, Western Australia. Alcheringa 5,

161-181 . ByerIy, G . R . , Lowe, D . R . & Walsh, M.M. 1986 . Stromatolites from the 3,300-3,500 Myr Swaziland Supergroup, Barberton Mountain Land, South Africa. Nature 319 , 489-

491 . Hofmann, H. J. 1976 . Precambrian microfiora, Belcher Islands, Canada: significance and systematics. Journal of Paleon­

tology 50, 1 040-1073.

A postscript on continuing microbial evolution

It is obvious that protis tan evolution did not grind to a halt at the end of the Proterozoic Eon. I t may be less obvious that continuing divers ification has als o been a characteris tic of Phanerozoic prokar­ yotes . On a broad s cale, maj or features of anaerobic metabolic divers ity were es tablis hed during the Archaean, and aerobic pathways were in place by the Early Proterozoic; however, evolving meta­ phytes and metazoans have furnis hed bacteria with a continuing s uccess ion of novel s ubs trates for metabolis m and enteric environments for coloni­ zation. Throughout Earth's his tory, rates of prokar­ yotic evolution have probably been a function of environmental evolution. F rom the pers pective of prokaryotes , then, the evolving multicellular biota can be viewed as a continually changing s eries of environments . Phanerozoic rates of bacterial evo­ lution may have been low in groups little affected by metazoan evolution, but for the many bacteria that depend directly or indirectly on metazoans , evolutionary rates were probably comparable to thos e of the animals thems elves .

Hofmann, H . ] . 1985 . Precambrian carbonaceous megafossils. In: D . F . Toomey & M.H. Nitecki (eds) Paleoalgology: contemporary research and applications, pp. 20-33. Springer-Verlag, Berlin. Hofmann, H.]. 1987. Precambrian biostratigraphy . Geoscience

Canada 14, 135-154. Jackson, M.] . , Powell, T . G . , Summons, RE. & Sweet, L P . 1986. Hydrocarbon shows and petroleum source rocks in sediments as old as 1 . 7 x 1 09 years. Nature 322, 727-729 . Knoll, A . H . 1985 . The distribution and evolution of microbial life in the Late Proterozoic era. Annual Review of Micro­ biology 39 , 391-417. Knoll, A.H., Strother, P.K. & Rossi, S . 1988 . Distribution and diagenesis of microfossils from the Lower Proterozoic Duck Creek Dolomite, Western Australia. Precambrian

Research 38, 257-279 . Schopf, J . W . (ed . ) 1983. Earth's earliest biosphere: its origin and evolution. Princeton University Press, Princeton. Schopf, J . W . & Packer, B . M . 1987. Early Archean ( 3.3 billion to 3 . 5 bilIion-year-old) microfossils from Warrawoona Group, Australia. Science 237, 70-73. Tyler, S. & Barghoorn, E . 5 . 1954. Occurrence of structurally preserved plants in pre-Cambrian rocks of the Canadian Shield. Science 119, 606-608 . Vidal, G. & Knoll, A . H . 1983. Proterozoic plankton. Memoir of

the Geological Society of America 161, 265-277. WaIter, M . R . (ed. ) 1976 . Stromatolites. Elsevier, Amsterdam . Woese, C R . 1984. Why study evolutionary relationships among bacteria? In: K.H. Schleifer & E. Stackebrandt (eds) Evolution of prokaryotes. FEMS Symposium 29, pp . 1-30. Academic Press, London .

1.3 Precamb rian Metazoans M. A. FEDONKIN

Although palaeontology as a science began more than 200 y ears ago, the first descriptions of Pre­ cambrian animals appeared relatively late, only in the first half of this century . This is explained by the rarity of Precambrian animal fossils. This rarity is due to the absence of mineralized skeletons and possibly because of a low biomass of metazoans in late Precambrian ecosy stems.

developing. The name most commonly used for the terminal Precambrian sy stem is the Vendian. Its ty pe area is the Russian Platform. In the upper half of this sy stem most of the soft- bodied fauna dis­ appears, though some trace fossils continue up to the top of the Vendian along with abundant Vendo­ taenian algae (Metaphy ta) and acritarchs (Sokolov & Ivanovski 1985 ).

Distribution in time and space

Origin of metazoans

Remains of the Precambrian fauna are now known from Australia, Africa, America, Europe, and Asia. The most representative localities are in the Nama Group, Namibia, the Pound Subgroup, South Australia, the Charnian Subgroup, U. K. , the Concepti on Group, Southeastern Newfoundland, the Valdai Series of Podolia, Ukraine, on the Onega Peninsula, and in the Khorbusuonka Series in Northern Yakutia (Glaessner 1984 ; Fedonkin 1987). Several thousand specimens assigned to more than 100 speci es have been found thus far in Pre­ cambrian deposits. Imprints and moulds of soft­ bodi ed ani mals are mainly preserved in terrigenous strata accumulated in marine shallow water en­ vironments. Less ty pically , fossils come from deeper water deposits, in turbidite and carbonate sediments. Unique taphonomic conditions in the Pre­ cambrian, due to a combination of special biotic and abiotic factors, resulted in excellent preservati on of non-skeletal i nvertebrates revealing fine details of their anatomy , i . e. external morphology and, in some cases, i nternal organs. The first unequivocal metazoan fossils appear strati graphically above tillites of the Laplandian (Varangerian) glaciation, which took place approxi­ mately 650 -620 Ma. The maxi mum geographical and stratigraphic distribution of Precambrian Metazoa occurs i n the lower half of the interval between these tillites and the base of the Tommotian Stage of the Lower Cambrian above. This i nterval more or less corresponds to the terminal sy stem of the Precambrian. Concepts of this stratigraphic sy s­ tem, known as the Sinian, Vendian, Ediacarian or Ediacaran accordi ng to different authors, are still

The large number and morphologi cal diversity of Metazoa in the first half of the Vendi an indicate that their phy logeneti c roots continue i nto older pre­ Vendian periods. This si tuation is indirectly sup­ ported by a comparative analy sis of amino acid sequences of globines of living i nvertebrates (Runnegar 1986 ) and by a decrease in the quantity and the diversity of stromatolites which began 1000 Ma and accelerated 700 -800 Ma (Walter and Hey s 1985 ). The possibility cannot be excluded, however, that the decline of stromatolites was promoted not only by Metazoa, which infl uenced, grazed upon and di sturbed bacterial mats and broke the stability of substrate, but by a series of glaciations whi ch took place 850 ± 50 , 740 ± 20 and 650 ± 20 Ma. Little i s known about this stage of Metazoan evolution, but it seems likely that the oldest animal communi ties, including Vendi an ones, were characterized by relatively low di versity i n compari son with Cambrian life. If diversity is consi dered to be a peculi ar mechanism for maintai ning the stabili ty of the biosphere, then the low diversi ty Precambrian biota was rather vulnerable to external abioti c fac­ tors as well as biotic innovations. A low diversity Precambrian fauna could not match the stability of later Metazoan communiti es. The possibility cannot be excluded that as soon as multicellular animals appeared, their communi ties were subjected to radical change, including mass extinctions as they approached the Phanerozoic level of differentiation. General characteristics

Late Precambrian animals have a wi de geographical

17

18

1 Major Events in the History of Life

distributio n with many identical fo rms o ccurring at distant lo calities. This indicates co smo po litanism, weak pro vincialism, and evidently lo w rates o f evo ­ lutio n after a rapid adaptive radiatio n at the begin­ ning o f the Early Vendian transgressio n. Altho ugh the systematics o f Precambrian animals are still pro blematic, o bvio us features include a co nsiderable diversity o f life fo rm and bo dy plan, a pro no unced do minatio n by Co elenterata, a lo w ratio o f the number o f species to that o f phyla, large size (even gigantism in many species, especially amo ng the mo st primitive o rganisms), the presence o f all majo r eco lo gical gro ups, co ncentratio n in shallo w marine enviro nments, a lo w activity o f vagile pre­ dato rs and scavengers, a relatively small bio mass o f infauna in benthic co mmunities, an eco lo gical o rganizatio n into sho rt tro phic chains, an abun­ dance o f suspensio n feeders and detritivo ro us animals, and an absence o f active filter feeders. Systematics

The traditio nal appro ach to the systematic po sitio n o f Precambrian invertebrates is based o n co mpari­ so n with yo unger Palaeo zo ic and even Recent animals (see also Sectio n 5 . 2 . 5 ). Fo r example, G laessner (1984 ) placed Precambrian animals in the fo llo wing taxa: phylum Co elenterata (classes Hydro zo a, Scypho zo a, Co nulata, medusae o f uncer­ tain affinity and pro blematic Petalo namae); phylum Annelida (class Po lychaeta); phylum Arthro po da (superclass Trilo bito mo rpha o r C helicerata o f uncertain class, and superclass C rustacea: class Branchio po da); phyla Po go no pho ra, Echiurida and so me fo rms o f uncertain systematic po sitio n. The classificatio n o f the Precambrian animals within the framewo rk o f living invertebrates pro duces many co ntradictio ns. Therefo re o ther appro aches and principles o f classificatio n have been develo ped. Fo r example, an attempt o f co mparative mo rpho ­ lo gical analysis o f Vendian Radiata and Bilateria has led to different results and a new classificatio n o f the o ldest M etazo a (Fedo nkin 1987), which is o utlined belo w. Radial animals (Radiata) . The co elenterate class C yclo zo a is characterized by a co ncentric bo dy plan, a vast disc-shaped gastral cavity, and a wide distri­ butio n o f metho ds o f asexual repro ductio n. So me fo rms have simple marginal tentacles. The repro ­ ductive o rgans are no t kno wn. This class co ntai ns predo minantly sedentary fo rms, and less co mmo nly animals living at the water-air interface and in

the plankto n. The fo llo wing genera are included: Nemiana (Fig lA), Cyclomedusa, Eoporpita (Fig. lE), Kullingia, Ovatoscutum (Fig. lE), Chondroplon, Medusinites, Ediacaria (Fig. l C ), Tirasiana, Nimbia and Paliella (Fig. 2H). The class Ino rdo zo a unites medusa-like o rgan­ isms with a symmetry o f uncertain o rder, which are characterized by a higher o rganizatio n than the C yclo zo a. Vario us co mplicated systems o f gastro ­ vascular channels, the presence o f repro ductive o rgans (go nads), and the do minance o f medusae in this gro up suppo rt this po int o f view. Asexual repro ­ ductio n is no t typical. The pattern o f gro wth in these animals is unusual co mpared to that in Recent co elenterates: new radial elements (antimeres) are fo rmed freely witho ut any regularity thro ugho ut life. Thus, they increase in number and o rder o f symmetry during o nto geny witho ut restrictio n. The co mbinatio n o f co ncentric and radial symmetry indicates a phylo genetic relatio nship between the Ino rdo zo a and C yclo zo a. The Ino rdo zo a includes Hallidaya, Lorenzinites, Rugoconites, Hiemalora (Fig. 2G ), Elasenia, Evmiaksia, and Pomoria. The class Tr ilo bo zo a is characterized by an un­ usual three-rayed symmetry, which o ccurs o nly as a terato lo gical pheno meno n amo ng recent Co elenter­ ata; amo ng o ther M etazo a it is kno wn o nly as a seco ndary feature. L ike the abo ve mentio ned classes o f Precambrian Co elenterata, representatives of the Trilo bo zo a are characterized by a mo de of gro wth unusual fo r recent Co elenterata. During o nto geny, instead o f co uples o f o ppo site antimeres being fo rmed, three antimeres o r identical radial elements in multiples o f three develo ped simul­ taneo usly. The do minatio n o f medusa life fo rms, co mplicated and regular systems o f gastro vascular channels, and a stable quantity o f repro ductive o rgans, indicate a high level o f o rganizatio n co m­ parable to that o f the Scypho zo a. Ho wever, Trilo bo zo a are characterized by different gro wth and symmetry, and an absence o f a circular channel and o ral aperture. It includes Skinnera, Tribra­ chidium, Albumares (Fig. IF), and Anfesta (Fig. 2D). Conomedusites (Fig. 2F), the o nly sedentary o rgan­ ism having a rather dense co nical theca and a fo ur­ rayed symmetry, is assigned to the class Co nulata. Other meduso ids with the same symmetry are do ubtfully co mpared with scypho zo an meduso ids; these include Ichnusina, Persimedusites and Staurinidia. It is no tewo rthy that as the symmetry o f the Precambrian Co elenterata is reduced, their o rgani­ zatio n beco mes mo re co mplicated: fro m primitive,

1 .3

Fig. 1

Precambrian Metazoans

19

Vendian metazoans. A, Nemiana simplex, x 0.5. B, Ovatoscutum concentricum, x 1. C, Ediacaria flindersi, x 1. D, Charnia x 1 . E, Eoporpita medusa, x 1. F, Albumares brunsae, x 4. Specimens in A and D are from the Khatyspyt Formation, Northern Yakutia, U.5.5.R. Specimens in B, C, E and F are from the Ust - Pinega Formation, southeast of the White Sea region, U.5.5.R.

masoni,

20

1 Major Events in the History of Life

Fig. 2 Vendian metazoans. A, Onega stepanovi, x 5 . B, Dickinsonia costata, x 1. C, Mialsemia semichatovi, x 1. D, Anfesta stankovskii, x 1 . 1. E, Bomakellia kelleri, x 0 . 7. F, Conomedusites lobatus, x 1. G, Hiemalora stellaris, x 1 . H, Paliella patelliformis, x 0 . 7 . 1, Pteridinium nenoxa, x 0.7. Specimens in A - E, G, I are from the Ust-Pinega Formation, southeast of the White Sea

Region, U . S . s . R . The specimen in F is from the Mogilev - Podolsk Series, Ukraine, U . S . s . R . , and the specimen in H is from the Khatyspyt Formation, Northern Yalutia, U . s . s . R .

1 . 3 Precambrian Metazoans dominantly sedentary Cyclozoa with a high order symmetry, through more advanced Inordozoa with a radial symmetry of variable (uncertain) order, to medusoid classes with a stable symmetry and the highest organization (Trilobozoa, Conulata, Scyphozoa). This sequence may reflect the early, pre-Vendian phylogeny of Precambrian Coelenterata. Precambrian colonial organisms are shaped like feathers, combs, fans, and bushes (Ford 1958 ; Glaessner & Wade 1966 ; J enkin & Gehling 1978 ; Anderson & Conway Morris 1982 ). Most forms were fixed to soft sediment by disc-shaped or sausage­ like organs of attachment, but rare, pelagic, freely swimming colonies are also known. The degree of integration and habit of these colonies suggest as­ signm ent to the Coelenterata, but it is impossible to determine their exact systematic position without evidence of the structure of individual polyps and the nature of sclerites or spicules that may have been present in some colonies. Functional differen­ tiation of polyps is not known. The possibility that the colonial organisms are representatives of the same coelenterate classes as the solitary forms can­ not be excluded. Colonial forms include Charnia (Fig. ID), Charniodiscus, Paracharnia, Pteridinium (Fig. 21 ), Rangea, Ramellina, Vaizitsinia, and Ausia. The Petalonamae is a special group of coelenterate grade described by Pfl ug (1970 ) as a group of high taxonomic rank that gave rise to many phyla of invertebrates. Most specialists now consider the Petalonamae to be a group of different, possibly unrelated Coelenterata of uncertain systematic position. Among them is the unusual class E rniet­ tomorpha which includes 27 species and 13 genera. However, some authors consider this diversity to be a taphonomic artifact, and reduce E rniettomorpha to five genera or even to one species, Ernietta plateau­ ensis. This sedentary organism had a multi-layered, sack-shaped body and lived with the base of its body partially buried in soft sediment. Bilateral animals (Bilateria). Among bilaterally sym­ metrical Precambrian animals, very few forms have a smooth, nonsegmented body. These are usually represented by only a few or even single specimens, and their interpretation is doubtful. Two monotypic genera, Vladimissa and Platypholinia, can be com­ pared with the turbellarians (Platyhelminthes). Protechiurus is considered to be the oldest echiurid. The overwhelming maj ority of Precambri an B ilateria have features resembling segmentation or metamerism. This initially suggested comparison

21

with annelids, arthropods, and other articulates. However, some so-called 'segmented' forms have an unusual structure: semisegments of the right and left sides alternate. This symmetry of glide reflection is not typical of younger bilaterians, but is known in the Precambrian among polymerous (consisting of numerous anatomically identical body parts) forms in the Dickinsoniidae as well as among oligomerous (consisting of few similar parts) forms in the Vend omiidae. The leaf-l ike Dickinsonia (Fig. 2B; up to 1 m body size) originally considered a coelenterate, or annelid worm or fl atworm, represents an independent branch of metazoans derived from the Radiata long before other bilaterians. This is indicated by the absence of a definite mouth and anus, an imperfect position of numerous semisegments, and relics of radial symmetry in early ontogeny. Dickinsoniidae could represent a separate class Dipleurozoa in the primitive phylum Proarticulata. The fam ily Ven­ domiidae also probably belongs to this phylum. These animals had a small, elongate discoidal body with a broadly arcuate anterior margin; a wide cephalic area is followed by a small number of segments or alternating sem isegments. The distal ends of the (semi)segments do not always reach the lateral m argins of the ovate fl at body. This family tentatively embraces Vendomia, Onega (Fig. 2A), Praecambridium, and Vendia . True segmented animals resembling annelids and arthropods did live in the Vendian oceans, and some of them can be compared to later Palaeozoic counterparts. For exam ple, Parvancorina has a shield-like, rather soft carapace with a faint mar­ ginal rim and elevated anterolateral and median smooth dorsal ridges. Approximately five pairs of stout anterior appendages are followed by up to twenty pairs of posterior fine appendages. The simi­ larity of Parvancorina to the Palaeozoic arthropods of the Marrellom orpha may indicate that it is close to the ancestors of Crustacea (Glaessner 1984 ). A rather unusual body plan is characteristic of the family Sprigginidae, which includes Spriggina and Marywadea. These animals, generally interpreted as annelids, have a horseshoe- shaped or half-moon­ shaped prostomium that resembles the head shield of primitive trilobites. The body segments, how­ ever, resemble those of rather pri mitive annelids. The same combination of a large head and a rather smooth body with long feather-like lateral append­ ages occurs in Bomakellia (Fig. 2E) and Mialsemia (Fig. 2C) - both united in the family B omakellidae. These animals seem to have had a ri gid carapace.

22

1

Major Events in the History of Life

Their body plan does not correspond to that of any group of living invertebrates. Recently it was sug­ geste d that both the Sprigginidae and Bomakellidae should b e assigned to the special class Paratrilobita, related to the phylum Arthropoda. Vendian - Cambrian evolutionary transition

One of the anomalies in the Precambrian Vendian fauna is an absence of evident ancestors of the important Cambrian invertebrate groups, including Archaeocyatha, Mollusca, Brachiopoda, and Echino­ dermata, all of which appear early in the Lower Cambrian as discrete phyletic lines. The low species diversity a nd prevalence of monotypic genera may indicate a relatively short interval between the rise of these invertebrate groups and their acquisition of the ability to build a skeleton. Skeletalization developed gradually during the Vendian (Section 1 .4 ). The first half of the period saw the appearance of Redkinia spinosa, an annelid­ like animal with chitinoid, comb-like jaws. Chiti­ noid tubes of sabelliditids appear at the same level, as well as the calcareous tubular fossils Cloudina. The end of the Vendian saw a wide distribution of tubular shells, sclerites and conodont-like fo rms. The small sizes and wide geographical distribution of the oldest shelly fossils could indicate that their Precambrian ancestors had small body sizes and a planktic mode of life. Trace fossils show that the majority of the vagile benthos lived in shallow-water marine environ­ ments. Dominant among them were deposit feeders and forms of detritivore which collected small food particles. These animals moved by various peri­ staltic methods. Precambrian trace fossils are not as diverse or deep as later examples. The biomass of Vendian infaunal communities was much smaller even in shallow-water environments. Sedentary epi­ faunal forms of the Vendian period (i. e. mainly primitive groups of coelenterates) were dominantly passive suspension feeders and, more rarely, pre­ dators. Active suspension feeders (filter feeders) are unknown. The activity of vagile predators and scavengers was low, at least in the first half of the Vendian. Coelenterata were domi nant in the plankton and nekton. The end of the Vendian Period was a critical moment in the history of life when biological pro­ cessing of sediments increased greatly and many new groups of invertebrates began to inhab it the sea floor. The body size of infauna, represented mainly by soft-bodied animals, also increas ed

at this time. All these phenomena, as well as the formation of a skeleton in other groups, may be adaptive and reflect increasing predation by vagile animals. Burrowing and the formation of skeletons had extremely important biological and evolution­ ary consequences that are not yet entirely understood by palaeontologists and zoologists. Recently Seilacher (1984 ) offered a new morpho­ logical a nd functional in terpretation of some Precambrian animals. Having noted that the Vendian fauna shows no close affinity with later invertebrates, he inferred that Precambrian organ­ isms do not have Recent analogues and have a unique organization. They are characterized by an extensive body surface, which has developed mainly because of their very complicated relief, and a low body volume by virtue of being relatively flat. The high surface-volume ratio of the body allowed the absorption of oxygen and organic matter dis­ solved in water by diffusion through the body sur­ face. Thus, neither a mouth and digestive organs nor respiratory organs were necessary. No less attractive is an older point of view, that the body of many Precambrian animals was favour­ able for harbouring photosynthesizing endosym­ bionts. This is supported by the leaf-like form of the body of many Vendian organisms, their occurrence in shallow water marine environments within the photic zone, and the large size of many of the most primitive form s. A certain correlation between the presence of algae-endosymbionts and large body size is noted, for example, in recent Cnidaria. The gigcm tism of many Precambrian inve rtebrates is especially striking when compared to the first very small shelly fossils which appear at the end of the Vendian and become numerous in the Tommotian Stage of the Lower Cambrian (Section 1 . 5 ). The larger body size of the Vendian Metazoa may reflect an adaptation of prey animals to in­ creasing predation pressure. The first half of the Vendian was characterized by rapid speciation under the conditions of the vast postglacial trans­ gression of the sea. The fauna rapidly reached its characteristic diversity, and rates of phyletic evo­ lution decreased. This is reflected in the large sizes of populations and the absence of provincialism in many groups. The middle of the Vendian saw a mass extinction of many groups (Section 2 . 13 . 1 ), especially those primitive animals which were characterized by a passive mode of feeding. One possible reason for extinction was the appearance of many small an­ cestors of Cambrian invertebrates, which had better

1 .3 Precambrian Metazoans developed modes of feeding and could considerably impoverish food resources in the pelagic zone. The passive feeding of many Vendian sedentary forms was relatively inefficient and may have led to their extinction. The collection of detritus from the surface of the sediment also became less effective. These circumstances, as well as the increasing population densities and growing predation, could direct natural selection to favour forms that began active colonization of bottom sediment with its new trophic peculiarities. The ecological niche of Vendian sedentary Coelenterata in the shallow marine environment became occupied by active suspension feeders (sponges, archaeocyathids, brachiopods) in the Early Cambrian (see Section 1 . 6 ). Possibly in parallel with the extinction of some groups, there was a decrease in body size in others in the second half of the Vendian. This could explain the sharp impoverishment, if not a gap in the fossil record, of invertebrates of the late Vendian. The decrease of body size may have led to the oligo­ merization of many primitive polymerous forms. This in its turn could have resulted in an increase in the level of organization and/or even in the specialization of some forms. From the middle of the Vendian, the increasing activity of predators and scavengers and the de­ structive activity of burrowin g organisms and perhaps the meiofauna inhibited the preservation of soft-bodied forms. Additionally, bioturbation led to more rapid biological oxidation of soft tissues of buried animals. When comparing the world of the Vendian with that of the Cambrian we are comparing two different categories of fossils. This makes it difficult to analyse the early evolution of invertebrates but to some extent explains the apparent absence of phylo­ genetic connections between the faunas of these two periods. The analysis of body plans of Vendian soft-bodied invertebrates reveals some previously unknown directions of morphological evolution in the Metazoa. The great abundance of Radiata in the Vendian refle cts the predominance of radially­ symmetrical animals of coelenterate grade in the early history of metazoans. The high diversity of symmetries reflects an early radiation of this phylum. The development of more complicated morphologies (i. e. the appearance of more complex systems of gastrovascular channels, reproductive organs, etc. ) while symmetry was reduced suggests an evolution from forms with a symmetry of infinitely

23

high order, through forms with an uncertain multi­ rayed symmetry, to forms with a stable order of symmetry. In the course of coelenterate evolut ion the archaic concentric body plan was replaced essentially by a radial one. The dominance of segmented form s am ong Vendian Bilateria possibly reflects a rel at ionship between processes leading to bilatera l sym metry and to metamerism in the phylogeny of early Metazoa. Bu t these processes did not alwa ys lead to coelomates. Unusual peculiarities of constru ctional morphology (from a neontologica l perspective), for example the plane of symmetry of glide reflection in some of the most primitive Vendian bilaterians, may indicate the early origin of bilateral quasi­ segmented forms from rather archaic R ad ia ta with an axis of symmetry of infinitely high or u ncertain order. The existence of a large quantity of short-live d phylogenetic branches in the Precambrian em pha­ sizes the importance of comparative-m orphological analysis at the Vendian chronological level in order t o discover major directions in the early evolut ion of multicellular animals.

References Anderson, M.M. & Conway Morris, S. 1982. A review, with description of four unusual forms, of the soft-bodied fauna of the Conception and St. John's Group (Late Pre­ cambrian), Avalon Peninsula, Newfoundland. Proceedings

of the Third North American Paleontological Convention 1, 1-8. Fedonkin, M.A. 1987. The non-skeletal fauna of the Vendian and its place in the evolution of metazoans. Nauka, Moscow (in Russian). Ford, T.D. 1958. Precambrian fossils from Charnwood Forest.

Proceedings of the Yorkshire Geological Society 31, 211-217. Glaessner, M.F. 1984. The dawn of animal life. A biohistorical study. Cambridge University Press, Cambridge. Glaessner, M.F. & Wade, M. 1966. The Late Precambrian fossils from Ediacara, South Australia. Palaeontology 9,

599-628. Harrington, M.J. & Moore, R.e. 1956. Dipleurozoa. In : R.e. Moore (ed.) Treatise on invertebrate paleontology. Part F: Coelenterata, pp. 24-27. Geological Society of America, Boulder and University of Kansas Press, Lawrence. Jenkin, R.J.F. & Gehling, J.s. 1 978. A review of the frond-like fossils of the Ediacara assemblage. Records of the South

Australian Museum 17, 347-359. Pflug, H.D. 1970. Zur Fauna der Nama-Schichten in Siidwest Africa. I. Pteridinia, Bau und systematische Zugehorigkeit.

Palaeontographica A134, 226-262. Runnegar, B. 1986. Molecular palaeontology. Palaeontology

29, 1-24. Seilacher, A. 1984 Late Precambrian and Early Cambrian

24

1 Major Events in the History of Life

Metazoa: preservational or real extinctions? In: H . D . Holland & A . F . Trendall (eds) Patterns of change in Earth evolution, pp. 159-168 . Springer-Verlag, Berlin . Sokolov, B . 5 . & Ivanovski, A . B . (eds) 1985 . Vendian System.

Historical- geological and paleontological substantiation. Paleontology, Vol. 1. Nauka, Moscow (in Russian) . Waiter, M . R . & Heys, C . R . 1985. Links between the rise of the Metazoa and the decline of stromatolites. Precambrian

Research 19, 149-174.

1.4 Origin of Hard Parts - Early Skeletal Fossils B . RUNNEGAR & S . BENGTSON

Introduction

Hard parts of organisms appeared almost instan­ taneously in the fossil record at the transition from the Proterozoic to the Phanerozoic. Biomineraliz­ ation (Section 4 . 4 ) may have evolved close in time to that event. Earlier records of biogenic minerals are spurious and involve either v ery small, isolated crystals (magnetite of possible bacterial origin) or carbonate encrustation of cyanobacterial sheaths that may have been induced indirectly by the photosynthetic activities of the organism. The earliest records of hard parts involve all major skeletal materials - calcite, magnesian calcite, aragonite, apatite, and opal. (About 40 minerals are known to be formed by modern organisms (Low enstam & Weiner 1983 ), but many of them are unstable under normal diagenetic conditions and they seldom form structures large or distinct enough to be recognized in the fossil record. ) All major types of skeletons are present - spicules, stiffened walls, shells, sclerites, and physiologically dynamic endoskeletons. The Early Phanerozoic skeleton­ forming biotas (Fig. 1 ) represent practically all major taxa of multicellular organisms known to produce mineralized skeletons today, some groups of biomineralizing protists, and a number of extinct groups of organisms, mostly metazoans (see also Section 5 . 2 . 5 ). The original mineralogy of the various groups of Late Precambrian and Cambrian fossils is not always well known. There are comparatively few studies on the diagenesis of early skeletal fossils. The com­ position of the sk eleton in most groups is only known from their gross mineralogy in various types of rock, or inferentially through comparisons with known related taxa. More detailed information has been derived from petrographic and geochemical studies of fossils and surrounding rocks (e. g. J ames

& Klappa 1983 ), and from studies of replicated crys­

tal morphologies (Runnegar 1985 ). This has been done in only a few cases, however, and further studies are needed. Carbonate fossils

Calcium carbonates, mainly calcite, magnesian cal­ cite, and aragonite, are the most common skeleton­ forming minerals today, and appear to have been dominant already among the first skeletal fossils. Whereas aragonite is unstable in diagenesis and is rarely preserved in the fossi l record, calcite and magnesian calcite may preserve their orig­ inal crystallographic structure given favourable circumstances. The tubular fossil Cloudina (see also Sections 1 . 3 , 5 . 2 . 5 ) is often considered to be the earliest known example of a mineralized skeleton, but its strati­ graphic position is somewhat uncertain, and it is not clear that it significantly predates the earliest more diverse assemblage of skeletal fossils. The tubular skeleton of Cloudina consists of stacked im­ bricating calcareous half-rings, suggesting that it was constructed by a secreting gland of an animal that was able to twist around in its tube. The wall was probably partly organic, stiffened by calcium carbonate impregnations. Other early carbonate tube-building animals include the anabaritids, first occurring in the c . 550 Ma Nemakit-Daldyn assemblage (see Fig. 1 ). Anabaritids attained a wide distribution before their disappearance in the Atdabanian. They were triradially symmetrical - an unusual feature sug­ gesting a possible phylogenetic relationship with triradial metazoans of the Ediacaran fauna - and appear to have been less mobile in their tubes than Cloudina. The original mineralogy of the tubes is

1 .4

25

Origin of Hard Parts

450

500

C-O

sso

PreC-C

Ma

� Calc i u m carbonate � Calcite � Aragon i te

� I 1 IQ22l D

Cal c i u m p h o s p h ate Opal i n e s i l ica Agg l u t i n ated skeleton Tentative range

Fig. 1

Temporal distribution of cIades of biomineralizing and agglutinating organisms in the Late Precambrian to Late Ordovician, compiled from various sources. Precambrian-Cambrian boundary (PreC- C) arbitrarily placed at the appearance of the Protohertzina-Anabarites assemblage and assigned an age of 550 Ma (see also Section 5. 10. 2). Clades defined as groups of taxa that appear to derive their biomineralizing habit from a common ancestor. (A few probably polyphyletic groups, such as 'calcareous tubes', have been retained due to their poorly known phylogeny.)

not known, but apparently ubiquitous recrystal li­ zation suggests that they may hav e been formed of aragonite. The succeeding Cambrian faunas c ontain more div erse types of tubular fossils. Some were cylin­ drical, resembling, for example, protectiv e struc­ tures built by certain modern annelids. Others, in particular the widespread and diverse hyoliths (see also Section 5 . 2 . 5 ), had more obtuse tubes and w ere closed by opercula. They were bilaterally symmetri­ cal animals with a U-shaped gut. The shell mineral was most probably aragonite, and a structure re­ sembling mo lluscan crossed-lame ll ar fabr ic has been observ ed in younger Palaeo zoic members of this group. Aragonitic shells are c haracte ris tic of early mol­ luscs (Runnegar 1985 ). The most pri mitiv e shell structure in Cambrian molluscs seems to hav e con­ sisted of a single layer of spherulitic aragonite prisms beneath an organic periostracum. This ty pe of structure may grow in an inorganic manner, and

the shape of the spherul itic 'prisms' is mould ed by surface forces rather than chemical bonds. These kinds of mineral deposits need not hav e been me­ dia ted by a protein substrate. Nacre ous lining s in prismatic sh ells had appeared by at least the Middle Cambrian and may hav e been present in Earl y Ca mbria n time. The fundamental differenc e between the ara gonitic fibres of spheru­ litic 'prisms' and the flat aragonitic tablets of nacre lies in the difference in the hab it of crystals; in nacre, growth on the (001 ) face is v ery slow, whereas in the fibres it is v ery fast. The result is a layered micr ostructure (nacre) which i s much stronger than fibrous aragonite. M ost of the common molluscan ultrastructures had ev olv ed by the Middle Cambrian. In addition to sp herulitic prismat ic aragonite an d nacre, these included tangentially arranged fibrous aragonite, crossed-lamellar aragonite, and foliated calcite. Various solitary and colonial animals among the earliest skeletal biotas built basal skeletons of

26

1

Major Events in the History of Life

calcium carbonate. Most of these are poorly known. The cup-shaped hydroconozoans and the probably colonial Bija and Labyrinthus may only questionably be referred to the cnidarians (Jell 1983 ). Others, such as Tabulaconus and Cothonion, have been studied in more detail and show certain similarities with corals, but their affinities nevertheless remain in doubt. Undoubted skeleton-forming cnidarians are not known until the Ordovician. The basic structural units in rugose and tabulate coral skel­ etons were spherulitic tufts (trabeculae) formed by fibrous calcite. Modem scleractinian corals form similar structures of aragonite fibres. As with the spherulitic 'prisms' of mollusc shells, the process of formation appears to involve little matrix-mediated control of crystal shape. However, nucleation of the fibrous trabeculae may be under more direct biochemical control. The sponge-like archaeocyathans constructed a supporting skeleton typically shaped like a double­ walled perforated cup. They are preserved as micro­ granular calcite, interpreted as representing original magnesian calcite (James & Klappa 1983 ). Calcium carbonate (aragonite or calcite) skeletons are also formed by several groups of sponges ('sclero­ sponges' and 'sphinctozoans') from the Middle Cambrian until the Recent (Vacelet 1985 ). The more common type of sponge mineralization is, however, the spicular skeleton (see below). All the skeleton types described above exhibit incremental growth, which occurs by addition of material to earlier formed growth stages. This type of growth puts strong geometrical constraints on morphology. Ways of avoiding this problem are : (1 ) periodical moulting of exoskeleton; or (2 ) con­ tinuous construction and destruction of the mineral phase by intimately associated living tissue. Trilobites, common in Cambrian rocks from the Atdabanian (c. 540 Ma; Fig. 1), are an example of animals that periodically moulted their exo­ skeletons . These were of calcitic composition and often show well-preserved crystallographic fabrics in their mineralized cuticle. Other examples are the coeloscleritophorans, uniquely Cambrian organisms with a complex exoskeleton consisting of hollow carbonate sclerites with a basal opening. Their orig­ inal mineralogy has not been definitely established, but the ubiquitous recrystallization and occasion­ ally preserved fibrous structure suggest that they were aragonitic. Echinoderms, first appearing in the Atdabanian and undergoing their first substantive radiation in the Middle Cambrian, developed a calcium

carbonate endoskeleton in which there was close interaction of mineral and living tissue. Modem echinoderms construct their skeletons of a mesh­ work (stereom) of almost pure magnesian calcite, in which each individual skeletal component is part of a large single crystal. All fossil echinoderms, in­ cluding the Cambrian ones, appear to have had an identical structure. Spicules - mineralized elements formed within living tissues - are widely distributed among Recent organisms. Spicules of magnesian calcite are characteristic of calcareous sponges and octocorals. In both groups the spicules are formed by special­ ized sclerocytes, sometimes originating intracellu­ larly and only later erupting from the cell membrane to be further enlarged by enveloping sclerocytes. Sponge spicules grow in crystallographic continuity, so that each spicule behaves optically as a single crystal of calcite. By contrast, octocoral spicules typically are composed of smaller acicular crystals. As the echinoderm plates, sponge and octocoral spicules are made of magnesian calcite, it has been suggested that magnesium is used to shape the crystals by selectively poisoning appropriate parts of the lattice (O'Neill 1981 ). Calcitic sponge spic­ ules have been found in the late Atdabanian (c. 535 Ma, Fig. 1 ), and possible octocoral spicules also appear in beds of the same age. Undoubted spi­ cules of octocorals are known from the Silurian. The fossil sponge and octocoral spicules have the same crystallographic properties as their modem counterparts. Although fossil spicules of various origins are common, they are rarely dealt with in scientific literature because they tend to be disarticulated and therefore difficult to identify taxonomically. Some spicular skeletons may fuse to form frameworks, as in hexactinellids, 'lithistid' demo sponges, and 'pharetronid' calcareous sponges, or the axial skel­ etons of pennatulacean and a few alcyonarian octo­ corals. Such structures are rare in the early history of these groups. Fossils resembling calcified cyanobacteria became common in the Early Cambrian. One group of such organisms, the helically coiled filamentous Obru­ chevella, is present as uncalcified filaments in rocks of Vendian age, but is frequently calcified after the beginning of the Cambrian. Calcified cyanobacteria have their mucilagenous sheaths impregnated with crystals, perhaps as a by-product of the photosyn­ thetic removal of CO2 from the water in which they lived (Riding 1977 ). Fossils that may be true cal­ carous algae occur in the c. 550 Ma Nemakit - Daldyn

1 .4

Origin of Hard Parts

beds of the northern Siberian Platform . More con­ vincing examples are first known from the Middle Cambrian. Phosphatic fossils

As a skeleton-forming mineral, apatite occurs today only in inarticulate brachiopods and vertebrates . Some recent organisms are also known t o produce amorphous calcium phosphate that may be crystal­ lized later into apatite . Among the earliest skeletal organisms, however, calcium phosphate appears to have been more widespread . Tubular fossils of phosphatic composition are a common constituent of Cambrian faunas . Most of them are referred to as hyolithelminths . The fine structure of hyolithelminth tubes has not been suf­ ficiently studied, but they appear to have grown incrementally by addition of lamellae . At least in some forms a systematic change in the orientation of fibrous elements in adjacent lamellae occurs, pro­ ducing a force-resistant structure similar to that of arthropod cuticles . The phosphatic tubes of the paiutiids had longitudinal septum-like structures on the inner surfaces . Conulariids had distinctly four-faceted cones built up of transverse phosphatic rods set in a flexible integument. Phosphatic conchs or shells were also widespread . In addition to phosphatic inarticulate brachiopods, there are also a number of problematic phosphatic shells, such as Mobergella and related fossils, char­ acterized by regularly placed paired muscle scars and a usually flattened shape . The brachiopods include a number of phosphate- and carbonate­ shelled clades, many of which were short-lived . One characteristic and diverse Cambrian group is the tommotiids - multisclerite-bearing animals presumably covered with more or less twisted coni­ cal sclerites built up of phosphatic growth lamellae . They vary in skeletal organization from the ir­ regularly shaped and frequently fused sclerites of Eccentrotheca to the highly organized scleritomes of Camenella and Tannuolina, in which each of the two asymmetric sclerite types had its mirror-image counterpart . Examples of periodically moulted exoskeletons of calcium phosphate are rare, but the valves of the ostracode-like bradoriids are commonly preserved as phosphate . Although some of them appear to have been flexible, they were most probably im­ pregnated to varying degrees with apatite crystal­ lites. Like most arthropod skeletons, they did not grow by accretion, but were periodically shed.

27

Whether or not the ecdysis involved resorption of mineral matter is not known, but resorption may explain the common occurrence of collapsed or buckled valves. The problematic fossil Microdictyon formed plate­ like structures with a more or less regularly hexag­ onal network of holes and intervening nodes. They were constructed of two or three distinct layers of apatite and show no evidence of incremental growth . Vertebrates, similar to echinoderms, have a plastic mode of skeleton formation as a result of a constant physiological exchange between mineral­ ized and cellular tissues . The phosphatic bone of vertebrates is intimately associated with fibrillar collagen, which does not seem to be the case in other phosphatic skeletons . Although undoubted vertebrate remains are not known until the Ordovician, certain Cambrian phosphatic fossils show a fine structure suggesting association with fibrous organic matter that may be homologous with vertebrate collagen . The small button-shaped sclerites o f the utah­ phosphans consist of a thin dense apatite layer covering a porous core; the latter has fine tubules or fibrils perpendicular to the lower surface . The 'buttons' are more or less densely set in an integu­ ment that is impregnated with smaller apatitic crys­ tallites. The tooth-shaped conodonts had a fibrous organic matrix in which the apatite crystallites were embedded (Szaniawski 1987) . In both these cases, a chordate affinity has been proposed using partly independent lines of evidence . Other suggested biomineralizing chordates (Palaeobotryllus, Ana to­ lepis) are even more problematic in their inter­ pretation. There are further examples of exclusively Cambrian fossils of phosphatic composition and unknown systematic affinity. Some of these are spine- or tooth-shaped objects, possibly reflecting the fact that apatite is a hard mineral suitable for the construction of wear-resistant structures . Siliceous fossils

Because of its non-crystalline, isotropic nature and intracellular method of formation, opal (a mineral gel consisting of packed spheres of hydrated silica) has had limited potential as a skeletal material ex­ cept in very small organisms . It is most widespread among protists . The only metazoans known to form it are hexactinellid sponges and demosponges, which use it for spicule formation . Most biogenic

28

1 Major Events in the History of Life

opal formed today is either dissolved in the water column before it is incorporated in the sediment or dissolved during early diagenesis, but under certain circumstances opaline skeletons may be preserved, usually as microcrystalline quartz or replacements by other minerals. The distribution of opal among the earliest skel­ etal fossils differs significantly from that of calcium carbonates and phosphates . Only four groups of silica-producing organisms are known from the time period under consideration (Fig . 1), hexacti­ nellids, demosponges, radiolarians, and chryso­ phytes(?) . All appeared during the Early Cambrian and all are still living . Whether this apparent im­ mortality of opal-producing lineages is a chance effect due to the small number of clades involved, or whether it has a more profound meaning, the pat­ tern differs considerably from that seen in the car­ bonate and phosphatic groups . In the latter two, the Cambrian radiation appears to have produced a large number of taxa of which only a few survived . Early history of skeletal biomineralization

Present knowledge of the fossil record confirms that mineralized skeletons of many different kinds and composition appeared very rapidly in a number of clades at the beginning of the Phanerozoic . Analysis of the precise pattern is still difficult, because in many cases the original mineralogy is insufficiently known and the taxonomic understanding of the various enigmatic early skeletal fossils is incomplete (see also Section 5 . 2 . 5 ). It is therefore difficult to know how many clades developed the ability to form mineral skeletons at this time . It seems clear, however, that this ability evolved independently a number of times. A current and widely held view is that those organisms that used phosphate rather than car­ bonate or silica were the first to diversify. Phos­ phate has been stated to be the dominant or even exclusive mineral of the earliest skeletal faunas . A phosphate - carbonate transition is said to have oc­ curred within clades such as the Ostracoda, Brachio­ poda, and Cnidaria, but also by the replacement through extinction of organisms with phosphatic skeletons by organisms with carbonate hard parts . Aragonitic materials are also postulated to have replaced calcitic ones throughout the remainder of the Phanerozoic. Available data, including the pattern of distri­ bution of clades of different biomineralizing habits through time (Figs 1, 2 ) and the phylogeny within

30 .-----�

'" QJ "'Cl '"

U

DO c

20

.!::!

�

QJ C

P h os p h ate s k e l etons

E tu

10

.D

E

:::J Z

S i l ica skeletons

O +-----��---.----r_--._--� 580

560

540

520

500

480

M i l l io n s of years ago Fig. 2 Cumulative curves of appearance of clades presumed to have independently evolved a biomineralizing habit. Based on the same data as Fig. 1 .

these clades, do not appear to support such views . 1 The relative amount of phosphate versus car­ bonate bound in biominerals in the Cambrian has been exaggerated by sampling biases (most early skeletal fossils are of millimetre size, and chemical extraction of microfossils is more likely to destroy carbonates than phosphates) and unrecognized cases of secondary phosphatization (the Cambrian was a time of extensive deposition of phosphatic sediments) . 2 Whereas phosphate skeletons were certainly more widely distributed among different clades in the Early Cambrian than they are today, the same may be said about carbonate ones. Among the clades shown in Fig. 1, 42% of the carbonate skel­ etons survive until the present, as compared to 25% of the phosphatic ones (protoconodonts are regarded as chaetognaths with mineralized grasping spines) . Both categories include clades that are today very successful and diverse. Thus the restriction of phosphate minerals to two major clades today may simply be the result of the different evolutionary success of various early lineages. Nothing in the history of vertebrates suggests that their skeletal mineralogy puts them at an evolutionary disadvan­ tage, and there is no reason to assume that the shell mineral was the particular factor that decided the survival of each of the early lineages. 3 The quoted examples of phylogenetic transition from phosphate to carbonate, or from aragonite to

1 .4

Origin of Hard Parts

calcite, appear to be suspect. For example, a sug­ gested evolutionary succession from phosphate to carbonate hard parts within the cnidarians depends upon the dubious taxonomic decision to place the extinct conulariids within the Cnidaria. The pro­ posed secondary origin of carbonate brachiopods from phosphate ones and the derivation of carbon­ ate ostracodes from pre-existing phosphate forms have the me rit of linking groups that are clearly closely related, but the proposal of a mineralogical transition is nevertheless weakly founded. In neither case has a strict phylogenetic analysis been able to demonstrate that the carbonate forms are in fact derived from the phosphate ones. The Early Phanerozoic radiation cannot be seen j ust as a radiation of biomineralizing taxa. The trace fossil record shows a similar rapid diversification of burrowing habits in non-biomineralizing organ­ isms, and the appearance at the same time of resist­ ant organic structures and agglutinating tubular fossils shows that the key event is not biomineral­ ization as such (see also S ection 1 .5) . To a certain extent, the appearance of mineralized skeletons may be seen as one of many aspects of the early radiation of multicellular organisms. Nevertheless, the apparent absence of biominerals in the Edia­ caran fauna and the nearly simultaneous 'skeletal­ ization' of cyanobacteria (notwithstanding reports of earlier sporadic cases of mineralized cyanobac­ terial sheaths), algae, heterotrophic protists (fora­ miniferans and radiolarians), and metazoans, seems to call for specific explanations. Attempts to explain the appearance of skeletons have often foundered on lack of universality. For example, models involving calcium availability or regulation do not explain the simultaneous appear­ ance of opaline skeletons, and the proposal that biomineralization began as a phosphate-excreting process at a time of high phosphate availability is not consistent with the pattern of appearance of v arious biominerals as discussed above. Models based on increasing P02 may have more explanatory power, as an increasing availability of oxygen would have made it easier for organisms to form skeletal minerals and proteins, and made outer mineralized skeletons less of a respiratory disad-

29

vantage. (There is a general but not perfect corre­ lation between distribution of mineralized skeletons and oxygen levels in modern marine faunas. ) A synecologically based explanation is that bio­ mineralization in animals and plants primarily arose in response to selection pressu res induced by grazers and predators. No evidence of grazers or predators is known from the Ediacaran fauna, whereas the first probable macrophagous predators (protoconodonts) appear with the first diverse skel­ etal biotas. Al though the various types of skeletons in the early Phanerozoic biota often had complex functions, most of them would have had the advan­ tage of at least passively deterring predators or grazers. S uch an explanation stresses the view of the early evolution of skeletons as a complex event, integrated with other aspects of the rapid biotic diversification at this period. It is not in conflict with physiologically and geochemically based models explaining how biomineralization became possible in the first place. References James, N . P . & Klappa, C . F . 1983. Petrogenesis of Early Cambrian reef limestones, Labrador, Canada. Journal of

Sedimentary Petrology 53, 1051-1096 . Jell, J . 5 . 1983. Cambrian cnidarians with mineralized skel­ etons . Palaeontographica Americana 54, 105 - 1 09 . Lowenstam, H . A . & Weiner, S. 1983. Mineralization b y or­ ganisms and the evolution of biomineralization . In: P . Westbroek & E . W . d e Jong (eds) Biomineralization and biological metal accumulation, pp . 191-203. Reidel, Dordrecht. O'Neill, P . L . 1981 . Polycrystalline echinoderm calcite and its fracture mechanics. Science 213, 646-648. Riding, R. 1977. Calcified Plectonema (blue-green algae), a Recent example of Girvanella from Aldabra Atoll. Palae­

ontology 20, 33- 46 . Runnegar, B . 1985 . Shell microstructure o f Cambrian molluscs replicated by calcite. Alcheringa 9, 245-257. Szaniawski, H. 1987. Preliminary structural comparisons of protoconodont, paraconodont, and euconodont elements. In: R . J . Aldridge (ed. ) Palaeobiology of conodonts, pp. 3547. Ellis Horwood, Chichester. Vacelet, J. 1985 . Coralline sponges and the evolution of the Porifera. In : S. Conway Morris, J . D . George, R. Gibson & H . M . Platt (eds) The origins and relationships of lower invertebrates. Systematics Association Special Volume 28, pp . 1-13. Oxford University Press, Oxford .

1.5 Late Precamb rian - Early Cambrian Metazoan Diversification S . C O N WA Y M O R R I S

Introduction

on pre-Ediacaran metazoans, much of it ques­ tionable . The second section then addresses the outlines of the adaptive radiation that is marked by the Ediacaran faunas and the succeeding Cambrian biotas .

Life on this planet is customarily divided into six kingdoms, the prokaryotic archaebacteria and eubacteria, and the four eukaryotic kingdoms of protoctistans, fungi, plants, and metazoans. Because the multicellular metazoans had their origins in unicellular eukaryotic ancestors, in principle the identification of such an organism in the fossil record would constrain the time of appearance of the metazoans . However, even the recognition of the first eukaryotes has proved problematic . It has been customary to regard eukaryotes as being de­ rived from prokaryotes, and given the profound differences between the two cell types such a dis­ tinction might seem to be readily identifiable in the fossil record. However, even these critical characters (e . g . presence of nucleus, cell wall composition) fail to survive fossilization, and the only guide is rela­ tive cell size . Thus, the search for the earliest eukaryotes has concentrated on evidence for either relatively large unicells (see also Section 1 .2) or, better, a more com­ plex multicellular organism, perhaps even with differentiated tissues . In terms of the former cri­ terion, the appearance of large cells in sediments dated at approximately 1300 - 1400 Ma is generally taken as the first reliable indication of eukaryotes . In similar aged strata, fossils composed o f large carbonaceous films probably represent multicellular protoctistans, perhaps brown algae . Nevertheless, given the overlap in cell diameters between eukar­ yotes and prokaryotes, it is not impossible that some cellular remains from yet older sediments are eukaryotes masquerading as prokaryotes . Given these problems, i t i s necessary to review first the generally agreed bench-marks leading to the appearance of metazoans . The earliest definitive metazoans are taken as the Ediacaran faunas (Glaessner 1984) that span the interval c. 550- 620 Ma . Allowing for considerable uncertainties the earliest eukaryotes may be as old as 1600 Ma, allowing a possible 1000 Ma for the development of metazoans. This article, therefore, is divided into two sections . the first reviews such slender evidence as is available

Pre-Ediacaran metazoans

The most compelling pre-Ediacaran evidence would be soft-bodied remains. Recently, structures inter­ preted as worms (Sun et al . 1986) have been reported from Northern China (Anhui and Liaoning prov­ inces) . In overall form some of these carbonaceous structures, known as Sinosabellidites and Pro to­ arenicola are very similar to a sausage-like mega­ scopic Precambrian alga known as Tawuia, but they differ in possessing fine annulations. Another sup­ posed worm, referred to as Pararenicola, also pos­ sesses annulations, but is somewhat smaller and stouter than Sinosabellidites . Nevertheless, their identification as metazoans is otherwise equivocal, not least because neither internal structures, such as a gut trace, nor cephalization are recognizable . In particular, claims for a so-called proboscis in Para­ renicola and Protoarenicola are dubious . Moreover, the quoted dates of between 850 and 740 Ma are based on questionable radiometric determinations and correlations with other regions in China, and the pre-Ediacaran status of these fossils is still open to doubt. With regard to trace fossils from pre-Ediacaran strata, there are numerous claims, but few have won acceptance . Supposed metazoan traces from the Medicine Peak Quartzite of Wyoming (Kauffman & Steidtmann 1981), dated at c. 2000 2400 Ma are remarkable in view of the current consensus that the seas were colonized by nothing higher than cyanobacterial mats . Another widely quoted example is a possible feeding trace (Brook­ sella canyonensis) from the Grand Canyon . This is ostensibly from the 1 100- 1300 Ma old Grand Canyon Series, but renewed searches appear to have been unsuccessful . While other specimens from a wide variety of localities provide a seemingly 30

1 .5

Metazoan Diversification

from a wide variety of localities provide a seemingly impressive roster of evidence for trace fossils, in nearly every case unresolved doubts remain . Even if some pre-Ediacaran traces prove genuine, their general scarcity is difficult to explain unless extrinsic factors (e . g . oxygen levels) prohibited the wide­ scale expansion of macroscopic metazoans into an effectively empty ecospace . While these relatively large trace fossils continue, therefore, to excite scepticism, it may be that more convincing evidence could be found at a micro­ scopic level . For example, possible faecal pellets have been reported from the c. 900 Ma Zilmerdak 'Series' of the Urals (Glaessner 1984), which, if con­ firmed, which, would indicate a grade of organiz­ ation above that of the turbellarians . Clearly, a more extensive survey in suitable lithologies is required . In particular, ultrastructural studies of sediments may show features diagnostic of bioturbation. For example, documentation of grain orientation and cation concentrations (e . g . iron, aluminium) around undoubted Phanerozoic trace fossils suggests a possible approach to establishing the biogenicity of some Proterozoic examples (Harding & Risk 1986) . Moreover, cherts that evidently formed at a very early stage of diagenesis, from the c. 700 Ma Doushantuo Formation in the Yangtze Gorges of Hubei Province, China, preserve narrow burrowlike structures that may represent the activities of a meiofauna. While the Precambrian fossil record is dominated by stromatolites, it has long been realized that they undergo a decline in diversity during the late Pre­ cambrian (see Fig . 2) . A recent reanalysis of the data (WaIter & Heys 1985) indicates that, in terms of both relative abundance and diversity, stromatolites began to decline in quiet, subtidal environments (where coniform varieties were especially abundant) from about 1000 Ma. This trend was established also in intertidal environments from c. 800 Ma, so that stromatolites were relatively unimportant by the beginning of the Cambrian. The traditional explanation links this pattern to the rise of grazing metazoans whose activities were detrimental to the formation of the microbial mats . Thus, the initial dip in stromatolite diversity at 1000 Ma may herald the rise of primitive grazers, while the accelerating process of decline after c. 800 Ma could represent the widespread distribution of metazoans . How­ ever, the development of disrupted stromatolitic fabrics (a thrombolitic texture) that may be a result of extensive burrowing by metazoans, only appears in the Cambrian. Further indirect evidence for the evolution of

31

metazoans at least one billion years ago comes from molecular studies . If it is demonstrated that the substitution of either nucleotides in nucleic acid chains or amino acids in polypeptides is stochasti­ cally constant and occurs at a known rate, then the differences between the sequences in any species pair should indicate their time of divergence . Using this assumption of the so-called molecular clock, existing data on haemoglobins (a group with a substitution rate that is appropriate for the time­ scales involved) have been used to suggest that the metazoans evolved between c. 800 and 1000 Ma. A related approach utilizes SS ribosomal RNA se­ quences, and such comparisons (Hori & Osawa 1987) suggest that Mesozoa might be the most primi­ tive metazoans (if they are not derived indepen­ dently from protoctistans) . Moreover, although the Mesozoa may have arisen before 1000 Ma, other metazoans such as the turbellarians and nematodes have divergence points at only c. 700 Ma . Ediacaran faunas

The evidence for pre-Ediacaran metazoans is mounting, but the view that the fossil record indi­ cates no metazoan older than c. 600 Ma is still respectable and it is the Ediacaran faunas that provide our first useful glimpse of metazoan evol­ ution (Glaessner 1984; Conway Morris 1985; see also Section 1 .3). Such faunas were described from Namibia, at that time Deutsch Sud-West Afrika, before the Second World War, and shortly after­ wards in Australia. At first regarded as Cambrian, their persistent occurrence beneath abundant shelly fossils soon led them to be consigned to the Pre­ cambrian, and continuing reports from numerous localities around the world have confirmed this observation. Until recently these faunas have been dated at c. 620- 680 Ma, with some claims of even 800 Ma. However, recent radiometric dating has cast major doubt on these estimates. High resolution uranium - lead dating of zircons from an ash fall that entombed an Ediacaran assemblage in South­ east Newfoundland (Fig. 1) yields a date of c. 565 Ma . Even so, the age range of the Ediacaran faunas may be considerable, and a span of c. 550- 620 Ma may not be unrealistic. The Ediacaran faunas are reviewed elsewhere (Section 1 . 3), and only a general survey in the present context is required . At present, there seem to be two broad assemblages. There are those of a shallow­ water type that are superbly represented in, the Flinders Ranges of South Australia, including the

32

1 Major Events in the History of Life

Fig. l

Ediacaran fossils from the Mistaken Point Formation (Conception Group) of Southeast Newfoundland, Avalon Peninsula. A, Pennatuloid with hold-fast. B, Pectinate organism. C, Stellate organism. D, Bedding plane with spindle organisms and medusoids. E, Branching organism with hold-fast. F, Pennatuloid with hold-fast and medusoid. Diameter of coin 23 mm.

Ediacaran Hills, and the closely similar fauna from the White Sea of northern U . S . 5 . R . In contrast, the faunas of the A valon Peninsula of southeast New­ foundland, which may be referred to as the Mistaken Point assemblage (Fig . 1), in reference to the spec­ tacular locality near Cape Race, appear to represent a deeper-water facies . Similar occurrences in Charnwood Forest, U.K. are one of the many lines of evidence that in the early Phanerozoic this area was joined to the Avalon area on one side of the Iapetus ocean . Possibly deeper-water faunas may

also occur in the Flinders Range, but as yet only preliminary reports are available . Despite the range of environments inhabited by these Ediacaran assemblages, they show several characters in common . Most typical are forms that the majority of workers would ascribe to the cnidarians . These include medusoids (Fig. ID, F), some of which may be placed with reasonable con­ fidence in cubomedusoids, chrondrophores, and perhaps scyphozoans . However, other jellyfish have a highly characteristic three-fold symmetry

1 .5

Metazoan Diversification

that finds no parallel amongst living cnidarians. Yet others lack sufficient characters to be assigned with confidence to any group . In addition, stalked forms with an expanded leaf-like body arising from a central rachis (Fig . lA, F) invite comparisons with the pennatulaceans (the sea-pens) . However, these similarities become increasingly tenuous amongst a variety of other foliate to bag-shaped organisms, and their cnidarian affinity is more questionable . Other organisms include a possible worm, the sheet­ like Dickinsonia, a medley of arthropod-like forms, and a possible echinoderm with penta-radial symmetry (Gehling 1987) . Although the Mistaken Point assemblages evi­ dently owe their preservation in most instances to being overwhelmed by volcanic ash, in many other cases the occurrences of these soft-bodied meta­ zoans as sandstone impressions above silts tone in­ tervals are difficult to explain, given the absence of such preservation in younger clastics . The problem is compounded by abundant trace fossils in some Ediacaran assemblages, most typically simple sinu­ ous trails, that cannot be linked to the activities of any of the known body fossils . It seems necessary to invoke a contrast between entirely soft-bodied organisms, such as the trace-producing worms that were possibly largely infaunal, and those with a tougher integument, many either epifaunal or pelagic and coming to rest on the sea-floor prior to burial . It was only this latter group that was suf­ ficiently tough to take impressions when immured by sediment. However, the explanation has not won universal approval. In a sweeping reappraisal Seilacher (1989) proposed that the Ediacaran organ­ isms represent an entirely separate group, possibly a distinct kingdom, that owe their preservation to a unique composition consisting of a sac-like body with a tough integument. While Seilacher has high­ lighted the taphonomic problems posed by this preservation, his ingenious proposal seems to be oversimplified and, while perhaps applicable to some of the sac-like ernietiids (see also Section 1 . 3) and Pteridinium, is difficult to reconcile with the bulk of the biota . Whatever disagreements surround the biological affinities of the Ediacaran fauna, it is clear that they lacked hard skeletal material, the widespread ap­ pearance of which was to usher in the Cambrian some 20 Ma later . However, one notable exception demands comment. In Namibia carbonate units, intercalated with clastics containing Ediacaran fossils, yield calcareous tubes referred to as Cloudina (Grant 1990) (see also Section 1 .4) . The tubes are

33

double walled with connecting partitions that give a cone-in-cone appearance, although the exact mode of secretion is not clear . There is evidence that originally the walls contained substantial amounts of organic matter, and this helps to fuel the specu­ lation that the origin of skeletal hard parts was as separate spicules or granules embedded in an organic matrix . The facies contrast in Namibia between the clastics bearing the Ediacaran fauna and the carbonates with Cloudina emphasizes the need for taphonomic judgements concerning orig­ inal faunal distribution . However, occurrence of Ediacaran faunas in dolomites in northern Siberia demonstrates that preservation is not governed simply by lithology . The role of the Ediacaran faunas in determining the origins of the Cambrian fauna at present is enigmatic. With the possible exception of the arthropod- and echinoderm-like forms, existing reports would indicate little continuity with either the shelly faunas or soft-bodied Burgess Shale-type assemblages . Descriptions of new finds from Siberia and Australia may go some way towards alleviating the problem, and it is likely that many of the putative ancestors are represented either by the unknown trace makers or animals too small to be preserved . The evident demise of the Ediacaran fauna has resulted in two alternative hypotheses that are not entirely exclusive . One appeals to a change in taph­ onomic conditions, in particular the rise of Cam­ brian predators and scavengers that militated against soft part preservation . It is, however, of considerable significance that a distinct gap separ­ ates the disappearance of the Ediacaran fauna from the debut of Cambrian assemblages, an interval that contains facies that otherwise would appear suitable for preservation (Narbonne & Hofmann 1987) . If indeed a substantial fraction of the Ediacaran fauna became extinct over a geologically brief period, then it may be that the subsequent Cambrian diversification was largely a response to the ecological opportunities presented . The evi­ dence for such an end-Ediacaran mass extinction (Section 2 . 13 . 1) at present is very tenuous . It is necessary to emphasize, however, that as yet no data point to any extra-terrestrial mechanism. If comparisons were to be drawn with other mass extinctions, then there are possible similarities with the end-Permian event (Section 2. 13.4) in which the formation of a super-continent and possibly devel­ opment of brackish oceans because of massive evaporite deposition are invoked as significant factors .

34

1 Major Events in the History of Life

Cambrian biotas

Whether or not there was an end-Ediacaran mass extinction, the ensuing diversifications of the Cambrian were a spectacular evolutionary event (Brasier 1979; Conway Morris 1987) . Most obvious is the appearance of abundant skeletal parts (Section 1 .4) composed of calcium carbonate, calcium phos­ phate or silica, which together provide for the first time in the history of Earth an adequate fossil record. Given that the bulk of the fossil record consists of shelly fossils, it is not surprising that the many ex­ planations offered for the Cambrian diversifications have focused on the origin of hard parts. While special explanations may be called for, soft-bodied organisms may have outnumbered greatly those with skeletons in the original Cambrian communi­ ties and the history of diversification of trace fossils during this interval is also an important component in documenting these adaptive radiations. Although the rise of the skeletal faunas is clear in outline, detailed resolution is hampered by uncertainties regarding inter-continental correlations, such that the exact sequence of events is still uncertain. Present evidence, however, suggests that (apart from Cloudina) the earliest skeletal fossils included anabaritids (elongated tubes with a highly charac­ teristic trifoliate cross-section) and the teeth of protoconodonts, a group probably related to the modem chaetognaths (arrow-worms). Shortly after­ wards they are joined by more shelly fossils, in­ cluding a distinctive monoplacophoran known as Purella, the gastropod Aldanella, and primitive hyo­ liths. The succeeding horizons record an abundance of additional shelly fossils (Bengtson 1977), many of enigmatic affinities (see also Section 5.2 .5) but also including additional monoplacophorans, the first gastropods, hyoliths, brachiopods, sponges, and, somewhat later, echinoderms. The majority of these fossils are relatively small (c. 1 - 2 mm), and are either composed of phosphate or are replaced sec­ ondarily by this compound. These small shelly fossils (see also Section 1 .4) are the subject of active study, with special interest in the more enigmatic taxa (Bengtson 1977) . Although for many species biological relation­ ships are entirely speculative, in others a natural classification is beginning to emerge. Three import­ ant groups include the tommotiids, which possessed a primary phosphatic skeleton, the coelosclerito­ phorans which comprise halkieriids, siphogo­ nuchitids, and chancelloriids, and the cambroclaves,

the last two having calcareous skeletons. In each group the skeleton is composite, being composed of a series of sclerites that disarticulated on death. This extraordinary array of small shelly fossils per­ sists during the early stages of the Cambrian, es­ pecially the Tommotian and Atdabanian, with some lingering into the Middle and even Upper Cam­ brian. It is noteworthy that the trilobites, which dominate the majority of Cambrian shelly faunas, are absent from the earliest assemblages. However, their appearance in different sections was probably not synchronous, and their debut was probably due to mineralization on pre-existing forms with only a chitinous skeleton, rather than an evolutionary event per se. The rise of these skeletal faunas has been inter­ preted in both ecological terms, especially the rise of predators conferring the need for protective structure, and in terms of changes in the physico­ chemical environment (Conway Morris 1987; see also Section 1 .4) . The evidence that many groups possessed either tightly interlocking sclerites that probably formed a coat over the exterior, or valves that enclosed or allowed the retraction of the soft parts, certainly supports a response to predation. In some specimens, especially tubicolous taxa, small boreholes occur. They probably represent predatory activity, but the nature of the attacker is speculative. It is also likely that the protoconodonts formed part of a predatory feeding apparatus, but in general it is necessary to infer that many of the early Cambrian predators were more or less entirely soft-bodied. Examples of Lagerstatten that might reveal the nature of such soft-bodied organisms are not known until the Atdabanian, and of these the Burgess Shale-like Cheng-jiang fauna in Yunnan Province, South China is by far the most important. This fauna has not yet received detailed analysis, and much of our information on the role of soft-bodied organisms in the initial Cambrian diversifications must continue to rely on evidence from trace fossils (Crimes 1987) . A general diversification that paral­ lels the skeletal record is now well known. In par­ ticular, Vendian traces typically are rather small and two-dimensional. Some ichnotaxa survive into the Cambrian, but a number (e.g. Harlaniella) are re­ stricted to this interval and therefore have a bio­ stratigraphic utility. The striking increase in trace fossil diversity near the Precambrian - Cambrian boundary (Section 5. 10.2) includes vertical burrows, scratch marks that generally are attributed to arthro­ pods, and other traces that often indicate increas­ ing behavioural complexity. It is also striking that

1 .5

35

Metazoan Diversification

ichnotaxa regarded as diagnostic of either shallow­ or deeper-water where they occur later in the Phanerozoic, are found together in shallow-water environments (Crimes & Anderson 1985) . Indeed, it has been proposed that the deep oceans were not colonized until later in the Palaeozoic, and that the displacement of some trace makers into deeper water was a result of competitive pressure in the shallows. While the role of ecological changes has domi­ nated discussion on the evolution of early meta­ zoans, it now appears that substantial alterations in the extrinsic physico-chemical environment were also taking place during this time (Conway Morris 1987) . The extent to which such changes influenced or even controlled evolutionary events is far from clear, although the near synchronous nature of them is certainly suggestive . Changes in the physico-chemical environment

Extrinsic changes are registered in several ways, including: palaeocontinental distributions, sea­ level curves, stable isotope variations (especially of carbon and sulphur), preference for either aragonite or calcite precipitation, and phosphate deposition . While the extent and nature of the late Precambrian super-continent is still under debate, there is clear evidence for major rifting episodes close to the Precambrian - Cambrian boundary that heralded its break-up . While the separation of continental

blocks would have encouraged the development of endemic faunas, the formation of hot, spreading ridges would have led to displacement of seawater and hence a major transgression . While the history of this Cambrian transgression is not well known in detail, it had the dual effect of increasing the habit­ able area for shallow-water marine life and pro­ viding an increasingly complete rock record as the facies belts migrated cratonward (Brasier 1979) . There is also evidence for substantial changes in ocean chemistry close to the Precambrian ­ Cambrian boundary (Fig . 2) . For example, measure­ ments of sulphur isotopes ({'\345) from very late Rrecambrian evaporites record a massive positive shift (the Yudomski event) that reflects the intro­ duction of substantial amounts of isotopically heavy seawater into areas of evaporite formation by some sort of upwelling. The shift is so significant that it probably represents long-term storage of deep­ water brines, where bacterial fractionation of sul­ phur led to accumulation of increasingly 'heavy' water. The sites of such storage may have been narrow 'Mediterranean-like' basins formed at an early stage of continental breakup, and the up­ welling episode may also be linked to continuing evolution of the basins. It is probably no coincidence that the Yudomski event overlaps with a major episode of phosphogenesis, that is now reflected in huge economic reserves of phosphate in China and elsewhere . It has been speculated that the influx of phosphorous raised levels of productivity and

Metazoan d ivers ificat i o n (fam i l ies)

Fig. 2

Changes in ocean chemistry as registered in O BC and 0345, inferred sea-level, and diversity of metazoans and stromatolites during intervals of the Riphean, Vendian and Cambrian. (Data for stromatolites from Waiter & Heys 1985; other data sources listed in Conway Morris 1987.)

Ri h ean 1 000

950

900

850

800

750

700

650

600

550

500 ( 1 06 yrs)

1 Major Events in the History of Life

36

helped to fuel the evolutionary radiations. Infer­ ences on ocean productivity have also been drawn on the basis of changes in carbon isotopic ratios (613q, which show a series of substantial shifts. However, in some instances storage of organic mat­ 2 ter (rich in photosynthetically sequestered 1 q, such as in anoxic basins, may be invoked as an explanation and could be linked to the formation and destruction of narrow marine basins alluded to above. Although somewhat less constrained in terms of timing, there is also evidence for a shift in inorganic precipitation (e.g. ooids) of calcium car­ bonate polymorphs, from aragonite in the late Pre­ cambrian to calcite in the Cambrian. The reasons for this shift are complex, but stem from processes of plate tectonics. These include hydrothermal metamorphism at spreading ridges that lower the Mg: Ca ratio of seawater, rise of partial pressure of CO2 by volcanic exhalations, and deposition of carbonates in shallow seas versus their weathering on exposed continents. Taken together, the shift towards calcite precipitation appears to be con­ trolled in part by continental breakup, growth of spreading ridges and subduction zones, and transgression of continental margins. Just how important extrinsic factors, most of which seem to stem ultimately from the processes of plate tectonics, were in controlling evolutionary events is still uncertain. Metazoan diversification may have had its roots far back in the Riphean but, as yet, the possible influence of extrinsic factors on biological evolution in this interval is largely specu­ lative. Nevertheless, the rise of skeletons near the Precambrian-Cambrian boundary can be linked with slightly more confidence to changes in ocean chemistry, and it is interesting that similar suggest­ ions have also been made in connection with skeletal evolution during the great Permo-Trias faunal turn­ over. Some workers have even suggested that en­ vironmental factors may have led to sequential mineralization, from aragonite to high magnesium calcite to phosphate to low magnesium calcite (Brasier

1986). The complexity of the processes and

the paucity of evidence in several critical areas, however, make this a challenging area for future palaeobiological research.

References Bengtson, S . 1977. Aspects of problematic fossils in the early Palaeozoic. Acta Universitatis Upsaliensis 415, 1 - 71 . Brasier, M . D . 1979 . The Cambrian radiation event. In : M.R. House (ed . ) The origin of major invertebrate groups. System­ atics Association Special Volume 12, pp. 103- 159. Academic Press, London. Brasier, M . D . 1986 . Precambrian-Cambrian boundary biotas and events. In: O. Walliser (ed .), Global bio-events . Lecture Notes in Earth Sciences No . 8, pp . 109 - 1 17. Springer­ Verlag, Berlin. Conway Morris, S. 1985 . The Ediacaran biota and early metazoan evolution. Geological Magazine 122, 77- 8 1 . Conway Morris, S . 1987. The search for the Precambrian­ Cambrian boundary. American Scientist 75, 156 - 1 67. Crimes, T.P. 1987. Trace fossils and correlation of late Pre­ cambrian and early Cambrian strata. Geological Magazine 124 , 97- 1 1 9 . Crimes, T.P. & Anderson, M.M. 1985 . Trace fossils from later Precambrian-early Cambrian strata and environmental implications. Journal of Paleontology 59, 310-343. Gehling, J . G . 1987. Earliest known echinoderm - a new Ediacaran fossil from the Pound Subgroup of South Australia. Alcheringa 11, 337-345 . Glaessner, M . F . 1984. The dawn of animal life. A biohistorical study. Cambridge University Press, Cambridge. Grant, S.W.F. 1990. Shell structure and distribution of Cloudina, a potential index fossil for the terminal Protero­ zoic. American Journal of Science 290A, 261 - 294. Harding, S . c . & Risk, M.J. 1986. Grain orientation and elec­ tron microprobe analyses of selected Phanerozoic trace fossil margins, with a possible Proterozoic example. Journal of Sedimentary Petrology 56, 684 - 696. Hori, H . & Osawa, S . 1987. Origin and evolution of organ­ isms as deduced from 5S ribosomal RNA sequences . Molecular Biology and Evolution 4 , 445-472. Kauffman, E . G . & Steidtmann, J.R. 1981 . Are these the oldest metazoan trace fossils? Journal of Paleontology 55, 923947. Narbonne, G . M . & Hofmann, H.J. 1987. Ediacaran biota of the Wemecke Mountains, Yukon, Canada. Palaeontology 30, 647- 676. Seilacher, A. 1984. Late Precambrian and early Cambrian Metazoa: preservational or real extinctions? In: H . D . Holland & A.F. Trendall (eds) Patterns of change i n Earth evolution, pp. 159 - 168 . Springer-Verlag, Berlin. Seilacher, A. 1989 . Vendozoa: organismic construction in the Proterozoic biosphere. Lethaia 22, 229 -239 . Sun Wei-guo, Wang Gui-xiang & Zhou Ben-he 1986. Macro­ scopic worm-like body fossils from the upper Precambrian (900- 700 Ma), Huainan district, Anhui, China and their stratigraphic and evolutionary significance . Precambrian Research 31, 377- 403. Waiter, M.R. & Heys, G . R . 1985 . Links between the rise of the Metazoa and the decline of stromatolites . Precambrian Research 29, 149 - 1 74.

1.6 Evolutionary Faunas J. J.

SE PKO SKI ,

Jr

Evolutionary faunas are sets of higher taxa (es­

Phanerozoic and the total number of classes has

pecially classes) that have similar histories of diver­

remained virtually constant since.

sification and together dominate the biota for an

The expansion of each evolutionary fauna is as­

extended interval of geological time. The expansion

sociated with the decline of the previously dominant

and decline of evolutionary faunas can be used to

fauna. The declines are much slower than the initial

describe large-scale variations in faunal dominance

diversifications, giving the faunas very asymme­

and to interpret temporal changes in global taxo­

trical histories. Such a pattern is difficult to simulate

nomic diversity in the fossil record. The concept

in 'random' models of diversification (Sepkoski

was introduced by Sepkoski

(1981),

1981)

who identified

but can be described with coupled logistic

equations of the form

three 'great evolutionary faunas' in the Phanerozoic marine record. These faunas were defined statisti­

dDJ d t

cally in a factor analysis of familial diversity within

Dj

=

rjDj (1 - L D/ Dj),

taxonomic classes, which grouped together classes

where

that attained their maximum diversities around the

is the diversity of the ith evolutionary

t, rj is its initial diversification Dj is its maximum or 'equilibrium' diversity, and LDj is the summed diversity of all faunas at time t (Sepkoski 1984; Kitchell & Carr in Valentine fauna at time

same time. The analysis permitted the histories of

rate,

the aggregate faunas to be traced from initial diver­ sification through dominance and into decline. This treatment of the faunas as units throughout their

1985) .

histories distinguishes the concept of evolutionary

fauna will diversify and replace the preceding fauna

faunas from that of 'dynasties',

only if its initial diversification rate is lower and

used by some

This equation states that an evolutionary

rj

authors for assemblages of dominant taxa during

equilibrium diversity is higher. If

specified intervals of geological time.

evolutionary fauna will expand so rapidly that the

is higher, the

preceding fauna will never appear to diversify; if

Dj

is lower, the evolutionary fauna will never be able

Marine evolutionary faunas

to expand and replace the preceding one. Thus, the

Characteristics .

The three evolutionary faunas iden­

bility in the sequential diversification of evolution­

tified in the marine fossil record are the Cambrian

ary faunas, although it does not specify their timing

Fauna, important during the Cambrian Period, the

or relative differences in maximum diversity.

coupled logistic equation suggests a certain inevita­

Palaeozoic Fauna, dominant from Ordovician to

Classes within evolutionary faunas tend to have

Permian, and the Modern, or Mesozoic-Cenozoic

similar mean rates of taxonomic turnover. Classes

Fauna, dominant in the post-Palaeozoic (Fig. lA).

in the Cambrian Fauna tended to have high turnover

The classes in each fauna share a number of charac­

rates, those in the Palaeozoic Fauna intermediate

teristics, or central tendencies, suggesting that they

rates, and those in the Modern Fauna comparatively

are not randomly assembled groups of taxa. The

low rates (with some exceptions in all cases). These

most striking characteristic is that the classes tend

differences are reflected in the responses of the

to diversify together, each successive fauna dis­

faunas to mass extinctions (Sepkoski

playing a slower rate of diversification but higher

Cambrian Fauna suffered large proportional re­

1984) :

the

level of maximum diversity than those preceding

ductions in diversity relative to the Palaeozoic fauna

it. These properties lead to a sequential expansion

during mass extinctions in the Ashgillian and

of evolutionary faunas and a resultant step-like

Frasnian, and the Palaeozoic Fauna suffered more

pattern of increase in global marine diversity (with

than the Modern at all major mass extinctions of the

the step between the Palaeozoic and Modern faunas

Phanerozoic. This differential reaction seems to

disrupted by the massive Late Permian extinction

have led to the great change in faunal dominance

event - Section 2 . 13.4) . This pattern is present even

associated with the Late Permian mass extinction

though most marine classes originated early in the

(Section

37

2 . 13.4) .

1 Major Events in the History of Life

38 900

.'!::

E � '0 '"

E

600

�

Q) .D

:::J Z

300

1 Diversity curves . A, Marine animal families. B, Terrestrial vascular plant species. C, Terrestrial tetrapod families . Each curve is divided into fields that illustrate the diversities of the constituent evolutionary faunas and floras . A, After Sepkoski (1984) ; Cm Cambrian evolutionary fauna, pz Palaeozoic fauna, Md Modern fauna; stippled field represents known diversity of families with rarely preserved members that lack heavily mineralized skeletons . B, After Niklas et al. (1983); numbered fields as in text. C, After Benton (1985); numbered fields as in text. Fig

600

400

200

600

·E � '0 '"

Q)

'"

u

Q)

'0

Q) Cl.

'"

0

400

E 200

200

=

=

�

�

Q) .D

Q) .D

:::J z

:::J Z

E

0

100

0 400

200

0

=

D

400

Geological time (106 yrs)

Evolutionary faunas also seem to have differing ecological characteristics . The Cambrian Fauna tended to be assembled into broadly intergrading communities that were dominated by generalized deposit feeders and grazers and had low epifaunal and infaunal tiering (Bottjer & Ausich 1986; see also Section 1 . 7. 1 ) . Communities of the Palaeozoic Fauna were dominated by epifaunal suspension feeders with complex tiering; many other ecological guilds were also represented so that the fauna as a whole seems to have occupied more 'ecospace' than the Cambrian Fauna (Bambach in Tevesz & McCall 1983) . Finally, the Modern Fauna is represented by yet more guilds and is characterized by large numbers of durophagous predators (Vermeij 1987) and mo­ bile deep infauna (Thayer in Tevesz & McCall 1983); epifaunal tiering is reduced in most communities . Sepkoski and Miller in Valentine (1985) demon-

200

0

strated that evolutionary faunas tended to form coherent assemblages within shelf environments throughout the Palaeozoic Era . Members of the Cambrian Fauna were spread across the entire shelf early in the Palaeozoic Era but became progressively restricted to deeper-water environments during the Ordovician as members of the Palaeozoic Fauna expanded across the middle and finally outer shelf. At the same time, early members of the Modern Fauna came to dominate inner shelf environments and later, deeper, low-oxygen environments . The Late Permian mass extinction eliminated dominance of the Palaeozoic Fauna from middle and outer shelf environments and led to expansion of the Modern Fauna across the entire shelf. It must be emphasized that none of these evo­ lutionary and ecological differences between the faunas is absolute . In a sense, the faunas are 'fuzzy

39

1 . 6 Evolutionary Faunas bounded sets' with their characteristics overlapping

2.

and some members of each fauna mimicking mem­

bites along with inarticulate brachiopods, mono­

bers of others. The characteristics thus represent

placophorans, hyoliths, and eocrinoids; most of the

nodes on a continuum. Major unsolved problems

problematical taxa of the so-called 'small shelly

are why such nodes should exist and why they

faunas' of the Tommotian are also included. Various

The Cambrian Fauna was dominated b y trilo­

of these classes are paraphyletic, with descendent

seem to change so little through the Phanerozoic.

monophyletic taxa belonging to other evolutionary

Composition and history.

The individual histories of

faunas; however, in most cases the paraphyletic

the marine evolutionary faunas are illustrated in Fig.

classes either declined long before their descendent

CAM B R I A N F A U N A

I n a rt i c u lata T r i l o b ita

Hyo lit h a

�

M o n o p lacop h o ra

' ,.

200E �

Eocr i n oidea

(J) �

������--���T� O

� Z

PA LAEOZO I C FAU NA

A rt i c u l ata A n t h ozoa

C e p h a l opoda

S t e n o l a e m ata

Ste l l e r o i d a

400 200 '0

O s t racoda

C ri no idea

M O DE R N FA U NA

�� Repti l i a

Osteic h t hyes

Biva lvia

600 Malacostraca

C h o n d ric h thyes

E 400 � '0

G a s t ropoda

2 Histories of the three great evolutionary faunas of the marine fossil record as represented by their familial diversities through the Phanerozoic. Representatives of the important classes in each fauna are illustrated above the diversity curves. (After Sepkoski 1984.) Fig

o

200 Gym n o l ae mata

600

400 G eo l o g i c a l t i m e

E

.D

D e m o s p o n g i a R h izopodea Ech i n o i d e a

200 (106 yrs)

o

i

40

1 Major Events in the History of Life

taxa diversified (e .g. the Monoplacophora) or con­ tained monophyletic subtaxa that diversified in parallel with the rest of the evolutionary fauna (e .g. the Inarticulata) . The Cambrian Fauna diversified very rapidly from the latest Vendian into the Early Cambrian and was the principal constituent of the 'evolutionary explosion' across the Precambrian ­ Cambrian Boundary (see also Section 1 .5) . Its maxi­ mum diversity was attained in the late Middle and early Late Cambrian . Beginning in the latest Cambrian, the fauna began a long, gradual decline, accentuated by the Ashgillian and Frasnian mass extinctions (Sections 2 . 13.2, 2 . 13.3) . The Palaeozoic Fauna initiated its expansion as the Cambrian Fauna began to decline; this combi­ nation resulted in nearly stable global diversity throughout the Late Cambrian. The Palaeozoic Fauna was dominated by articulate brachiopods with important contributions from crinoids, corals, ostracodes, cephalopods, and stenolaemate bryo­ zoans . These groups were major components of the Ordovician radiations, which tripled global taxo­ nomic diversity over a 50 million year interval . The Palaeozoic Fauna attained its maximum diver­ sity from the Late Ordovician to Devonian and then began a long decline . During the Carboniferous and Permian, this decline was matched by a slow ex­ pansion of the Modem Fauna so that again global diversity remained nearly constant . The Palaeozoic Fauna was severely reduced by the Late Permian mass extinction (Section 2. 13.4) but in the Mesozoic underwent two radiations : one in the Triassic, ter­ minated by the Norian mass extinction (2 . 13 .5), and a second, slower expansion in the Jurassic . The Jurassic expansion was reversed in the Cretaceous when global diversity exceeded Palaeozoic levels, and the remnants of the Palaeozoic fauna again went into decline . The Modem Fauna is dominated by gastropod and bivalve molluscs, osteichthyan and chond­ richthyan fishes, gymnolaemate bryozoans, mala­ costracans, and echinoids . Most of these classes appeared during the Cambrian and Ordovician Periods but diversified only slowly through the Palaeozoic Era . They suffered minor extinction rela­ tive to the Palaeozoic fauna during the Late Perrnian and became the dominant fauna in the Triassic. Through the Mesozoic and Cenozoic, the Modem Fauna continued the rather slow and steady diver­ sification initiated earlier, producing the long post­ Palaeozoic increase in global taxonomic diversity. Throughout their histories, the three 'great' evo­ lutionary faunas experienced considerable internal

turnover, with continuous change in ordinal and lower-level taxonomic composition . This was par­ ticularly true of the Cambrian Fauna, which under­ went very rapid changes during its initial radiation. It may prove useful to subdivide this fauna and define two more evolutionary faunas : an Ediacaran Fauna, encompassing the distinctive soft-bodied animal fossils of the Vendi an (Sections 1 . 3, 1 .5), and a Tommotian Fauna, comprising the mostly prob­ lematical skeletal taxa of the earliest Cambrian (Sec­ tions 1 .4, 1 .5, 5.2 .5) . These possible faunas seem to fit into the general progression of evolutionary rates and diversity levels observed for the three great evolutionary faunas. The Ediacaran and especially Tommotian taxa appear to have had higher diversi­ fication rates and more rapid evolutionary turnover than the remainder of the Cambrian Fauna, and seem to show successive increases in diversity lead­ ing into the Cambrian Period . Further analysis of diversity patterns and faunal change in the Vendian and Early Cambrian are needed to assess whether such additional evolutionary faunas are useful for describing the early metazoan radiation .

Terrestrial biotas The concept of evolutionary faunas has proved use­ ful for organizing faunal turnover and diversity change in the marine record and has been extended with varying success to other evolutionary systems, specifically terrestrial vascular plants and tetrapod vertebrates . Niklas et al. (1983) identified four major plant groups, which can be termed evolutionary floras, in species-level data on tracheophyte diversity (Fig. lE). These are: (1) an initial Silurian-Devonian flora of early vascular plants that radiated and then disappeared during the Devonian; (2) a pteridophyte­ dominated flora, including ferns, lycopods, sphenop­ sids, and progymnosperms, that diversified in the Late Devonian and Early Carboniferous and domi­ nated plant communities to the end of the Palaeozoic Era; (3) a gymnosperm-dominated flora of seed plants that appeared in the Late Devonian and rose to domi­ nance in the Mesozoic; and (4) an angiosperm flora that originated in the Early Cretaceous and rapidly radiated to dominance thereafter, replacing the pre­ ceding gymnosperm flora . As in the marine system, each of these floras (excepting the angiosperms) originated early in the history of vascular plants and radiated sequentially to produce step-like increases in global tracheophyte diversity. Three 'assemblages' of terrestrial tetrapod families

1 . 7 Diversification of Marine Habitats have been identified by Benton (1985) in the ver­ tebrate fossil record (Fig . Iq. These comprise : (1) the labyrinthodonts, anaspids, and synapsids, which appeared during the Middle Palaeozoic and completely dominated the terrestrial vertebrate record to the end of the Palaeozoic; (2) the early diapsids, dinosaurs, and pterosaurs, which arose in the Triassic, attained maximum diversity in the Late Jurassic and Cretaceous, and disappeared at the terminal Cretaceous mass extinction (Section 2 . 1 3 . 7); and (3) the lissamphibians, turtles, croco­ diles, lizards, birds, and mammals, which originated in the Triassic and Jurassic, expanded through the Cretaceous, and then diversified to very high levels in the Cenozoic. Although these assemblages have some similarities to marine evolutionary faunas, there are some important differences : the assem­ blages do not all appear early in the history of tetrapods and their sequential diversifications are not all associated with step-like increases in global diversity . It remains to be seen whether such pat­ terns could be identified if more terrestrial taxa (e . g . the arthropods) were included and analyses per­ formed at lower taxonomic levels . If so, evolutionary

41

faunas and floras would appear to be a general property of the development of Phanerozoic biotas .

References Benton, M.J. 1985. Patterns in the diversification of Mesozoic non-marine tetrapods and problems in historical diver­ sity analysis . Special Papers in Palaeontology 33, 185 - 202 . Bottjer, Dj. & Ausich, W.!. 1986. Phanerozoic develop­ ment of tiering in soft substrata suspension-feeding com­ munities . Paleobiology 12, 400-420. Niklas, K.J . , Tiffney, B . H . & Knoll, A.H. 1983. Patterns of vascular land plant diversification . Nature 303, 614- 616. Sepkoski, J.J., Jr. 1981 . A factor analytic description of the Phanerozoic marine fossil record . Paleobiology 7, 36-53. Sepkoski, J . J . , Jr. 1984. A kinetic model of Phanerozoic taxo­ nomic diversity. III . Post-Paleozoic families and mass ex­ tinctions. Paleobiology 10, 246-267. Tevesz, M .J . 5 . & McCall, P . L . (eds) 1983. Biotic interactions in recent and fossil benthic communities. Plenum Press, New York. Valentine, J.W. (ed . ) 1986. Phanerozoic diversity patterns: pro­ files in macroevolution . American Association for the Advancement of Science and Princeton University Press, Princeton. Vermeij, G.J. 1987. Evolution and escalation. An ecological history of life. Princeton University Press, Princeton .

1 . 7 Early Diversification of Major Marine Habitats

1 . 7. 1 Infauna and Epifauna

w. I . A U S I C H & D . J . B O T T J E R

Introduction Benthic marine habitats and the organisms that populate them represent an intricate and diverse ensemble . Much of the initial development and diversification of metazoans was for life in this realm. Marine benthos have invaded most types of substratum at depths ranging from the supertidal to abyssal . This array of habitats, with concomitant physical and chemical limiting factors, has probably been relatively constant through most of the Phan-

erozoic . Similarly the general trophic strategies for exploitation of marine benthic habitats has been constant. Both infaunal and epifaunal organisms developed, including suspension feeders, deposit feeders, predators, scavengers, grazers, and others . However, through eustatic changes in sea-level and plate motion in the lithosphere, the habitat location has been constantly changing. The great diversity in this benthic system is con­ tributed by organisms . At any one time organisms

42

1 Major Events in the History of Life

differentially adapt to a plethora of physical, chemi­ cal and biological limiting factors . The development of simple to complex ecological structuring within habitats is variable; and, of course, through evo­ lution and extinction, the organisms populating benthic habitats have been continually in flux .

The benthic habitat

The infauna. In modern environments particulate organic material is abundant immediately above and below the sediment - water interface and de­ creases in quantity both up into the water column and down into the sediment (Fig. 1 ) . Both suspen­ sion feeders and deposit feeders exploit this re­ source . Infaunal deposit feeders mine particulate organics within the sediment, whereas infaunal suspension feeders typically feed from the water that is immediately above . The primary physical constraints on depth of burrowing are the position of the redox boundary, and sediment stiffness, which increase with depth . Phylogenetic constraints on the development of specialized structures (e . g . fused siphons) have also been important i n the history of the infauna . Infaunal suspension feeders are largely stationary . They all feed as active suspension feeders from water immediately above the sediment surface, and particulate food in that water moves past them horizontally . In contrast, infaunal deposit feeders are mobile, and they feed on a stationary food

Vl Vl «

�

� f-«

u.J

�

13:I

Vl 0 .D

'"

<J.l 0 .D .S '" � - <J.l ..c-

�3 <J.l E I

<J.l

Ic

"0 <J.l <J'

Relat ive c u rre n t ve l o c i ty

T

B e n t h i c hyd rodyna m i c b o u n d a ry layer

Sed i m e n t­ water i n terface Fig. 2 Velocity profile of the benthic hydrodynamic boundary layer.

I

sion feeding benthos on soft substrata (Bottjer & Ausich 1986) .

Tiering. Spatial separation and structuring is a com­ mon biological method of resource partitioning within communities . Vertical community structure has been documented for epifaunal and infaunal suspension feeding communities (Bottjer & Ausich 1986) and infaunal deposit feeding communities (Levinton & Bambach 1975) . Bottjer & Ausich (1986) called this spatial arrangement of organisms tiering . They developed a history of tiering complexity through the Phanerozoic for suspension feeding palaeocommunities in soft substrata, deposited in subtidal shelf and epicontinental sea settings at depths greater than several metres below fair­ weather wave base (Fig . 3) . A comprehensive Phanerozoic history of tiering for deposit feeding palaeocommunities in these environments has yet to be compiled . Evidence for such a history, which must come primarily from studies on various fea­ tures of bioturbation (cross-cutting relationships of trace fossils, burrow depths, extent of reworking) is currently being developed (e .g. Crimes & Anderson 1985; Wetzel & Aigner 1986; Droser & Bottjer 1988) . The suspension feeder tiering history (Fig . 3) displays the maximum heights and complexity of tiering in a characteristic benthic palaeocommunity at various times . Physically dominated settings are unlikely to support a biota with this maximum development of tiering complexity . The tiering his­ tory is of primary tier feeders (Bottjer & Ausich 1987), which are organisms whose body or burrow intersects the sea floor . Although detailed tiering

1 Major Events in the History of Life

44 7ii � .J. Q; c Q;

E- E :7:Q;

-0 u Q; u ..c '" �'t:

E�

g .� Q; u c

�

0

Pa l aeozo i c

100 50 0

M esozo i c

....... .. . .- . . . . 1---- .-- .... •

•

•

..

•

• •

u

'0 N 0

c Q;

• •

U

-50

-100

100 50 0

-50

-100

Fig. 3 Tiering in soft-substrata suspension-feeding communities through the Phanerozoic. The heaviest lines represent the maximum level of tiering above or below the substratum at any time. Other lines represent levels of tier subdivision. Solid lines represent data, and dotted lines are inferred levels. (From Bottjer & Ausich 1986 .)

histories have not been compiled for other major environmental settings, such as reefs or hard­ grounds, they should reflect the relative changes in suspension feeder tiering.

Faunal histories and ecological structure Faunal diversifications and the history of benthic faunal ecological structure can be understood best in the context of temporally distinct faunas, which include the following : Vendian Fauna, Tommotian Fauna, Cambrian Fauna, Palaeozoic Fauna, and Modern Fauna . Sepkoski (1984; Section 1 .6) defined the Cambrian Fauna, Palaeozoic Fauna, and Modern Fauna based on familial diversities .

Vendian Fauna. Fossils o f the first benthic 'metazoans' are known from the Vendian (c. 620 570 Ma) . This fauna (Sections 1 .3, 1 .5) was initially best described from the Ediacaran Hills in South Australia, but is now recognized world-wide (Glaessner 1984) . Glaessner (1984, page 52) recog­ nized 31 named species from the vicinity of Ediacara and assigned these fossils principally to modern metazoan groups, including Hydrozoa, Scyphozoa, Conulata, colonial Cnidaria, Polychaeta, and Arthro­ poda . Seilacher (1984) offered a sharply contrasting interpretation for the Vendian Fauna . He argued that many fossils interpreted as medusoids are actually trace fossils and that the non-medusoid fossils represent a clade distinct from all extant metazoans . Clearly the zoological affinities and autecol ogy

of Vendia n taxa must be understood before com­ munity ecological structure can be reconstructed. However, whatever the trophic habit of members of the Vendian Fauna, it is apparent that Vendi an communities displayed some ecological structure . Charniodiscus and Glaessnerina sp ecies apparently attained variable heights above the sea floor . Maxi­ mum preserved heights of individuals include the following : G . grandis, 16 cm; C. longus, 25 cm; C. arboreous, 60 cm . Other organisms lived directly on the bottom . It is possible that this height distinction among members of the Vendi an Fauna may rep­ resent an ecological structuring analogous to epi­ faunal tiering. The widely distributed Vendian Fauna apparently suffered major extinction (if not total extinction; Seilacher 1984) at the end of the Proterozoic (Section 2 . 1 3 . 1) . The Phanerozoic record of benthic faunas has always been significantly different from that present during the Vendian . Trace fossils from the Vendian are Palaeozoic in affinity and indicate that a worm-like fauna of shallow-burrowing deposit feeders existed during this time (Glaessner 1984) . Vertical dwelling bur­ rows are generally lacking, indicating that infaunal suspension feeders were rare or had not yet devel­ oped. Thus, at most a shallow infaunal tier of deposit feeders existed, up to several centimetres below the sediment- water interface, in soft sub­ strata Vendian environments .

Tommotian Fauna. The first major occurrence of fos­ silized metazoan hard parts was during the Tommotian at the base of the Cambrian (c. 570 Ma) . The Tommotian Fauna preceded the first occurrence of trilobites, which was approximately at the base of the Atdabanian (c. 560 Ma) (Conway Morris 1987) . This fauna (Sections 1 .4, 1 .5) is recorded by a variety of very small, principally phosphatic skel­ etons . Characteristic taxa include small conical shells such as Protohertzina and Anabarites, inarticulate brachiopods, the sclerites of Lapworthella, archaeo­ cyathids, and trace fossils (e . g . McMenamin 1987) . Like the Vendian Fauna, the Tommotian Fauna has recently been documented to occur worldwide . More autecological study on elements of the Tommotian Fauna is necessary before the palaeo­ ecological structure of these early Phanerozoic communities can be fully understood . Problems include ( 1 ) which of the component taxa are skeletal remains of single organisms and which are sclerites of some larger creature (for example Halkieria; Conway Morris 1987) ; and (2 ) the autecology and

1.7

45

Diversification of Marine Habitats

functional morphology of Tommotian organisms

+5 cm, and +5 to + 10 ern (see Fig. 4) . The +5 to + 10

that have no clear living counterpart.

cm Cambrian tier included eocrinoids, edrioaster­

Regardless

of

shortcomings

in

the

detailed

oids, crinoids, archaeocyathids, and sponges (Figs

understanding of the Tommotian Fauna, it is clear

4, 5) .

that it represents the initial establishment of the

among others, a variety of echinoderms, sponges,

basic benthic ecological structure, albeit simple and

archaeocyathids, and inarticulate brachiopods.

composed of small organisms, that would character­ ize

the

remainder

Tommotian forms,

Fauna

simple

of

the

Phanerozoic.

includes

sessile

suspension

and

feeders

The

mobile

such

as

The

0

to

+5

cm suspension feeders included,

Infaunal suspension feeders were also close to the sediment-water interface during the Cambrian. Only the

-6

cm tier was occupied in environments

&

below fairweather wave base (Bottjer

1986) .

as archaeocyathids and inarticulate brachiopods,

Cambrian inner and middle shelf carbonate deposits

and predators such as

Protohertzina

(McMenamin

1987) .

Droser

(1988)

Bottier

6

cm. If these results are typical for

such Cambrian environments, they indicate the

0

typically smaller than one centimetre. Epifaunal

continued presence of the

suspension feeders were confined to the lowest

deposit

levels within the benthic boundary layer and were

Cambrian. In contrast, deeper

probably characteristically within the tier of Bottier

&

Ausich

(1986) .

0

to

+5

reported that in

of western U.5.A. bioturbation occurs at depths no greater than

Tommotian skeletons and skeletal elements are

&

Ausich

Sinotubulites, more complex suspension feeders such

and

suspension

to

-6

cm tier for both

feeders

through

SkolitllOs,

the

possibly

cm

made by deposit feeders, is abundant in Cambrian

Trace fossils associ­

strata deposited in nearshore settings above fair­

ated with Tommotian faunas indicate that the initial appearance of vertical burrows

4-5

curred during this time (McMenamin

weather wave base, forming the typical pipe-rock.

cm deep oc­

1987)

in near­

Palaeozoic Fauna.

The Palaeozoic Fauna (Sepkoski

shore settings above fairweather wave base. In

1984;

general, though, trace fossils formed in soft substrata

from the Ordovician to the Permian and was

Section

1 . 6)

characterized benthic habitats

settings below normal wave base appear to pen­

dominated

etrate depths no greater than several centimetres in

anthozoans, ostracodes, cephalopods, stenolaemate

the substrate (e.g. Crimes the

0

to

-6

&

Anderson

cm tier of Bottier

&

Ausich

Whether driven by ecological pro­

cesses, general laws of size increase, or intrinsic diversification after approximately the

Tommotian

Fauna

was

10

million years,

replaced

by

the

Cambrian Fauna. The Cambrian Fauna represents a diversification of metazoans and an increase in body size of benthos, both of which resulted in more complex benthic communities. From analysis of familial diversities, dominant faunal elements in the Cambrian Fauna include trilobites, inarticulate brachiopods,

hyolithids,

monoplacophoran mol­

luscs, eocrinoid echinoderms, and archaeocyathids. The Cambrian Fauna dominated the benthic habitat for approximately

55

million years.

Typical preservation of a Cambrian benthic com­ munity reveals a simple tiering structure; a relatively simple structure is also evident in the Burgess Shale fauna despite preservation of the soft-bodied faunal component (Section

3.1 1 .2) .

Tiering levels for both

epifaunal and infaunal suspension feeders remained quite low (Bottier

&

Ausich

1986) .

articulate

brachiopods,

crinoids,

1985); thus (1986) was

present for both suspension and deposit feeders.

Cambrian Fauna.

by

Two tier levels

are defined for epifaunal suspension feeders:

0

to

E � Cl! U

� .c:: c

2:: co 3: .'.

0 51 20J 10 •

Alcyonarians Sponges

•

Alcyonarians

A r c h aeocyat h i ds

_

C o ra l s

G ra ptol i t e s

c Cl!

E

B ryozoans

:;;

S ponges

u

5

Cl! > 0 ..0 co Cl! u c

�

0

•

A l cyona rians

A r c h a e ocyat h i d s G rapto l i t es

Co ra l s B ryozoans

0

Fig. 4

S ponges

1£ 10 ISI D Ic,IC21 P �rl J I K IPgIN� Q P a l ae ozo i c

Mesozoic Cenozo i c

Tiering history o f Phanerozoic colonial suspension

feeders on soft substrata from non-reef, shallow subtidal shelf, and epicontinental sea settings. Vertical distribution within each tier is arbitrary and only implies occupation in a tier for the duration indicated. (From Bottjer

&

Ausich 1986.)

1

46

Major Events in the History of Life

C r i n o i d ea

•••

�.:::::.::;: ....... _ •

­ u o Q) '" .c u

800

The Precambrian Reefs extend well back into the Precambrian (Copper 1974; Hecke1 1974; James 1983) . The earliest unquestioned stromatolites (see also Section 1 . 2) are from the Fig Tree Group of the Barberton Mountain Land, South Africa, 3300 - 3500 Ma . They are regarded as the product of a photoresponsive microbial community in a shallow, evaporitic environment, and associated microfossils include filamentous forms . There is no certainty that the microbes are true cyanobacteria but already a range of stromatolitic forms from linked domes to pseudo columns is present . The stromatolitic car­ bonate ecosystem became widespread some 2500 2300 Ma (Fig. 1 ) . Our knowledge of the evolution of the microbial communities responsible is vague but increasing . Most were probably eubacteria, but true cyanobacteria may not be very old . Eighteen mor­ photypes are known from the 1600 - 2000 Ma Gunflint Chert in Canada; most are of blue-green appearance but many are of unknown affinity. In Australia, the 1600 Ma Paradise Creek microbiota is morphologically comparable, and the 800 - 1000 Ma Bitter Springs community is extremely similar to that of modern stromatolite communities (Walter 1 976) . Early Proterozoic stromatolites ranged from non­ marine to deeper-water environments . Extensive shelf biostromes, mainly of linked domes, grade into shelf-break bioherms with several metres relief of branching columnar morphs in the Slave Province of Canada . Individual build-ups reach 100 m di­ ameter and 20 m thickness . They are cut by channels draining the shelf and die out rapidly down slope into the basin . Although extremely simple ecosys­ tems, these are regarded as positionally and func­ tionally comparable to modern reef ecosystems. Elsewhere stromatolitic masses on Proterozoic shelves < 60 m thick and 1 - 2 km long are reported. Precambrian stromatolites tend to be larger than

U

u N

0

Z «

.:;:

I Cl... u.J

�

�

0 Iu.J

Ma

� Cl...

0

2000

3000

I U � «

C u m u l a t i v e t h i c kness I

m

!

Fig. 1 Distribution in time of the major reef building biotas of the Precambrian and Early Phanerozoic.

Phanerozoic examples, and up to the early Palaeo­ zoic, tall, narrow, erect, branched columnar mor­ phologies dominated . This is in contrast to the broader unbranched forms which subsequently dominated. Columnar stromatolite diversity in­ creased up to the Middle - Late Riphean but a sharp decline in abundance and diversity set in about 800 Ma . With the earliest claimed eukaryote being from the 1200 - 1400 Ma Beck Springs Dolomite of California, and traces of metazoans, although poorly documented and often doubtful, known from about 1000 Ma, the decline in stromatolite build-ups seems likely to reflect the rise of grazing heterotrophs . However, Copper (1974) pointed out that the first well preserved metazoans of the Ediacara fauna (Sections 1 . 3, 1 . 5) appeared not to include algal grazers, although this fauna was not recovered from carbonate facies . He timed the decline of stromato­ lites as slightly later and implicated the widespread 675 - 570 Ma Late Precambrian glaciation .

54

1 Major Events in the History of Life

The Cambrian

The Ordovician

The first metazoan reefs date from the earliest Cambrian . In clastic facies, skolithid reefs are com­ mon world-wide . They form extensive masses of agglutinated sand grains, an early equivalent of modern sabellid buildups, < 80 cm thick and hundreds of kilometres long as fringing reefs in the breaker zone cut by surge channels . Accumulated thicknesses of skolithid sands often precede the first development of archaeocyathid reefs in the Lower Cambrian, the earliest of all skeletal frame­ work reefs . Archaeocyathids, with small, usually cup-shaped, mainly solitary, porous carbonate skeletons, have been considered a distinct phylum but modern opinion tends to favour their classifi­ cation as a subgroup of the Porifera . They form mainly small patch reefs, mostly < 3 m thick and 10 - 30 m in diameter, although larger fringing or barrier reefs are claimed in the later Lower Cambrian with accumulated thicknesses < 100 m . Bioherms are dominated b y three t o four struc­ turally different genera, some rooted and func­ tioning principally as bafflers . Renalcis, Epiphyton and Girvanella are frequently associated as over­ growths and may form the bulk of the skeletal material . About 30 - 50% of the build-up is fine carbonate mud with some bioclastic debris, little pore space, and few cavities . There may be pockets or lenses of shelly material but generally fauna of the adjacent facies is rare . There is no evidence of biological destruction by borers or grazers . No obvious reef zonation is reported for archaeo­ cyathid build-ups . They are the first of a range of skeletal organisms to form patch colonizations of the sea floor with minor relief, which persist in time to form biohermal masses in the rock record . They declined at the end of the Lower Cambrian and became extinct in the early Middle Cambrian, initiating a period which, in the absence of suitable skeletal organisms, lacked significant reef growth . Algal stromatolite build-ups made a brief comeback, possibly with a decline in grazers, as gastropods are scarce in the later Cambrian (Copper 1974) . Lithistid sponges occur in some of these stromatolite masses and skeletal algae are not uncommon . Stromatolite build-ups with or without a sponge contribution persist into the early Ordovician but an explosion in diversity of grazing gastropods correlates with the effective disappearance of stromatolites as major components of build-ups on open shelves .

The early Ordovician sees a rise in small bioherms constructed of lithistid sponges, particularly Archaeoscyphia (somewhat archaeocyathid in ap­ pearance), and skeletal algae . Locally, the recep­ taculitid alga Calathium, or Pulchrilamina of dubious affinities but possibly a stromatoporoid, may be important biohermal components . Again, these mounds show no zonation and little relief, no borers but common burrowers, and increasingly diverse associated biotas including echinoderms, trilobites, brachiopods, crinoids, early bryozoans, and rich pockets of gastropods and cephalopods . Build-ups reach cumulative thicknesses of 20 m and lengths of 87 m. Larger examples may show simple suc­ cession (James 1 983), climaxing in encrustations of Pulchrilamina. In addition, the early Ordovician has the earliest examples of mud mounds dominated by the cavity structure stromatactis (variously con­ sidered as of organic or purely physical origin) and lacking any (other) sign of organic framework, < 76 m thick and 300 m across . Similar structures are recorded sporadically through the rest of the Palaeozoic, whilst stromatactis is frequently a com­ ponent of build-ups dominated by (other) metazoans . There was a great expansion in benthic marine life in the early Middle Ordovician . The stromatopo­ roids, with doubtful Cambrian representatives, the bryozoans, and the tabulate corals had all evolved and the rugose corals appeared for the first time . These groups, including the major components of the most successful Palaeozoic reef communities, diversified rapidly and non-stromatoporoid sponges declined as reef builders . However, it was almost another 100 million years before these new components realized their full potential . Initially, bryozoan reefs dominated, constructed of small encrusting, domed, massive, plus erect bifolial and cylindrical colonial morphologies trap­ ping and binding lime mud . A few small sponge reefs were bound by bryozoans and stromatopo­ roids, with blankets of shell coquinas and pelma­ tozoan debris . These mainly small, unzoned build-ups may have had as much as 1 m relief and formed accumulations up to 4 m thick, but in the later Middle Ordovician, large shelf-break carbon­ ate masses, < 250 m cumulate thickness and 60 km long, are dominated or largely constructed by bryo­ zoans (Webby 1984) . Associated faunas included crinoids, brachiopods, together with blue-green (Girvanella, Sphaerocodium) and red (Solenopora)

1 . 7 Diversification of Marine Habitats algae, some sponges and, in some of the larger build-ups, stromatactis . Tabulate coral and bryozoan build-ups coexisted briefly, with later Middle Ordovician Labyrinthitos patch reefs, but by this time the stromatoporoids were beginning to diver­ sify . From the later Ordovician until the end of the Devonian, major build-ups were dominated by stromatoporoids, with corals and skeletal algae as major contributors, whilst bryozoans and other sponges were reduced to minor roles. However, corals alone and less commonly bryozoans con­ tinued to contribute patch reefs, forming bioherms and sometimes extensive biostromes, whilst sponges sometimes dominated build-ups in deeper water. Upper Ordovician build-ups range from small patches dominated by Tetradium, fasciculate Rugosa, Receptaculita and other skeletal algae, through small algal and stromatolitic pinnacle reefs < 30 m high and 0 . 8 km in diameter, to zoned and unzoned coral - stromatoporoid build-ups and large stroma­ tactis mounds < 100 - 140 m high and 1 km in diameter. A shelf-break complex of patch reefs, individually < 15 m high and 50 m in diameter, grades from talus flanked domical stromatoporoid mounds at the margin, through communities of laminar and domical stromatoporoids, to patches of diverse corals, algae, and ramose bryozoans in the

55

shelf interior . By the late Ordovician, there is in­ creasing evidence of borers and skeleton-breaking organisms at work. The development of reef communities suffered another set-back with the late Ordovician extinc­ tions (Section 2 . 1 3 . 2) . Build-ups are few and small until mid Llandovery times . Thereafter, patch reef development becomes widespread, particularly in the later Llandovery and Wenlock, with individual examples developing < 5 m relief on the sea floor, < 60 m cumulative thickness, and 100 m or more in diameter. Succession may be well developed with pioneering faunas of syringoporids, favositids, spheroidal stromatoporoids, halysitids, or crinoid groves . In the diversification stage, stromatoporoids of various morphologies, colonial rugose corals, and tabulate corals (particularly heliolitids) may be prominent, with a rich associated fauna of brachio­ pods (often in nests), bryozoans (some cryptic), crinoids, microfauna, and stromatactis . Algae are not so prominent. Stromatoporoids, with or with­ out tabulate corals, form the domination stage . Most build-ups show little lateral differentiation internally. However, among the hundred or more patch reefs of Middle Silurian age in the Great Lakes area, the largest structures show greater com­ plexity . The 15 km2 Marine Reef of Illinois has a core largely constructed of stromatactis, with a cen-

N

60��____�____�______+-____�____��____�60

o r-------+---��+_��-,����--�-+_� O

5 Fig. 2 Devonian continental reconstruction showing the distribution of organic build-ups (reefs and bioherms) and their latitudinal limits. (After Heckel & Witzke 1979 . )

56

1 Major Events in the History of Life

tral lagoonal facies, an externally fringing com­ munity of corals and stromatoporoids and a flanking apron composed largely of skeletal debris.

The Siluro - Devonian These Silurian build-ups were the forerunners of the spectacular development of reef growth in the Devonian, representing the first major peak in reef diversification and possibly the all time acme for reef ecosystems (Fig . 2) . Major reef complexes, per­ sisting over tens of millions of years, resulted in cumulative thicknesses of reef and perireefal car­ bonates < 2 km thick and stretching for hundreds of kilometres along shelves . Fringing, barrier, and shelf based atolls (faros) are represented. Reef edge, fore reef, and back reef zones are clearly differ­ entiated with detailed palaeoecological zonation comparable in complexity and in variation of con­ stituent faunas and floras to modern major reef complexes . Principal constructors of the reef mar­ gin were stromatoporoids and the blue-green algae Renalcis . Stromatactis is often present, and corals play a subsidiary role although they were more important on the reef flat and in areas where a reef rim was poorly developed or missing . Back reef facies are characterized by distinctive lithologies and assemblages, in particular by the stromatopo­ roid Amphipora. In some places, for example the Canning Basin of Western Australia, talus aprons and pinnacle reefs can be demonstrated on fore reef slopes descending to basinal facies < 180 m below contemporary sea level . Compared with Recent reefs, those of the Devonian show much less evidence of the activity of borers, grazers and scrapers; much of the break­ down of the rapidly cemented reef rock appears to have been physical .

This episode of reef building was terminated by the collapse of shallow-water ecosystems and the extinction or near extinction of the principal frame­ building organisms near the end of the Frasnian (Section 2 . 13 . 3 ) . In the Canning Basin, reef growth locally continued into the Famennian almost totally dominated by skeletal and non-skeletal algae . In the succeeding Carboniferous major build-ups are rare, although mud mounds are common, reflecting the relative paucity of suitable constructors among the skeletal organisms in the re-established level bottom communities . It was almost another 100 million years before large scale reef complexes were again developed, and then not on the scale of those of the Devonian .

References Copper, P. 1974. Structure and development of early Palaeo­ zoic reefs . Proceedings of the 2nd International Coral Reef Symposium 6, 365 - 386. Heckel, P.H. 1974. Carbonate buildups in the geologic record . In: L . F . Laporte (ed . ) Reefs in time and space, Special Publication of the Society of Economic Paleontologists and Mineralogists, No . 18 pp . 90 - 1 54. Tulsa, Oklahoma. Heckel, P.H. & Witzke, B .J. 1979 . Devonian world palaeo­ geography determined from distribution of carbonates and related lithic palaeoclimatic indicators. Special Papers in Palaeontology 23, 99- 123 . James, N . P . 1983 . Reefs . I n : P.A. Scholle, D . G . Bebout & CH. Moore (eds) Carbonate depositional environments . Memoir o f the American Association o f Petroleum Geo­ logists, No . 33, pp. 2346-2240 . WaIter, M.R. (ed. ) 1976. Stromatolites . Developments in Sedi­ mentology, No. 20. Elsevier, Amsterdam. Webby, B . D . 1984. Ordovician reefs and climate : a review. In : D . L . Bruton (ed . ) Aspects of the Ordovician System . Palaeontological Contributions from the University of Oslo, No . 8, pp. 87-98.

1 . 8 Terrestrialization

1 . 8 . 1 Soils

daily wetting and drying and to salinity variations . A s such they were preadapted t o life on land . Some silicified Precambrian forms can be compared di­ rectly to extant cyanobacteria found in subaerial settings (Campbell 1979) . In present day environ­ ments, too hostile for higher plants (such as deserts or at high altitude), primitive microbial communi­ ties are dominated by cyanobacteria, both filamen­ tous and coccoid, and chlorophytes . If such forms are capable of widely colonizing modern deserts, it would be naive to doubt their ability to colonize the ancient land surfaces . Golubic and Campbell (1979) have compared the mid-Precambrian microfossil Eosynechococcus moorei with the extant cyanobac­ terium Gloeothece coerulea, which is a sub aerial form, providing a suggestion of the earliest terres­ trial microbiota . Biogenically influenced terrestrial to supra tidal phosphates have been recorded from the Middle Cambrian of the Georgina Basin of Northern Australia (Southgate 1986) . In these examples very well preserved phosphatized microbial tubes, identical to calcified fungal tubes in present day calcrete soils, occur in phosphate horizons associ­ ated with shallowing-upwards peritidal deposits . The exact setting for their formation (supratidal or fully terrestrial) is uncertain but the remarkable similarities between these phosphatic fabrics and those of present day microbial soil carbonates must place this discovery as the strongest candidate for the earliest biologically active soil . The 'greening' of the land surface, albeit by a microbial sludge, would have begun a series of wide reaching changes in weathering and sedimen­ tary processes . Land surfaces, lacking any biological cover, are prone to erosion by wind and runoff. Even simple microbial mats on the surface would have provided some binding of weathered materials (CampbeU 1979), although roots provide a much more effective binding agent . As a result of binding, rates of erosion may have decreased and weathered materials would have had a longer residence time in the soil, allowing greater decomposition . The biological cover might also have increased levels of carbon dioxide in the soil, and would have added organic acids; both factors would have promoted chemical weathering in the soil . All these effects

V . P . WRIGHT

Introduction The soil is probably the most studied and best understood ecosystem on Earth, yet very little is known of its origins or the timings of each develop­ mental stage in its evolution . This situation arises both because of the low preservation potential of soils and through a lack of study . A variety of soils have been recognized in Pre­ cambrian sequences ranging back to over 3000 Ma . During the latter half of the Precambrian and through the Phanerozoic a gradual diversification of soil types occurred (Retallack 1986), reflecting both atmospheric evolution and biological diver­ sification, especially since Middle Palaeozoic times . Although many, i f not most, details o f the evo­ lution of soil communities and their interactions remain conjectural, several major stages can be de­ fined. The evidence, circumstantial at best, suggests that biologically active soils have existed since at least Middle Cambrian times (Fig . 1 ) .

Abiotic soils No direct evidence has been found for biologically active soils during the Precambrian, although a variety of weathering profiles and structural palaeo­ sols have been discovered (Retallack 1986) . Organic­ rich palaeosols apparently occur in the 2400 Ma Blind River Formation of Ontario (Campbell 1979) . High levels of radiation, adverse temperatures and atmospheric conditions must have prevented colonization of the land surface, even though micro­ bial life existed in the contemporaneous seas . The soils which developed during the Precambrian were the products of purely physical and physico­ chemical processes .

Microbial soils Cyanobacteria were abundant in the Precambrian, including intertidal forms which were adapted to

57

1 Major Events in the History of Life

58 Cz U

0 N

0 u.J :2

V1



<J>

�

<J>

� .D 2 .!::! :2

�

�::J

<J>

U >£ 0.. 0 C

u <J>

? -0 '"

'l)

CCl

'6 0 �

C!2

.�

o

E 'l) -0 0

::J 0 00 '" £ 0.. 0 u <J>

E

'"

� 0.. -0 'l) $

g

::J .D c le

'0.0 £ V> -
-

�.

C!2 0 <J> 0 'l)

0 00 C ..2 'l) a; u

cle £ u � 'l) <J> 0 0..

0 u <J> le $ ..2 -0 ::J

0 u 'l) :s � 'l) -0 'l) 2

E

>-

E

"0 'l) ::J C 0 � <J>

-'

E

t: <J>

.8 � .!::! c 'l) -0

� ..s

.� .D �

C le C

::J £ c le

�.

�

;.;:

<J>

E

::J

<J>

<J>

"l!

u...

:>

�

<J>

...L 1 900 M a Abiotic soils

Fig. 1 Geological ranges of major soil categories and first known appearances of important soil ecosystem components .

must have increased many-fold with the adven t of rooted vegetation . On newly exposed surfaces cyanobacteria are usually the first colonizers, followed by lichens . Fungally produced oxalic acid in lichens is a major factor in rock decomposition, but the timing of the appearance of the fungal - cyanobacterial ass oci­ ation is unclear. The vast majority of lichen-forming fungi belong to the ascomycetes, but the earliest record of these is from the Ludlow of Gotl and (Sherwood-Pike & Gray 1985) . Present day microbial soils are best developed in restricted settings which do not provide guides to

the sorts of soils possible in the past . Under suitable conditions relatively thick microbial mats may have developed, especially in humid climatic regimes . Such soils must have provided suitable micro­ climates for the first terrestrial invertebrates (Rolfe 1985; Section 1 . 8 . 3), even if the bare landscapes were still too hostile . However, no records of such microbial soil faunas are known with confidence . Highly bioturbated palaeosols have been recorded from the Late Ordovician (Ashgillian) Juniata Formation of Pennsylvania, U . S . A . (Retallack 1986) . These consist of burrows 3 - 16 mm in diameter, extending to depths of 50 cm in the now compacted argillaceous palaeosol . The burrows occur in fluvial overbank deposits, and it is often very difficult to determine if such burrows are truly those of soil dwellers or the result of burrowing during temporary subaqueous phases caused by flooding. As yet no attempt has been made to detect organic carbon isotope signatures in pedogenic carbonates in Lower Palaeozoic palaeosols, but this might prove a fruitful avenue of investigation .

Bryophyte soils By the Late Ordovician bryophytic-like terrestrial vegetation had appeared (Section 1 . 8.2). Such a veg­ etation cover, although relatively thin, would again have provided opportunities for faunal colonization . The nearest possible present day analogues for such biotas are to be found associated with lichens or moss cushions . They are characterized by a com­ munity of microarthropods, such as mites and springtails, but these only have a geological record back to the Early Devonian (Siegenian) Rhynie Chert . Could it really have taken over 40 million years for invertebrates to have colonized the bryo­ phyte 'felt' covering the land surface, a land surface which probably already had a long history of micro­ bial cover? The earliest known terrestrial faunas of the Early Devonian (Section 1 . 8 . 3) were already diversified and contain representatives of the major soil ecosystem components . The earliest evidence of a terrestrial biota, although tentative, consists of faecal pellet-like ovoid and cylindrical bodies of hyphal fragments from the Ludlow of Gotland (Sherwood-Pike & Gray 1985) . These may provide evidence of myco­ phagous feeders, and the presence of associated ascomycete fungal remains indicates that the de­ composer subsystem of the soil ecosystem had already evolved.

1 . 8 Terrestrialization Rooted soils The next major step was the development of a rooted plant cover . This happened progressively with the diversification of the vascular plants from Early Silurian times, with a further major step in late Devonian times when true forests first ap­ peared. The final stage in this series of events, at least to date, was the rise of the grasses in the Tertiary (Section 1 . 1 1 ) . The consequences of a rooted plant cover were far greater than those of a simple microbial or bryophytic one . The increased stability of the soil, and increased biomass, would have resulted in thicker soils and thicker humus. The degree of biochemical and biophysical weathering would have increased dramatically, and from Devonian times on soil-types diversified in re­ sponse to these changes (Retallack 1986) . The advent of a prominent rooted zone would have been associated with the development of the rhizosphere, with its own complex biotic inter­ actions. A critical event would have been the in­ itiation of symbiotic fungal - root relationships (mycorrhizal associations), in which the fungal component acts as a nutrient supplier to the roots . These fungal associations occur either internally within the root (endomycorrhizae) or as sheaths around the roots (ectomycorrhizae) . Occurrences of actual fungal remains with roots have been re­ corded from the Rhynie Chert and also abundantly from early Carboniferous soils, as calcification pro­ ducts of basidiomycete fungi around root tubes . In such cases, however, it is difficult to categorically establish that the fungi were not simply saprophytic forms.

Ecology The soil is an essential component of the terrestrial ecosystem, and one of its most critical functions is to decompose organic matter, making plant nutri­ ents available for recycling. The primary producer subsystem must, by all reasonable considerations, have been present from Cambrian times or earlier. The possible occurrence of fungal tubes in middle Cambrian terrestrial phosphorites of Australia, and the presence of ascomycete remains from the Ludlow of Gotland suggest that by the Middle Silurian, if not much earlier, the decomposer sub­ system had also developed . Thus recycling became possible . Possible microarthropod faecal pellets in the Silurian suggest the presence of consumers (mycophagous forms) . Some 20 million years later,

59

as revealed in the Siegenian Rhynie Chert, a fauna of spring-tails, mites, spiders, and trigonotarbid arachnids had appeared, representing many of the important components of the ecosystem (Section 1 .8 . 3) . By early Carboniferous times the soil ecosystem had evolved to a point where it produced a variety of humus fabrics identical to those found in present day soils (Wright 1987), which must reflect the action of the same types of complex biogenic processes . The evidence is frustratingly incomplete, and further work is required especially to integrate the occurrences of the early soil faunas with their associated soils . The effort needs to be made to search for evidence of biofunction in early Palaeo­ zoic terrestrial deposits, since such soils were prob­ ably organically active . What can be said, with growing confidence, is that the first vascular plants must have colonized a land surface which already had a long history of biological activity. Studies of microbial or bryophytic soils today will provide us with some clues as to the possible forms taken by these earliest soils .

References Campbell, S . E . 1979 . Soil stabilization by a prokaryotic desert crust: implications for Precambrian land biota. Origins of Life 9, 335 - 348 . Golubic, S. & CampbeII, S . E . 1979 . Analogous microbial forms in Recent subaerial habitats and in Precambrian cherts : Gloeothece coerulea Geitler and Eosynechococcus moorei Hofmann . Precambrian Research 8, 201 -217. RetaIIack, G.J. 1986. The fossil record of soils . In: V . P . Wright (ed . ) Paleosols: their recognition and interpretation, pp . 1 - 57. Blackwell Scientific Publications, Oxford . Rolfe, W.D.I. 1985 . Early terrestrial arthropods: a fragmentary record . Philosophical Transactions of the Royal Society of London B309, 207-218. Sherwood-Pike, M.A. and Gray, J . 1985 . Silurian fungal re­ mains: probable records of the Class Ascomycetes. Lethaia 18, 1 -20. Southgate, P.H. 1986 . Cambrian phoscrete profiles, coated grains, and microbial processes in phosphogenesis: Georgina Basin, Australia. Journal of Sedimentary Petrology 56, 429-441 . Wright, V.P. 1987. The ecology of two early Carboniferous soils. In : J. Miller, A . E . Adams & V.P. Wright (eds) European Dinantian environments. Special Publication of the Geological Journal No . 12, pp . 345 - 358. John Wiley & Sons, Chichester.

1 Major Events in the History of Life

60

1 . 8 . 2 Plants D . E D W A R D S & N . D . B U RG E S S

Introduction Land plants encounter problems relating to water stress, uptake, and transport, and to aerial dispersal of propagules . Survival in such habitats is associated with three major strategies : 1 Drought avoidance via opportunism and ephemeral life cycles completed under favourable conditions . 2 Extreme desiccation tolerance involving the capacity of cytoplasm to rehydrate and then function normally (poikilohydry) . 3 Maintenance of an internally hydrated environ­ ment by biochemical and anatomical modifications (homoiohydry) . Extant land vegetation includes representatives of all major groups; cyanobacteria, algae, bryo­ phytes and tracheophytes. The last are usually considered most successful and are homoiohydric possessing xylem (with lignin) for water transport, a waxy cuticle (cutin) for reducing evaporation, sto­ mata and an intercellular space system for gaseous transport (Raven 1984) . The poikilohydric life style of cyanobacteria, algae, and bryophytes is usually considered more primitive, is of particular signifi­ cance in the colonization of unstable environments, and hence would have been important in pioneering land plants . The preservation potential of land plants is linked to these strategies in that cutin and lignin are dur­ able and may persist, albeit modified, in fossils, but in poikilohydric forms, the only parts which might be expected to be fossilized are resting stages and/or dispersal units such as spores. The latter, impregnated with sporopollenin, a complex fatty polymer, also occur in tracheophytes . Thus although there is no direct record of thallophytes (cyano­ bacteria and algae) colonizing moist land surfaces in the Early Palaeozoic, it seems likely that they were present. A possible limiting, physical factor may have been high ultraviolet (UV) radiation cor­ related with low atmospheric oxygen . Indeed it has been postulated that lignin evolved from precursors involved in UV absorbance, and that cutin and sporopollenin initially had a similar role in UV reflectance . With regard to higher plants, attempts to demon­ strate the vascular status of megafossils, thus pro-

viding unequivocal evidence for land vegetation, have traditionally dominated research . However, more recently the affinities of Ordovician and Silurian microfossils have been rigorously appraised in the search for alternative pioneering colonizers . The first records and ranges of all fossils thought relevant to terrestrialization are documented in Fig . 1, and numbers below refer t o that figure .

Sporomorphs

1, 2 . Cryptospores . (lacking trilete ( Y ) or monolete

( I ) marks; after Richardson & Edwards 1989. )

Obligate permanent tetrads (1), so named because they do not split into four spores (monads) on dispersal, possess durable, smooth, unornamented walls thought to be impregnated with sporopol­ lenin, although this has not been chemically proven. They are thus considered to derive from land plants . Those characterizing Upper Ordovician spore as­ semblages are smaller, smooth walled, and often lack the enveloping smooth or sculptured 'mem­ brane' typical of most early Silurian forms . Its ab­ sence may result from poorer fossilization potential. Such tetrads increase in numbers and diversity, dominating assemblages until the end of the Llandovery. Thereafter they become relatively less common and occur only rarely in Lower Devonian sediments, where they are probably reworked . Gray (1985) has argued most persuasively that as com­ parable tetrahedral tetrads (sometimes membrane enclosed) occur in certain living liverworts, they thus derive from poikilohydric plants with bryo­ phyte physiology and life histories . It is also pos­ sible that they belonged to freshwater or marine algae for which there are no modern analogues, or that they were shed by intermediate extinct forms that lived in ephemeral water bodies producing spores when these dried up . Membrane enclosed monads and obligate dyads (2) have similar ranges to tetrads and were probably of similar derivation. 3 . Dyads. Habitually lacking a membrane, and be­ lieved to split into the consistently associated alete (lacking trilete or monolete marks) spores with thinner proximal faces and identical distal features, these are distally smooth walled in earliest records (Rhuddanian) and sculptured from the Homerian . They persist throughout the Silurian and are rela­ tively common in basal Devonian assemblages . Dyads occur in Salopella-like sporangia in the Pffdoli . The affinities of that genus remain

STRAT l G RA P H Y

2

Emsian

et

a;

o > UJ Cl

Cl

Z

c Q) OD

,,!!

'"

c

�

c

'50

,�

z

o

3:

Q)

...J

Ashgi l i

Caradoc

I

4J 1

rf\ \J0

Aero n i a n

------1 g � L--

(J)

0-

We n l oc k

Telyc h i a n

I Z

'

�

o -'" ..c u o

I

H I G H E R PLANT M EGAFO S S I LS

CUTICLES

I I

,

� ---"

\

I

,,

'r

,

I -' ,

I ,

1

N E MATO P H YTA L E S

l

ALGAE

1

FUNGI

5 6 7 8 9 1 0 1 f '1 � { 1 5 1 6 1 7 1 8 1 9

4

3

,�

c '"

c

I

S PO R O M O R P H S

,. >w. -Uil

1

� v

VI I 11 1

'

I

I

" . I 61

I

I

f-.-I

00

� ....,

�...... ....,

N'

� ...... ::::.

c' ;::

1

Fig. 1

g 1

F i rst appearances and strat i graph ic ranges of m i c ro- a n d m e gafoss i l s relat i n g t o colon izat i o n o f l a n d b y p l a n t s , R , Z a n d T refe r t o i n i t i a l m a j o r rad i a t i o n s of t h e vasc u l a r p l a n t s u bd iv i s i o n s Rhyn i o p hyt i n a , Zoste ro p h y l i o p hyt i n , and T r i m e rophyt i n a respective l y , F o r m o re co m p l ete explanation of see text.

1 -1 9 ,

0\ .....

62

1 Major Events in the History of Life

obscure, as vascular tissue has never been demon­

from sterile axes occur in the northern hemisphere

strated .

Ludlow . The earliest

direct

evidence for stomata

(with two guard cells) comes from sterile axes at the

4, 5. Monads with triradiate marks.

The earliest

base of the PHdoli, with examples on

Cooksonia and

monads not derived from obligate tetrads are

Zosterophyllum

smooth walled, usually with equatorial thickening,

ever, since stomata were described on Canadian

e.g.

Ambitisporites Cooksonia pertoni

recorded in the Gedinnian . How­

which is later recorded in PHdoli

Emsian

sporangia.

ably also present on the earlier Australian Ludlow

Sculptured monads

occur continuously from the Homerian, showing

they were prob­

Baragwanathia abitibiense

B. longifolia,

where cells are not preserved.

rapid diversification and increases in numbers in the Ludlow and PHdoli. For the most part their origins, be they bryophyte or tracheophyte, are unknown,

but

their

presence

Gedinnian

Cooksonia pertoni

in

Silurian

(Fanning

et al.

and 1988)

indicates that some of the producers may have been rhyniophytoid

(10) .

Higher plant megafossils

9. Sterile axes.

It is usually accepted that those with

dichotomous branching and peripheral support tissues (sterome), the earliest being Llandovery

Eohostimella,

are derived from erect land plants .

Late Silurian and early Devonian examples (e . g .

Cutic1es

6. 'Nematothallus'.

Hostinella) Associated with the earliest tet­

tracheids

possess a central strand composed of

(14), but whether or not pre-Ludlow repre­

rads are small cuticular fragments, thought ana­

sentatives were vascular is unknown. Some may

logous to vascular plant cuticles . They lack stomata

derive

and are usually imperforate, with smooth ?outer

tracheids had not yet evolved, while others may

from

plants

of

small

stature

in which

surface . Ornamented forms appear in the Wenlock

possess conducting tissues of bryophytic nature (cf .

and continue into the Gedinnian . A reticulate pat­

Lower Devonian

tern represents the outlines of surface cells in the

exhibits many homoiohydric characters, and would

underlying tissue . In

Nematothallus,

such cells were

± isodiametric in tangential section, while in the

cuticles of higher plants they are elongate vertically . Lang (1937) suggested that they covered

thallus,

Nemato­

a thalloid plant composed of tubes which he

placed in the Nematophytales

(15, 1 6),

a taxon for

plants with organization neither algal nor higher

Aglaophyton (Rhynia) major

which

be assigned to the Tracheophyta but for the moss­ like conducting tissues) .

1 0 - 1 3 . Fertile tracheophytes. Wenlock Cooksonia (Rhyniophytina: 1 0) is generally accepted as the

earliest erect pteridophyte-like plant (Edwards

et al.

1983) . Reservations as to its affinity stem from a complete lack of anatomy . Spores occur in PHdoli,

plant . While there may be some doubt that all durable

and stomata and sterome in Gedinnian

C. pertoni.

spores derive from land plants, this is not the case

Tracheids have never been demonstrated in central

for cuticles, although being imperforate and hence

strands .

relatively impermeable to gases, their function and

Steganotheca

even composition in Nematothallus might have been different from that in tracheophytes, e . g . primarily

to later examples, e . g .

as UV screens, facilitating runoff, or in defence . It is

Thus although

Cooksonia, Salopella

and

are usually assigned to the Rhynio­

phytina because of general morphological similarity

Rhynia gwynne-vaughanii,

they are better called 'rhyniophytoid' to emphasize

unlikely that they belonged to the tetrad producers

our ignorance . A major radiation is recorded in the

because, although the first records are coincident,

early Gedinnian, but they then became insignificant

cuticles persist into the Emsian and are sometimes

constituents of land vegetation (Edwards & Fanning

quite common constituents of Lower Devonian

1985) .

assemblages .

7, 8. Higher plant cuticles.

Baragwanathia

longifolia

(1 1 )

in

Australian

Ludlow strata is morphologically similar to Lower Homerian fragments with

Devonian examples, with sufficient anatomical as

larger, more strongly demarcated and aligned cells

well as morphological characters to indicate lyco­

are interpreted as sporangial from comparison with

phyte affinity. Thus, even in the absence of anatomy

dispersed and

in Silurian representatives, its vascular status is un­

in situ

Gedinnian rhyniophytoid

examples (1 0) . Cuticles without stomata deriving

questioned .

The earliest lycophyte with typical

1 . 8 Terrestrialization sporangium/sporophyll organization is the late Emsian Leclercqia . Zosterophyllum myretonianum (12) is the earliest fertile member of the Zosterophyllophytina, al­ though there are records of its characteristic branch­ ing (K- and H-shaped) in sterile Pffdoli axes. The first major zosterophyll radiation is recorded in the late Gedinnian of south Wales. Dawsonites sp . (13), a fragment of a fertile truss of Psilophyton in the south Wales Siegenian, marks the beginnings of the Trimerophytina. The Ludlow Australian record is less convincing. The trimerophytes diversified rapidly in the Emsian and are considered ancestral to ferns s.l., progymnosperms and sphenopsids .

Nematophytales

15. Microfossils of tubular organization, either as iso­ lated tubes or wefts, are recorded from the Telychian into the Lower Devonian . The most conspicuous tubes are internally sporadically thickened ('banded'), broadly resembling tracheids in their ornament, but there is no direct evidence that they were lignified . The source plants are problematic: they occur with smaller tubes in Nematothallus (Lang 1937) and have been found in plants with organi­ zation otherwise typical of Prototaxites (16). The habitats of such organisms, be they freshwater or terrestrial, remain as conjectural as their affinities . In that some tubes (but not banded forms) have been recorded attached, rather than just adpressed to cuticles of Nematothallus (6) type, they may well derive from land plants . Further isolated examples include tubes with smooth thick or thin walls, or filaments (occasionally branched) composed of elongate, narrow cells . The latter frequently occur in monotypic wefts or may be associat� d with wider smooth or banded tubes . Some of the associations may belong to Nematothallus or Nematoplexus . 1 6 . Prototaxites (Wenlock- Upper Devonian) is included because it is sometimes cited as a land plant largely due to its occurrence in tracheophyte assemblages in freshwater sediments . Its organi­ zation, in which narrow filaments enclose wider smooth tubes, is unique, and hence in the absence of reproductive organs its affinities, possibly algal or fungal, remain unknown, and speculation on the functions of its tissues unrewarding. 1 7. Parka, best known from the Scottish Gedinnian, a possible epiphyte in lacustrine habitats, may have

63

some relevance to the ancestry of higher plants in that it has been compared with the charophycean Coleochaete, although the latter lacks the cavities with numerous alete spores found in Parka. Com­ parative biochemical and ultrastructural studies suggest that among the green algae the Charo­ phyceae show closest similarities with bryophytes and tracheophytes while Coleochaete, with its parenchymatous organization, and protection, nutrition, and prolonged retention of the zygote, possesses the greatest number of advanced features .

18. Pachytheca i s exceedingly common in certain marginal fluviatile and lacustrine facies in the Lower Devonian . Its frequent association with Prototaxites has led to the suggestion that it was involved in its vegetative reproduction . However, the fossils suggest that the organism comprised a sphere of a mucilage-like substance in which filaments of cyanobacterial dimensions were embedded. Its habitat is interpreted here as freshwater, possibly littoral lacustrine . 19. Fungi. Although not considered plants, fungi are included here because it has been suggested that initial terrestrialization was possible only after the development of a symbiotic association between a semiaquatic green alga and an aquatic oomycete fungus, and that in the colonization of nutrient­ poor environments the fungus would have ex­ ploited large volumes of substrate for minerals (cf. mycorrhiza today) . Resting spores of presumed mycorrhiza in some Rhynie Chert axes are fre­ quently cited as supporting evidence, but the abundant spheres and hyphae may just indicate saprotrophism (i . e . decomposition of dead organ­ isms) in peat development. Further evidence for terrestrial fungi is the record of ascomycetes remains (hyphae, probable conidia, and ascospores) from the Ludlow of Gotland (Gray 1985), and similar, but more poorly preserved, material from the late Llandovery . Terrestrial vegetation It is postulated that moist land surfaces in the early Palaeozoic would have been coated with a green scum, perhaps initially of cyano- and eubacteria, later joined by filamentous and unicellular algae . Such an encrusting layer would have both physically stabilized and chemically broken down the sub­ strate, releasing nutrients and, in stable environ­ ments, resulting in the build-up of humus (see also

64

1 Major Events in the History of Life

Section 1 . 8 . 1 ) . From the middle Ordovician onwards microfossils morphologically convergent with those from later tracheophytes suggest a novel vegetation, possibly with thalloid organisms covered by cuticle and spore producers with liverwort life-style; aerial dispersal indicates the attainment of some stature . The appearance of Ambitisporites in the Llandovery heralded a new phase - the advent of pteridophyte­ like plants with axial organization, possibly forming a 'turf' just a few centimetres high . The larger size permitted by homoiohydry, the concomitant main­ tenance of turgor and hence a hydrostatic skeleton, conferred potential superiority over poikilohydric forms in terms of wind dispersal of propagules and in shading, thus limiting the productivity of smaller forms . Throughout the late Silurian there is an increase in axis diameter and length of fragments : sprawling Baragwanathia probably formed thickets . Lower Devonian assemblages suggest that many of the tracheophytes grew in monotypic stands, exten­ sive cover resulting from prolonged rhizomatous activity. Such plants would have provided mutual support - some of the Emsian trimerophytes attained a height of over 1 m. As to habitats, the best direct evidence comes from the Rhynie Chert, but as all these early pteridophytes were homo­ sporous (i . e . with spores of one size), the free-living gametophyte would have required moist conditions both for vegetative growth and reproduction . With regard to route of terrestrialization for higher plants, physiological considerations support transmigration from fresh water on to land .

References Edwards, D . 1980 . Early land floras . In: A.L. Panchen (ed . ) The terrestrial environment and the origin of land vertebrates . Systematics Association Special Volume 15 pp. 55 - 85 . Academic Press, London. Edwards, D. & Fanning, V. 1985 . Evolution and environment in the late Silurian-early Devonian: the rise of the pteri­ dophytes . Philosophical Transactions of the Royal Society of London B309, 147 - 165 . Edwards, D . , Feehan, J . & Smith, D . G . 1983 . A late Wenlock flora from Co. Tipperary, Ireland. Botanical Journal of the Linnean Society 86, 1 9 - 36 . Fanning, V . , Richardson, J.B. & Edwards, D. 1988. Cryptic evolution in an early land plant. Evolutionary Trends in Plants 2, 13-24. Gray, J . 1985. The microfossil record of early land plants : advances in understanding of early terrestrialization, 1970- 1984. Philosophical Transactions of the Royal Society of London B309, 167-195. Lang, W.H. 1937. On the plant remains from the Downtonian of England and Wales. Philosopical Transactions of the Royal Society of London B227, 245 - 291 .

Raven, J.A. 1984. Physiological correlates of the morphology of early vascular plants. Botanical Journal of the Linnean Society 88, 105 - 126. Richardson, J.B. & Edwards, D . 1989 . Sporomorphs and plant megafossils. In: CH. Holland & M . G . Bassett (eds) A global standard for the Silurian System, Geological Series No . 9, pp . 216-226. National Museum of Wales, Cardiff.

1 . 8 . 3 Invertebrates P . A . SELDEN

Introduction The diversity of invertebrate species on land greatly exceeds that in the sea; this is almost entirely due to the terrestrial insects which form 70% of all animal species alive today . However, of over 30 invertebrate phyla, only the arthropods, the molluscs, and the annelids have significant numbers of macroscopic terrestrial representatives . A greater number of phyla include very few terrestrial species, crypto­ biotic representatives, or internal parasites on terrestrial organisms . The body plans of some highly successful marine phyla have apparently precluded their terrestrialization; these include the sipunculid, echiuroid, and priapulid worms, cnidarians, lophophorates, chaetognaths, pogonophores, hemi­ chordates, and echinoderms . No phylum originated on land, and no major terrestrial taxon has become extinct, as far as is known . Outstanding questions on terrestrialization are : what physiological mechanisms enabled inver­ tebrates to emerge onto land; did each taxonomic group use similar mechanisms; were their routes onto land the same; did they all come onto land simultaneously, suddenly or gradually, or in dif­ ferent invasions? The hardest evidence comes from comparative physiology, but palaeontology has the power to test theories based on living material, and uniquely adds the dimension of time . Invertebrates moving from seawater to land ex­ perience profound changes in all aspects of life (Little 1983) . On land, water supply is at least vari­ able, and commonly seasonal . To invertebrates, whose air breathing mechanisms utilize diffusion to a far greater extent than ventilation, oxygen is more available in air than in water because the diffusion coefficient (partial pressure per unit length, in ml/[min x cm2 x cm x atm]) of oxygen in

1 . 8 Terrestrialization water is 0 . 000034, but in air is 1 1 . 0 . Support is more problematical in the less viscous aerial medium than in water, but once overcome, locomotion is easier and faster. The difference in refractive index between air and water poses a problem for visual sense organs in transition, but high frequency vibrations can be perceived more easily in air, resulting in a greater use of sound by terrestrial invertebrates . On land, internal fertilization is the norm, and greater protection (e . g . from drought) is afforded to the developing embryos . Changes in nutrition, ion balance regulation, and excretion are also necessary for terrestrialization . Some land animals avoid the difficulties of water supply by living in soil interstitial water; strictly, such animals (e . g . protozoans, ostracodes, and nematodes) should not be regarded as terrestrial . Poikilohydry i s used only b y small terrestrial animals, such as protozoans, tardigrades, nema­ todes, and rotifers, whose habitat is subject to seasonal drought periods . Many soil, litter, and crevice dwellers are able to take advantage of the high humidity in such habitats, and though they are often able to foray in drier situations (e .g. wood­ lice across the kitchen floor), retreat to the. humid home base is essential to prevent desiccation . In addition to woodlice, the centipedes, millipedes, flatworms, leeches, and earthworms are included in this group . Some animals, such as many land snails, can withstand desiccation during dry periods by aestivation, but require water or high humidity for activity at other time s . Finally, the true invertebrate conquerors of the terrestrial habitat, not requiring a humid environment in which to flourish, but active in dry, and even desert, conditions, are the majority of insects, many arachnids, and a few crustaceans . All terrestrial arthropods have waterproofing in the cuticle, but the form this takes differs in each arthropod group and is not always well studied. The differences may be important for palaeontology, however, since the preservation potential for differ­ ent cuticles is not the same .

The fossil record The fossil record of terrestrial invertebrates is shown in Fig. 1 (Rolfe 1980; Chaloner & Lawson 1985) . There is no fossil record of terrestrial flatworms, nemerteans, or nematodes, although fossil examples of parasitic and aquatic nematodes are known (Conway Morris 1981) . Oligochaete annelids are known from the Carboniferous . Their traces, in-

65

cluding burrows and faecal pellets, occur in palaeo­ sols from the Carboniferous onwards . They may have emerged onto land with the first humic soil (Section 1 . 8 . 1 ) . Land snails, both helicinid prosobranchs and stylommatophoran pulmonates, are recorded from the Upper Carboniferous, indicating that they had already become significant members of the land fauna by that time . The earliest basommatophoran pulmonate is Late Jurassic in age; this contradicts evidence from comparative morphology, which suggests that basommatophorans were ancestral to the other pulmonates . Possibly the development of ground shade and deciduous leaf litter (probably Lower Carboniferous) was necessary before land snails could be assured of the damp conditions necessary for colonization (Solem 1985) . All extant insects are terrestrial or secondarily aquatic, and there were no terrestrial trilobites, as far as we know. The record of Onychophora, which includes the Recent Peripatus, appears to begin with Aysheaia from the marine, Middle Cambrian Burgess Shale . Terrestrial uniramians (myriapods and insects) were thought to have evolved from land-living onychophorans, but there is new evi­ dence that the earliest myriapods were marine . This comes from myriapod-like fossils in marine sedi­ ments from the Silurian of Wisconsin and the Middle Cambrian of Utah . By the Devonian, milli­ pedes, centipedes, and arthropleurids had appeared in terrestrial faunas, and some reached giant pro­ portions in the Carboniferous forests . The earliest apterygote insects occurred in the Devonian, but the first pterygotes were Carboniferous in age . Eurypterids ranged from Ordovician to Permian and were predominantly aquatic animals, but from the Silurian onwards some were amphibious, as evidenced by their accessory lungs . They illustrate a failed attempt at terrestrialization using a method now being tried by the Crustacea. Their close rela­ tives, the scorpions, succeeded however, by changing their gills into lungs . All other arachnids are primarily terrestrial today, and the evidence from comparative morphology suggests that each arachnid group emerged onto land independently. The oldest are the trigonotarbids : extinct, close rela­ tives of spiders, with good terrestrial features, from the Lower Devonian of Rhynie, Aberdeen . In the Devonian are also found mites, pseudoscorpions, and possibly spiders, and by the Carboniferous there were more arachnid orders than today; only the spiders have radiated more dramatically in later periods .

V) u.J l-

< I

u.J

STRAT I G RA P H Y

0

U

0 z N

U

Q u a t e r n a ry Neogene

0 �

U

\

0 ....J

I

0 I CL 0 I U V) >- z z< 0 0�

I

I

V) 0 -

0 are certain to be significantly different but full data are unavailable. Vertical spacing between sections (BC, etc. see Sheldon 1987) is not to scale . Reversible shifts in means are common within sections (not shown subdivided) . Sections in the Teretiusculus Shales span c. 3 million years. (From Sheldon 1987.)

lution appears to be fairly common, particularly in Tertiary mammals (e . g . Gingerich 1985; Godinot in Cope & Skelton 1985) . Bown & Rose (1987) saw no sign of stasis in Eocene primates from Wyoming, reporting both gradual anagenesis and cladogenesis in sharp contradiction to the predictions of the punctuated equilibrium model . They highlighted the problems that gradualism causes for systematic palaeontology and biostratigraphy . Bell et al. (1985), in a multicharacter study of a Miocene stickleback lineage, found that taxonomically significant mor­ phological change was accomplished by protracted trends and by rapid bursts of evolution, without tight synchrony of change among characters (mosaic evolution) . Although there are some well attested examples of gradual cladogenesis in the fossil record (Gingerich 1985), the great rarity of branching points where nodes are known is consistent with common patterns of change in which cladogenesis is rapid and/or involves small, isolated populations .

Random change and trend reversals There has been much interest in the possibility that some morphological trends seen in the fossil record may be the result of processes other than natural selection . The genetic drift hypothesis predicts that

the morphology of selectively-neutral characters will vary through time as a random walk. Indeed, it has been argued that evolutionary rates exist only when the hypothesis of a symmetrical random walk can be refuted . However, Sheldon (1987) argued that temporary reversals of variable characters probably occur in all evolutionary lineages, and so many trends driven by selection may be indistinguishable from random walks . Reversals probably reflect times when some other attribute, genetically uncorrelated with the one under consideration, was selected . It would be unreasonable to expect that the one feature chosen for plotting was consistently the only one favoured by selection, or that it was always linked to every other favoured trait . In fact, long-sustained net trends in a single character may reflect genetic coupling to other characters having negative effects on fitness (see also Section 2 .2) . The widespread tendency not to expect reversals, or to interpret them as ecophenotypic change or random drift, led to the unrealistic portrayal of phyletic gradualism as unidirectional change . Re­ versals have many consequences . For instance, they complicate the theoretical arguments (Fortey 1988) concerning differentiation between cases of gradu­ alism and punctuated equilibria; they should not be automatically taken to indicate that the observed change is only ecophenotypic; and jumps in mor-

110

2 The Evolutionary Process and the Fossil Record

phological trends cannot be used to estimate the amount of time missing at diastems .

Patterns of evolution in different environments It is still inappropriate to estimate the relative im­ portance of particular patterns of evolution in dif­ ferent environments . Given the immense range of attributes of living organisms (e . g . complex life cycles and reproductive strategies) it would not be surprising to find different patterns emerging from broadly similar environments . Benthic inver­ tebrates, for instance, have a wide diversity of larval dispersal modes and these early stages, although rarely preserved, might profoundly influence pat­ terns . There is some evidence, as might be expected, that abrupt speciation and extinction are commonly associated with benthic species living in shallow marine settings . However, Sheldon (1987) suggested that, almost paradoxically, stasis seems to prevail in these more widely fluctuating, rapidly changing environments, whereas species living in, or able to track, narrowly fluctuating, slowly changing environments show persistent phyletic evolution rather than stasis . Some of the perceived punctuations in shallow benthic settings may simply reflect higher rates of short-term deposition and more hiatuses (less completeness) than in offshore, pelagic environ­ ments . But, although the most reliable evolution­ ary patterns will come from the most complete sequences, the depositional conditions promoting completeness might in themselves encourage grad­ ual phyletic evolution, especially of benthos.

Conclusion Studies of the fossil record have revealed a wide spectrum of microevolutionary patterns, from which can be inferred a variety of evolutionary processes. Punctuated equilibrium and phyletic gradualism should be viewed as just two theoretical versions of many possible evolutionary patterns and the temp­ tation to force poorly documented cases to fit one or other of these models must be resisted . Often there is simply too little data to assess patterns of change as, for example, with the genus Homo (Section 1 . 12) . We are not yet in a position to assess accurately the relative frequency of particular patterns and the domain of their expected settings .

Individual taxa probably exhibit different pat­ terns at different times, and different morphological characters in the same species may evolve at differ­ ent rates . Episodic changes need not be associated with branching events and demonstrating stasis in a species is not the same as demonstrating punctu­ ated speciation . Most geneticists believe that a punctuated appear­ ance of species is consistent with neo-Darwinian theory . In many ways it is explaining stasis which is certainly more prevalent than would have been predicted from studies of living organisms that represents the greater challenge .

References Bell, M . A . , Baumgartner, J.V. & Olson, E . c . 1985 . Patterns of temporal change in single morphological characters of a Miocene stickleback fish. Paleobiology 11, 258 -271 . Bown, T.M & Rose, K.D. 1987. Patterns of dental evolution in early Eocene anaptomorphine primates (Omomyidae) from the Bighorn Basin, Wyoming. Memoir of the Paleontological Society 23, 162 pp. Cheetham, A.H. 1987. Tempo of evolution in a Neogene bryozoan: are trends in single morphologic characters misleading? Paleobiology 13, 286-296 . Cope, J . C .W. & Skeiton, P.W. (eds) 1985 . Evolutionary case histories from the fossil record . Special Papers in Palae­ ontology 33, 1 - 203. Eldredge, N . & Could, S.J. 1972. Punctuated equilibria; an alternative to phyletic gradualism. In: T.J.M. Schopf (ed .) Models in paleobiology, pp. 82 - 1 15. Freeman, San Francisco . Fortey, R.A. 1988. Seeing is believing: gradualism and punc­ tuated equilibria in the fossil record . Science Progress 72, 1 - 19 . Cingerich, P.D. 1985 . Species i n the fossil record : concepts, trends, and transitions . Paleobiology 11, 27-41 . Hallam, A. 1982. Patterns of speciation in Jurassic Gryphaea. Paleobiology 8, 354-366 . Malmgren, B . A . & Kennett, J . P . 1981 . Phyletic gradualism in a Late Cenozoic planktonic foraminiferal lineage; DSDP Site 284, southwest Pacific. Paleobiology 7, 230-240. Malmgren, B . A . , Berggren, W.A. & Lohmann, C . P . 1983. Evidence for punctuated gradualism in the Late Neogene Globorotalia tumida lineage of planktonic foraminifera. Paleobiology 9, 377-389 . Sheldon, P.R. 1987. Parallel gradualistic evolution of Ordovician trilobites. Nature 330, 561 -563. Stanley, S . M . & Yang, X. 1987. Approximate evolutionary stasis for bivalve morphology over millions of years: a multivariate, multilineage study. Paleobiology 13, 1 1 3 - 139. Williamson, P.C. 1981 . Palaeontological documentation of speciation in Cenozoic molluscs from Turkana Basin. Nature 293, 437- 443.

2.4 Heterochrony K . J . McNAMARA

Introduction Heterochrony is the phenomenon of changes through time in the appearance or rate of develop­ ment of ancestral characters . While the recognition of a close relationship between ontogeny and phylo­ geny has a long history it was not until the late nineteenth century that it was formalized by E . Haeckel i n his 'Biogenetic Law' (ontogeny recapitu­ lates phylogeny) . This involved a change in the timing of developmental events; but only in one direction - by terminal addition . Phylogenetically this meant that ancestral adult forms were encapsu­ lated in the juvenile stages of their descendants . This became known as recapitulation . Exceptions to this rule (known as 'degenerate' forms) were noted by the leading protagonists of recapitulation, particularly palaeontologists such as A. Hyatt (ammonites), R. Jackson (echinoids and bivalves), and C . E . Beecher (trilobites and brachio­ pods) . With the awareness that these so-called de­ generate forms were at least as common as examples of recapitulation, the Biogenetic Law began to slide into oblivion . I n the nineteen-twenties W. Garstang recognized that ontogeny did not always recapitulate phy­ logeny - it created it. Garstang believed that the retention of ancestral juvenile characters by de­ scendant adults, which he termed paedomorphosis, was the key to understanding the evolution of many major groups of organisms, in particular the evol­ ution of vertebrates from tunicate larvae . However, recent research has shown that both paedomor­ phosis and 'recapitulation' play important roles in evolution (Gould 1977; Alberch et al. 1 979; McNamara 1986a; McKinney & McNamara 199 1 ) .

the timing of onset or cessation of morphological development and size change can also produce heterochrony . If size alone changes between ances­ tor and descendant, dwarfs or giants are produced. If the rate of shape change is increased, or its period of operation is extended, the descendant adult passes morphologically 'beyond' the ancestor: this is peramorphosis (this equates, to some degree, with the Haekelian 'recapitulation') . Conversely, if the rate of shape change is reduced, or its period of operation is contracted, the descendant adult passes through fewer growth stages, so resembling a juven­ ile stage of the ancestor: this is paedomorphosis . These terms can be applied not only to the appear­ ance of meristic characters (in other words, individ­ ual structures produced during an organism' s ontogeny) but also to subsequent changes of shape of these structures during ontogeny . Thus not only may the rate of induction of structures vary, but the structures which are produced may show phylo­ genetic changes as they vary their rate of shape change . These two basic forms of heterochrony are known respectively as differentiative heterochrony and growth heterochrony (Figs 1, 2) . The relationship between size and shape is known as allometry. If the relative size and shape of a structure remain the same relative to overall body size during ontogeny, growth is isometric. However, if a particular structure increases in size relative to the whole organism, as well as changing its shape, growth shows positive allometry. Should a structure decrease in relative size, growth shows negative allometry. Increasing the degree of allometry is expressed phylogenetically as peramorphosis . Reducing it produces paedomorphosis . Similarly, extending or contracting the period of allometric growth produces peramorphic or paedomorphic descendants respectively . Paedomorphosis and peramorphosis are morpho­ logical expressions of heterochronic processes. Pae­ domorphosis can occur by progenesis, neoteny, or post-displacement (Fig . 1) . Peramorphosis can occur by hypermorphosis, acceleration, or pre-displacement (Fig. 2) .

Nomenclature Heterochrony involves the decoupling of the three fundamental elements of growth: size, shape, and time, or the extension or contraction of these el­ ements . Temporal changes of size and shape relative to one another produce heterochrony, when either size or shape, or both, are affected by changes in their rate of ontogenetic development . Changes to

111

2 The Evolutionary Process and the Fossil Record

112

Pr o g e n e s i s c

1

Progenesis often occurs by precocious sexual maturation . Consequently morphological and size development is prematurely stopped, or severely retarded . The resultant adult paedomorph will be smaller than the ancestral adult . The prematuration morphological history of both the progenetic form and its ancestor will be identical . Progenesis is often global, affecting all structures, but it may also affect local growth fields . Some characters, how­ ever, are likely to have a more distinctly juvenile appearance than others . Thus, in the fossil record, it is generally possible to deduce the operation of progenesis : the morphotype is smaller than its presumed ancestor and resembles a juvenile stage of the ancestor. It will, however, be appreciably larger than the corresponding ancestral juvenile stage .

1

Post-d i s p l acement A

B

00 8

c

Neoteny is reduction in rate of morphological devel­

1

N eo t e ny

A

B

c

o H

Post-displacement occurs by changes to the timing of

11

A n cestor B

08

C

G

opment during juvenile growth . If maturity occurs at the same age in both ancestor and descendant, they will be the same size . Often onset of sexual maturity is delayed in neotenic forms, consequently the neotenic forms attain a larger adult size. Neoteny may be global in its effects on the organism, or dissociated when it affects only certain morphologi­ cal structures . Reduction in degree of allometry of specific structural elements will result in neoteny. Unlike progenesis, where the juvenile ontogenetic trajectories of ancestor and descendant are alike, juvenile growth trajectories are different between ancestors and descendants .

D

onset of growth of particular morphological struc­ tures . Thus, by comparison with the ancestor, a structure commences development at a later stage, compared with other parts of the organism . Should subsequent development and cessation of growth be the same in the descendant as in the ancestor, the displaced structure will attain a shape at maturity resembling that found in a juvenile of the ancestral form. The displaced structure is also likely to be smaller than in the ancestor .

Hypermorphosis occurs by extending the juvenile 1 The relationship of the three paedomorphic processes to the ancestor. Progenesis occurs by precocious maturation, post-displacement by the delayed onset of growth, and neoteny by reduced rate of morphological development. Differentiative paedomorphosis is shown by the spine production, growth paedomorphosis by the central spot. (From McNamara 1986a . ) Fig.

growth period, by a delay in the onset of sexual maturation . Early juvenile development will progress at the same rate as in the ancestor. By extending growth allometries to a larger size, the hypermorphic adult can attain morphological charac­ teristics quite distinct from those of the ancestral adult . Like progenesis, hypermorphosis is often global in its effects, but it too can affect only local growth fields .

2 . 4 Heterochrony

1

Hype rmorphos i s B

c

E

D

Prp.ri i , n l .::ll rpmpnt D

A"C' D

Acceleration of rate of morphological development during ontogeny will produce a peramorphic descendant . In allometric terms, acceleration is an increase in the degree of allometry . For meristic characters it is an increase in the rate of production of structures . If acceleration is operating only on specific structures, then there need be no overall increase in body size . However, the particular struc­ ture is likely to be larger . As with neoteny, juvenile ontogenetic growth trajectories will be different in the ancestor and descendant.

Pre-displacement involves the earlier onset of growth of a specific structure . This allows a longer period of growth and development . Ancestral allometries will therefore, in effect, be extended . The resultant struc­ ture will be more advanced morphologically and larger than the equivalent structure in the ancestral adult, so long as cessation and rate of growth are identical in ancestor and descendant .

E

Acce l e rat i o n

c

1 13

''11

The identification of heterochronic processes in the fossil record is generally based on the precept that these processes can be characterized by the study of size and shape alone . The assumption is made that size is a proxy for time : the larger the organism, the longer period of time it took to reach that size. This assumption may not always be valid . She a (1983) has suggested that two forms of pro­ genesis and hypermorphosis can be recognized . In the first, time and size are not dissociated; thus smaller size correlates with shorter time, larger size with longer time . This Shea calls 'time hypomor­ phosis ( progenesis)' . The corresponding pera­ morphic process is time hypermorphosis . In the second case the progenetic form attained its reduced size and shape in the same amount of time that the ancestor took to attain maturity. This occurred because the rates of size and shape change were equally reduced through ontogeny compared with the ancestor . This Shea termed 'rate hypomorphosis ( progenesis)' . Time and rate progenesis or hyper­ morphosis can theoretically be distinguished in the fossil record . Early ancestral and descendant onto­ genies will be the same when time progenesis has occurred, whereas they will differ in rate progenesis . Future emphasis on the study of growth lines in suitable invertebrate groups, such as molluscs, corals, and echinoids, will allow the true rates of growth of fossil organisms to be ascertained (McKinney 1988) . =

=

Fig. 2 The relationship of the three peramorphic processes to the ancestor . Hypermorphosis occurs by delayed sexual maturation, pre-displacement by earlier onset of growth, and acceleration by increasing the rate of morphological development. The spines and central spot demonstrate differentiative and mitotic peramorphosis, respectively. (From McNamara 1986a . )

114

2 The Evolutionary Process and the Fossil Record

Heterochrony at different hierarchical levels While most of the literature dealing with hetero­ chrony as a factor in evolution concentrates on its role at the specific or supraspecific level, it needs to be stressed that much recognized intraspecific morphological variation in populations is, in fact, engendered by heterochronic processes . These act upon both meristic and allometric traits . For instance, intraspecific variation in ammonites often involves variation in the numbers of ribs or tuber­ cles generated at a certain size . Similarly, in echinoids intraspecific variation often involves dif­ ferences in the rate of production of meristic charac­ ters, such as the number of coronal plates and spines . Variation in numbers of these structures between two individuals of the same size may be accounted for either by variations in rates of devel­ opment (neoteny or acceleration) or by onset and offset of growth (pre- or post-displacement and progenesis or hypermorphosis) . However, variations in rate of size increase may also produce such intraspecific differences . Thus if two individuals from a single population each 20 mm in length possess, in one case, six spines, and in the other eight, this may reflect a variation in rate of spine development (neoteny or acceleration), if both attained 20 mm in the same period of time . Alternatively, the individual with six spines may have increased in size at a faster rate through onto­ geny, and thus only have had sufficient time to generate six spines . It is possible to test whether this latter mechanism has occurred by analysing the developmental patterns of other structures . For instance, if one of these organisms reached a length of 20 mm faster than the other, then all of its structures should appear relatively paedomorphic. However if, as is often the case, intrapopulational variation shows some characters to be paedomor­ phic and others peramorphic, then rates of structural development will have changed . Selection o f heterochronic morphotypes, and the resultant morphological evolution of a new species, is reflected in substantial shifts in the mean values of heritable phenotypic variation of shape or size of morphological structures . These occur by pertur­ bations to the developmental programme . These may be under strong directional selection pressure (see below) . Evolution of a substantial new hetero­ chronic morphology may result in the evolution of new adaptive structures . These allow either geo­ graphical or ecological separation from the ancestral stock, and subsequent genetic isolation and estab-

lishment of a new species (see also Section 2 . 2) . In recent years documentation of heterochrony in the fossil record at the interspecific level has been undertaken in particular on ammonites (see McKinney 1988), echinoids (McNamara 1988), and trilobites (McNamara 1986b) . It has been suggested (McNamara 1982) that heterochrony may be one of the factors responsible for rapid speciation events. This is particularly so where progenesis or hyper­ morphosis have occurred . However, gradual, phy­ letic changes may equally well be engendered by small modifications in growth rates between popu­ lations, resulting in subtle shifts in morphology through time. Heterochrony has been proposed as a major factor in evolution at the supraspecific level . For instance, the orthodox view of the origin of vertebrates is that they may have arisen from the pelagic larva of a tunicate-like deuterostome invertebrate . This would have occurred by progenesis from an early larval stage . The free-swimming tunicate larva possesses all the fundamental chordate characters : a noto­ chord, dorsal hollow nerve cord, gill slits, and post­ anal propulsive tail. Attainment of precocious sexual maturation would have caused the retention of such ancestral larval characters into the adult phase and a consequent major adaptive breakthrough . The earlier that perturbations to the embryo­ logical developmental system occur, the more profound the morphological consequence . Taxo­ nomically, this is likely to be expressed at a high level. For instance, it has been suggested (McNamara in McKinney 1988) that progenesis at early developmental stages has been instrumental in the evolution of a number of higher taxa: saleniid, tiarechinid, neolampadoid, and clypeasteroid echinoids; edrioasteroids; baculitid ammonites; thecideidine and craniacean brachiopods; and branchiosaurid amphibians (Fig. 3) . Other heterochronic processes have also been instrumental in the evolution of higher taxa . For instance, it has been proposed that birds may have evolved from theropod dinosaurs . The very large orbits of birds, their inflated braincase, retarded dental development, and overall limb proportions indicate that early birds may have been paedomor­ phic theropods. Feathers are thought to have been present on juvenile theropods . The paedomorphic processes were probably neoteny and post­ displacement .

2 . 4 Heterochrony

115

include the evolution of a number of anagenetic paedo- and peramorphoclines in spatangoid echin­ oids, such as Schizaster, Hemiaster, Lovenia, Pericosmus, and Protenaster (Fig. 4) . All show evolution from coarse to fine-grained sediments (probably shallow to deep water) . Conversely, the Cenozoic brachiopod Tegulorhynchia evolved along a paedomorphocline from deep to shallow water into the genus Notosaria (Fig . 5) . Similarly a number of trilobite lineages are thought to have evolved by heterochrony along the same environmental gradient (McNamara 1986b) . In the marine environment changing water depth and sediment type are frequent environ­ mental gradients along which paedo- and peramor­ phoclines develop .

Ecological causation of heterochrony

Fig. 3

Reconstruction o f a paedomorphic branchiosaurid amphibian.

Heterochrony and directed speciation The pattern that is emerging from studies of hetero­ chrony in the fossil record is one of frequent directed heterochronic speciation . The direction of morpho­ logical evolution is strongly constrained by the nature of the organism's own ontogeny. Thus a number of characters in a lineage may show progressive paedomorphosis or peramorphosis . Provided that the descendant morphotypes are suitably adapted along an environmental gradient, a phylogenetic trend, in the form of a paedomorpho­ dine or peramorphodine, may develop . The environ­ mental and morphological directionality may be induced by the effects of either competition or predation. With induction of the heterochronic morphological gradient by competition, the persist­ ence of the ancestral form constrains selection to one direction: along the environmental gradient away from the ancestral species . The phylogenetic pattern generated will be one of cladogenesis . Selec­ tive pressure from predation in one environment may induce the evolution of a paedo- or peramor­ phocline . In this case the phylogenetic pattern is one of anagenetic speciation . Recent studies of echinoids, brachiopods, bi­ valves, ammonites, graptolites, and ammonites (see McNamara in McKinney 1988) indicate that the anagenetic pattern is common . Specific examples

While many of the examples of directed hetero­ chronic evolution have been interpreted as having arisen by selection of morphologically adaptive characters, it has also been argued that other factors, such as life history strategies, which affect elements such as size and time of maturation, may also be targets of selection. McKinney (1986) has suggested that for a suite of Tertiary echinoids selection favoured large forms along an environmental gradi­ ent from shallow to deep water (equating with unstable to stable environments) (Fig. 6) . He argued that any subsequent morphological changes were incidental allometric by-products of the size change . The larger size was attained either by slower, neo­ tenic growth or by extended, hypermorphic growth . This indicates that the target of selection was repro­ ductive timing and/or body size . Such size increase along lineages (Cope's Rule; Section 2 . 10) may reflect K-selective pressure (large body size, delayed reproduction and development, and longer life spans in a stable environment) . While analyses of other echinoid lineages does not provide unequivocal corroboration of this pattern, there is ample evidence that many pro­ genetic species are, conversely, r-selected (small body size, early maturation, high fecundity, and short life span in an unstable environment) . Many so-called 'dwarfed' faunas may be r-strategists, inhabiting unstable, fluctuating environments . The small body size of progenetic Late Cretaceous oys­ ters and ammonites may have been an adaptation to a soft, unstable substrate . The same is true for many progenetic brachiopods . High fecundity of progenetic species has been documented in edrioas­ teroids and trilobites (McNamara in McKinney

2 The Evolutionary Process and the Fossil Record

116

Ontogeny

== ==> = :::== ========= c:::==

Periporal a rea

R EC ENT -T'"" l E!>! I S=� C:'= P-:-: O'-> E N E --I

B

?��

• • •

"C,

d

T.

PI�iio;m P.

doederleini

australis

PLlOCENE

MIOCENE

U

Z

w

. .

..

.

o S:2

--... ...P.antia us tralis

o

. . . . _--_. P. philipi

T.

coela ta

T.

U 8

OLlCOCENE

Z

squamosa

w

P.

prea ustralis

U 8 ;f: Z

w

E O.C E N E

...J

Se d i m e n t g ra i n size

", '

•

,, ' •

. .. Q •

•

•

= :> == = === :::== c:::=== E n v i ro n== m e ntal ==== grad i e== nt

Fig. 4 Peramorphocline in the evolution of the periporal area in the Cenozoic spatangoid echinoid Protenaster. Later species underwent greater morphological change in this character than earlier species. This led to the ability of later species to feed from progressively finer grained sediments . (From McNamara 1985 . )

1988) . Although in these cases small size and pre­ cocious maturation may have been the principal targets of selection, unless the resultant progenetic morphology was also adaptively significant in the new environment, selection would not have occurred . Most heterochronic changes occur a s a result of changes to the internal developmental regulatory system. However, certain changes may actually be induced by environmental perturbations . The effect of changing environmental pressures may lead to facultative heterochrony within populations . For instance, the frequency of development of pae­ domorphs in living populations of salamanders is directly influenced by the population density . When low, a large proportion of individuals develop as neotenic paedomorphs, attaining maturity in their larval form, so remaining and reproducing in the aquatic environment . At high population densities

T.

�

= > =====::::=== =� > ::::=.::=::

O NTOG E N Y

boongeroodaensis ( j u ve n i l e )

T.

/"�'"

.�

boongeroodaensis ( ad u l t )

Fig. 5

Paedomorphocline in the evolution of the umbonal region of the Cenozoic rhynchonellid brachiopods Tegulorhynchia and Notosaria, showing paedomorphic retention of larger foramen (and hence thicker pedicle) and smaller umbonal angle. The environmental gradient along which the paedomorphocline evolved was from deep to shallow water. (From McNamara 1983 . )

few individuals are paedomorphic, the high density levels inducing metamorphosis to the terrestrial form . There is also indirect evidence from the fossil record (McNamara 1986b) that changes in water temperature at different water depths in the marine environment may have been a factor in inducing progenesis in a number of lineages of Cambrian trilobites . Experimental work has demonstrated the effect of higher temperatures in inducing premature maturation in some living arthropods .

2 . 4 Heterochrony Proge n e s i s/ acce l e rat i o n

OIl

r OIl

LOW

..

•

S e l ectio n

•

N eot eny/ hype r m o r p h o s i s

Stab i l ity

•

HIGH

H eteroch rony

O rgan i s m s i z e

•

k

LARG E

Water depth

Fig. 6

Diagrammatic summary of McKinney's (1986) proposed relationship between water depth, ecological stability, selection, organism size and heterochronic process, based on a suite of Tertiary echinoids from southeastern U.S.A.

Frequency of heterochrony in the fossil record Any attempt to assess the frequency of heterochrony or of the heterochronic processes is fraught with problems, not the least of which are historical prejudices . While the Haeckelian school were blink­ ered to the existence of paedomorphosis, the Garstang/de Beer school were equally contemptuous of peramorphosis . In a recent survey of palaeon­ tological literature from 1976 to 1985, McNamara (in McKinney 1988) documented 272 examples of heterochrony; of these, 179 were of paedomorphosis, the remaining 93 were of peramorphosis . The most comprehensive recent analyses of heterochrony in the fossil record have centred on trilobites, echinoids, ammonites, bryozoans, and graptolites . These studies have shown both pae­ domorphosis and peramorphosis to be important factors, but paedomorphosis still predominates . The greater frequency of paedomorphosis, if true, may occur because existing developmental programmes are utilized . Peramorphosis always requires the pro­ duction of novel bauplans by extending the pre­ existing developmental pathways . Trilobites show a changing relative frequency of paedomorphosis and peramorphosis . During the Cambrian paedomorphosis, particularly that in­ duced by progenesis, was predominant . Post­ Cambrian forms, however, show a marked decline in the incidence of progenesis and a greater fre­ quency of peramorphosis . It has been argued (McNamara 1986b) that this change may reflect an improvement in regulation of the developmental system in later trilobites .

117

Recent analysis of heterochrony in irregular echinoids (McNamara 1988) has highlighted the complex activity of heterochrony in single lineages, some characters being paedomorphic, others pera­ morphic. Furthermore, one paedomorphic structure might have evolved by neoteny, whilst another might have formed by post-displacement . The operation of a complex array of heterochronic pro­ cesses has been termed mosaic heterochrony. With each structural element of an organism essentially following its own ontogenetic trajectory, and each being potentially subject to changes in develop­ mental regulation, there is the possibility of the evolution of a multitude of heterochronic morpho­ types . Any one of these may potentially form a new species, with the target of selection being the result­ ant morphotypes, size, or life history strategies . Ammonites have featured prominently in studies of heterochrony for over 100 years . They were used initially as examples of 'recapitulation' , by Hyatt and co-workers; while to O. Schindewolf and other workers in the first half of this century they showed evidence only of paedomorphosis . Recent research (see McKinney 1988) has shown the ubiquity of both phenomena, but peramorphosis, particularly of the septa, appears more common than paedomor­ phosis . Colonial organisms, such as bryozoans and grap­ tolites, sJ;tow a two-tiered heterochronic pattern . Both the individual animals and the colony as a whole may be affected by heterochrony . The former is known as ontogenetic heterochrony, the latter as astogenetic heterochrony. This two-tiered structure is comparable with the two-tiered structure of dif­ fentiative and mitotic heterochrony present in non­ colonial organisms . Astogenetic heterochrony has been reported more often than ontogenetic hetero­ chrony (McKinney 1988), perhaps because astoge­ ne tic changes reflect developmental modification of ontogenetic characters, so reflecting the individuality of the colony as a whole . Heterochrony in colonial organisms may have been important in macroevo­ lution . Ontogenetic heterochrony in highly inte­ grated colonies may result in large morphological differences between ancestor and descendant . It would appear that periods of reef building cor­ respond to periods of high integration in colonial animals . It is likely, therefore, that astogenetic heterochrony will predominate during periods of reef building . Many of the examples of heterochrony involving vertebrates occur in amphibians . Most of these show paedomorphosis, akin to that seen in living

118

2 The Evolutionary Process and the Fossil Record

salamanders . Many of the interspecific and inter­ generic differences in allometry of skull plates in Palaeozoic fishes are the result of mitotic hetero­ chrony, though few studies have actually couched it in these terms . Similarly, phylogenetic changes in limb allometries in mammals are due to hetero­ chrony . The limited evidence from studies of mammals seems to suggest a predominance of peramorphosis over paedomorphosis . This may occur because of the frequent operation of Cope's Rule in mammal lineages, suggesting size as being an important target of selection . For example, exten­ sion of ancestral allometries by increased size in horses through the Tertiary resulted in peramorphic descendants by hypermorphosis . However, some characters, such as development of the foot, show paedomorphic reduction in some digits .

Developmental processes underlying heterochrony Changes to the onset, offset and rate of growth of morphological characters are essentially under three interactive levels of control: genetic, hormonal, and cellular. Perturbations to the genetic regulation of hormonal and cellular development, particularly at early embryological stages, are likely to be critical factors in heterochrony . Developmental regulation is not simply a matter of discrete entities called 'regulatory genes' acting upon 'structural genes' . It involves a complex inter­ action between active sites or structural components of proteins, combined with cell - cell interactions (Campbell & Day 1987) . Developmental processes are controlled by highly organized, dynamically structured multigene families . The manner in which the genome is encoded and expressed in develop­ ment is far from clear, although it would seem that only a small area of the highly dynamic, constantly changing genome is occupied by genes for development . The region involved in regulation in a typical eukaryote gene is the promotor region . This contains DNA binding proteins specific to the gene, and capable of controlling the level of transcription . The role of the promoter sequence in gene control, and its effect on growth, highlights the activity of hor­ mones in growth, and how perturbations to the genetic control of hormone production can have a strong phenotypic expression . Growth, moulting, and sexual reproduction in arthropods, for instance, are all under hormonal control. It has been suggested (Campbell & Day

1987) that hexapods evolved from a myriapodous ancestor by progenesis : a small change in the genetic control of the hormone responsible for the incep­ tion of maturation, and of the hormone controlling post-larval development, had a profound effect on the phenotype . Even within fossil lineages the activity of genes controlling hormone production can be inferred . Two forms of progenesis in trilo­ bites have been identified : sequential and terminal (McNamara 1986b) . Terminal progenesis is likely to have occurred by a premature cessation in pro­ duction of a juvenile hormone . Sequential pro­ genesis, where each intermoult period is shortened, is thought to have occurred by premature pro­ duction of an ecdysone-like moulting hormone during each intermoult period. This premature hormonal activity will have been under direct genetic control . The third factor in the developmental processes that cause heterochrony is activity at the cellular level . Hall (1984) has stressed the importance of the number and mitotic activity of the cells in the initial skeletal condensation in vertebrates . Thus onset of growth is determined by the number of stem cells that start condensation, the proportion that divides, rate of cell division, and amount of cell death . These parameters all act early in development and deter­ mine the time of onset of growth . The rate of growth of skeletal elements is influenced by adjacent tissues, hormones, and allometric factors . Muscle action, tendon insertion, blood flow, innervation, and growth of adjacent tissues modulate the growth rate . Cessation of growth is partially determined very early in development by the number of growth plate cells and the number of times they divide . Timing of development of secondary ossification centres also affects the offset signal . Metabolic inhibition by production of a growth inhibitor to suppress cell proliferation and protein synthesis also stops growth .

References Alberch, P., Could, S .J . , Oster, C . F . & Wake, D . B . 1979 . Size and shape in ontogeny and phylogeny. Paleobiology 5, 296 -317. Campbell, K.S.W. & Day, M . F . (eds) 1987. Rates of evolution . AlIen & Unwin, London. Could, S.J. 1977. Ontogeny and phylogeny. Belknap Press, Cambridge . Hall, B . K . 1984. Developmental processes underlying hetero­ chrony as an evolutionary mechanism. Canadian Journal of Zoology 62, 1 - 7. McKinney, M.L. 1986. Ecological causation of heterochrony:

2 . 5 Red Queen Hypothesis test and implications for evolutionary theory. Paleobiology 12, 282 -289 . McKinney, M . L . (ed . ) 1988. Heterochrony in evolution: a multi­ disciplinary approach . Plenum Press, New York. McKinney, M . L . & McNamara, K.J. 1991 . Heterochrony: the evolution of ontogeny. Plenum Press, New York. McNamara, K.J. 1982 . Heterochrony and phylogenetic trends. Paleobiology 8, 130 - 142 . McNamara, K . J . 1983 . The earliest Tegulorhynchia (Brachio­ poda : Rhynconellida) and its evolutionary significance. Journal of Paleontology 57, 461 -473 . McNamara, K.J. 1985 . Taxonomy and evolution of the Caino­ zoic spatangoid echinoid Protenaster. Palaeontology 28,

119

31 1 - 330 . McNamara, K.J. 1986a. A guide to the nomenclature of hetero­ chrony. Journal of Paleontology 60, 4 - 1 3 . McNamara, K.J. 1986b . The role o f heterochrony i n the evol­ ution of Cambrian trilobites . Biological Reviews 6 1 , 121 - 156 . McNamara, K . J . 1988 . Heterochrony and the evolution of echinoids. In: e . R . e . Paul & A . B . Smith (eds) Echinoderm phylogeny and evolutionary biology. Oxford University Press, Oxford . Shea, B.T. 1983. Allometry and heterochrony in the African apes . American Journal of Physical Anthropology 62, 275 - 289 .

2 . 5 Red Queen Hypothesis M . J . BENTON

Introduction Palaeontologists have long argued that the distinc­ tive features of the evolution of life were produced by changes in the physical environment. Changes in climate, or in sea-level, for instance, might explain why certain groups died out, or why an adaptive radiation took place at a particular time . This trend has continued in recent research into mass extinc­ tions (Section 2 . 12), whether their cause is said to be changes in the earthbound physical environ­ ment, or the impact of asteroids . O n the other hand, many ecologists have viewed the large-scale aspects of evolution (macroevolution) as simply a scaled-up version of microevolution . Evolutionary change, they argue, can be produced by competition between organisms, and by inter­ actions between predators and prey . This ecological view stresses the influence of the biotic environ­ ment, that is, other plants and animals, on evolution .

Van Valen's Law The ecological view of macroevolution was codified by Van Valen (1973), who presented palaeontologi­ cal and ecological evidence for a model of evolution that depended on the biotic environment, and termed the model the Red Queen Hypothesis . The palaeontological evidence was based on a study of

the rates at which different groups of plants and animals go extinct through time . Van Valen used plots of species survivorship (Fig. 1) which showed the proportions of an original sample of organisms that survive for various intervals . He found, contrary to his expectations, that the probability of extinction within any group remained constant through time - his Law of Constant Extinction. For example, families or species of modern mammals are just as likely to become extinct as were their Mesozoic ancestors living 200 Ma. A species might disappear at any time, irrespective of how long it has already existed . Evolutionary biologists might have intuit­ ively expected species within any group to become longer-lived over time on average . Van Valen's start­ ling discovery seemed to deny some basic assump­ tions of evolution . If evolution is taken to mean improvement in the adaptation of a species to its environment through time, why is it that modern mammals are not better at surviving than their Mesozoic forebears? Van Valen's explanation for the Law of Constant Extinction was that the various species within a community maintain constant ecological relation­ ships relative to each other, and that these inter­ actions are themselves evolving . Thus, the antelope on an African savanna, for example, evolves greater speed in order to escape from the lion, but the lion

120

2 The Evolutionary Process and the Fossil Record

A

•

( ,,)

1 000 -

•

also evolves greater speed in order to catch its dinner. The status quo is maintained . If biotic inter­ actions did not follow a pattern of ever-moving dynamic equilibrium, the community would be shattered . Were all antelopes to achieve a quantum improvement in their running speed, lions would starve and populations of antelope might outstrip the carrying capacity of the environment. This balance is the Red Queen Hypothesis. [In Lewis Carroll's Through the Looking Glass, the Red Queen told Alice, 'Now here, you see, it takes all the running you can do, to keep in the same place' . ] After 1 973, many biologists accepted the Red Queen model, while others were critical . The main problem was simply the counter-intuitive claim that species do not improve their chances of survival through time . If organisms are continuously evol­ ving and adapting, why do they not get any better, on average, at avoiding extinction? There were prob­ lems also with Van Valen' s particular formulation of the Red Queen model . He explicitly made a zero­ sum assumption : that there were fixed amounts of energy available to communities, and that any gain by one species was exactly offset by equal losses to others . It is not at all clear, however, that the amounts of energy, or resources in general, have remained constant through time . It is equally prob­ able that the total global biomass has increased markedly many times as major new habitats were exploited (e . g . the move onto land (Section 1 . 8), the evolution of 'trees', the origin of flight (Section 1 . 9), and the evolution of deep-burrowing habits by various marine groups (Section 1 . 7. 1 ) ) . The increase in biomass is possible by pulling more of the global carbon into the biotic part of the biogeochemical cycle, and/or by speeding up the rate at which carbon, and other essential elements, are cycled through the system (Benton 1987) .

.

"""

•

Ext i n ct

" Living

1 00 -

•

•

. .

"' '''

••

-

1 0 --= -

-

'"

•

•• ...

.._.

I

I 200

I

o

(1 06 y r s ) B •••••

1 000 c:

•

•

•

••

••

1 00 -:

• •

•

•

•

•

•

•

", ,,, ,,, ,,, ,,,

1 0 -: •

.. ••• " "'.

E xt i n ct

'" L i v i n g

o

•

I

10

••••

T 30

Fig. 1 Taxonomic survivorship curves for A, pelecypod (bivalve) and B, mammalian genera (major Therian orders). These graphs show roughly linear trends, and indicate constant probabilities of extinction for genera of any duration . (After Van Valen 1973 . )

Problems with 'progress', and the Stationary Model Although the Red Queen model does not predict improvements in the ability to avoid extinction, it does explicitly assume that, within any lineage, later members will be competitively superior to earlier ones: that a present-day antelope can run faster than its Pliocene forebear, that modern mammals are clearly competitively superior to their Palaeo­ cene ancestors . This notion of progress is frequently assumed by biologists and palaeontologists, but is probably impossible to test directly . Nevertheless, simple assumptions of progress of this kind have

2 . 5 Red Queen Hypothesis been criticized recently (Benton 1987) . There seems to be no adequate way yet of demonstrating 'pro­ gress' in macroevolution, least of all competitive improvement. The evolution of horses can be taken as a well known example of an adaptive trend, or record of improvement through time . The early small leaf-eating horses of Eocene times gave way to larger animals with fewer toes (greater running speed) and deeper teeth (for grinding up the new silica-bearing grasses) in the Miocene (Section 1 . 1 1 ) . However, i f the fossil record were reversed, we could equally well demonstrate how the horses adapted to the diminishing grasslands by becoming smaller forest-dwellers, living a cryptic life and switching to a diet of tree leaves. Where is the progressive improvement of competitive ability? The whole question seems to hinge on how macro­ evolution is viewed . If organisms are generally very well adapted, finely tuned by natural selection, and if the physical environment has only minor effects, the Red Queen Hypothesis has to hold . If, on the other hand, organisms are viewed as only moder­ ately well adapted, natural selection as only a spor­ adic force for evolutionary change, and the physical environment as an important influence through local and global extinction, and radiation events, then the Red Queen Hypothesis cannot be correct. In 1984 Stenseth and Maynard Smith formalized an alternative to the Red Queen model, termed the Stationary Model. This model assumes that evo­ ution is driven mainly by abiotic factors, and that it will cease in the absence of changes in the physical environment . The two models make very different predictions and, as Stenseth and Maynard Smith (1984) wrote, 'the choice between the Red Queen and Stationary Models will have to depend primar­ ily on paleontological evidence' . The Red Queen model predicts that the rates of speciation, extinction, and phyletic evolution will remain constant in ecosystems, even when the diversity of species has reached equilibrium so that the numbers of species do not change . The Station­ ary Model, however, predicts that at equilibrium no evolution will occur . Bursts of evolution, extinction, and speciation will happen only in response to changes in the physical environment. These two models can be visualized by plots of species sur­ vivorship over time, which gives a measure of the rate of extinction (Fig . 2A, B) .

Testing the models Hoffman and Kitchell (1984) applied a palaeonto-

121

logical test. The first problem they encountered was to find an example spanning several million years in which no environmental change had occurred . Such a case i s highly unlikely, and it proved neces­ sary to make allowances for episodic perturbations in the physical environment. The modified patterns are still distinctive (Fig . 2C, D) . The Red Queen model predicts an approximately regular decline in the number of species surviving (that is, constant extinction), with occasional changes of slope that correspond to major environmental perturbations . The Stationary Model predicts a distinctly stepped pattern, with constant numbers of species at equilibrium, and sudden extinctions at times of environmental change . Hoffman and Kitchell (1984) also examined the records of microfossils (coccoliths, foraminiferans, radiolarians, diatoms, and others) from 1 1 1 deep­ sea boreholes through the past 50 million years of sediments of the Pacific Ocean floor. The species survivorship curves obtained from these data (Fig . 3) are more or less smooth, rather than stepped, and they seem to support the Red Queen model . An analysis of the cumulative appearance of new species also gave general support to the Red Queen model, although there was some evidence of stepping . Further analysis shows there to be considerable variation in the probability of extinction over geo­ logical time : for example, there seem to have been particular periods in which all the microfossil groups had high extinction rates. These indicate plankton extinction events which would normally be attributed to sharp changes in the physical environment . When Hoffman & Kitchell (1984) made allowances for these events, the various analyses again pointed to the Red Queen model. Another test, also using the plankton record, was carried out by Wei & Kennett (1983) . Their study was based on the fossil record throughout the world of 149 species of foraminifera over the past 24 million years. They found that major changes in rates of extinction and speciation corresponded to palaeoceanographic perturbations (Fig . 4), and they regarded their data as consistent with the Stationary Model . These two studies illustrate some of the practical difficulties involved in testing the Red Queen model . One serious problem is in separating biotic from abiotic factors in order to assess their relative significance : it is probably impossible to pigeon­ hole both kinds of phenomena as independent factors . Secondly, in many real situations, and pos-

2 The Evolutionary Process and the Fossil Record

122

C o n stant e n v i r o n m e n t B

A 1 00 bO C

;;:

.� .

:::J '" '" Q)

'u

Q)

1 00 ,-----Stat i o n a ry model

Red Q u ee n model 10

10

a. Vl

B

A 5

5

10

10

Species d u rat i o n

S p e c i e s d u rat i o n

Pe r i o d i c a l l y d i st u rbed e n v i r o n m e n t

C

D

1 00 bO C

;;:

.� .

:::J

'" Q)

'u

'"

1 00 bO C

;;:

.�

Red Q u e e n model

.

:::J '" '" Q)

10

'u

Stat i o n a ry model 10

Q) a. Vl

Q)

a. Vl

C

Fig. 2

D 5

10

S p e c i e s d u rat i o n

sibly including the two described above, the test would be inconclusive . For example, it would be hard to distinguish between two predicted patterns if both curves were stepped as a result of rapid changes in the physical environment. At the other extreme, continuous small changes in environ­ mental conditions might give similar gently sloping graphs for both models . Furthermore, as Hoffman and Kitchell (1984) pointed out, both the Red Queen and Stationary models assume that ecosystems tend towards a diversity of species that is at equilibrium . Both hypotheses predict that the total number of species is constant: the addition of one species to the system causes the loss of another . Yet models that do not assume such an equilibrium could also account for the Law of Constant Extinction and for the other palaeontological data . In these models, there would

5 S p e c i e s d u rat i o n

10

Predictions o f the Red Queen and Stationary models for species survivorship in constant (A, B), and periodically disturbed (C, D) environments. (After Hoffman & Kitchell 1984 . )

be no necessary limit to diversity, and the rates at which species arise or go extinct would not be correlated with each other, nor with total diversity. The data so far are equivocal on these points . Other research seems to count against the Red Queen Hypothesis . Kitchell et al. (1989) studied a simple predator - prey relationship : naticid gastro­ pods and bivalves . The naticid gastropods prey on bivalves by boring through their shells and extract­ ing the flesh . Both groups are plentiful as fossils, and such predation has left identifiable borings in fossilized bivalve shells . Kitchell and her colleagues modelled the predator -prey system mathematically and found that, whatever the starting point, the system tended to a static position . The bivalves evolved either to reproduce early (before they were eaten) or to devote all of their energy to building a thick shell to minimize the chance of successful

123

2 . 5 Red Queen Hypothesis

00 c

.�

.:;

:J

A

'" 2 . 1 6 '� ,,0 "f", 1 .92 ".b", "k , 00 , ,, 1 . 68 " 2 .40

"

'u Q) Of>

Q) 0..

00

Of>

-'0

C

1 . 92 1 . 68

1 . 68

1 . 44

1 .44

I

'..,cm "l'"" "'-,()

1 . 20

'?ro '0

1 . 20

0 0

'" 2 . 1 6 ,,.00 ,� ,0000 1 . 92

2 .40

2.16

0 <xl

1 . 44

Of>

B

2 . 40

.96 . 72

_ 0 ,11""" aDID

.96

-

.24

,

0 '-:----:.':---'"--="'---"'--=0::-- ' '& 50 40 30 0 20 0 10

.48 .24

1 . 20

..,

. 96

o

. 72

IIIGllDI

.48

,

. 72

0

.48

=

o

10

20

30

40

50

.24 60

°0

,, 00 ,. ,

,i'

,a.o """0 0 ,

, a>

'';?

0 10

,d 30

I 40

I 50

I 60

Du rat i o n (1 06 yrs) Fig. 3 Species survivorship curves for A, coccoliths; B, radiolarians; and C, planktic foraminifera. Each graph contains data for three cohorts - taxa arising in the earliest Pliocene, in the latest Pliocene, and in the Holocene, and each is represented by a different symbol. (After Hoffman & Kitchell 1984 . )

Tempe rate

Tra n s i t i o n a l 0.5

0 P l e i stoce ne 2.0 4.0

Pl iocene

-''"

8.0 10.0

>'f::, 1 2 . 0 s Q)

.§ I-

1 4 .0

Q) C Q) U 0

'i

�

"0 "0

'i

20.0

N 21

N 16 N 15

N 14 N 12 N ll

N 10 N9

16.0 18.0

N 22

.'l:!

6.0

-;;;

0 . 1 0 .2 0 .3 0.4 0.5 0 .6

NB

�

>-

u.J

N7

N6 N5

22 . 0 24 . 0

Latest O l i gocene

Fig. 4 Extinction rates for species of Neogene foraminifera for the whole of the oceans ('total'), and for tropical, transitional, and temperate assemblages, for the past 25 million years (latest Oligocene to Recent) . The 'N' zones are standard Neogene planktic foraminiferal zones. Peaks in extinction are correlated globally, and linked to major changes in the physical environment. (After Wei & Kennett 1983 . )

124

2 The Evolutionary Process and the Fossil Record

boring attacks . This result speaks against the Red Queen model, which would require constant evo­ lution in a particular direction . Other biologists have argued that species prob­ ably do not keep running towards unattainable goals, as the Red Queen Hypothesis predicts . Each species is faced with the need to make compromises . Many bivalves, for example, have to balance the need for a strong shell against the costs of a heavy shell . The compromise solution is to have a thin corrugated shell. 'Constant running' in one direc­ tion is often not possible in a lineage, and the simplistic view of the Red Queen model as con­ tinuous and endless evolution in one direction may be denied by the limitations of genetic variation, development, and mechanical design factors .

References Benton, M.J. 1987. Progress and competition in macroevo­ lution. Biological Reviews 62, 305- 338. Hoffman, A. & Kitchell, J.A. 1984 . Evolution in a pelagic planktic system: a paleobiologic test of models of multi­ species evolution. Paleobiology 10, 9 - 33 . Kitchell, J . A . , DeAngelis, D . L . & Post, W . D . 1989. Predator­ prey interactions on the ecological and evolutionary time scale . In: N . C . Stenseth (ed . ) Coevolution in ecosystems and the Red Queen Hypothesis. Cambridge University Press, Cambridge. Stenseth, N . C . & Maynard Smith, J. 1984. Coevolution in ecosystems: Red Queen evolution or stasis? Evolution 38, 870 - 880. Van Valen, L . M . 1973. A new evolutionary law . Evolutionary Theory I, 1 -30. Wei, K-Y . & Kennett, J.P. 1983 . Nonconstant extinction rates of Neogene planktonic foraminifera . Nature 305, 218-220.

2 . 6 Hierarchy and Macroevolution N . ELDREDGE

Introduction Evolution is the scientific explanation of the design apparent in organismic nature . Natural selection is generally seen to be the principal cause of determin­ istic modification of the phenotypic properties of organisms through time . Macroevolution most commonly connotes the degree of such modification, and thus in its most general sense is simply 'large­ scale genotypiclphenotypic change' . Microevo­ lution, in contrast, refers to the relatively slight amount of change that occurs on a generation­ by-generation basis through natural selection and genetic drift . A connotation of elapsed time is often implicit in the distinction between micro- and macroevolution: microevolution takes place in relatively short amounts of time (e . g . 'ecological time,' over a few generations), while macroevolution is generally held to occur in geological time . Yet some theories of macroevolution (e . g . saltation theories of the geneticist R. Goldschmidt, or the palaeontologist O . Schindewolf) invoke brief (even single generation) genotypic and phenotypic trans­ formation . A further distinction commonly drawn

between 'microevolution' and 'macroevolution' sees the former as a within-species or perhaps within­ genus level phenomenon, in contrast with the degree of change associated more typically with the emergence of taxa of higher categorical rank (i . e . in the Linnaean hierarchy - families, orders, classes, etc . ) . The 'evolutionary synthesis', dating from the mid nineteen-thirties, forms the core of modern evo­ lutionary theory . The synthesis followed the succes­ sful fusion of Darwinian selection with an emer­ ging understanding of the principles of heredity (achieved primarily through the efforts of geneticists R.A. Fisher, J . B . S . Haldane and S. Wright) . This neo-Darwinian paradigm of drift- and selection­ mediated dynamics of genetic stasis and change was then integrated with the data of systematics, palaeontology, and other biological subdisciplines to form what was widely heralded as a unified theory of evolution . It is the general position of the synthesis that 'macroevolution' is simply microevo­ lution summed over geological time . Specifically, generation-by-generation stability and transform­ ation, mediated by natural selection and genetic

2 . 6 Hierarchy and Macroevolution drift, were held to be both necessary and sufficient to account for all aspects of the evolutionary history of life . Though the geneticist Dobzhansky (1937) and the systematist Mayr (1942) both sketched versions of macroevolutionary theory, it was left primarily to the palaeontologist Simpson (1944, 1953) to formu­ late the principles of macroevolution within the synthetic theory. The assumption that microevolution yields a complete account of the evolutionary process when projected over evolutionary (geological) time restricts the study of evolutionary mechanics to laboratory and field investigations of living organ­ isms . The role of palaeobiology in such a scheme, however, is by no means thereby rendered trivial: as Simpson (1944), for example, endeavoured to show, the integration of evolutionary theory with patterns of evolutionary events drawn from the fossil record is no simple matter. In particular, Simpson was concerned to show that it is the task of palaeontology to determine the relative intensities, and importance, of various microevolutionary pro­ cesses (e .g. mutation rate, selection, population size, etc . ) required to explain various evolutionary patterns of the fossil record . In that spirit, Simpson developed his model of 'quantum evolution' (rapid, 'all-or-nothing' modification of adaptive features of organisms in relatively small populations) to explain the relatively abrupt appearance so typical of many higher taxa . Recent years have seen an alternative view emerge on the relationship between palaeontologi­ cal data, geological time, and theories of the evo­ lutionary process. In traditional evolutionary biology, it is the phenotypic (and underlying gen­ etic) properties of organisms that are of central interest, and which 'evolve' . Organisms vary in these respects within local populations; populations are aggregated into species (Section 2.2). Natural selection 'sorts' the phenotypic attributes of organ­ isms within populations to yield (1) stasis or change in phenotypes, and (2) the emergence of new species (and, by simple extension, higher taxa) . 'Hierarchy theory' accepts the neo-Darwinian paradigm of within-population variation, selection, and drift, but seeks to extend the list of evolutionary entities beyond genes, organisms, and populations . Specifically, species, monophyletic (higher) taxa, and ecosystems have come to be viewed as having real existence, and are variously termed 'systems', 'entities', or even 'individuals' . The goal of hierarchy analysis is to elucidate the nature of each kind of large-scale entity, and thus to determine their pos-

125

sible role(s) in the evolutionary process. If large-scale systems such as species, higher taxa, and ecosystems are real entities, they exist on a spatiotemporal scale which is too large to be encom­ passed in laboratory and field experimental studies of the Recent biota. It is the fossil record that reveals the actual dimensions of such systems, and thus it falls in large measure to palaeobiology to examine how they can be integrated with existing theories of the evolutionary process. Such work has two aspects : (1) the determination of any relevance of such large-scale systems to the original problem of evolution - that is, the origin, maintenance, and further transformation of adaptive phenotypic fea­ tures of organisms; and (2) the recognition of other effects on the general history of life that may result from the existence of such larger-scale entities . Specifically, the concept that large-scale systems such as species, taxa, and ecosystems are them­ selves entities, not merely epiphenomena or simple (and perhaps arbitrarily delineated) collectivities of organisms, has led to several palaeobiological theor­ ies that allege a degree of additional process to macroevolution, over and above - and in some instances 'decoupled' from - microevolutionary processes.

Hierarchies in evolutionary biology Several meanings of the term 'hierarchy' are in general use in biology (Grene 1 987) . In the context of evolutionary theory, however, only two hierarchi­ cal systems are generally recognized: the genealogi­ cal and ecological (economic) hierarchies (Table 1 ) . Both are thought to b e implicated in the evolution­ ary process (Eldredge 1985, 1986, 1989; Salthe 1985), though some authors recognize one hierarchy and not the other. Both hierarchies consist of nested sets of entities forming distinct levels . Each level consti­ tutes a class (or category - e . g . 'species'), specific examples of which are entities or 'individuals' (e . g . Archaeopteryx lithographica) . The entities o f any given level have as parts the entities of the adjacent lower level and form, in turn, parts of the adjacent higher level: demes have organisms as parts; in turn, demes are parts of species . The entities a t each level interact o r behave in specific ways that unite them to form the entities of the next higher level . In the genealogical hierarchy the activity is 'reproduction' in the most general sense; thus, speciation is seen as the production of more entities (i . e . species) of like kind - an activity ultimately responsible for the ongoing existence of

126 Table 1.

2 The Evolutionary Process and the Fossil Record The genealogical and ecological hierarchies .

Genealogical hierarchy

Ecological hierarchy

Monophyletic taxa Species Demes Organisms Germ linea

Biosphere Ecosystems Avatars (populations) Organisms Somab

Composed of hierarchically nested chromosomes, genes, codons and base pairs. b Composed of hierarchically nested organ systems, organs, tissues, cells and proteins. a

higher taxa . In the economic hierarchy, direct inter­ action among entities of any given level cohere the entities of the adjacent higher level; thus it is the interaction among local populations of non­ conspecifics (as in predator -prey interactions) that unites them into local ecosystems . The two hierarchies arise out of the two types of organismic activity, that is reproduction, on the one hand, and processes related to matter - energy trans­ fer on the other. Viewed in this light, Darwin's distinction between sexual and natural selection is clear. In sexual selection, relative reproductive success arises strictly from among-population vari­ ation in some aspect of reproductive behaviour, physiology, or anatomy . In natural selection, an organism's relative success in economic (matter ­ energy transfer) activities has an effect on that organisms's probability of successful reproduction . The two hierarchies are direct outgrowths of these two distinct categories of adaptation that arise under sexual and natural selection. In sexual organisms, reproduction implies a local pool of suitable partners - a 'deme' . In most instances, there will be pools of suitable partners elsewhere; thus local demes form regional 'species' . Most modern treat­ ments of species recognize them as reproductive communities, within which mating occurs, outside which it does not. Paterson (e . g . 1985) recently suggested that species are reproductive com­ munities composed of organisms sharing a particu­ lar set of reproductive adaptations, or 'specific mate recognition system' ('SMRS') . His concept obviates the ambiguity of disjunct distributions, where potential mates never meet. Moreover, because the SMRS is an adaptive system subject to (sexual) selection (favouring mate recognition in isolation) speciation minimally must entail (presumably allo­ patric) divergence of the SMRS . Speciation is seen

as an outgrowth simply of continued reproduction in isolation, leading to modification of the SMRS . Because new (sexual) species arise in this fashion as a matter of course, higher taxa are maintained (as long as speciation rate exceeds extinction rate) . Higher taxa are seen strictly as lineages of species; they are recognized (just as are clones of strictly asexual organisms) only when new adaptations ('synapomorphies' of phylogenetic systematics; Section 5 . 2 . 2) arise and serve as markers for the lineage . As such, monophyletic taxa do not 'repro­ duce', that is, they do not produce additional entities of like kind . Genera do not give rise to new genera, the way that new species arise from old . The economic activities of organisms of a species lead them to form local populations ('avatars') which may, but need not, be coextensive with local demes of the same species . But above this level (Table I), a crucial distinction between the genealogical and economic hierarchies arises . Whereas the repro­ ductive adaptations of organisms are shared by organisms in other demes elsewhere, the economic adaptations of organisms lead to cross-genealogical interactions between local populations belonging to different species . Local ecosystems interact with other such systems on a regional scale, but maps of genealogical systems and economic systems simply do not coincide . It is especially significant that species are not parts of economic systems . Thus, by sheer dint of the existence of two classes of organ­ ismic activity - hence adaptations - organisms are simultaneously parts of two separate, hierarchi­ cally arranged systems . And in particular, inter­ action within and between entities of the two different hierarchies is of the greatest importance in elucidating a full causal theory of the evolutionary process .

The evolutionary process: role of the genealogical and economic hierarchies Discussions of macroevolution traditionally empha­ size the origin of higher taxa in the context of large­ scale adaptive change . Under this synthesis, linear trends are often said to be generated by 'orthoselec­ tion', i . e . long-term, predominantly directional natural selection, as distinct from 'orthogenesis', or linear phyletic change through unspecified causes internal to organisms . In general, the accumulation of significant amounts of adaptive transformation within a lineage has been termed anagenesis, which is commonly, if not invariably, held to be a process distinct from cladogenesis, or lineage splitting . Thus

2 . 6 Hierarchy and Macroevolution much, if not all, macroevolutionary change has tra­ ditionally been considered to occur without any (or any significant) degree of speciation . A major excep­ tion to this generalization is the theme of adaptive radiations, in which morphological transformation proceeds rapidly and independently in several or many different directions, and lineage-splitting is directly invoked as part of the process . Simpson's (1944) earliest formulation of 'quantum evolution' also invoked lineage splitting (though not expressly termed 'speciation'); later (Simpson 1953) modified in favour of a purely phyletic conceptualization of quantum evolution . The hypothesis of punctuated equilibria (Eldredge & Gould 1972) is based, in part, on the empirical claim that most species exhibit relative morphologi­ cal stability throughout the bulk of their strati­ graphic ranges (see also Section 2 . 3) . Thus most anatomical change appears to occur along with speciation . Such species stability facilitates recog­ nition of species as spatiotemporally-bounded enti­ ties; it further leads to the postulate that linear trends in macroevolution may reflect processes of species sorting in addition to directional natural selection . In general terms, such a model proposes that actual transformation of morphology occurs via directional natural selection (plus, perhaps, genetic drift) on a standard generation-by-generation basis . But the linearity of the trend through long periods of time - when the species remain morphologically stable and vary among themselves with respect to the evolving trait - arises through sorting of vari­ ation among species through a variety of potential causes . The term 'species selection' itself embraces a number of variant conceptualizations . As devel­ oped as an outgrowth of punctuated equilibria (Eldredge & Gould 1972); the term itself was intro­ duced by Stanley 1975; see also Stanley 1979), 'species selection' was virtually synonymous with the more general term 'species sorting' used here . Subsequent authors, seeking more precise parallel usage between organismic and higher-level selec­ tion, contend that 'species selection' is applicable only to species-level properties of species (cf. Vrba 1984) . This argument holds that phenotypic (and underlying genotypic) properties of organisms are the focus of organismic selection. True species selection should be invoked only to explain species­ level adaptations; it cannot logically be applied to the situation in which species differ merely in the frequencies of one or more organismic phenotypic traits . Williams (1966) was the first to argue this

127

point, claiming that 'group selection' can pertain only to group-level adaptations . Jablonski (1987) argued that geographical ranges are species-level properties, and show high heritability in his data on Cretaceous molluscs; he concluded that species ranges are therefore subject to true 'species selection' . Hull (1980) has discussed two components that must be present for selection to occur at any level; these two components serve in addition as criteria for evaluating claims of species-level selection . According to Hull (1980), among entities involved in any instance of selection, there must be an inter­ actor as well as a replicator. The relative success of interactors is recorded in the subsequent represen­ tation of their underlying replicators . Thus, in natural selection, relative economic success of organisms will affect their relative reproductive success, and hence the frequencies of the underlying genotypes . Organisms in this instance are both interactors and 'reproducers', with replicative fid­ elity supplied by their genes . Hull's (1980) selection criteria imply that species selection cannot be directly analogous to natural selection . If species are genealogical entities (if, in other words, it is the reproductive activities of organisms that lead to the formation and continued existence of species), species are causally connected to the replicative activities of genes; but if it is further true that species, as whole entities, do not play direct roles in ecological systems, species cannot be said to be interactors, and Hull' s criteria for selection are not met by definition for species . 'Species selection' appears t o b e most analogous to Darwin's sexual selection - because factors affecting rates of speciation and species extinction are involved . Species sorting i s a function of differential extinc­ tion and origination of species within a mono­ phyletic clade . It is the goal of macroevolutionary theory to specify the causal processes underlying such sorting. In addition to processes at work within a given level, the entities above and below in the hierarchy provide constraints (initial, or boundary conditions) on processes occurring within any given level - the 'upward and downward causation' of many hierarchy theorists . Vrba's (1980) 'effect hypo­ thesis' is an example of upward causation within the genealogical hierarchy . The effect hypothesis postu­ lates that macroevolutionary patterns (for example, linear trends in one or more morphological attributes within a clade over geological time) may arise simply as an outgrowth (side effect) of the

128

2 The Evolutionary Process and the Fossil Record

biology of the organisms themselves. Nothing more - specifically, no selection at the species level - need be invoked as an explanation of such patterns . Palaeontologists have sought links between characteristic speciation and extinction rates in lineages (macroevolutionary patterns), on the one hand, and aspects of organismic biology on the other - at least since Williams (1910) noted the apparent correlation between variation, niche width, and stratigraphic duration. Williams claimed that broadly niched (eurytopic) species, in addition to their characteristically wider geographical (habitat) occurrence, tend to display greater morphological variability (both within and certainly among populations) and longer stratigraphic dur­ ations than more narrowly niched stenotopes . Focusing especially on aspects of niche-width, macroevolutionary theorists have attempted to account for rates of both speciation and extinction. It seems, for example, that lineages comprised pre­ dominantly of eurytopic species show lower rates of species extinction and origination than lineages comprised of predominantly stenotopic species (see Eldredge & Cracraft 1980) . The contrast is especially clear in sister-lineages . Indeed, Vrba (1980) used the Miocene - Recent sister lineages of Aepycerotini (impalas) and Alcelaphini (wildebeests, hartebeests, topis, etc.), the former species-poor and eurytopic, the latter speciose, with short-ranging stenotopic species, to illustrate one possible cause underlying the 'effect hypothesis' (see also Section 2 . 10) . Vrba postulated that the trends in alcelaphine evolution were simply aggregates of higher rates of speciation and the accumulation of adaptive modification in the lineage of stenotopes - while little signifi­ cant evolutionary transformation accumulated within the co-ordinate lineage of eurytopes, the aepycerotines .

Interhierarchic interaction and macro evolution Darwin's (1859) original formulation of natural selection (where relative economic success affects relative reproductive success within a local popu­ lation of a species) serves as a model of the mech­ anics of interaction between the two hierarchies in the evolutionary process. Organisms, as members simultaneously in both the economic and genealogi­ cal hierarchies, patently stand as the prime causal link between the two (although some hierarchy theorists (notably Salthe 1985) see direct causal

interaction between entities at various levels of the two hierarchical systems) . Under the synthesis, species and higher taxa are generally depicted as having niches (or 'adaptive zones' in the case of higher taxa); further, in a widely used extension of Wright's (1932) metaphor of 'adaptive peaks', species and higher taxa are generally depicted as occupying peaks, or series of adjacent peaks (i . e . in an 'adaptive range') . Thus the most general approach to macroevolution under the synthesis holds that species and higher taxa are distinctly economic entities - effectively collapsing the dual hierarchy system into a single scheme . Yet, following arguments outlined above, it has seemed to recent theorists that species and higher taxa are different sorts of entities from those that form complex biotic economic systems . Species are aggregates of local demes, all of which share a common fertilization system. From an ecological point of view, species are typically integrated into a variety of different ecosystems . Yet organisms with­ in a species, as a rule, retain sufficient similarity in terms of economic adaptations that local populations are, to a great degree, redundant from one another. That is to say, the actual ecological role played by species is to serve as a reservoir of genetic infor­ mation. Local populations are notoriously ephem­ eral; local extinction, on several geographical and temporal scales, is often counteracted by recruitment from neighbouring demes. An important conse­ quence of the mere existence of species is that local parts of ecosystems are continually replenished from demes elsewhere . Recent studies of larval recruit­ ment in intertidal communities - after events that range from slight to total disruption - amply bear out the role that species play as reservoirs of genetic information . Darwin (1871) called species 'permanent varieties' . The expression is apt in the context of macro­ evolution, because the complexion of ecosystems is forever modified upon the final extinction of a species; the possibility of replacing local populations with conspecifics is forever lost. In general, just as species display a within-species pattern of supply of organisms to replace local populations, following extinction events that result in the loss of many higher taxa, the identities of the surviving taxa determine the natures of the sub­ sequently founded ecosystems . Disruption of eco­ systems results in extinction - the more severe the disruption, the higher the characteristic level of disappearance of taxa, from species on up; and the higher the average level of taxonomic extinction,

2 . 6 Hierarchy and Macroevolu tion the greater the change in economic systems . Theories (e . g . the 'Red Queen Hypothesis', Sec­ tion 2.5) often depict evolution as a process of inexorable adaptive change . Recent empirical and theoretical work in palaeobiology suggests rather a different picture : the ecological systems of which all organisms are parts, are formed from whatever organisms are extant at any given moment . With normal, small-scale fluctuations in composition and relative abundance of organisms, ecological systems appear to be quite stable . Speciation and extinction do occur, and so affect the composition of eco­ systems . Some phyletic modification may accrue within species, but, because most demes are ephemeral, little net change typically accumulates within species throughout most of their histories . Little in the way o f concerted evolutionary change, either within species or among species within lineages, tends to occur unless and until external perturbation disrupts ecosystems to the point where entire species - and higher taxa - become extinct, rendering impossible the resumption of ecosystems of the same composition as before . Thus, although the presence of genealogical entities (species and higher taxa - as packages of genetic information) are indispensable to the formation and ongoing existence of ecosystems, it appears that it is pri­ marily the disruption of such economic systems that leads to significant amounts of change within entities of the genealogical hierarchy . Hence mass extinction appears to be an important causal cornerstone of macroevolution .

References Darwin, C . 1 859 . On the origin of species . J. Murray, London . Darwin, C . 1871 . The descent of man, and selection in relation to sex. J. Murray, London. Dobzhansky, T. 1 937. Genetics and the origin of species. Columbia University Press, New York. Eldredge, N. 1985 . Unfinished synthesis . Oxford University

129

Press, New York. Eldredge, N. 1986. Information, economics and evolution. Annual Review of Ecology and Systematics 17, 351 - 369 . Eldredge, N. 1989 . Macroevolutionary dynamics: species, niches and adaptive peaks . McGraw-Hill, New York. Eldredge, N. & Cracraft, J. 1980. Phylogenetic patterns and the evolutionary process . Columbia University Press, New York. Eldredge, N. & Gould, S.J. 1972. Punctuated equilibria: an alternative to phyletic gradualism. In : T.J.M. Schopf (ed . ) Models i n paleobiology, pp . 82 - 1 15. Freeman, Cooper & Co . , San Francisco. Grene, M. 1987. Hierarchies in biology. American Scientist 75, 504-510. Hull, D.L. 1980. Individuality and selection. Annual Review of Ecology and Systematics 11, 31 1 - 322. Jablonski, D . 1987. Heritability at the species level : analysis of geographic ranges of Cretaceous mollusks . Science 238, 360 -363 . Mayr, E. 1942 . Systematics and the origin of species . Columbia University Press, New York. Paterson, H . E . H . 1985 . The recognition concept of species. In: E . S . Vrba (ed. ) Species and speciation, pp . 21 - 29 . Transvaal Museum Monograph No. 4. Transvaal Museum, Pretoria. Salthe, S.N. 1985 . Evolving hierarchical systems. Columbia University Press, New York. Simpson, G . G . 1944. Tempo and mode in evolution . Columbia University Press, New York. Simpson, G . G . 1953 . The major features of evolution . Columbia University Press, New York. Stanley, S . M . 1975 . A theory of evolution above the species level. Proceedings of the National Academy of Science 72, 646-650 . Stanley, S . M . 1979 . Macroevolution: pattern and process. W.H. Freeman, San Francisco. Vrba, E . 5 . 1980. Evolution, species and fossils: how does life evolve? South African Journal of Science 76, 61 - 84. Vrba, E . 5 . 1984. What is species selection? Systematic Zoology 33, 318-328. Williams, G . c . 1966. Adaptation and natural selection . Princeton University Press, New Jersey. Williams, H . S . 1910. The migration and shifting of Devonian faunas. Popular Science Monthly 77, 70 - 77. Wright, S. 1932. The roles of mutation, inbreeding, cross­ breeding and selection in evolution. Proceedings of the VIth International Congress of Genetics 1, 356- 366.

2 . 7 Patterns of Diversification P . W . SIGNOR

Introduction

fluence the composition of the fossil record . Sea­ level, which largely controls epicontinental marine deposition and preservation of fossils therein, has varied throughout the geological past. Low sea stands are usually represented in the stratigraphic record as diastems, disconformities, or unconform­ ities, and lack any fossil record of shelf faunas . Other time-dependent biases include monographic effects (Raup 1972) and the distribution of systematists' labour (Sheehan 1977) . There are significant time-independent biases. For example, terrestrial environments (and the organisms that inhabit them) are not well rep­ resented in the stratigraphic record, in comparison to marine habitats (e . g . Padian & Clemens in Valentine 1985) . Among marine organisms, heavily skeletonized forms are preserved far more fre­ quently than lightly or non-skeletonized forms . Palaeobiologists often presume that the ratio of heavily skeletonized to non-skeletonized species has been approximately constant, at least since the early Phanerozoic, but no data or arguments to support that contention have been advanced . On the contrary, there is some evidence that skeletons have become more robust in time in response to newly evolving predators (Section 4 . 13) . The net result of these biases is quite severe, amply justifying the ancient laments about the incompleteness of the fossil record . Only approxi­ mately 10% of the skeletonized marine species of the geological past and far fewer of the soft-bodied species are known (Sepkoski et al. 198 1 ; Signor in Valentine 1985) . No doubt whole clades and com­ munities of the past remain to be discovered. More importantly, these biases continue to obscure all but the most fundamental patterns in the history of diversification. A brief aside on the semantics of diversity might prevent confusion . The term diversity has been used in two senses . Unfortunately, the two usages are rather different, and treating the term carelessly confounds an important concept . In the palae­ ontological literature, diversity is often used to mean richness, or the number of taxa present. Diversity also has a second meaning, incorporating both rich-

The past 3 . 5 billion years have witnessed substantial change in the numbers of protist, animal, and plant taxa on Earth . The magnitude of that net change is evident from comparison of the lush biological diversity present in so many modem habitats with Archaean sediments seemingly barren of fossils . But reconstructing the geological history of organic diversity has proved difficult. Biases in the preser­ vation, collection, and study of fossils have com­ bined to obscure patterns of change in diversity. Despite the difficulties, a variety of different patterns of diversification has now been documented at scales ranging from local communities to the entire biosphere . These patterns indicate that the net accumulation of taxa through time has been quite unsteady .

Biases in the fossil record The geological history of taxonomic and ecological diversification is obscured by a variety of time­ dependent and time-independent filters . Most of these are various sorts of sampling biases, which cause the observed fossil record to differ from the actual history of the biosphere (see also Section 3 . 1 2) . The most severe of the time-dependent biases is the loss of sedimentary rock volume and area with increasing age (Raup 1976b) . Both sedimentary rock area and rock volume correlate strongly with the numbers of animal species described from that stratigraphic interval (Raup 1972, 1976b) . Rock volume and area affect apparent species richness by influencing the likelihood that a given species is preserved, discovered, and described (Raup 1976b) . Similar biases have been documented in the fossil record of vascular plants on land (Knoll et al. 1979) . The quality of preservation of fossils within sedi­ mentary rock also tends to deteriorate with increas­ ing age, because of extended exposure to diagenesis (Raup 1972) . The kinds of sedimentary rock and, by implication, the environments preserved in the stratigraphic record have varied greatly through time. Variability in the representation of palaeo­ environments in the stratigraphic record must in-

130

2 . 7 Patterns of Diversification ness and evenness of distribution. In this second sense, a community composed of three equally common species would be regarded as more diverse than a community of three species where one species far outnumbered the remaining two . The ecological literature generally restricts usage of diversity to the latter meaning. Most papers on the history of diversity, however, have treated diversity as synonymous with number of taxa present, and that is the approach used here .

Taxonomic diversity Tabulations of classes, orders, or families, instead of species, are commonly employed to minimize sampling bias in palaeontological estimates of bio­ logical diversity . One need only find a single species to document the presence of a higher taxon, whereas every species must be discovered to provide com­ plete documentation of species richness. Therefore, researchers have employed classes, orders, or families as surrogates for species in estimates of biological diversity in situations where our ability to sample species is hindered . Higher taxa have also been employed as metrics of morphological or ecological diversification (e.g. Erwin et al. 1987) . The utility of higher taxa as first order metrics of species richness is dubious (see Sepkoski 1978 for a contrary view) . Tabulations of marine orders and families (Sepkoski 1978, 1979, 1982) and estimates of species richness (Sepkoski et al. 1981; Signor in Valentine 1985) are not congruent, indicating that numbers of higher taxa do not parallel changes in underlying species richness (Fig . lA, B) . Generic diversity is rather similar to estimated patterns of species richness, but patterns of the diversity of families, orders, or classes are increasingly dissimi­ lar . Similarly, Raup's (1979) analysis of the Permo­ Triassic mass extinction (Section 2 . 13 .4) indicates that 17% and 52% reductions in the number of marine orders and families, respectively, represent approximately a 96% reduction in the number of species . Likewise, patterns in the numbers of ter­ restrial vertebrate orders are rather dissimilar from patterns in the numbers of genera (see Padian & Clemens in Valentine 1985) . Higher taxa are buffered from fluctuations in numbers of species and con­ sequently are poor metrics of changes in species richness. Higher taxa are more or less artificial constructs that are not defined by species richness . Indeed, most higher taxa incorporate relatively few species (Sepkoski 1978) . Therefore, the lack of concordance

131

200

A O rders

B

800 Fam i l i es

c

1 00 Genera

50

1 00

D

50

Fig. 1 Diversity of marine animals. A, numbers of marine orders through time . (Data from Sepkoski 1978 . ) B, numbers of marine families through time . (Data from Sepkoski 1979 . ) C, percentage change i n the number o f marine genera through time . (Data from Sepkoski et al. 1981 . ) D, estimated percentage variation in the number of skeletonized marine invertebrate species. (Data from Signor in Valentine 1985 . )

132

2 The Evolutionary Process and the Fossil Record

between the number of species, families, orders, or classes through time is not surprising . In contrast, biological species can be defined and recognized through patterns of reproductive isolation : they are real biological units . As an entity, the species possesses biologically significant characteristics lacking in higher taxa . But change in the numbers of species through time is a more difficult problem to attack than change in higher taxa, because of the inherent deficiencies of the fossil record (Section 3. 1 2) . The numbers o f Phanerozoic marine orders and families have been tabulated by Sepkoski (1978, 1979, 1982) (Fig. lA, B) . The number of orders in­ creased rapidly until the Late Ordovician, and then remained approximately constant for the remainder of the Phanerozoic. The number of families also increased rapidly in the Cambrian and Ordovician, reaching a plateau of about 400 families for the remainder of the Palaeozoic . Following the Permo­ Triassic mass extinction (Section 2 . 13 .4), the number of families has increased more or less continuously to the present . A preliminary tabulation of the number of genera shows a pattern generally similar to the change in numbers of families through time (Sepkoski et al . 1 981) (Fig . 1C) . Compilations of the number of described marine species through time show low numbers through the Palaeozoic and Mesozoic, followed by a sub­ stantial increase in the Cenozoic (Raup 1976a) . These tabulations are undoubtedly skewed by sampling biases, as discussed above (see Raup 1976b) . Attempts to infer patterns of species richness from changes in the numbers of higher taxa have produced patterns generally similar to Raup's tabulation, but show lower numbers of species in the Palaeozoic and Mesozoic. Analytical calculations to remove the effects of sampling bias result in a similar pattern (Signor in Valentine 1985; Fig . ID) . The history of diversification of terrestrial vertebrates produces a quite different pattern . Compilations of the numbers of tetrapod orders through time show no longstanding equilibrium (Padian & Clemens in Valentine 1985; Fig . 2A) . There was a steady increase through the Middle Palaeozoic, reaching a Mesozoic plateau that began in the Late Triassic . Following the Cretaceous ­ Tertiary mass extinction (Section 2 . 1 3 .6), the number of orders increased briefly and then began to decline (Fig. 2A) . The Tertiary adaptive radiation of birds is superimposed upon this diversification, and nearly doubled the number of terrestrial vertebrate orders (Fig . 2A) . The pattern at the generic level is similar,

but more exaggerated (Padian & Clemens in Valentine 1985) . The number of genera rose quickly through the Palaeozoic to a peak in the Permian . Following a severe reduction in generic diversity at the end of the Permian, the number increased, regaining Permian levels in the Cretaceous . In the Cenozoic the number of genera increased nearly tenfold . The history of the diversification of vascular plants forms still a third pattern (Fig . 2B) . In the Northern Hemisphere, there was a gradual increase in species richness to a peak of over 40 species early in the Late Devonian (Niklas et al. in Valentine 1985) . Following a slight decline in the Late Devonian, the number increased rapidly to over 200 species by the Middle Carboniferous . With the exception of a brief decline at the end of the Permian, the number of species increased gradually through the remainder 50 Terrestrial ve rteb rate o rd e rs

30

I' .., r , I I I j I , 1 1 , 1 1 1 , 1 1 1 , 1 I I , t 1 1 , 1 J I I I 1 1 1 1 1 ' I ' 1 I I 1 , 1

A

10

600

Vasc u l a r p l a n t s p e c i e s

B

300

Fig. 2 Diversity of terrestrial organisms. A, numbers of terrestrial vertebrates through time . Solid lines indicate changes in the number of amphibian, reptile, and mammal orders; dotted lines indicate the number of avian orders. (Data from Padian & Clemens in Valentine 1985 . ) The Cenozoic is subdivided into the five epochs of the Tertiary plus the Pleistocene . B, numbers of terrestrial plant species (mostly in the Northern Hemisphere) through time . (Data from Niklas et al. in Valentine 1985 . )

2 . 7 Patterns of Diversification of the Palaeozoic, the Triassic, Jurassic and Early Cretaceous. Following the origin of angiosperms (Section 1 . 10) in the Late Cretaceous, species rich­ ness increased rapidly to over 600 in the Quaternary . Patterns of global taxonomic richness differ at various levels of the taxonomic hierarchy. Unbiased species data would most accurately reflect changes in biological complexity through time, but species­ level data are the most susceptible to sampling bias. The species richness of marine animals, vascular plants, and terrestrial vertebrates have quite dif­ ferent histories, but all indicate significant re­ ductions in diversity at the Permo-Triassic, Norian, and Maastrichtian extinction events (Section 2 . 1 3) . All three patterns also share a tremendous diversification of species beginning in the Cretaceous .

Local diversity An important component of the history of diversity is the temporal pattern of species richness within individual communities . Bambach (1977) compiled counts of species present within 386 previously described ancient marine communities . He assigned the communities to one of three generalized habi­ tats : nearshore high stress, nearshore variable, and open marine environments . Alpha diversity, or within-community diversity, remained constant in the high-stress environment communities throughout the Phanerozoic, but increased twofold in the variable nearshore and open marine environ­ ments during the Mesozoic (Bambach 1977; Fig . 3C) . The Mesozoic increase in alpha diversity appar­ ently was accommodated through trophic diversifi­ cation of the major clades of marine animals in shelf communities (Bambach 1983) . In the Cambrian, there were relatively few clades and each clade had a limited range of roles . The number of clades increased in the Palaeozoic, an increase that was paralleled by a limited diversification of trophic roles . The Mesozoic increase in diversity was accompanied by a much larger diffusion of taxa into new trophic roles, especially into infaunal life­ modes . The expansion of marine animals into infaunal life-modes is one component of the pattern of increasing tiering (Section 1 . 7 . 1 ) . Tiering, the spatial development of communities both above and below the sediment surface, had increased through the Phanerozoic. This increase has been attributed to a number of physical and biological processes, but

133

the net result is undoubtedly an increase in local habitat complexity and organic diversity .

Controls on diversity The nature of the processes controlling species rich­ ness is the subject of considerable speculation. At the level of communities, such processes are not well understood, even in the modem world . Area, habitat complexity, environmental stability, physi­ cal disturbance, and other factors may well be important controls on alpha diversity . What tran­ spired to bring about a twofold increase in within­ habitat species richness in the late Mesozoic is equally unclear . A better understanding of the processes regulating species richness in the Recent is probably a prerequisite to resolving this question . Area is a primary influence on diversity at local, regional, and global levels . Area appears to regulate diversity primarily through variation in rates of extinction . Reduction in habitable area decreases population size, which increases the chance of extinction . For organisms dwelling among the benthos on continental shelves, change in diversity appears to be related to variation in shelf area (Sepkoski 1976; but see Flessa & Sepkoski 1978) . Severe reductions in shelf area have also been im­ plicated as the cause of a mass extinction (Sections 2 . 1 2 . 1 , 2 . 1 3 . 4) . Mass extinctions severely reduce the number of taxa present in the biosphere . Diversity generally rebounds following extinction events and often increases to surpass previous levels, but that re­ bound requires geologically significant intervals of time . The time necessary for recovery is, in part, proportional to the magnitude of the extinction (Sepkoski 1 984) . During the intervening time be­ tween extinction and recovery, diversity is reduced . I f extinctions are spaced more closely than the necessary recovery time, the biosphere will remain relatively impoverished (e . g . Hansen 1988) . At the level of the biosphere, plate tectonics is undoubtedly the most potent control on the diver­ sity of organisms (Valentine et al. 1 978; Signor in Valentine 1985) . Marine organisms often share com­ mon range boundaries, and geographical regions with relatively homogeneous faunas and distinct boundaries are termed provinces (see also Section 5.5) . The boundaries of these provinces are defined by the joint limits of distribution of common species, and are controlled by patterns of climate and oceanic water circulation. In turn, climate and oceanic circulation are determined largely by the

134

2 The Evolutionary Process and the Fossil Record

distribution of continental land masses. Continu­ ously changing continental configurations have thus regulated the number and distribution of provinces through time . Changing levels of provinciality have no apparent impact on biological complexity within individual communities, but alter global diversity through adding or subtracting whole provinces . The number of provinces has increased greatly since the late Mesozoic breakup of Gondwana and Laurasia (Valentine et al. 1978), a change undoubt­ edly responsible for much of the increase in species richness of marine animals . The species richness o f vascular plants and terres­ trial vertebrates is also heavily influenced by plate tectonics (e . g . Padian & Clemens in Valentine 1985) . Tertiary isolation of the terrestrial faunas of South America, Australia, Africa, and Madagascar permit­ ted the evolution and persistence of unique faunas, while other clades dominated the Hcilarctic continents of Asia, Europe, and North America . Isolation and interspersed periods of faunal inter­ change have contributed greatly to the diversity and taxonomic composition of the terrestrial vertebrate faunas of the different continents . In summary, local and regional patterns in the number of taxa vary semi-independently and are compounded at regional and global scales . Increas­ ing alpha diversity within marine communities resulted in a comparable increase in global diversity in the Late Mesozoic . Similarly, changes in provin­ ciality through time have altered global diversity . Changes at each level of the ecological hierarchy, from the community to the biosphere, influence trends in global diversity . Such trends therefore represent complex interactions of physical and biological processes operating on many scales.

Diversity within clades A systematic pattern of temporal change in diversity within individual clades has recently been recog­ nized (Gilinsky & Bambach 1987) . Clades appear to contain more subtaxa early in their histories than later on . The primary cause of this trend appears to be a systematic decline in rates of origination within established clades through geological time, although rates of extinction also increase somewhat through time (Gilinsky & Bambach 1987) . The obvious inference from this statistical generalization is that clades are established during brief adaptive

A

. . .. . . · . ... .. . · . . . . . . . . · · . . . . . . . . · · . . . . . · . . ·

::::: :: :::::::: . . . .. ::: :::: ::: :: :: : ::::: :::::: :::::::: . . .. .

. .. ... . .. . . . . .. . . . . .. .. .. . .. .. .... . .. .. . . . . . .. ..

. . . .

.

::::: ::: ::: :::: : : :: :::: : . . .

. . .. . . .

Speci es/Myr

. . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . .

· ·

· ·

. .

. .

. .

· . . · .

.

A l p h a d i versity

·

.

.

.

. . . .

· . · . · . · . · . · . · . · . · . · . · . · . · . · .

. . . . . . . .

r"T"l......,,..-I;I��� I ��I� m�t D

G e n e r i c d i ve r s i ty

E

F a m i l ies

.

.

.

.

.

���t�

:..:..:..:..

. .. .. .. . .. .. .. .. . . . . . . . . . . . .. . . . .

. .

.. .. · . . . · . . . · . . . . · . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .

. .. . .. .. .. .. . ... .. .. . . .. . .. .

· . . · . .

.. .. .. · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . . · . . .

· . . · . . · . . · . . · . . · . .

:::: :: ::

.. . . ...... ... .. . .. .. . . . . . . . .... .. .... .. . .... .. . . .. . . .. . .: . : . : . : . ·

o

·

· . . · . .

·

. .. .. . . . . . .. �;-o"Im;... .... .. . : . : . : . . . . .. . . . . . .

.

ff

·

·

C

· . . . . · . . · . . · . . · . .

·

·

B

. .. .. .. . ·. . .. . .. . . . . .. . .. . .. . . . ... . . .... .. .. ... .... .

Trace fos s i l s

. .. ... . .. . .. . . .: . .

......... . .. .. . . ... .. . .. . . . . . . :. : . : .

. . . . .... .. ... ... . . . . .. .. .. . .. . . . . .. . . . .. . . : .: . . . .

...... ... . . . .. . . .. . . . . . . .. . . . .:

:

. .. . . . . . .. . . . . : .:

5

D

C

P

Tr

. . . .. . . . .. . . . . .: . . . . . . : .: . :

. ·.. ...... .. . .. . . . . . . . . .. . . . . . . .. .. . . . . .. . . . . . . . .. . . : . .. .. . K

.

.

.

· . . . · . . · . . · . .

Ce n

Fig. 3 Trends in marine animal diversity through time . (After Sepkoski et al. 1981 . ) The five trends in diversity accepted as evidence against an equilibrium model of Phanerozoic diversity. A, number of the trace fossil genera within individual habitats . B, numbers of described species, normalized for length of the geological period. C, alpha diversity of marine invertebrates . D, E, numbers of genera and families respectively . The original figure was presented without vertical axes, as shown here .

radiations and subsequently begin a long decline to eventual extinction . Speciation, not extinction, might be the dominant factor in clade diversity (Gilinsky & Bambach 1987) .

2 . 7 Patterns of Diversification Modelling change in diversity through time Sepkoski (1978, 1979, 1984) applied quantitative models of population growth to the history of taxo­ nomic diversity of marine organisms . These models were developed to describe and predict the insrease in numbers of individuals within single popu­ lations . The mathematical assumptions of the simplest model, the logistic model of population growth, are : (1) there is a maximum number of individuals that can be supported in the environ­ ment (the carrying capacity); and (2) population growth is exponential and declines linearly as the population size approaches the carrying capacity . These assumptions may also reasonably apply to taxonomic diversification at the level of the bio­ sphere . Sepkoski's (1978) successful application of the logistic model to describe the increase in numbers of marine orders suggests that the model describes the behaviour of change in the numbers of orders rather well . Similar results have been obtained in analyses of the radiation of angiosperms (Lidgard & Crane 1988) . Further applications of more complex models from population biology treat different aggregates of clades as faunas, comparable to populations competing for resources (Sepkoski 1979, 1984) . These models also produce good fits with the available data .

Faunal equilibrium? Over the past 20 years, a number of theorists have questioned the empirical pattern of increasing species richness through time, suggesting instead that the diversity of marine animals has been at equilibrium through much of the Phanerozoic (see Signor in Valentine 1 985 for review) . In this view, changes in species richness reflect only biases and not biologically significant trends . The equilibrium itself could be absolute, with constant numbers of species through time, or dynamic, with an equilibrium shifting in response to the changing physical world. However, the similar patterns observed in a variety of separate metrics of diversity (Fig. 3) provide convincing evidence that the ap­ parent pattern of species richness through time is not artifactual (Sepkoski et al. 1981 ) . But important questions about historical patterns of local and global species richness, and the ultimate controls on those patterns, remain to be resolved.

135

References Bambach, R.K. 1977. Species richness in marine benthic environments through the Phanerozoic. Paleobiology 3, 152- 167. Bambach, R.K. 1983. Ecospace utilization and guilds in marine communities through the Phanerozoic. In : M.J.5. Tevesz & P . L . McCall (eds) Biotic interactions in Recent and fossil benthic communities, pp. 719 - 746. Plenum Press, New York. Erwin, D . H . , Valentine, J.W. & Sepkoski, J . J . , Jr. 1987. A comparative study of diversification events : the early Paleozoic versus the Mesozoic. Evolution 41, 1 1 77 - 1 186. Flessa, K.W. & Sepkoski, J.J., Jr. 1978. On the relationship between Phanerozoic diversity and changes in habitable area. Paleobiology 4, 359- 366. Gilinsky, N.L. & Bambach, R.K. 1987. Asymmetrical patterns of origination and extinction. Paleobiology 13, 427-445 . Hansen, T .A. 1988. Early Tertiary radiation of marine molluscs and the long-term effects of the Cretaceous ­ Tertiary extinction. Paleobiology 14, 37-51 . Knoll, A . H . , Niklas, K.J. & Tiffney, B . H . 1979 . Phanerozoic land-plant diversity in North America. Science 206, 1400- 1402 . Lidgard, S. & Crane, P.R. 1988 . Quantitative analyses of the early angiosperm radiation. Nature 331, 344-346. Raup, D . M . 1972. Taxonomic diversity during the Phanerozoic. Science 177, 1065 - 1071 . Raup, D . M . 1976a . Species richness in the Phanerozoic: a tabulation. Paleobiology 2, 279 - 288. Raup, D.M. 1976b . Species richness in the Phanerozoic: an interpretation . Paleobiology 2, 289 -297. Raup, D . M . 1979 . Size of the Permo-Triassic bottleneck and its evolutionary implications. Science 206, 217-218. Sepkoski, J.J., Jr. 1976 . Species diversity in the Phanerozoic: species- area effects . Paleobiology 2, 298 -303. Sepkoski, J . J . , Jr. 1978. A kinetic model of Phanerozoic taxo­ nomic diversity, I. Analysis of marine orders. Paleobiology 4, 223 -251 . Sepkoski, J . J . , Jr. 1979 . A kinetic model of Phanerozoic taxo­ nomic diversity, 11 . Early Phanerozoic families and multi­ ple equilibria. Paleobiology 5, 222 - 251 . Sepkoski, J . J . , Jr. 1982. A compendium of fossil marine famil­ ies. Milwaukee Public Museum Contributions to Biology and Geology No . 51 . Milwaukee Public Museum, Milwaukee . Sepkoski, J.J., Jr. 1984. A kinetic model of Phanerozic taxo­ nomic diversity, Ill. Post-Paleozoic families and mass extinctions . Paleobiology 10, 246- 267. Sepkoski, J . J . , Jr. , Bambach, R.K., Raup, D . M . & Valentine, J.W. 1981 . Phanerozoic marine diversity and the fossil record . Nature 293, 435 -437. Sheehan, P.M. 1977. Species diversity in the Phanerozoic: a reflection of labor by systematists? Paleobiology 3, 325 - 328 . Valentine, J.W. 1985 . Phanerozoic diversity patterns . Princeton University Press, Princeton. Valentine, J . W . , Foin, T . e . & Peart, D. 1978 . A provincial model of Phanerozoic diversity. Paleobiology 4, 55 - 66 .

2 . 8 Coevolution S . C O NWAY M O RRIS

Introduction

survivorship curves by L.M. Van Valen indicated that the rate of extinction is stochastically constant, i . e . the probability of extinction is constant irres­ pective of the duration of a particular taxon . Van Valen explained this pattern by the now well known Red Queen Hypothesis (Section 2 . 5), arguing that the environment in which evolution occurs is largely defined by biotic interactions that operate so that the improvement in fitness of any one species auto­ matically reduces the fitness of all others (given that the sum of all fitnesses remains unchanged) . In this sense ecological units consisting of numerous inter­ acting species, which exhibit unceasing evolution­ ary change as species attempt to restore their fitness in the face of a constantly deteriorating biotic en­ vironment, may be said to show coevolution . Whether the Law of Constant Extinction is empiri­ cally demonstrable, and whether the Red Queen Hypothesis is the appropriate explanation, have both been extensively debated . Recent analyses of planktic species (mostly coccoliths, foraminifera, and radiolarians) from Cenozoic ocean deposits from mid-to-Iow latitudes give some support to the Red Queen Hypothesis, although the data on species survival have to be considered in the context of an environment that is not effectively constant . There are three other areas in the fossil record that may be explained in the broad context of co­ evolution . These are claims for reciprocal patterns between: (1) predators and prey; (2) plants and animals, especially insects; and (3) phylogenetic congruence between symbiotic taxa, especially parasites and their hosts .

Ever since the first species was joined by a second one, the potential for some sort of coevolution has existed. However, the possibility of documenting coevolution in the fossil record depends on the scope of the definition that is accepted . In a broad sense, coevolution has been taken to include almost any biological interaction, with emphasis often placed on mutualistic associations . Stricter defi­ nitions emphasize reciprocal responses between individuals of two species where each exerts, either sequentially or synchronously, an influence on the other's heritable characters . Whether such coevolutionary oscillations are stable over geo­ logical periods of time is not certain, and it seems questionable whether coevolution in a strict sense has ever been recognized in the fossil record . Accordingly, evidence for more broadly based in­ teractions that seem in some sense to represent responses by one taxon or group to changes in another is presented here .

Gaia On the grandest scale there has been considerable interest in the concept of Gaia, whereby a system of organically mediated feedbacks maintains the Earth's surface in a state of homeostasis that is largely independent of external vicissitudes that otherwise would imperil the continuation of life . However, while it is accepted that biological activi­ ties can mediate geochemical, and probably geo­ physical, cycles, there has been less enthusiasm for the notion that life in toto could act as the primary regulator of Gaia . This is because individual species, rather than the entire biosphere, must be accepted as the units for evolutionary selection, with life exploiting those opportunities offered by changing environments .

Predators and prey Much interest has been expressed in the possibility of an arms race between predator and prey, with a spiralling escalation of attack and defence . There are, however, reasons to doubt that a long-term oscillation would persist . In particular, Vermeij (1982) pointed out that: (1) as most predators feed on at least several species, if confronted by an increasingly well-defended prey they will switch to one of greater vulnerability; and (2) the predator itself will be prey for other species, so that selective

The Law of Constant Extinction While not as grandiose in its scope, the so-called Law of Constant Extinction may have implications for interactions between members of entire ecologi­ cal groupings of taxa . Analyses of numerous taxon

136

2 . 8 Coevolution factors that favour the predator' s own survival, as against its ability to obtain a meal, will predomi­ nate . Indeed, the only cases where reciprocal evolution of prey and predator may show consistent trends are where a victim can on occasion maim or even kill its attacker. Notwithstanding the potential problems in documenting predator- prey arms races, a number of attempts have been made to demonstrate reci­ procity in the fossil record on the basis of broad­ scale trends. One of the best known analyses concerns changes in the brain mass of Cenozoic ungulate herbivore and carnivore mammals. From a study of brain sizes Jerison (1973) concluded that the carni­ vores maintained proportionally larger brains than the ungulates, although both showed a persistent increase during the Cenozoic. This pattern was explained by a type of coevolutionary feedback whereby selection pressure exerted by the larger­ brained carnivores forced a corresponding increase in the herbivores, which in turn fuelled further increases in the carnivores . The validity of this analysis, however, has been questioned . Radinsky (1978) pointed out that: (1) many of the comparisons involve carnivores and ungulates of different stratigraphic age; (2) the estimates of body weight (needed as part of the calculation of the relative brain size) may require revision; and (3) some samples may be too small to provide reliable comparison . He concluded that, where the data are adequate, supposed differences between ungulates and carnivore brain sizes can­ not be demonstrated . Moreover, other studies of mammalian evolution in the Cenozoic (Bakker in Futuyma & Slatkin 1983) have argued that while both carnivores and herbivores show trends towards greater efficiency (e . g . for running) the so-called adaptive gaps may widen, especially when replace­ ment faunas arrive following a mass extinction. Changes in predator- prey interactions have also been identified in the marine record (see also Section 4 . 1 3 . 1 ) . The rise of predators in the Cambrian is followed by an episode of increasing predatory activity in the Middle Palaeozoic and finally a major reorganization of prey and predatory ecologies during the Jurassic and Cretaceous (the so-called Mesozoic Marine Revolution) . However, apart from the parallel rise of offensive and defensive adap­ tations, it has not been possible to demonstrate specific series of reciprocal changes, and it seems that evolutionary responses may have been diffuse . Amongst invertebrates, recent research has investigated possible coevolution between preda-

137

tory gastropods, especially naticids, and their prey of bivalve molluscs, which they attack by drilling through the shell (Kitchell in Nitecki & Kitchell 1986) . While evidence for naticid attacks may extend back to the Triassic, it first became widespread in the Cretaceous . Study of drilling behaviour demon­ strates a remarkable stereotypy over geological time in terms of both position on the prey and ability to resume attack after interruption . However, in terms of possible coevolutionary responses between predator and prey, the only persistent trend that can be documented in the fossil record is a mutual increase in size.

Plants and animals The widespread inference of coevolution between plants and arthropods, especially insects, in modem biotas has often been extended into the geological past. However, despite some classic examples, such as between figs and fig-wasps, there is serious reason to doubt whether many Recent plant­ animal interactions can be regarded as strictly coevolutionary. Nevertheless, there is a widespread assumption that the diversification of plants and insects in the fossil record has been governed by coevolutionary forces . At present only the growing evidence for plant- animal interactions can be documented, leaving open the question of whether any of the examples fall into the domain of strict coevolution . Evidence for such interaction can be traced to the early Devonian, both in the form of direct associ­ ations (e . g . trigonotarbid arachnids lurking in sporangia of Rhynia from the Rhynie Chert), and more generally in plant morphology that ostensibly either promoted (e . g . spore sculpture) or hindered (e . g . stem spines) arthropod interactions . In these latter cases caution must be exercised, as specific features of plant anatomy may have had multiple functions including resistance to water loss, shielding from ultraviolet radiation, and so forth . By the Carboniferous there is considerable evidence for plant- animal interactions, both direct, such as spores in insect guts or various trace fossils (coprolites, borings, chew marks), and indirect, from insect mouth parts or plant anatomy, especially of spores and seeds . However, in no case has it been demonstrated that either partner was exerting reciprocal selective pressure over a period of geo­ logical time . Evidence of responses to arthropods in younger floras (Crepet 1979) includes study of reproductive structures, such as those of cycads,

138

2 The Evolutionary Process and the Fossil Record

whose cone anatomy appears to trend towards excluding insect attack. Particular attention has been given to the activities of insects in pollination, to which has been linked the rise of the bisporangiate condition. In the Jurassic gymnosperms, for example, coleopterans (beetles) may have been of particular importance, but with the rise of the angiosperms in the Cretaceous (Section 1 . 10) the role of dipterans (flies) and hymenopterans (bees and wasps) is regarded as crucial . In particular, links between insect group and flower or inflor­ esence anatomy may allow inferences on potential pollinators . However, in many examples the assumptions are based on uniformitarian premises and it is also important to realize that strict coevolution has not been demonstrated . Indeed, in many cases it is likely that evolution was sequential, the insects following plant diversificati0t:t rather than acting as primary mediators . Although the greatest interest in plant- animal coevolution has concerned the role of arthropods, speculation has extended also to vertebrates . Stebbins (1981) argued for coevolution between Cenozoic mammalian grazers (e . g . horses) and the grasses (see also Section 1 . 1 1 ) . The development of hypsodont teeth to cope with the siliceous grasses (specifically the secretion of opalines in the plant epidermal cells), and of running ability in more open savanna, would seem to be linked intimately with the spread of grasslands . However, reciprocal connection between degree of hypsodonty and silica content or distribution, that could be taken as strict coevolution, has not been demonstrated .

Phylogenetic congruence The final area where the fossil record may contribute to the documentation of coevolution is in the identification of congruent phylogenies where mutualistic associations, especially between para­ sites and their hosts, are reflected in their respective histories of cladogenesis (Mitter & Brooks in Futuyma & Slatkin 1983) . This pattern is often refer­ red to as Fahrenholz's Rule, but to date the evidence for congruence has varied widely and in very few clades of parasite and host is strict congruence evident . Where parasites possess limited abilities for dispersal then host - parasite congruence may occur. However, if a parasite species is pursuing a particular feature, in what is known as resource tracking, then typically it will occur in those taxa that happen to share the particular resource . Thus, in the Mallophaga (chewing lice) the limited dis-

persal of those infesting some mammals (e . g . pocket gophers) contrasts to the distribution of those para­ sitizing birds, where feather type seems to be of particular importance (Timm in Nitecki 1983) . Where Fahrenholz' s Rule appears to be applicable, then in principle the fossil record of a well skeletized host could give insight into the evolutionary history of the parasites, which are almost invariably soft­ bodied and unknown as fossils . While the fossil record may throw light on times of divergence in such instances, the relatively few well documented lineages still only provide a broad indication of evolutionary events, and tightly constrained his­ tories do not appear to be available . The numerous commensal associations that have been documented may prove a more fruitful area for establishing phylogenetic congruence between symbionts in the fossil record . These include host specific epizoans, e . g . cornulitids, spirobids and other 'worms', and more intimate associations such as those between stromatoporoids and corals . However, in no case does it appear that strict coevolution has occurred, and in at least some cases there is evidence that the host has evolved (at least morphologically) at a substantially faster rate than its partner.

Conclusion While evidence of species interaction is manifest in the fossil record, examples of strict coevolution have yet to be documented . This may reflect problems of resolution and insufficient study, but it seems more likely that long term associations only rarely fall into the category of coevolution as it may be usefully understood .

References Crepet, W.L. 1979 . Insect pollination: a paleontological per­ spective. BioScience 29, 102 - 108. Futuyma, D.J. & Slatkin, M . 1983 . Co-evolution . Sinauer, Sunderland, Massachusetts. Jerison, H.J. 1973 . Evolution of the brain and intelligence. Academic Press, New York. Nitecki, M.H. (ed . ) 1983. Co-evolution . University of Chicago Press, Chicago. Nitecki, M . H. & Kitchell, J . A . (eds) 1986 . Evolution of animal behavior: paleontological and field approaches . Oxford University Press, New York. Radinsky, L. 1 978. Evolution of brain size in carnivores and ungulates. The American Naturalist 112, 815 - 831 . Stebbins, G . L . 1981 . Coevolution of grasses and herbivores . Annals of the Missouri Botanical Garden 68, 75 -86. Vermeij, G.J. 1 982. Unsuccessful predation and evolution. The American Naturalist 120, 701 - 720 .

2 . 9 Adaptation P . W . SKELTON

adequacy in the face of competition becomes the expected rule . Persistent competition may eventu­ ally perfect some adaptations, but there are many reasons, ranging from environmental change to the inherent constraints of bodyplans, why others are not perfected . B y limiting the notion o f function t o the effect of any given feature on the lives of its possessors, and by placing natural selection in the creative driving seat of evolution, Darwinism avoids the teleology, and thus the unacceptable mystery of earlier explanations of adaptation . Evolutionary thinking requires a distinction be­ tween adaptation as a process of gradual modifi­ cation in a population, and as a state of being in individuals, in relation to prevailing circumstances . With natural selection, the state o f being adapted, in respect of some feature or complex of features, also takes on two aspects: first, there is the element of function - the way in which the feature or complex operates - and, second, there is the selec­ tive benefit to the possessors of the feature or com­ plex, in terms of preferential survival and/or fecundity, deriving from its operation . Varying emphasis on one or other aspect in different usages of the term 'adaptation' has led to much confusion and misunderstanding; so it is worth teasing them apart somewhat. Darwin himself still used the term in an essentially vernacular fashion, primarily stres­ sing functional suitability: the 'best adapted' were simply those individuals possessing the most 'useful' variations for the operation of various life functions, such as feeding, locomotion, and seed dispersal. These, he repeatedly postulated, would tend to be favoured by natural selection in the 'struggle for existence', so fuelling the continuing process of further adaptation in populations . This Darwinian sense of adaptation, still recognizable independently of selective effects, though assumed to be both the product of, and producing them, is still widely used today, especially by palaeontol­ ogists, for reasons discussed below . Evolutionary biologists, on the other hand, have considerably refined the theory of natural selection and, in so doing, have subtly redefined the meaning of adaptation precisely in terms of selective effects .

Introduction Natural historians have long admired the ways in which the construction and activities of living organisms seem to be so well suited, or 'adapted', to the natural circumstances in which they live . Especially striking is the extent of co-operation of their features that renders them so adapted . Illustrations readily spring t o mind, such a s the stiffened tail, the two backward-pointing toes on each foot, the stout beak, and the extendible tongue of the woodpecker, which together enable it to perch on tree stems and probe them for insects . To most pre-Darwinian thinkers such adaptive traits were the essentially static attributes of fixed species, perfectly fitting them, perhaps by divine appointment, to their places in nature . Certain pre-Darwinian transformists, such as J-B . Lamarck and Erasmus Darwin, in contrast, postulated that the inheritance through many generations of habitually acquired developmental modifications (such as, say, the building up of well exercised muscle) was a means of evolutionary adaptation . But there has never been any satisfactory evidence for this . Both the static and the Lamarckian views of adaptation were teleological; that is to say they appealed to final causes in placing the prospect of a function (either in God's mind, or in the 'needs' or 'strivings' of organisms) prior to the appearance of the feature adapted for it. Such explanations, though still rife in the popular imagination, are categorically denounced by most biologists today because of the inherently untest­ able character of the final causes . The DarwinIWallace theory o f natural selection, on the other hand, asserted that adaptations be­ came established in evolving populations through the preferential survival and reproduction of indi­ viduals possessing naturally occurring, heritable variations, which conferred advantage in the 'strug­ gle for existence' arising from the excessive fec­ undity of the populations in relation to resources . This means that adaptation need not b e perfect, as earlier naturalists had tended to opine . Rather,

139

140

2 The Evolutionary Process and the Fossil Record

Adaptation in evolutionary biology The 'neo-Darwinian synthesis' of the nineteen­ thirties to the nineteen-fifties attributed all evo­ lutionary adaptation to the operation of natural selection upon the phenotypic manifestations of genes in populations . Darwin's (and others') speculations on the additional operation of pro­ cesses other than selection (such as the effects of use and disuse, and other somatic influences on the germ line) were brushed aside . Statistical models of population genetics allowed the operation of natural selection to be quantified . 'Fitness', which to Darwin had been a somewhat vague expression of relative adaptedness of individuals, now became rigorously defined as the proportional survival and fecundity of a given genotype (usually simplified in scope so as to refer to all carriers of a specified pair, or pairs of alleles) relative to that genotype in the population which has the most descendants . (Unfortunately this tends to be referred to as 'Darwinian fitness' . Perhaps 'neo-Darwinian fitness' would be better, to distinguish it from Darwin's vaguer usage . ) Since the effect of any truly 'useful' feature on some life function may be assumed to contribute to the fitness of its possessors, the term adaptation, too, came to be defined by neo-Darwinians in terms of the pro­ motion or maintenance of fitness . Dobzhansky (1970), for example, cited the expression 'adaptive value' as a synonym for 'Darwinian fitness' . Or again, more recently, Ridley (1985) has stated 'Adaptation means good design for life . To understand how any particular property of an organism is adapted, it is necessary to think how it enhances its bearer's chances of survival and reproduction' . This subtle shift of emphasis in the definition from mere operational suitability to selective effect means that, in order to demonstrate adaptation, it is not good enough simply to show the effectiveness of a feature in the service of some function, however impressive that may be; it has to be shown that the feature thereby confers greater fitness on its possessors relative to alternatives . Williams (1966) further qualified this selective criterion . A given feature may accidentally benefit its possessors in special circumstances without any prior adaptation for that particular effect. For example, quick reac­ tions clearly promote the fitness of car drivers, though they obviously did not evolve by virtue of that effect . Williams considered such an 'effect' an inadequate criterion for recognizing true adap­ tation; his definition of the latter also requires evi­ dence for prior moulding of a feature by natural

selection in the service of its recognized junction(s)' . However, this creates an awkward grey area for the practical consideration of the origin of adaptations, when chance 'effects' are transformed by natural selection to established 'functions' (discussed below), and so the distinction between the two cannot always be recognized . The neo-Darwinian formulation has the effect, worrying to some, of making Darwin's charac­ terization of natural selection, 'the survival of the fittest', explicitly tautologous (as 'the survival of the survivors' ) . However, far from trivializing the theory, as might at first seem to be the case, this conclusion represents the logical outcome of purging it of teleology; selection simply acts on what organisms actually do, not what any meta­ physical agent thinks they 'ought' to do . Evolution is thus seen to be drawn in unpredictable directions by the transient effects of a myriad of immediate causes, making adaptation highly conditional . Clear illustration of this is provided by the banded snail, Cepaea nemoralis (Linne), common in parts of Britain and continental Europe . Most populations of this species are strikingly polymorphic, showing dif­ ferences in both the colour of the shell (shades of brown, pink, or yellow) and its patterning (with, or without, a variable number of longitudinal bands) . Cain and Sheppard (1954) were able to link marked differences in the relative frequencies of these variants in different habitats with preferential predation by thrushes (as estimated from broken shells around the birds' 'anvil' stones) . Effective camouflage was found to be the guiding principle, with, for example, unbanded pink and brown shells dominant in the browny leaf litter of beech woods, unbanded yellow shells on shortgrass downs, and banded yellow shells in hedgerows and longer grass . So, features adaptive in one setting were found to be demonstrably maladaptive in another, in many instances only a short distance away . Nor, indeed, can one even generalize to the extent of saying that colour and pattern, although variable, are in principle adaptive for the single broad func­ tion of camouflage; subsequent work has shown that other factors, such as response to temperature change, are of overriding importance in some per­ ipheral populations, in which different shell types acquire differing fitnesses because of their greater or lesser tendency to absorb solar heat ( Jones et al. 1977) . The intimate linkage of adaptation with natural selection raises the issue of what, precisely, natural selection acts upon, and what is thus the focus for

2 . 9 Adaptation adaptation. In cases like that of the Cepaea poly­ morphism cited above, particular variations can be attributed to the action of single gene loci (e . g . for shell colour, for presence or absence of bands, and so on) . For a given population, simple com­ binations of alleles can be assigned fitness values directly derived from, say, the live and broken shell counts . From such data one may model changes in the relative frequencies of the phenotypic variants, and thus the course of adaptation, in the population . Such exercises are the stock-in-trade of population genetics (Dobzhansky 1970) . But as Dobzhansky himself has pointed out, this apparent focus on genes as units of selection is illusory, since the fitness values assigned to them are statistical abstractions derived from the fate of those genes in many different total genotypes . Any given gene, no matter what its notional fitness value, will stand or fall according to the fate (survival and reproduction) of the genotype it finds itself in . Moreover, the phenotypic expression of any gene usually depends greatly upon interactions with other genes in the genotype and with the environment. So selection really only operates directly upon the phenotypic expression of whole genotypes . But then again, in sexually reproducing organisms, only the genes survive intact from one generation to the next . So we are here dealing with a hierarchical effect: selection on individuals in a population has the effect of altering gene frequencies in the gene pool of that population, and the latter changes in turn alter the genetic complexion of individuals in future generations, in adaptive ways . This means that adaptations, at no matter what organizational level within the individual, from the coadaptations of regulatory genes in chromosomes, through those of organelles in cells, and of tissues in organs, to the modifications of whole components of morphology, physiology and behaviour, must involve a net benefit to the individuals possessing them . The only case where this is not literally so is where there is 'kin selection' in favour of close relations, who are genetically similar if not identical . In such cases altruistic behaviour of individuals may promote their own eventual 'inclusive fitness' (effective genetic representation in future generations) through the reproductive efforts of the kin they assist, albeit at personal cost. So, for example, worker bees literally have an individual fitness of zero, being sterile, but are (painfully) well adapted to defend their genetically similar sisters due to become queens, who reproduce for them. The study of such adaptations has grown enormously from

141

seminal work done in the nineteen-sixties, particu­ larly by W.D. Hamilton, and the theme is succinctly reviewed by Grafen (1984) . Other kinds of 'group selection' arguments - involving the notion of adaptations arising 'for the good of the species' require somewhat unrealistic circumstances to work, and so have found little favour with evol­ utionary biologists (Williams 1966) .

Adaptive diversity The previous section showed that in any consider­ ation of function and selection, organism and environment are inseparable . Lewontin (1983, p. 280) further stressed that ' . . . the environments of organisms are made by the organisms themselves as a consequence of their own life activities . How do I know that stones are part of the environment of thrushes? Because thrushes break snails on them. Those same stones are not part of the environment of juncos who will pass by them in their search for dry grass with which to make their nests . Organisms do not adapt to their environments; they construct them out of the bits and pieces of the external world . ' While the claim that 'organisms do not adapt to environments' is perhaps a little over­ enthusiastic, Lewontin's point about organisms defining their environment is important. The con­ stant dynamic interplay between the niche that each species so defines for itself (in Lewontin's terms) and the selection imposed on the individuals of the species by the changing constraints of that niche is one major reason for the bewildering adaptive diversity of life . Organisms of different sizes experience different environmental constraints, because of physical scaling effects, and are correspondingly diversely adapted : so, for example, the construction of an elephant has much to do with coping with gravity, while that of a pond-skater has more to do with surface tension; and bacteria (if they could think!) would probably find the notion of gravity about as abstruse as an elephant would find their experience of being jostled by molecules and ions . Then, again, there are the differences between media : stream­ lining is hardly an overriding design factor in the terrestrial mammalian carnivores, yet it is clearly a vital adaptation for their marine cousins, the seals . And even in the same circumstances, differences in habits create different experiences of the world : zebras see grass as food, and have the jaws and teeth to cope with it; lions see it as useful cover on the way to the zebras, for whom their jaws and

142

2 The Evolutionary Process and the Fossil Record

teeth are suited . Adaptation breeds diversity, as Darwin rightly emphasized . Yet there is order in this diversity . The physical laws and functional specifications of habit men­ tioned above demand analogous adaptations for similar circumstances, leading to convergence: the streamlined shape of the seal is broadly repeated in other fast swimmers of similar size, such as penguins, porpoises and sharks, though with dif­ ferences in detail, of course, reflecting inherited differences in basic architecture . Moreover, the in­ herited bodyplans and the modes of growth of organisms limit the adaptive possibilities open to them. For example, gas exchange in insects takes place via branching tubes reaching into the body from spiracles along their sides . The main branches of this 'tracheal' system are ventilated tidally by contractions of the body, but in the finer tubules supplying the muscle fibres, diffusion alone suf­ fices . Muscle fibres must therefore lie close to axial tracheae, and so muscles cannot exceed a few mil­ limetres in diameter . This in turn constrains body size, the largest known insects being some Carbon­ iferous dragonflies with a wingspan not much greater than that of a crow . Considerations of such architectural constraints on evolution, and hence the mapping out of the adaptive potentialities that remain, are the business of constructional morphology (Section 4 . 1 ) .

Adaptation in palaeontology The neo-Darwinian definition of adaptation is prob­ lematical, to say the least, for palaeontologists - as the mere contemplation of trying to gauge relative survival and fecundity of genetically differing vari­ ants in fossil assemblages should suggest. However, this is not impossible . Using an ingenious argument, which deserves wider exploitation by palaeontologists, Sambol & Finks (1977) documented natural selection in a population of the Cretaceous oyster, Agerostrea mesenterica. A bulk collection from an undisturbed assemblage of their markedly plicate, arcuate shells was made from a single locality in the Maastrichtian of New Jersey, U . 5 . A . From the annual growth increments of the shells it was possible to deter­ mine the age at death of some two and a half thousand individuals . Censuses of selective mor­ tality could thus be carried out in relation to four morphometric parameters of the shells (Fig . 1 ) . The censuses showed that older individuals clustered more tightly than younger ones around mean

x

(

y

)

Fig. 1 Agerostrea mesenterica. Morphometric parameters measured by Sambol & Finks (1977) : shell arc length (AL); maximum plical height (PH); number of anterior plicae (in this case, 8); and curvature index (Y/X) . Exhalent flow would have issued around the concave posterior part of the shell .

values for the number of plicae and, with some unavoidable bias from ontogeny, for plical size as well as the overall arc length of the shell, indicating centripetal (stabilizing) selection on these features . The arcuate shape o f the shell, in contrast, was subject to differential mortality favouring maximum curvature; i . e . directed selection had operated . The features investigated all had well established functional linkages with gill suspension feeding, detected from comparisons with living oysters . In particular, the selection for increased curvature of the arcuate shell would have maximized the velocity of the exhalent current, thereby reducing the chances of recycling the processed water through the gills . This then is a clear 'snapshot' record of adaptation by natural selection in a fossil popu­ lation . As stressed by Sambol & Finks, however, the data only show the time-averaged pattern of selection on the several generations of oysters comprising the assemblage, which probably accumulated over some 200 years . Although the selection for increased curvature is consistent with the morphological trend shown by successive species of the oyster's inferred phylogenetic lineage, only a small fragment of the history of natural selection operating in this case has been sampled . Indeed, because o f the virtually insurmountable practical difficulties attached to linking longer-term evolutionary changes in the fossil record with measurable natural selection, palaeontologists have continued to use the term 'adaptation' in the sense generally adopted by Darwin, stressing functional

2 . 9 Adaptation suitability, rather than in direct reference to effects on neo-Darwinian fitness values . This distinction is important in that it lays some palaeontological per­ ceptions of adaptation open to deserved criticism . The danger is that the morphology of a fossil organism can too easily become atomized in the mind of the palaeontologist to so many discrete components, to each of which a function is imagin­ atively assigned according to the apparent suit­ ability of its morphology . The implicit assumption is that every feature must serve some function, or it would not be there . So, if one story is found wanting, another can be slipped into its place . This reductionist approach, branded as 'the adaptationist programme' , has been critized by Gould & Lewontin (1979) for proliferating adaptive hypotheses ('Just So Stories') where none may be warranted . Many features are simply the geometrical con­ sequences of the way organisms grow, and need no functional explanation per se (see also Section 4 . 1 ) . For example, any given point o n the aperture o f a Nautilus shell traces a near perfect logarithmic spiral with growth . One could devise all manner of specious arguments for how this might be 'adaptive', and the precision with which this geo­ metry is maintained might then be considered evidence for stabilizing selection . However, a brief consideration of the way the shell grows demolishes such arguments . If shell incrementation proceeds at fixed rates around the growing aperture of an expanding coiled tubular shell, logarithmic spiral growth is the geometrical consequence . That aspect of the Nautilus shell needs no adapting in order to arise, and so functional explanations for it are redundant; indeed, modification of the growth mechanism itself would be necessary to escape from such a geometry . It is thus imperative, as stressed above, always to consider the whole organism as a developing entity, in its environmental context. Bits and pieces can­ not be interpreted in isolation . Nevertheless, within that 'holistic' framework it is not only legitimate, but pragmatic to consider how particular components might have contributed to the overall conduct of an organism's life, by virtue of adaptive modification from some constructional groundplan (Mayr 1983) . Having dispensed with the erroneous re­ ductionist demand for a function for every feature, we must now ask: how is original function to be detected at all in fossil organisms (see also Section 4 . 1), and by what means can adaptation for it (in the operational sense) be diagnosed? Three steps are

143

necessary: (1) from a consideration of what is known of the organism's affinities, construction, and autecology, a plausible function, or alternative functions, may be proposed for a given feature or set of features, where a constructional argument alone seems inadequate . Such hypotheses may be suggested by comparisons with similar living organisms, the design attributes of analogous machines, or even simply from theoretical con­ siderations; (2) the suitability of the feature's construction (within the constraints determined by the organism's bodyplan and mode of growth) for the proposed function must be tested . This is done by comparing it with an idealized model (paradigm) designed for that function, to see how effective the feature would be in its service; and (3), crucially for the confirmation of adaptation, evidence that the feature has indeed been modified from some dif­ ferent ancestral condition, so as to approach the form of the paradigm, must be discovered, to show that the feature in question probably did perform the function attributed to it. Testing for adaptive convergence with a paradigm really only requires an evolutionary sequence of specific modifications to be established, and this can be derived even from an outline phylogeny . There should also be reasonable evidence that the feature(s) in question consistently served the same broad function . Not all features lend themselves to such a broad-brush approach, of course, as the discussion of Cepaea polymorphism above illus­ trated. Others, however, usually concerned with such basic operational functions as feeding and loco­ motion, can be relied upon with greater confidence . A good example of such a test is Chamberlain's (1981) study of streamlining and static stability in ammonoids . Several lines of evidence suggest that the smooth-shelled ammonoids which he studied maintained the swimming habit. From an exper­ imental study of accurately constructed models he was able to draw contours of drag coefficient and static stability values on a graph of possible shell shapes, and so to identify two 'adaptive peaks' where these factors were most favourable for ef­ ficient swimming . Real ammonite data show an impressive migration to the higher of these peaks with time, providing strong circumstantial evidence for adaptation.

The origin of adaptations Ultimately, however, one is faced with the question of how an adaptation arose in the first place . Natural

144

2 The Evolutionary Process and the Fossil Record

selection certainly provides a mechanism for adap­ tation once a functional effect has become apparent, but biologists can only speculate about how a feature might have become involved in this process, from comparative studies of living forms . Here palaeon­ tologists come into their own, for the fossil record furnishes the only concrete record of evolutionary history . But the virtual impossibility of charac­ terizing the role of selection in much of this means that arguments about the origins of adaptations have to be cast in terms of structural change and likely functional consequences . Two modes o f origin are conceivable : either the feature newly appeared, or it was derived from some precursor . The former implicates a 'hopeful monster' , presumably generated by some macro­ mutation. Quite apart from the vanishingly small probability of such an extreme mutation yielding a fitter genotype, the main problem with the hopeful monster model is its untestable nature . Recognition of a homologous precursor to an adapted feature, on the other hand, allows testing for functional co­ option . The point to be established is that the precursor did not originally serve the eventual function, but that with some slight modification (for whatever reason) or change in environment, ' it fort .1itously manifested effects similar to the eventual function, for which it thus became adapted . Such a precursory feature is conventionally termed a preadaptation and such, for example, would have been the grasping hands of human ancestors, for tool use, when liberated by bipedalism . Some biologists instinctively recoil from the term, believing it to smell of teleology . Its literal am­ biguousness (it might be construed, erroneously, as referring to some mysterious process of adaptive priming prior to the acquisition of a function) is a trivial problem, for which the remedy is simply learning the correct definition, with its crucial refer­ ence to fortuitous co-option . A second complaint, that it does nevertheless seem to suppose some end-directed evolution in its reference to eventual function, is a misunderstanding rooted in the essentially different working methods of many biologists and palaeontologists . Biologists investi­ gating microevolution can directly analyse pro­ cesses, but frequently stress the unpredictability of the longer term outcomes of evolution because of the myriad influences at work . Palaeontologists are in the opposite situation, knowing (some of) the outcomes of evolution, but with little direct evidence for the processes involved . The apparent 'end­ directedness' of a preadaptive hypothesis is simply

the benefit of hindsight. While the danger of sup­ posing history to have been inevitable must be avoided, it is legitimate to try to determine in retrospect at least some of the more prominent factors which made it take the unique course that it did follow. Any explanatory hypothesis must be tested against other historical models (in much the same way that Sherlock Holmes might have reconstructed the true nature of a crime) . In testing a preadaptive hypothesis, it is necessary to predict (strictly, retrodict) in detail the probable historical outcome of that model, beyond what has already been established, and to show how the retrodictions of alternative models significantly dif­ fer . Closer inspection of the fossil record can then point to one or other model being the more prob­ able, perhaps on a statistical basis . Skelton (1985) adopted this approach in testing a preadaptive hypothesis for the evolution of rudist bivalves (Fig. 2) . Constructional analysis had suggested that the spirogyrally coiled primitive forms were con­ strained by their growth geometry from exploiting the 'adaptive zones' (broad styles of adaptive morphology) occupied by their uncoiled tubular descendants . Shortening and eventual invagination of the external ligament in some spirogyrate forms had been identified as the preadaptive step which allowed the constructional changeover to uncoiled growth . The retrodiction of this model was that uncoiled taxa should have undergone an initially exponential diversification, focusing on the incep­ tion of ligamentary invagination, unmatched by their contemporaneous spirogyrate cousins . Other historical models (including a null hypothesis of random speciation and extinction) gave different retrodictions . An analysis of stratigraphical range data yielded the pattern given by the preadaptive hypothesis, at the generic level (though the species data were less enlightening, probably because of preservational bias), and this was taken to confirm the novel adaptive exploitation of the preadapted condition in the uncoiled clade .

Terminology Could and Vrba (1982) have expressed dissatis­ faction with this terminology for discussing the origin of adaptations . Noting the literal connotation of the word 'adaptation' to imply that something has been progressively 'fitted to' (ad + aptus means 'towards a fit') the execution of some function, they followed Williams (1966) in restricting the use of that term to those features which can be shown to

2 . 9 Adaptation

Fig. 2

Synoptic evolutionary history of uncoiling and its consequences in the rudists . (After Skelton 1985 . ) In primitive forms such as Diceras (lower left), the external ligament (eL) constrained the shell to grow spirogyrally, limiting its adaptive scope . Shortening and invagination of the ligament (iL) in Monopleura (centre) allowed 'uncolled' growth. Adaptive diversification ensued (e.g. clockwise from top left, Durania, Hippurites, and Pachytraga) as uncoiled taxa entered new adaptive zones.

have been shaped by natural selection for their current use . Other features, which have some useful effect by virtue of their construction, but which show no clear evidence of having been produced by natural selection through the expression of that effect, they termed exaptations ('fit (aptus) by reason of (ex)') . However, they conceded that exaptations may undergo 'secondary adaptation', so enhancing their effectiveness . As an example, they cited the useful role of the skull sutures in young mammals in aiding parturition; these are also to be found in young birds and reptiles, where they obviously have no such role to play . Together, they designated (their) adaptations and exaptations as aptations simply meaning features fitted to some function or effect. Exaptations were seen as being co-opted (,co-optation') either from pre-existing adaptations for other functions, or from constructional elements with no previous functional effect ('nonaptations') . Both constitute forms o f what has been labelled earlier, here, as 'preadaptation', though Could and Vrba criticized this term . They argued that : (1) it fails to distinguish the two kinds of exaptation, and

145

so appears to have an adaptationist bias in hinting at an earlier adaptive role for the precursory feature; and (2) it is misconstructed, implying prior fitting towards some subsequent function (the teleological odour detected by some biologists) . Time will tell if these neologisms are adopted . However, their practical value has to be queried . To deal first with the more trivial aspect of etymological correctness, if the 'fitting-to' of adaptation is simply taken to mean 'suitability for', as is the implication both in common neo-Darwinian and Darwinian (,useful variations') usage, then preadaptation literally connotes no more than 'prior suitability for' ( . . . some fortuitous functional effect) . More im­ portantly, Williams' (1966) distinction between 'effects' and 'functions' (discussed above), upon which Could and Vrba base their 'exaptations' and 'adaptations', breaks down when origin of adap­ tations is considered. In so far as any adaptation is derived from a precursory feature (whether some preadaptive trait or even a mutational novelty), then the latter must have passed through the stage of being an exaptation - i . e . exhibiting fortuitous beneficial effects - to have become subject to the selection that produced the adaptation . But the very moment that such 'exaptive' benefits were expressed, fitness would have been affected and selection would have started adapting the feature . Thus, although skull sutures are indeed 'exaptive' for mammalian parturition, the extended delay in their closing up is clearly adaptive for that process . Could and Vrba might term this a 'secondary adap­ tation', but surely this is no different from any 'primary adaptation' if we accept that all are founded on exaptations . The only reason we may choose to call one thing an exaptation and another an adap­ tation relates to the degree of modification shown . For example, it is easy to think of the skull sutures mentioned above, with the slight adaptive delay in their closure, as an exaptation, but the hooked beak of an eagle would be branded by most people as a clear adaptation for tearing flesh . Yet it is only the shape of the beak which is thus adaptive; the beak itself was, again, an exaptation for the role . In other words exaptation and adaptation are really just two aspects of the same thing, the former emphasizing derivation and the latter, destiny . To attempt to distinguish them as separate entities (which is implicit in any statement that some feature is an exaptation and not an adaptation or vice versa) seems to be as illogical as classifying the 'arrivals' and 'departures' at a railway station as two funda­ mentally different kinds of train .

146

2 The Evolutionary Process and the Fossil Record

Again, in view of the difficulties attached to de­ tecting natural selection in fossil materiat how practicable is it to attempt to distinguish between an 'exaptation' derived from a previous adaptation and one derived from a 'nonaptation'? How could one show that some feature of a fossil organism (even if dearly a product of the mode of con­ struction) did not somehow adapt the organism to its niche? The conventional toolkit of terms - adaptation and preadaptation - appear to suffice for the nature of the material to be studied, with the proviso that they are used with well understood and precisely defined (preferably explicitly stated) meanings :

Adaptation. In neo-Darwinian usage, this is a feature or complex of features in an organism which promotes or sustains the (neo-Darwinian) fitness of its possessors; or, in a palaeontological context, which has some identifiable functional effect pre­ dicted to have been of selective benefit to its pos­ sessors (a prediction which if borne out by some means of analysis would render the feature a neo­ Darwinian adaptation as well) ; or, in both cases, the associated historical process of modification of features in an evolving population .

Preadaptation . This is a feature or complex of features of an organism, whether already serving some func­ tional role or merely a constructional product, which, by virtue of its fortuitous suitability for novel functional effects, becomes co-opted as a new adaptation (in the senses given above) in descend­ ants of the organism . It should be dear that, despite the apparently simple meaning of adaptation as a vernacular term, and its fundamental importance in evolutionary theory, it actually opens onto a terminological

minefield . Safe routes across can only be picked out by adhering to dear definitions and thinking very carefully about their practical applications.

References Cain, A.J. & Sheppard, P.M. 1954. Natural selection in Cepaea. Genetics 39, 89 - 1 16. Chamberlain, J.A., Jr. 1981 . Hydromechanical design of fossil cephalopods. In: M . R House & J.R. Senior (eds) The Ammonoidea. Systematics Association Special Volume, No. 18, pp . 289 - 336 . Academic Press, London. Dobzhansky, T. 1970 . Genetics of the evolutionary process . Columbia University Press, New York. Gould, S.J. & Lewontin, R.e. 1979 . The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proceedings of the Royal Society of London, B205, 581 - 598. Gould, S.]. & Vrba, E.5. 1982. Exaptation - a missing term in the science of form. Paleobiology 8, 4 - 1 5 . Grafen, A. 1984. Natural selection, kin selection and group selection. In: ] . R Krebs & N . B . Davies (eds) Behavioural ecology: an evolutionary approach, pp. 62- 84. Blackwell Scientific Publications, Oxford . Jones, J . S . , Leith, B . H . & Rawlings, P. 1977. Polymorphism in Cepaea: a problem with too many solutions? Annual Review of Ecology and Systematics 8, 109 - 143 . Lewontin, R e . 1983 . Gene, organism and environment. In: D.S. Bendall (ed . ) Evolution from molecules to men, pp. 273-285. Cambridge University Press, Cambridge . Mayr, E. 1983. How to carry out the adaptationist program? The American Naturalist 121, 324 - 334. Ridley, M. 1985 . The problems of evolution, p. 26. Oxford University Press, Oxford . Sambol, M. & Finks, R M . 1977. Natural selection in a Creta­ ceous oyster. Paleobiology 3, 1 - 16 . Skelton, P.W. 1985 . Preadaptation and evolutionary inno­ vation in rudist bivalves. In: J.e.W. Cope & P.W. Skelton (eds), Evolutionary case histories from the fossil record. Special Papers in Palaeontology No. 33, pp. 159 - 1 73 . Williams, G . e . 1966. Adaptation and natural selection - a critique of some current evolutionary thought . Prince ton University Press, Princeton .

2 . 10 Evolution of Large Size M. J . B E N T O N

body weight is proportional to volume (a three­ dimensional measure) . Thus, bone cross-sectional area has to increase relatively faster than body weight, which is why elephants and dinosaurs have legs like tree trunks (Fig . 2) . Under high stress, leg bones can buckle, or they can break without bending much . The strength of muscles also limits the size of an animal . A large animal has to be able to pull itself up from a lying position, and the heavier the animal is the more massive its muscles must be . So, muscle dimensions and muscle strength also limit the maximum size of a land animal . Locomotion is yet another limiting factor . A hypothetical animal weighing 140 tonnes could stand safely enough, but if it walked its legs would break. This is because, in walking, the force of the weight of the animal is expressed at an angle through the leg bones. Even if a giant animal could stand safely with its legs positioned vertically beneath it, it might not be able to walk because the breaking force of the bone is relatively greater . Hokkanen (1986) concluded that the heaviest pos­ sible animal able to walk on four legs would have weighed no more than 100 tonnes. The largest dinosaurs have estimated weights in the range of 80 - 140 tonnes, but the larger forms are poorly known . The 78 tonne weight of Brachiosaurus is the greatest generally accepted weight known for a terrestrial animal . The strength of bone and muscle, as described above, would have limited Brachiosaurus to a sedate walking pace of about 1 m/s with strides of only 2 . 5 m or so (quite short for an animal with 3 m legs) (Alexander 1985) . In land plants, the continuously growing supporting tissues (lignin-lined xylem cells) within a tree trunk allow vast heights and weights to be achieved . The maximum height is probably limited in part by the ability of a plant to raise sap . Water has to be 'pumped' from the ground and raised up the trunk, against the force of gravity, by means of osmosis (the sap has a higher salt content than the ground water), and the hydrostatic effect of tran­ spiration (water loss through leaves exposed to the air) . There are also mechanical constraints imposed by

Introduction Many plants and animals of the past and present are very large compared to the human scale . In particu­ lar, vertebrates, gymnosperms, and angiosperms achieved giant dimensions on occasion, and ap­ parently several times independently in each group (Table 1; Fig . 1 ) . The focus here, however, will be on truly large organisms on the human scale . The key macroevolutionary questions to be asked are : 1 Why do certain groups achieve giant size while others do not? Is it simply chance, or are there historical and mechanical reasons? 2 Why do some groups never produce giants? 3 Does evolution always go from small to large, or can it reverse? 4 How long does it take for large size to evolve in a lineage? 5 Are large organisms better adapted than small ones?

Giants and mechanical constraints The bony internal skeleton of vertebrates is ideally suited to supporting great weights in terrestrial giants . The acquisition of a fully upright posture in both dinosaurs and mammals, where the limb bones are tucked immediately beneath the body, permitted giants to evolve . The major constraints on large size in a terrestrial vertebrate are limits to the strength of bones and to the power of muscles . As animals become larger, the bones and muscles in the legs come under increasing strain, and there have to be modifications in their shape and design . Hokkanen (1986) made simple biomechanical calculations of bone and muscle strengths in order to determine the size of the largest feasible terrestrial tetrapod . Each leg bone must be strong enough to support one-quarter of the total body weight, or more if the weight is concentrated at the back, as is often the case, and there has to be a fairly large safety factor in order to allow the animal to walk or run . The strength of a bone is proportional to its cross­ sectional area (a two-dimensional measure), while

147

148

2 The Evolutionary Process and the Fossil Record

A selection of large organisms, giving some key dimensions . Fossil forms are preceded by t, and the weights quoted for these are estimates (a question mark implies that estimates are very uncertain because complete skeletons are unknown).

Table 1

Max. length (m)

Organism

Plants Algae Macrocystis, Pacific giant kelp Gymnospermophyta Sequoiadendron, Giant sequoia Pseudotsuga, Douglas fir Angiospermae Eucalyptus, Mountain ash Animals: Vertebrata Class Placodermi Dunkleosteus Class Chondrichthyes Cetorhinus, Basking shark t Carcharodon Rhincodon, Whale shark Class Reptilia Suborder Squamata Eunectes, Anaconda snake Python tKronosaurlls, Pliosaur Suborder Crocodylia t Deinosuchus Suborder Pterosauria t QlIetzalcoatlus Suborder Dinosauria t Brachiosaurus t Diplodoclls t Antarctosallrlls t 'Supersaurlls' t 'Ultrasallrus ' t 'Seismosaurlls' Class Mammalia Order Perissodactyla t Indricotherillm (= Balllchitherillm) Order Artiodactyla Giraffa, Giraffe Order Proboscidea Loxodonta, African elephant Elephas, Indian elephant Order Cetacea Balaenoptera, Blue whale Physeter, Sperm whale t Basilosallrus

Max. height (m)

Max . weight (t)

84- 1 12 126 . 5

c. 2500

60

114.3

9 10.5 13 12.6

15

0 . 23

8.4 10 15.2 16

wing span

1 1 - 12 23-27 27 30 ?24- 30 ?30- 35 ?30 - 36

11.3

0 . 09 12

?IS ?16 - 1 7

c. 6

40 - 78 18.5 80 ?75 - 1 00 ?100- 140 80+

20

5-6 7-10 6 33 . 5 20 . 7 21 . 3

the vast weight of a tall tree and the possible strength of its trunk. The weight acts vertically down the trunk, but winds can cause tremendous stresses as the crown of a tree is pushed from side to side . Experiments show that winds with speeds of 60 - 65

3-4.4 3

2-10 4 190

kmlh exert a lateral force on the tree equal to its weight (Fraser 1962) . The girth of the tree then increases in proportion to the weight (i . e . relatively more rapidly than the height increases) . At 100 m tall, a tree may be as much as 30 m in circumference

2 . 1 0 Evolution of Large Size

Fig. l

149

A selection of large animals drawn to scale . Measurements are given in Table 1. (Drawing by Elizabeth Mulqueeny. )

(Table 1), and a t much greater heights, the circum­ ference would tend to approach the height .

Why so few giants? Most other groups of organisms appear to be re­ stricted from achieving large size by mechanical and physiological constraints . For example, arthro­ pods have an external skeleton which has to be moulted frequently as the animal grows . After each moult, the animal is soft-bodied for a while, and hence vulnerable . The shed skeleton also represents a loss of body materials that have to be replaced . To achieve giant size, an arthropod would suffer the cost of moulting dozens of times . A more important constraint on large size is probably the respiratory system of tubes in the exoskeleton that allow air to

diffuse throughout the body passively . At moderate to large size, this technique would not allow all body tissues to receive an adequate supply of oxygen . There are similar constraints on large size in most other invertebrates - e . g . the respiratory system of annelids and nematodes (simple diffusion into the body); the filter-feeding habits of brachiopods, most molluscs, coelenterates, bryozoans, graptolites, and some echinoderms; and mechanical constraints of the exoskeleton of brachiopods, most molluscs, and most echinoderms . It is assumed that filter-feeding by means of exposed cilia cannot sustain a large organism. The shells of brachiopods and molluscs can reach large sizes (e .g. the giant clam, Tridacna, 1 m across), but as body size increases, shell thick­ ness has to increase in proportion to body weight to

150

2 The Evolutionary Process and the Fossil Record

Fig. 2 The pillar-like skeleton of the forelimb of A, Elephas, the Indian elephant and B, Diplodocus, a sauropod dinosaur, showing convergent graviportal (weight-bearing) adaptations : columnar arrangement of shoulder girdle (sc= scapula) and limb bones, relatively long humerus (h), large separate radius (r) and ulna (u), block-like carpal bones (c), and relatively short finger bones spreading out over a cushioning pad .

maintain the strength of the shell . The potential weight of the shell, and the amount of particulate calcium carbonate to be extracted from the seawater, tend to prevent huge size . The same is probably true for echinoids .

Cope's Rule In 1887, E . D . Cope presented a new principle of evolution, that organisms always tend towards large size. He could find no examples in which a lineage or clade of plants or animals evolved towards smaller size. Although Cope never explicitly defined this as a 'law' of evolution, it has since come to be known as Cope's Rule . In considering Cope' s Rule, many authors have focused on particular advantages of evolving large size (see below) . However, Stanley (1973) argued that Cope's Rule had general application, not be­ cause of any particular advantages of large size, but since groups tend to arise at small body size relative

to their ecological optimum. Amongst mammals, for example, the original members of most clades in the Cretaceous and Palaeocene were small carni­ vores or insectivores . On the other hand, large forms are unlikely ancestors for major new lineages since they tend to be specialized to particular habi­ tats, often by virtue of the physiological demands imposed by large size . Stanley (1973) surveyed a range of animal taxa, and found that the ancestors of a clade were generally smaller, on average, than a random sample of their descendants . Histograms of body size tended to be concentrated initially at small sizes and to be rather symmetrical . Through time, the histograms developed longer and longer tails to the right as larger body sizes arose (Fig. 3) . Size decrease also does take place in many lin­ eages, but it is rare . For example, modern horse tails and clubmosses are midgets in comparison with their Carboniferous tree-like ancestors . Certain ver­ tebrate groups have also shown reductions in size since the Pleistocene, but some of the former giants (e . g . mammoth, aurochs, giant kangaroo and wombat, giant ground sloth, glyptodon, moa) may have suffered because of human influence (see also Section 2 . 13.8).

Evolution of large size The evidence of the fossil record is that giant size can evolve very quickly in certain groups . For example, the first (small) dinosaurs of the late Triassic date from the Carnian . By mid-Norian times, 5 - 10 million years later, prosauropods such as Plateosaurus had reached body lengths of 5 m . The sauropodomorph line then achieved a length of 12 m with Melanorosaurus in the Early Jurassic, and sizes continued to increase rather slowly until the Late Jurassic when the largest known dinosaurs occurred (Table 1 ) . This last phase of size increase towards giantism - a leap from body lengths of about 12 m and weights of 10 tonnes to maxima of 30 m and 80 tonnes or more, occurred between the Bathonian and the Kimmeridgian, a time of about 20 million years . Mammals achieved large size just as rapidly, if not more so . From a maximum of cat size just before the end of the Cretaceous, rhinoceros-sized uintatheres and astrapotheres are known 10 million years later in the Late Palaeocene and Early Eocene . The largest land mammal of all time, the rhinoceros Indricotherium, was in existence by the Early Oligocene, 30 million years after the radiation began . Whales achieved large size even more

2 . 1 0 Evolution of Large Size

o

n

=

85

Pl iocene

n

=

58

Miocene

n

=

47

10

Early & Middle Eocene

12

14

Anterior - posterior length of fi rst lowe r molar ( m m )

Fig. 3

The size ranges of North American rodents - an early group, and two later groups - to show the shift from small sizes to a broad range of body sizes including many large ones. The index of size is the length of the first lower molar, which varies directly with overall body size. (After Stanley 1973.)

rapidly - the Late Eocene Basilosaurus was 21 m long, after 15-25 million years of evolution . Stanley (1979) noted that, in contrast, 'large' molluscs took much longer to evolve . The first large free-swimming clam was Megalomoidea which ap­ peared after nearly 100 million years of radiation . The first large epifaunal bivalves, the inoceramid rudists of the Jurassic and Cretaceous, took nearly 400 million years to appear . Amongst land plants, large size arose at the end of the Devonian, and especially in the Carbon­ iferous, with the first tree-like clubmosses (Lepidodendron, 45 m high) and horsetails (Calamites, 16 m high) . This had taken 50 - 60 million years of land plant evolution . Really giant gymnosperms (Sequoia and other redwoods) are known from the Jurassic, as much as 250 million years after the origin of land plants, and 150 million years after the origin of gymnosperms .

Advantages and disadvantages o f large size Numerous advantages of large size have been postulated (Stanley 1973) : improved ability to cap­ ture prey or escape from predators, greater repro­ ductive success, increased intelligence (large bodies have large brains), better stamina, expanded size range of possible food items, decreased annual

151

mortality, extended individual longevity, and in­ creased heat retention per unit volume . Protection from predation would seem to be a great advantage . Adult elephants and rhinoceroses have no regular threat from carnivores today . However, thick­ skinned mammals of the Oligocene to Pleistocene of the Northern Hemisphere and South America were subject to attacks by specially adapted sabre­ toothed cats - the Machairodontidae in North America, Europe, Africa, and Asia, and the Borhyenidae in South America. The sauropod dino­ saurs are assumed to have been immune from attack since the largest predatory dinosaurs could only have tackled very young sauropods, or dying adults . A disadvantage of large size may be greater proneness to extinction . This is not simply an attri­ bute of large size, but rather an expression of specialization . Large animals are often more re­ stricted in their niches, in their scope for adaptation, than smaller relatives . Their need for large amounts of food, or for particular environmental conditions, may make them more likely to suffer when habitats change . Also, the fact that large animals tend to have small population sizes, and hence small gene pools, makes their hold on life seem more precari­ ous . The death of a few more individuals than normal may precipitate species extinction . Bakker (1977) showed that terrestrial tetrapods surviving mass extinctions in the Late Palaeozoic and Mesozoic tended to be of small body size . Thus, the large dicynodonts and dinocephalians of the Late Permian died out, leaving smaller dicynodonts and cynodonts to cross the system boundary. A similar explanation has also been given for selec­ tivity in the Cretaceous- Tertiary event on land (Section 2 . 1 3 . 7) . In more general terms, Stanley (1979) suggested that species longevity varies with the reciprocal of body size: small species tend to survive longer than large species. This is supported by evidence from the Pliocene and Pleistocene mammalian fossil record . The only modern species that can be tracked back before 3 Ma are small mammals . All the large ones arose after that, and this is probably not an artifact of a poor fossil record since such forms are more readily fossilized than small ones . Within any clade, lineages of large organisms may be expected to display shorter taxon durations, lower rates of speciation, and higher rates of extinction (Stanley 1979), and hence greater vola­ tility in the face of environmental stress. These ideas have yet to be tested thoroughly. They are of added interest since they could be seen as charac-

152

2 The Evolutionary Process and the Fossil Record

teristics that are subject to species selection (since these are not organism-level features) . They could also potentially be interpreted as examples of the 'effect hypothesis' (Vrba 1983; see also Section 2 . 6) . This hypothesis suggests that species-level charac­ teristics, such as species duration or broad ecologi­ cal adaptation, may be incidental effects of individual characters, such as dietary or habitat preferences. Natural selection, acting on organisms, might select for large body size, which in turn might produce higher extinction rates within a lineage . These higher rates could be interpreted as an incidental effect of natural selection, rather than as a result of species-level selection . These ideas are still highly controversial .

References

some large dinosaurs . Zoological Journal of the Linnean Society 83, 1 -25. Bakker, R.T. 1977. Tetrapod mass extinctions - a model of the regulation of speciation rates and immigration by cycles of topographic diversity. In : A. Hallam (ed . ) Patterns of evolution as illustrated by the fossil record, pp. 439-468 . Elsevier, Amsterdam. Fraser, A.I. 1962. Wind tunnel studies of the forces acting on the crowns of small trees. Reports on Forest Research 1962, 178 - 1 83 . Hokkanen, ] . E . I . 1986. The size of the largest land animal . Journal of Theoretical Biology 118, 491 -499. Stanley, S . M . 1973 . An explanation for Cope's Rule . Evolution 27, 1 -26. Stanley, S.M. 1979 . Macroevolution: pattern and process . W.H. Freeman, San Francisco. Vrba, E . S . 1983. Macroevolutionary trends : new perspectives on the roles of adaptation and incidental effect. Science 221, 387-389 .

Alexander, R.McN. 1985 . Mechanics of posture and gait of

2 . 11 Rates of Evolution - Living Fossils D . C . FISHER

Introduction The study of rates of evolution encompasses a wide variety of approaches to characterization of the amount of evolutionary change within particular groups of organisms, over specified time intervals . The high level of interest that palaeontologists and evolutionary biologists have shown in this subject is not surprising, since rates are a common focus in the analysis of any process . The importance of rates, however, is often only marginally attributable to intrinsic interest in 'how rapidly' or 'how slowly' a process operates . Rather, information on rates tends to be used as a means of investigating the under­ lying dynamics of the process in question, or some­ times as input for analysing the dynamics of a related process . Much of the work on rates of evo­ lution has thus been directed toward a better under­ standing of the dynamics of evolutionary change . Studies have been designed with the intent of com­ paring rates of evolution in a variety of ways within and between particular taxonomic groups, ecological settings, and lineage geometries (e . g . ,

ancestor -descendant sequences that include lin­ eage splitting versus ones that do not) . While inter­ esting generalizations are emerging, a greater appreciation is also being gained of the difficulties of quantifying rates of evolution . 'Living fossils' i s a term frequently used t o denote extant representatives of groups of organisms that have survived with relatively little change over a long span of geological time . Such groups are im­ plicitly recognized as having displayed unusually low rates of evolution . In both professional and popular literature, living fossils collectively appear to have attracted more attention than have groups displaying unusually high rates of evolution . This may be partly because, in keeping with the inherent paradox of the term 'living fossil', evolutionary his­ tory is expected to involve conspicuous change, and it is surprising when it does not . In addition, evolutionary rate statements are commonly (though not exclusively) framed in terms of putative ances­ tor - descendant pairs, and it is easier to recognize these when the total amount of change has been small than when it has been large . Instances of

2 . 1 1 Rates of Evolution living fossils are thus more likely to be accepted on prima facie grounds than are instances of higher evolutionary rates . In any event, living fossils have frequently provided a focus for discussions of evol­ utionary rate and have helped to clarify some of the factors that may be involved in promoting or inhi­ biting evolutionary change .

Three ranges of values for evolutionary rates G . G . Simpson was one of the early major contribu­ tors to the quantitative study of evolutionary rates. He proposed that rates be classified by their abso­ lute value as 'low', 'medium', or 'high' . Although this might be considered trivial, Simpson (1944, 1953) argued that frequency distributions of evo­ lutionary rates for sufficiently inclusive sets of taxa typically contain three discrete modes, allowing low, medium and high categories to be re�ognized on non-arbitrary grounds . This empirical claim sug­ gests some degree of disjunctness in the operation of the processes and/or constraints that interact to pro­ duce evolutionary change . Simpson coined the term 'bradytely' to refer to the phenomenon of supra­ specific taxa that have shown consistently low rates of evolution . Bradytely thus encompasses the same general concept implied by 'living fossil' , but with­ out the arbitrary stipulation that a representative of the group be alive today . Simpson also suggested 'horotely' to refer to taxa comprising the middle mode in the spectrum of observed evolutionary rates and 'tachytely' to refer to supraspecific taxa showing consistently high rates of evolution . Although Simpson's (1953) demonstration of the multimodality of evolutionary rates has sub­ sequently been shown to be flawed (Gingerich 1983; Stanley 1985), the terms denoting these rate cate­ gories (especially bradytely and tachytely) have had considerable heuristic value . They are now commonly used to refer to ranges of rate values regardless of whether multimodality has been de­ monstrated independently . For instance, in a study applying the terms in this latter fashion, Raup and Marshall (1980) showed that rates within several orders of mammals were significantly higher (e . g . Cetacea and Rodentia) o r lower (e . g . Perissodactyla and Carnivora) than the mean for all mammalian orders . However, whether evolutionary rate distri­ butions (at a given rank, within some more inclusive group of organisms) tend to show some 'typical' form and, if so, whether that form is multimodal, unimodal but non-normal, or unimodal and nor­ mal, are presently open questions .

153

Qualitative categories of evolutionary rates Evolutionary rates may also be categorized by the aspect of evolutionary change that is measured. Three commonly discussed categories are genetic, morphological, and taxonomic rates . However, various subdivisions of each of these are also sig­ nificant . For instance, genetic rates include rates of DNA nucleotide substitution and rates of gene re­ arrangement, among others . These two kinds of rates refer to different processes of genetic change, acting at different levels in the hierarchy of genetic struc­ ture . Each offers its own perspective on the general phenomenon of evolutionary change, and it is con­ ceivable that each will show a different frequency distribution, even over the same large group of taxa . In the same way, morphological rates are sometimes subdivided into 'size' rates and 'shape' rates, since these two factors are commonly treated as different, though not unrelated, aspects of mor­ phology . Finally, taxonomic rates include various approaches to measurement of the longevities and rates of origination and extinction of taxa . Termi­ nology for categories of taxonomic rates varies somewhat among authors, and each category may be further subdivided according to the taxonomic rank treated . In each case, the meaning of such rates depends critically on the underlying taxonomic phil­ osophy . The type of taxonomic rate that will be focused on here is the rate of origination of new taxa of specified rank, since this corresponds most closely to a 'rate of evolution' (i . e . without intro­ ducing aspects of extinction rate) . In this categorization of evolutionary rates, gen­ etic and morphological rates refer to changes in the genotype and phenotype, respectively . An alterna­ tive convention is to distinguish between molecular and morphological rates of evolution . This retains all aspects of genotypic change within molecular evolution, but adds to it components of protein evolution that would ordinarily be considered changes in the phenotype, albeit at a molecular level. Although molecular data are usually available only for living organisms, increasing effort is being focused on extraction of some molecular data from appropriately preserved fossil material (Section 2 . 1 ) . Still, except for the success of such efforts, molecular rates can only be measured directly over relatively short timespans . Alternatively, they may be com­ puted from the cumulative divergence of contem­ poraneous taxa . In this case, some parsimony assumption is used to partition change between or among the separate lineages involved . Although

154

2 The Evolutionary Process and the Fossil Record

this approach may seem to remove molecular rates from the domain of palaeontology, we must still relate measured divergence to the time interval over which it has developed - the time since the most recent common ancestor of these taxa . Tectonic or palaeogeographical data suffice for this in certain instances, but palaeontological data provide the most commonly applicable constraints on the time of splitting of lineages . For this reason, and because of their common focus on analysis of the pattern and process of evolution, palaeontology and studies of molecular evolution are closely related (Section 2 . 1 ) . Measurements o f morphological rates may also be based on comparisons among contemporaneous taxa for which the divergence history is relatively well known . However, when morphological features can be sampled in a succession of stratigraphic intervals, we have the option of calculating rates 'directly' from the fossil record . Since any source of morphological disparity between samples will con­ tribute to perceived evolutionary rate, it is im­ portant to be aware of, and if possible control for, non-evolutionary components of variation (e . g . dif­ ferential ontogenetic representation, differential taphonomic biases, or range shifts in clinally vary­ ing populations) . If it can be argued that consecutive samples represent a series of ancestors and their descendants within a species-level lineage - an ideal situation that approximates 'tracking' mor­ phology through time - the resulting rate is referred to as a 'phyletic' rate . However, if the phylogenetic context of consecutive samples is more complex or unresolved than this, the rate is better referred to as a 'phylogenetic' rate (Raup & Stanley 1978) . Phylo­ genetic rates imply a disclaimer recognizing that increments of change may have been measured between samples that do not bear a direct ancestor ­ descendant relation to one another . Depending on the history of morphological change and the pattern of phylogenetic relationships linking consecutive samples, phylogenetic rates may be either greater or less than corresponding phyletic rates (i . e . the phyletic rates that might be measured if an arguably ancestor-descendant sequence were available) . Both of these types of rate represent transformation within a 'lineage' (broadly construed, possibly at a supraspecific level), but they differ in the degree of resolution with which the lineage can be traced . Taxonomic origination rates are likewise de­ signed to quantify change through time, but they differ fundamentally from the rates discussed thus far . To the extent that new taxa are erected to recognize some increment of morphological change

within lineages, origination rates incorporate a transformational component comparable to that as­ sessed by molecular and morphological rates . How­ ever, origination rates also include a component representing the cladogenetic (or lineage splitting) aspect of evolutionary change . The relative contri­ butions of these two components - lineage trans­ formation and lineage splitting - are difficult to quantify and rarely reported. They vary from group to group depending both on taxonomic practice and on the actual evolutionary history of the group under study .

Units of measurement for evolutionary rates Genetic or molecular rates are sometimes quantified in terms of the number of events involving a par­ ticular type of change, per time interval . Com­ parisons of molecular rates may be normalized for the number of entities 'at risk' for change (e . g . number o f nucleotide substitutions per site, per million years), but this is not practical in all in­ stances (e . g . computing the number of potential gene re arrangements ) . Molecular rates based on distance measures (e . g . DNA - DNA hybridization, immunological distance) are given in units appro­ priate to the distance measure utilized . Morphological rates may be expressed as change in the value of some morphological variable (any appropriate units of measurement), per time inter­ val . However, variables of different dimensionality (e . g . lengths versus areas) must be divided by an appropriate factor before they can be properly com­ pared . Moreover, we are usually interested in proportional rather than absolute changes in mor­ phology . Given the scaling relationships of most morphological variables (and their variances), a convenient solution is to measure morphological rates in terms of differences in the logarithm of the value of the variable of interest. A difference of a factor of e (base of natural logarithms, 2 . 718) per million years was defined by Haldane (1949) as a morphological rate of 1 darwin (d) . Rates of taxonomic origination may be measured as the number of new taxa (within a given higher taxon) per time interval . This is often expressed as a percentage increase, normalized for the length of the time interval . Rate of origination may also be calculated from the rate of change in total diversity at a given taxonomic level and the rate of extinction at that level. In interpreting origination rates, it is important to consider such possible complications as differential effects of taphonomic and mono-

2 . 1 1 Rates of Evolution graphic biases, and differential application of taxo­ nomic practice within and between groups being compared (Raup & Marshall 1980) . From an evo­ lutionary standpoint, however, a more fundamental issue with rates of taxonomic origination is that they lump together information on lineage trans­ formation and lineage splitting . Given the current unevenness of our detailed phylogenetic knowledge of most groups, this may be an unavoidable com­ promise, and indeed, it offers some benefits of convenience and succinctness in the representation of evolutionary history . However, it is to be hoped that more phylogenetically discriminating ap­ proaches to studying diversification will be developed in the future .

The effect of measurement interval on evolutionary rates Measured rates are commonly treated as indepen­ dent of the interval length over which they are measured. For processes occurring at approximately constant rates, this characterization is acceptable . However, for any variable-rate process, the measured rate is an average and may be influenced strongly by rate fluctuations during the measure­ ment interval . Depending on the temporal structure of rate fluctuations and the range of intervals being considered, measured rates will be more or less susceptible to biasing effects from interval length . Some molecular rates appear to behave in 'stochastically constant' fashion, at least over certain time-spans (commonly of the order of tens of millions of years) . The relative constancy of these rates (with both rate and constancy varying from one molecular system to another) has led to the proposal of the 'molecular clock' hypothesis (see also Section 2 . 1 ) . According t o this hypothesis, molecular difference, once calibrated to reflect rate of change, can be used as a measure of time since lineage divergence (Fitch 1976) . However, even for molecular clocks that are relatively 'well behaved' over a particular time inter­ val within a given group, there is growing evidence that observable change has either accelerated or decelerated at other times during the history of that group (Goodman et al. 1982; Gingerich 1986) . For divergence times that span periods of significant rate change, systematic biases can be anticipated . The factors thought to influence morphological rates (see below) are known to fluctuate on a variety of time-scales . Because neither the highest nor the lowest rates are likely to be maintained over pro­ tracted periods of time, the largest range of variation

155

should be observed in comparing rates measured over the shortest time intervals . For the same reason, there should be a tendency toward intermediate values, which are due to averaging of rate fluctu­ ations, when measuring over longer intervals . Since morphological rates are typically expressed in terms of net change in the value of some morphological variable, changes in the direction of morphological change, as well as in the rate of change per se, contribute to the moderation of rates measured over longer time intervals . This interaction is partly responsible for the decline in maximum observed morphological rates with increasing measurement interval (Fig . 1) . However, as Gingerich (1983) pointed out, the lower, and to some extent the upper bounds of the distribution of observed rates in Fig . 1B are also influenced by factors unrelated to evolutionary process . The lower bound corresponds to a practical limit of measurement precision, beyond which earlier and later forms would not usually be recognized as different, yielding a rate of zero . The upper bound, on the other hand, rep­ resents an effective limit beyond which pairs of earlier and later forms differ so strongly that their relationship, and hence their appropriateness for a rate calculation, is likely to be questioned . The result is a tendency for longer measurement inter­ vals to yield lower rates . Because of these biasing factors, comparison of rates measured over very different time intervals is a non-trivial problem . Many comparative studies of evolutionary rates have not adequately dealt with this issue . Taxonomic rates are also affected by measurement interval, but not in all the ways noted above . As with morphological rates, rates of origination cal­ culated over longer intervals are likely to be damped by averaging a range of shorter-term values . How­ ever, rates of origination are not moderated by changes in the 'direction' of evolution; 'new taxa' are new taxa, even if they show reversals in certain attributes . In addition, with rates of origination, low values do not suffer an interval-related bias based on measurement precision, nor do high values necessarily engender suspicion of lack of relationship .

The effect of stratigraphic completeness on evolutionary rates Stratigraphic completeness (see also Section 3 . 12) could in principle affect the precision of palaeon­ tologically documented divergence times, but in practice, phylogenetic uncertainties and disconti-

2 The Evolutionary Process and the Fossil Record

156

B

A 4.5

v; c:

.�'"

� V

�

>-

;;; c:

.g::J

(3 >

L.U

4.0

3.5 3.0

1 2.0

f, ,, ,

1 .5

1 .0

c:

V

�

C '" c: o

.�

EXT R E M E C HAN G E

(3

, 1 1 ,'", " , ,

40 . 20

-'

o

I

0

'

I

EXTRE M E C H A N G E

o ,

04 o o

�'

, , " 1 1 33 2 ' " " , , I

·4.0

I I

'------'-

· 2 .0

�, , ,

o , o

6 .0

c:

,

0

o o

8.0

L.U

2 I I l 2 2 3 15 2 2 ' 2 1 5 2 5 5 5 2 '2 I [ 5 :5 4 X 6 X I X '2

�

1 0.

>

,

21 1 12 I 423 2 1 4 1 3:':5 4XXS 23XX'X o 42XX9

0.5

�

�

0

"' �

.� ,

2. 5

2.0

v;

J

o

4 16

I

� I II : 1 2 1 1 4 1

2'0

I

30

t

'2

40

Meas u re m e n t i n te rval ( 1 0 6 yrs)

5'0

.6

.0

NO MEAS U RA B L E C H A N G E

" f

, , , 1I I

I

I I

I 13 :5 '2 I 1 1 21 1 2 2 212 2 1 '2 33551 2 1 1 '2 23�(' I 1 1 85 I I I 1 21 534 19 2 2 34 1 33 2 I " !�i ' '2

I 1 1 24 2 t 1215 31 I 31 2 41 1 1 1 1 3 3 . 33 I 2 2 1 32 1 1 I 3 1 1 1 32 4 1 2 � 1�4 �J k3 1 ' I t 1 86 I 412 1

'"

4.'"'" 6 r---:c --= . 2'"'"" 6.-=. 8'"'"" .o --' . 6'"'"" � .o --c " . 0 --=' rc .'O .0-"""' .1 � .0 -.-' 0 4c:.0 -02c:2 . 0-1 4.o--

Ln Meas u re m e n t i n te rval ( 1 0 6 y r s )

Fig. 1 Inverse relationship between morphological rates of evolution and the time intervals over which they are measured (after Gingerich 1983), illustrating some of the biasing effects discussed in the text. A, Subset of rates shown in B, plotted on linear axes. B, Logarithmic transformation of rate distribution; time intervals range from 1 . 5 to 350 million years. Rates plotted in domain I (open squares) represent laboratory selection experiments; domain 11 (open circles, and digits for multiple cases; X > 9) represents historical colonization events; domain III (digits) represents post-Pleistocene events; and domain IV (digits) is drawn from the pre-Holocene record of invertebrates and vertebrates .

nuities i n the preserved record o f taxa (even within intervals that have a sedimentary record) are more important sources of error in these estimates . Stratigraphic completeness increases in importance, however, when morphological or taxonomic rates read 'directly' from the fossil record are considered . I n a relatively incomplete section, the actual age difference between two samples may be either much greater or much smaller than their estimated age difference based on linear interpolation from dated levels. This translates into substantial im­ precision in rate measurements . In order to mini­ mize this problem, Dingus & Sadler (1982) suggested that rates only be measured at levels of resolution for which stratigraphic sections can be considered complete (i . e . for which each included interval of given magnitude is likely to be rep­ resented by sediment) . Following this recommen­ dation, relatively incomplete sections limit us to longer time intervals for rate measurement and thus, through the biasing effect of interval length, tend to yield lower rates than might be seen in more complete sections . Relatively incomplete sections also tend to reduce measured rates of origination .

Factors affecting actual rates of evolution Having explored some of the factors that tend to distort perceptions of evolutionary rates, the sources of real variation in such rates will now be discussed . Among the more conspicuous of these are controls of the rate of transformation within established, species-level lineages. These include : mutation rate; generation time; degree of resource specialization; and the nature, amount, and distribution of varia­ bility within populations . Population size may also be important but is probably overshadowed by population structure - the pattern and scale of subdivision of populations and the degree of re­ productive interaction between those subdivisions . Other factors are a t least partly extrinsic t o the species in question : rate of environmental change; ecological factors such as the level of interspecific competition; and, in general, the intensity of selec­ tion (assuming selection and fitness are defined so that intensity of selection is not trivially equivalent to rate of evolution) . Another group of controls overlaps somewhat with the first but may be distinguished as operating at a different level in the genealogical hierarchy . It

2 . 1 1 Rates of Evolution

157

consists of factors that determine the rate of in­ itiation of new species-level lineages . Speciation rate assumes particular importance in a punctuated view of evolution, but its role in influencing evo­ lutionary rate is not dependent on the predominance of a punctuated mode of evolutionary change . In­ trinsic controls on speciation rate include such factors as dispersal ability (also relevant as a deter­ minant of population structure) and degree of resource specialization . There are also extrinsic con­ trols, such as rate or incidence of habitat fragmen­ tation by geomorphic or tectonic processes.

Living fossils - alternative definitions Living fossils figure in discussions of evolutionary rates as a conspicuous and yet potentially tractable case in which the relationship between a large-scale evolutionary pattern and its underlying causes may be explored (Eldredge & Stanley 1984; Schopf 1984) . As noted above, the central concept in the definition of living fossils is survival over long periods of time with minimal morphological change . Auxiliary cri­ teria have been appended by various authors and do indeed apply to certain cases traditionally recognized as living fossils . However, they are much less applicable to others . For instance, a relict geo­ graphical distribution and greatly diminished pre­ sent (relative to past) diversity characterize Sphenodon (a rhynchocephalian) and Nautilus (a nautiloid cephalopod), but not Limulus and related genera (horseshoe crabs) . Likewise, Latimeria (a coelacanth) and Neopilina (a monoplacophoran) rep­ resent clades once thought to be extinct, but Lepisosteus (a gar) and Lingula (an inarticulate brachiopod) have long been known from both fossil and Recent biotas . Living fossils are some­ times referred to as 'species' that have persisted for inordinately long periods of time, but few if any instances are actually founded on well docu­ mented species-level identity. The most generally useful definition therefore focuses on supra specific taxa that have shown unusual morphological conservatism . One of the most commonly cited living fossil groups is the Xiphosurida, or horseshoe crabs . Fig . 2 provides some sense of the morphological conservatism that can be seen within this group, comparing the extant species Limulus polyphemus with the Triassic Limulus vicensis . While the generic identity of these two species may be questioned (Fisher in Eldredge & Stanley 1984), their overall anatomical similarity is evident . Other species with-

Fig. 2 Horseshoe crabs, a commonly cited living fossil group . A, Dorsal aspect of a juvenile Limulus polyphemus, Recent, distributed along much of the eastern coast of North America; c. one half actual size. B, Dorsal aspect of a specimen of Limulus vicensis, Triassic, France; c. actual size . The tail spine is not preserved on this specimen, but it was presumably present originally. (From Bleicher 1897.)

158

2 The Evolutionary Process and the Fossil Record

in the group show greater morphological diver­ gence, but the reputation for bradytely has focused on comparisons such as that given here .

B RADYT E LY

TAC H YT E LY

Bradytely - alternative explanations The problem posed by living fossils is to explain the general phenomenon of bradytely . Simpson's (1944, 1953) interpretation was that the low rates of long­ term morphological evolution shown by bradytelic lineages are a consequence of unusually low rates of intraspecific phyletic transformation (Fig . 3A) . This appears to be a testable proposition, but it has thus far received little direct, empirical evaluation (per­ haps because few bradytelic groups have a suf­ ficiently continuous fossil record) . However, some of the factors that have been suggested as respon­ sible for low rates of phyletic transformation (e . g . unusually low levels o f morphological o r genetic variability) have been assessed within bradytelic groups and found not to differ significantly from values typical of nonbradytelic taxa (e . g . Selander et al. 1970) . Other factors that could in principle be responsible (e . g . extreme habitat stability, or strongly canalized development) are difficult to test. Some factors do seem to hold for a wide range of bradytelic groups and have been thought to contrib­ ute directly to low rates of intraspecific change (e . g . ecological generalization and broad physio­ logical tolerance; Simpson 1953) . Nevertheless, consideration of alternative explanations is clearly warranted . Another approach to interpreting bradytely steps up a level in the hierarchy of evolutionary pro­ cesses - from intraspecific interactions to the circumstances surrounding speciation events (cladogenesis) . It depends, furthermore, on the proposition (associated with the concept of punc­ tuated equilibrium) that most morphological change is accomplished during and driven by cladogenesis, and that the subsequent history of species tends to be dominated by morphological stasis . Under this characterization of evolution, a low rate of intra­ specific transformation would be the norm and would not be seen as a sufficient cause of bradytely . However, bradytely might be due to unusually low rates of speciation within bradytelic lineages (Fig. 3B); according to this interpretation, low speciation rate would allow few opportunities for morphologi­ cal change and would thus restrict a lineage to a relatively low rate of change averaged over the long term (Eldredge 1979) . As long as speciation is understood as a process that is not itself dependent

A

� Morpho l ogy

B

C Fig . 3

Schematic representation of three explanations of controls on long-term rate of morphological evolution. (After Fisher in Eldredge & Stanley 1984.) A, The contrast between bradytely and tachytely may be due to differences in the rate of intraspecific morphological transformation. B, The same contrast may be due to differences in rate of speciation. C, Bradytely and tachytely may also reflect higher-order patterns of differential survival and cladogenesis .

on morphological change, this interpretation represents a novel perspective on bradytely . Yet there is still a question as to why certain taxa, retrospectively recognized as bradytelic, show such a low rate of speciation . One possible answer has been suggested by the observation that a number of bradytelic taxa also show tendencies toward eurytopy - i . e . they have, at least in many respects, relatively broad, generalized ecological require­ ments . In this, they contrast with stenotopic taxa, which have relatively narrow, specialized require­ ments . It has been suggested that, relative to stenotopic taxa, eurytopic taxa are less subject to di­ rectional selection, often have broader geographical

2 . 11

Rates of Evolution

ranges, and tend to have populations that are less susceptible to range disruption and consequent reproductive isolation Gackson 1974; Eldredge 1979; Vrba in Eldredge & Stanley 1984) . This may result in a lower rate of speciation and a lower likelihood of morphological divergence during speciation . The case studies of bradytely in Eldredge & Stanley's (1984) compendium offer qualified support for the association of bradytely, low speciation rate, and eurytopy, but rigorous evaluation of this pattern is difficult because of the lack of quantitative indices of morphological conservatism or eurytopy . In addition, measurements of speciation rate are sub­ ject to significant sampling problems, such that even an evaluation of the relationship between speciation rate and subjective assessments of bradytely and eurytopy would be complicated . A third interpretation o f bradytely i s that i t arises at an even higher level in the genealogical hierarchy, as a result of differential survival of relatively primitive and relatively derived lineages within a clade (Fig . 3C) . Treated simply as a phylogenetic pattern, bradytely may or may not have any single lower-level cause, but whether it does or not, it could be independent of any systematic difference in intra specific rates of transformation or rates of speciation (Fisher in Eldredge & Stanley 1984) . While any of these three explanations of bradytely might operate in isolation from the others, they are not mutually incompatible . Nor can the possibility be ruled out that different instances of bradytely are traceable to different mixes of factors operating at a variety of levels . Although there are thus no simple answers, the investigation of bradytely has led to an expanded appreciation of the possible controls of long-term evolutionary rates.

159

References Bleicher, M. 1897. Sur la decouverte d'une nouvelle espece de limule dans les marnes irisees de Lorraine. Bulletin des Seances de la Societe des Sciences de Nancy 14, 116- 126 . Dingus, L. & Sadler, P. M. 1982. The effects o f stratigraphic completeness on estimates of evolutionary rates. System­ atic Zoology 31 , 400-412. Eldredge, N . 1979 . Alternative approaches to evolutionary theory. Bulletin of the Carnegie Museum of Natural History 13 , 7 - 1 9 . Eldredge, N. & Stanley, S.M. (eds) 1984 . Living fossils . Springer-Verlag, New York. Fitch, W.M. 1976 . Molecular evolutionary clocks. In: F.J. Ayala (ed .) Molecular evolution . Sinauer, Sunderland, Mass . Gingerich, P.D. 1983 . Rates of evolution: effects of time and temporal scaling . Science 222, 159 - 1 61 . Gingerich, P.D. 1986. Temporal scaling of molecular evolution in primates and other mammals. Molecular Biology and Evolution 3, 205 - 221 . Goodman, M . , Weiss, M . L . & Czelusniak, J. 1982. Molecular evolution above the species level : branching pattern, rates, and mechanisms . Systematic Zoology 3 1 , 376-399. Haldane, J . B . 5 . 1949 . Suggestions as to quantitative measurement of rates of evolution. Evolution 3, 5 1 - 56. Jackson, J.B.C. 1974. Biogeographic consequences of eurytopy and stenotopy among marine bivalves and their evo­ lutionary significance . American Naturalist 108, 541 - 560. Raup, D.M. & Marshall, L . G . 1980 . Variation between groups in evolutionary rates: a statistical test of significance . Paleobiology 6, 9 - 23. Raup, D.M. & Stanley, S.M. 1978. Principles of paleontology, 2nd edn . Freeman, San Francisco . Schopf, T.J.M. 1984. Rates of evolution and the notion of living fossils. Annual Review of Earth and Planetary Sciences 12, 245-292. Selander, R.K., Yang, S . Y . , Lewontin, R. c . & Johnson, W.E. 1970 . Genetic variation in the horseshoe crab (Limulus polyphemus), a phylogenetic "relic" . Evolution 24, 402-414. Simpson, G.G. 1944. Tempo and mode in evolution . Columbia University Press, New York. Simpson, G . G . 1953 . The major features of evolution . Columbia University Press, New York. Stanley, S.M. 1985 . Rates of evolution. Paleobiology 11, 13-26 .

2 . 12 Mass Extinction: Processes were destroyed in the Late Devonian (Section 2 . 1 3 . 3) and end-Triassic (Section 2 . 1 3 . 5) episodes and the calcareous plankton (foraminifera and coccolitho­ phorids) drastically reduced at the end of the Cretaceous (Section 2 . 1 3 . 6) . The biggest event of all was at the end of the Permian (Section 2 . 13.4), when many important Palaeozoic groups went completely extinct, including fusulinid foraminifera, came rate and inadunate crinoids, trepostome and crypto­ stome bryozoans, rugose corals, and productid brachiopods . All but the first of these extinction episodes have subsequently been accepted by palaeontologists as the most significant extinction events in Phanerozoic history (Raup & Jablonski 1986) . The correlation between major sea-level falls and Newell's mass extinction events is indeed striking (Fig . 1; Jablonski 1986) . On a smaller scale, there is an equally striking correlation between the extinction of environmentally sensitive groups such as am­ monoids and other episodes of widespread re­ gression, probably correlating with sea-level fall, in both the Palaeozoic and Mesozoic (e . g . Hallam 1987a) . Following ecological research on island bio­ geography, it is clear that smaller habitat areas can accommodate fewer taxa, so reduction in area must lead to lower diversity as the extinction rate in­ creases . Whether the extinction is due to reduced habitat diversity, increased competition, crowding effects, or whatever, the basic empirical relation­ ship appears to be well established . Critics have pointed out that inferred episodes of significant marine regression do not always correlate with notable mass extinctions of marine organisms . This is most obviously true for eustatic falls of sea­ level in the Quaternary and Middle Oligocene, the latter being probably the largest in the Tertiary (Haq et al. 1987) . At least two explanations can be put forward, both of which take into account the phenomenon of biological adaptation. Quaternary regressions were followed by equally rapid trans­ gressions after geologically short time intervals, limiting the effect of reduced habitat area and per­ mitting a sufficient number of organisms to survive and expand their populations during the succeeding transgressions . Quaternary faunas are likely to have been relatively eurytopic, or environmentally toler­ ant, because they represent survivors of environ­ mentally stressful Late Cenozoic times . The same

2 . 12 . 1 Earth-bound Causes A . H A L L AM

Introduction The idea that mass extinctions could be caused by strictly Earth-bound phenomena is an old one, dating back to the so-called heroic age of geology in the early part of the nineteenth century . Following the pioneering extinctions research of his compatriot G . Cuvier, the French geologist Elie d e Beaumont pro­ posed that catastrophic, virtually instantaneous upheavals of mountain ranges at infrequent inter­ vals through geological history caused drastic envi­ ronmental changes leading to the destruction of a high proportion of the Earth' s biota . The correlation between episodes of diastrophism and times of major organic turnover was also noted by the American geologist T . e . Chamberlin at the begin­ ning of this century, and by European geologists such as E. Suess and J . F . Umbgrove (Hallam 1981a) . Modern research on tectonic activity suggests, how­ ever, that it is too localized geographically and insufficiently 'catastrophic' in time to account satis­ factorily for mass extinction events . Attention must be confined to phenomena global in scale that can give rise to drastic changes in the physical environ­ ment . The only plausible contenders are changes in sea level and climate, and episodes of increased volcanicity.

Sea-level The American palaeontologist Newell (1967) was the first person to make an explicit correlation be­ tween mass extinction episodes among Phanerozoic marine invertebrates and eustatic falls in sea-level, attributing the extinctions to increased environ­ mental stress consequent upon substantial re­ duction of habitat area of shallow epicontinental seas. He distinguished six such episodes : end­ Cambrian, end-Ordovician, Late Devonian, end­ Permian, end-Triassic and end-Cretaceous . The first two are especially well marked by trilobite extinc­ tions and the last three by ammonite extinctions . Extensive communities o f reef-dwelling organisms

160

-

2 . 1 2 Mass Extinction: Processes -¥

-¥

P l i oc e n e- P l e i stoce n e M i oc e n e

S e a l evel

\"

..J

O l i gocene

Pa l aeoce n e & Eocene

�*

� \� *

Cretaceo u s

J u ra s s i c

Triassic

Pe r m i a n

Carbon ifero u s

Devo n i a n

S i l u rian

O rd o v i c i a n

Cambrian

�

(* \

"\

* --

(*

\*

�

Fig. 1 Phanerozoic sea-level curve with Newell's (1967) six extinction events shown as asterisks . High sea-level to left. (After Hallam 1984 . )

consideration may apply also to the Middle Oligocene regression, which followed closely on a significant increase in marine extinction rates across the Eocene - Oligocene boundary . It is likely that for long periods of Phanerozoic time most organisms became so well adapted to conditions of relative environmental stability, including equable climate, that even modest changes of sea-level could have had a striking effect on so-called 'perched faunas' in extensive and extremely shallow epicontinental seas . Such palaeogeographic phenomena cannot be closely matched at the present day, which marks an unusually regressive episode in Earth's history. The latest Exxon sea-level curve from the Triassic

161

to the present (Haq et al. 1987) does not show unusually large falls at the times of the two greatest marine extinction episodes during this interval, the end of the Triassic and the end of the Cretaceous . The Exxon curve is based largely, however, on seismic stratigraphy, and should not be treated as more than a tentative model to be subjected to testing by other evidence . There is indeed con­ siderable evidence of a major end-Triassic regression (Hallam 1981b), and some strong indi­ cations that the extent of the end-Cretaceous re­ gression has been underestimated by Haq et al. (Hallam 1987b) . For several events, namely the end-Permian, end­ Triassic and end-Cretaceous, mass extinctions in the marine realm appear to correlate closely with mass extinction of some terrestrial vertebrates, notably those large in size, which, because of their relatively low population numbers and reproductive rates, would be more vulnerable to environmental disturbance than smaller organisms (see also Section 2 . 10) . Obviously such extinctions cannot be ac­ counted for by reduction in land area, and a more likely explanation is bound up with the increased continental seasonal temperature contrasts induced by regression of epicontinental seas . While much attention has been paid to regression as a promoter of extinctions it should be noted that there is a strong association between inferred sea­ level rises that follow directly after falls and the spread of anoxic water in epicontinental seas, as recorded for instance by widespread laminated black shales . Habitable areas can be as severely reduced by this means as by regression, with a mass extinction event ensuing . For many extinction events, both major and minor, a clear correlation exists with extensive deposits of black shales . Among the major events the best examples are the basal Silurian and basal Famennian (Devonian), effectively equivalent to the end-Ordovician (Sec­ tion 2 . 1 3 . 2) and end-Frasnian (Section 2 . 13.3) ex­ tinction events . Among minor events the clearest examples are the Cenomanian- Turonian boundary and Early Toarcian (Hallam 1987a) . The spread of anoxic bottom waters may possibly also be impli­ cated as a contributory factor in the end-Permian (Section 2 . 13.4) and end-Triassic (Section 2 . 13 . 5) events . For much of Phanerozoic history the ocean might have been poorly stratified, in marked contrast to the present-day situation (Wilde & Berry 1984) . In consequence the deeper ocean would be more or less anoxic and could not have served as a refuge for

162

2 The Evolutionary Process and the Fossil Record

shallow-water organisms at times of regression, or if they were outcompeted by other organisms . It is more than possible that the great bulk of the modern deep-sea fauna, which contains representatives of most phyla, is no older than Tertiary . Since the Late Eocene there has evidently been a system of strong currents induced by Antarctic glaciation, which have served to aerate bottom water in the deep ocean (Hallam 1981a) . Lack of a deep water anoxic zone could help to explain why there is no signifi­ cant extinction recorded for the major Middle Oligocene regression . The cause of sea-level changes is bound up either with the melting and freezing of polar icecaps or with tectonics, such as the uplift and subsidence of ocean ridges and the splitting or collision of conti­ nents . The end-Ordovician event might well have had a glacioeustatic cause, associated with growth and disappearance of the Saharan ice sheet, but for the other major extinction events the most likely cause is tectonoeustatic . This poses a problem, be­ cause the rates of sea-level rise and fall produced by plate tectonics are approximately three orders of magnitude lower than for glacioeustasy, thereby allowing more time for organisms to adjust to a changed environment and hence avoid extinction . Unfortunately there are as yet insufficient data from the strati graphic record, on amount and rate of sea­ level change, to resolve this problem satisfactorily . There remains another possibility, that rapid re­ gressions and transgressions on a regional rather than a global scale could be produced as a result either of changes in the pattern of lateral stresses in the crust (Cloetingh et al. 1985) or by the rise of mantle plumes to cause epeirogenic uplift, with related volcanism associated with subsidence (Loper & McCartney 1986) . The fact that such changes would not strictly come under the category of eustatic is irrelevant as far as the organisms are concerned, provided that the changes in question are both geographically extensive and rapid, thereby leading to drastic changes in the environment .

Climate Changes of sea-level could have, as a by-product, some climatic consequences, but climate could of course fluctuate with time independent of eustasy . Stanley (1984, 1987) has been the strongest advocate of the view that temperature changes in the marine realm have been the dominant causal factor in Phanerozoic mass extinctions . This interpretation involves a gross extrapolation from his detailed

studies of Plio-Pleistocene molluscan extinctions off the Atlantic and Gulf coasts of the U . S . A . Whereas there is a high rate of species extinctions in this region, there is negligible evidence of contemporary extinctions around the Pacific margins, or the Mediterranean . Stanley maintained that, because the extinctions are regional not global in extent, eustatic changes cannot be invoked . Instead he argued for a more pronounced lowering of tempera­ ture on the American east coast than elsewhere, as a result of palaeogeographical factors . Extending back through time, the next major marine extinction event for which temperature decline can plausibly be invoked is across the Eocene - Oligocene boundary . This 'event' is de­ cidedly not sudden in geological terms and is marked more by a pronounced increase in extinction rate rather than a drastic change over a narrow time interval . There is good independent evidence from oxygen isotopes of a fall in both surface and bottom water temperatures, but no indication from the curve of Haq et al. (1987) of sea-level changes sig­ nificantly larger than at other times in the Tertiary . For pre-Tertiary times, however, the evidence im­ plicating temperature as a causal factor is weak to non-existent, forcing Stanley to resort to some special pleading (though it could be argued that the end-Ordovician event (Section 2 . 13.2) had an ulti­ mate climatic causation, if the glacioeustatic in­ terpretation is accepted) . For example, the largest extinction event of all, at the end of the Permian (Section 2 . 1 3 .4), took place during a period of clima­ tic amelioration, marked by the Middle Permian disappearance of the Gondwana ice sheet. It is con­ ceivable, of course, that the end-Permian event was induced by an episode of temperature rise, but no plausible case has been made for this . One of the points that Stanley cited in favour of his temperature control hypothesis is that the most extinction-vulnerable organisms, such as reef dwellers, were tropical in distribution throughout Phanerozoic history . While this may be true, it does not necessarily establish temperature as the key control, because tropical organisms tend to be gen­ erally stenotopic, as they are relatively sensitive to a variety of environmental factors . A really extensive overturn of deep anoxic water at the beginning of episodes of climatic change has been suggested as a possible contributing factor to mass extinction events in the oceans (Wilde & Berry 1984) . As discussed above, the rise and spread onto continental shelves of anoxic water is often associ­ ated with marine transgressions, so that it may be

2 . 1 2 Mass Extinction : Processes unnecessary to invoke climatic change as well . As regards changes in air temperature, the only satisfactory record comes from Late Cretaceous to Recent terrestrial plants . No striking extinction event has been recorded among these organisms for the Cenozoic, but at the end of the Cretaceous there were significant extinctions in the North Temperate Realm of western North America and Eastern Asia . Whereas the palaeobotanical consensus has related such extinctions to gradual temperature decline through the Late Cretaceous, the most recent re­ search in the North American Western Interior suggests a temperature rise in the Maastrichtian and no significant change across the Cretaceous ­ Tertiary boundary (Wolfe & Upchurch 1987) . Further back in time the evidence from terrestrial plants is more obscure, and has so far not been adequate to establish a convincing picture of climatic change .

Volcanism The end-Cretaceous extinction event is the one that has received by far the most attention (see also Sections 2 . 1 3 . 6, 2 . 1 3 . 7) . Notwithstanding the claims made for extra-terrestrial impact, there is strong evidence for marine regression at this time, suggesting that this phenomenon is involved in the extinctions . Sea-level change cannot account, how­ ever, for the drastic extinctions at the Cretaceous ­ Tertiary boundary of calcareous plankton, nor for such physico-chemical evidence as an anomalous enrichment on a global scale of iridium, and the presence locally of quartz grains with shock­ metamorphic laminae, in Cretaceous-Tertiary boundary layers (see also Section 2 . 12.2) . Evidence of this sort has been claimed as conclusive for bolide impact, but in fact a case of at least equal plausibility can be made for terrestrial volcanism on a massive scale (Hallam 1987b) . It is known that aerosols enormously enriched in iridium compared with crustal rocks can be expelled from the mantle during flood basalt eruptions . Eruptions of this kind on a sufficient scale over several 100 000 years could produce the observed global enrichment of the element . The Deccan Traps of India, erupted during the magnetic zone that embraces the Cretaceous -Tertiary boundary, are the most ob­ vious candidate . There is good evidence of contem­ porary explosive volcanism in other parts of the world, and reasonable grounds for believing that such volcanism can generate the pressures required to produce shock-metamorphic laminae in mineral grains.

163

Massive volcanism over an extended period would have deleterious environmental conse­ quences . It is known that flood basalt fissure erup­ tions that produce individual lava flows with volumes greater than 100 km3 at very high mass eruption rates are capable of injecting large quan­ tities of sulphate aerosols into the lower strato­ sphere, with potentially devastating atmospheric consequences . Such volatile emissions on a large enough scale would lead to the production of im­ mense amounts of acid rain, reduction in alkalinity and pH of the surface ocean, global atmospheric cooling, and ozone layer depletion . Atmospheric cooling would be reinforced by ash expelled into the atmosphere by contemporary explosive volcanicity . Thus for the end-Cretaceous extinctions a com­ pound scenario seems to be required, involving both sea-level fall and volcanicity on an exception­ ally intense scale, with associated climatic changes (there is as yet, however, no evidence to support the notion that volcanicity was a direct causal factor for other mass extinction events) . Loper and McCartney (1986) noted that increased end-Cretaceous volcan­ ism correlates with a significant change in the geo­ magnetic field, with a long Cretaceous reversal­ free period coming to an abrupt end in the Maastrichtian. They proposed a model involving periodic instability of the thermal boundary layer at the base of the mantle . This layer accepts heat from the core and transmits it upward by way of mantle plumes. As it thickens by thermal diffusion it be­ comes dynamically unstable and hot material erupts from it. Heat is extracted from the core at a greater rate, increasing the energy supply and hence the magnetic reversal frequency of the dynamo in the fluid outer core . Hot material rises through mantle plumes to the surface to give rise to volcanic activity . Both non-explosive and explosive volcanism can be produced, depending on the condition of the litho­ sphere, which varies regionally . Increased mantle plume activity has the potential for causing uplift of extensive sectors of continents and hence regression of epicontinental seas . Present-day hotspots are as­ sociated with regional topographic bulges, so it is reasonable to infer that most epeirogenic uplifts reflect hot, low density regions in the astheno­ sphere, derived from plume convection . Epeirogenic subsidence on the continents and marine trans­ gression might be expected to follow episodes of substantial volcanic eruptions . Fischer (1984) put forward a general hypothesis that relates changes of sea-level, climate, and volcan-

164

2 The Evolutionary Process and the Fossil Record

icity to produce two supercycles during Phanerozoic time . Times of high rates of ocean floor spreading and oceanic volcanicity correlate with buoyant ocean ridges and consequently high sea-level stands . Less carbon dioxide is removed from the atmosphere by terrestrial weathering because of reduced continental area, and the volcanicity brings more of the gas to the Earth's surface . Thus the carbon dioxide content of the atmosphere is high, and because of the greenhouse effect the climate is equable, with no polar ice caps . The converse tec­ tonic situation gives rise to low sea-level stands, low atmospheric carbon dioxide, and stronger climatic differentiation between the tropics and the poles. The rates of change involved in such processes appear, however, to be too low to account for mass extinction events . The most promising line of ap­ proach in generating terrestrial models is probably a closer investigation of the relationship between sea-level change, continental uplift, volcanism, and mantle plume activity, as has been proposed for events across the Cretaceous - Tertiary boundary (Sections 2 . 1 3 . 6, 2 . 1 3 . 7) . The end-Permian extinction episode (Section 2 . 13 .4) is an especially promising candidate for this type of investigation.

References Cloetingh, S . , McQueen, H . & Lambeck, K. 1985 . On a tectonic mechanism for regional sea level variation. Earth and Planetary Science Letters 75, 157- 166. Fischer, A.C. 1984 . The two Phanerozoic supercycles . In: W.A. Berggren & J . A . van Couvering (eds) Catastrophes and Earth history, pp. 129 - 150. Princeton University Press, Princeton. Hallam, A. 1981a. Facies interpretation and the stratigraphic record. W.H. Freeman, Oxford . Hallam, A. 1981b . The end-Triassic bivalve extinction event. Palaeogeography, PalaeoC/imatology, Palaeoecology 35, 1 -44. Hallam, A. 1984. Pre-Quaternary sea-level cha�lges. Annual Review of Earth and Planetary Sciences 12, 205- 243 . Hallam, A. 1987a . Radiations and extinction in relation to environmental change in the marine Lower Jurassic of northwest Europe . Paleobiology 13, 152 - 1 68. Hallam, A. 1987b . End-Cretaceous mass extinction event: argument for terrestrial causation . Science 238, 1237-1242. Haq, B . U . , Hardenbol, J . & Vail, P.R. 1987. Chronology of fluctuating sea levels since the Triassic . Science 235, 1 1 58- 1 167. Jablonski, D . 1986 . Causes and consequences of mass extinc­ tions . In: D.K. Elliott (ed . ) Dynamics of extinction, pp. 183-229 . Wiley, New York . Loper, D . E . & McCartney, K. 1986. Mantle plumes and the periodicity of magnetic field reversals . Geophysical Research Letters 13, 1525 - 1528. Newell, N.D. 1967. Revolutions in the history of life . Special Papers of the Geological Society of America 89, 63-91 .

Raup, D . M . & Jablonski, D. (eds) 1986 . Patterns and processes in the history of life. Report of Dahlem Workshop, 1985 . Springer-Verlag, Berlin, Heidelberg. Stanley, S.M. 1984. Marine mass extinction: a dominant role for temperature . In: M.H. Nitecki (ed . ) Extinctions, pp. 69- 1 17. University of Chicago Press, Chicago. Stanley, S.M. 1987. Extinction . Scientific American Books, New York. Wilde, P. & Berry, W.B.N. 1984. Destabilisation of the oceanic density structure and its significance to marine extinction events . Palaeogeography, Palaeoclimatology, Palaeoecology 48, 143 - 162. Wolfe, J.A. & Upchurch, C . R . 1987. North American non-marine climates and vegetation during the late Cretaceous . Palaeogeography, Palaeoclimatology, Palaeo­ ecology 61, 33- 78 .

2 . 12.2 Extra-terrestrial Causes D . JABL O N SKI

Introduction Extra-terrestrial causes have long been invoked for mass extinctions, but only in the past decade has the general scientific community taken the idea seriously . Geochemical, sedimentary, and other signals in the stratigraphic record are sufficient to suggest that it is impossible to ignore extra­ terrestrial impacts as potential explanations for the biotic crises that punctuate the fossil record . The case is not fully proven for any single mass extinc­ tion, although it is strongest for the end-Cretaceous event (W. Alvarez 1986; L . W . Alvarez 1987; see Hallam 1987 and Officer et al . 1987 for different views; see also Sections 2 . 12 . 1 , 2 . 1 3 . 6, 2 . 1 3 . 7) . In any event, the initial discovery of iridium and other geochemical anomalies at the Cretaceous- Tertiary boundary has sparked an immense amount of inter­ disciplinary research on the problem of mass ex­ tinctions and potential extra-terrestrial forcing agents .

Potential mechanisms Proposed extra-terrestrial causes for mass extinc­ tions have included variation in solar heat output, massive solar flares, sudden influx of cosmic rays owing to a nearby supernova or the Solar System's crossing of the Galactic plane, and collisions with comets, asteroids, or other extra-terrestrial objects

2 . 1 2 Mass Extinction : Processes (collectively termed bolides) . Until recently such factors were at best subject to only the weakest verification based on approximate correlations in timing, and at worst simply reflections of desper­ ation in the face of seemingly inexplicable biotic upheavals . New lines of evidence for possible bolide impacts at one, and perhaps as many as five, extinc­ tion events have shifted these speculations into the realm of testability . Earth-crossing asteroids (asteroids whose orbits cross that of the Earth or could cross as a result of long-range gravitational perturbations) are suf­ ficiently common that significant bolide impacts must have occurred in the geological past. The Earth should suffer impacts by c. six 1 km asteroids per million years, and by c. two asteroids of 10 km ' or more per 100 million years, i . e . about a dozen large impacts since the beginning of the Phanerozoic (Shoemaker 1984) . Effects of 1 km objects are uncer­ tain but, as discussed below, most workers believe that impact by a 10 km bolide would have severe, global consequences . The average collision rate for comets is almost certainly lower than that for asteroids . Cometary impact rates could occasionally be raised, however, by perturbing the Oort cloud of comets that surrounds the Solar System far beyond the outer­ most planets (inner edge about 104 Astronomical Units (AU) from the Sun, where 1 AU is the distance from the Sun to the Earth) . Passage through the higher stellar densities in the spiral arms of the Galaxy might raise collision rates by about 10% (Shoemaker 1984) . This low-frequency modulation of cometary impacts would be punctuated ap­ proximately once per 100 million years by short­ lived bursts (1 - 3 million years) triggered by close passage of individual stars (Hut et al. 1987) . Evidence for periodic extinctions, still hotly de­ bated, suggests (but does not prove) a more regular and frequent perturbation of the Oort cloud . Hy­ pothesized mechanisms include : oscillations around the Galactic plane, where encounters with stars and molecular clouds would be most probable; a tenth planet in a highly eccentric orbit beyond Pluto (at c. 100 AU); and a dim solar companion star, christened Nemesis in advance of discovery (at distances variously estimated in the order of 104 - 105 AU) . Debates on the astronomical plausibility of these mechanisms, with Nemesis maintaining a slight edge, are reviewed by Shoemaker & Wolfe (1986) and Hut et al. (1987) (see also Section 2 . 1 2 . 3) . The magnitude and geographical scale of an im­ pact's effects depend on bolide size and velocity but

165

thresholds have not been determined . An asteroid 10 km in diameter was estimated for the end­ Cretaceous event on the basis of global iridium levels, and although potential effects are still poorly understood they would probably have been severe . W. Alvarez (1986), L . W . Alvarez (1987) and Prinn & Fegley (1987) emphasize the following possibilities : 1 Darkness caused by the global cloud of fine dust particles generated by the impact. For 2 - 1 1 months, this darkness may have been sufficiently profound to halt photosynthesis, thereby causing the collapse of marine and terrestrial food chains. 2 Cold would accompany the darkness, with tem­ peratures dropping below freezing in continental interiors . Maritime climates would be less severely perturbed, owing to the thermal inertia of oceanic waters . 3 Greenhouse effects and global warming could fol­ low the cold-temperature excursion if the bolide(s) struck in the ocean . After dust grains coagulated and settled from the atmosphere, the remaining burden of water vapour could trap infrared energy reflected from the Earth and raise global tempera­ tures by as much as 100e . The duration of this greenhouse episode is uncertain, with estimates ranging from months or years to much longer spans than the immediate cold, dark aftermath - perhaps as long as 1000 years (Prinn & Fegley 1987) . 4 Nitric acid rain might result from shock heating of the Earth's atmosphere during impact (see Prinn & Fegley 1987, whose calculations are followed here) . Energy from atmospheric entry and, especially, the supersonic plume ejected upon impact would pro­ duce very large amounts of nitric oxides . These compounds would undergo a series of reactions and ultimately rain out as nitric and nitrous acid . On land this would severely damage foliage (and, presumably, animals) both directly and through mobilization of trace metals . In the ocean, within a decade or less, the acid rain could lower the pH of the mixed layer (especially the upper 30 m) to 7 . 5 - 7.8, sufficient to dissolve calcite and thus severely stress calcareous organisms . Further, in­ jection of so much strong acid into the atmosphere would elicit a significant exhalation of oceanic CO 2, which, combined with the accumulation of CO2 in the atmosphere owing to depressed activity of marine phytoplankton, would yield greenhouse warming over thousands of years . This impressive menu of impact-driven pertur­ bations could be expected to cause mass extinc­ tions of the observed magnitudes . Indeed, a

166

2 The Evolutionary Process and the Fossil Record

number of palaeontologists have argued that the hypothesized perturbations are too severe for the observed extinctions, even at the Cretaceous ­ Tertiary boundary (e . g . Hallam 1987) . However, the impact-effect models are very poorly constrained and require extrapolation far beyond hard obser­ vational data; a new generation of more realistic and sophisticated models may provide an improved basis for critically comparing hypothesized causes with observed extinction patterns .

Biological evidence The initial impetus for seeking extra-terrestrial im­ pacts was of course the biological pattern of extinc­ tion in the fossil record, whether perceived as peaks in global extinction rates or as disap pearances of taxa or biomass in local sections . Unfortunately, the biological consequences of impacts, massive vol­ canism, and other alternatives are not sufficiently understood or sufficiently unique to provide critical tests . Complex biological upheavals enacted on scales of months, years or decades, as postulated by impact scenarios, are extremely difficult, often im­ possible, to resolve in single stratigraphic sections, and challenge the limits of global correlation . Short­ term events are superimposed on more protracted patterns in the expansion and contraction of taxa, owing to Earth-bound physical and biotic factors, so that the effect of a given boundary event on a particular taxon (particularly a waning one) is de­ batable . At present, the strongest constraints that palaeontological data can provide involve consist­ ency between a given mechanism and the biological pattern observed in an imperfect fossil record .

Onset and aftermath. For extra-terrestrial impacts, biological responses include abrupt onset of extinc­ tion with an extremely brief crisis period, and a re­ latively short-lived reorganization and rebound during return to pre-impact conditions . In end­ Cretaceous impact models, for example, most en­ vironmental perturbations would last only 1 - 10 years, an interval impossible to correlate among distant localities, and within which events are vir­ tually unresolvable in the geological record . Geo­ logically abrupt onset of mass extinction is a requirement but not a unique prediction of impact hypotheses : even such gradual processes as marine regression or transgression could in principle carry threshold effects that would produce sudden ex­ tinction pulses on stratigraphically-resolvable time­ scales.

Hypothesized greenhouse warming, and possibly other palaeoceanographic anomalies, would persist for some thousands of years beyond the impact itself. Some palaeontological (and geochemical) evidence supports a geologically brief - but eco­ logically protracted - recovery period, particularly in terrestrial plants (reviewed by Wolfe 1987) and marine plankton (reviewed by Zachos & Arthur 1986), although, again, these would not be unique to extra-terrestrial events . Extinction patterns observed at critical bound­ aries cannot be taken at face value . Seemingly abrupt extinction can result from erosion or non­ deposition of sediments during the critical time interval, so that biological events are compressed into single beds . At the same time, artificially gradational extinction patterns result when sam­ pling deteriorates, or is simply uneven, in the in­ terval approaching the boundary (a phenomenon termed backwards-smearing, or the Signor- Lipps effect - see Jablonski 1986a; Raup 1987) . Step wise patterns of extinction, with pulses of extinction arrayed around a mass extinction boundary, have been claimed to reconcile the re­ quirements of abrupt extinction with observations that seemed to suggest gradual loss of taxa . Such stepwise patterns - with up to 12 discrete extinc­ tion events claimed near the Cretaceous- Tertiary boundary - are also taken as the geologically rapid succession of extinction events expected during cometary bombardment . These stepwise patterns are distinct from prolonged patterns of decline such as suggested for Late Cretaceous ammonites, and are recorded near the Cenomanian - Turonian, Cretaceous- Tertiary, and Eocene - Oligocene boundaries (Hut et al. 1987) . Unfortunately, such patterns cannot yet be taken at face value, because they can also be generated by sampling effects, local ecological changes, and/or minor breaks in sedi­ mentation imposed on either abrupt or gradational extinction . Lazarus taxa (which seem to suffer extinction but then reappear later in the stratigraphic record; Jablonski 1986a; Raup 1987) provide one means of partially controlling for unevenness in sampling and preservation : the proportion of Lazarus taxa, i . e . of observed last appearances that represent artificial extinction, permits a rough quantitative assessment of the reliability of extinction data with­ in and around critical time intervals . Most stepwise extinction sequences contain some Lazarus taxa, suggesting that sampling effects are indeed a factor. More rigorous and comprehensive approaches are

2 . 1 2 Mass Extinction : Processes required to place confidence limits on bed-by-bed extinction patterns . Detailed studies of critical time intervals are urgently needed, but the plea for more centimetre by centimetre sampling near extinction events is somewhat misguided. At that scale, local ecological effects, the vagaries of sampling, and even biotur­ bation are likely to overwhelm the fine structure of global events . Careful sampling of relatively long geological sequences that encompass extinction events would be especially valuable, so that absences as well as presences could be recorded throughout, to provide some statistical control . Consistency of extinction patterns among widely separated localities also should be sought in a criti­ cal fashion; caution is necessary, particularly for apparent stepwise patterns, because different taxa - say, ammonites and benthic gastropods have different sampling characteristics, even on broad geographical and temporal scales (see Jablonski 1986a on the biology of Lazarus taxa) .

Selectivity has been claimed for most mass extinc­ tions : large-bodied taxa, reef-dwellers or tropical organisms in general, and endemic taxa all appear to suffer preferential extinction (Jablonski 1986a, b) . Critics (and some supporters!) of impact hypotheses have claimed that impact-driven extinction would be random rather than selective, so that any observed taxonomic or ecological selectivity would be contrary evidence . This claim seems inappro­ priate, however: taxa differ in their vulnerability to environmental change, so that any given pertur­ bation, regardless of scale, should affect some groups more severely than others . Survivorship of widespread taxa, non-tropical taxa, small-bodied taxa, members of detrital food chains, freshwater taxa, deciduous plants, and plankton whose life cycles include resting cysts, has been claimed for the end-Cretaceous extinction (Jablonski 1986a, b; Hallam 1987) . All are consistent with, but not ex­ clusive to, impact hypothese s . Similarly, the pos­ sibility that mass extinctions are qualitatively different from background extinctions in their vic­ tims (e .g. see Jablonski 1986a, b) does not require impact events - any perturbation of sufficient magnitude could, for example, cross a threshold of extinction effects so that broad geographical range could determine survivorship but species richness was no longer important. Periodicity. The apparent periodicity of post­ Palaeozoic extinction events has sparked much re-

167

search and speculation on extra-terrestrial forcing factors (see also Section 2 . 12.3) . The periodicity it­ self, however, is not an adequate test for extra­ terrestrial causes, although few alternatives have been advanced (Hallam 1987 reviewed a hypothesis of endogenous periodicity in mantle plumes; see also Section 2 . 1 2 . 1 ) . Clearly, the critical role for palaeontological data in testing for extra-terrestrial causes of mass extinctions lies in the degree of correspondence between biological events and independent physico-chemical evidence for impacts or other extra-terrestrial forcing mechanisms . As discussed below, however, assembling such evi­ dence is not as straightforward as was once hoped.

Physical evidence Several physico-chemical phenomena have been proposed as independent evidence for extra­ terrestrial impact . Although each has its critics, and some may not be strictly diagnostic, taken together the data make a strong case for the end-Cretaceous and Late Eocene extinctions, with weaker but suggestive evidence for several other post­ Palaeozoic events (Raup 1987) . The strongest Earth­ bound alternative at this time appears to be volcanism (Hallam 1987; Officer et al. 1987; see also Section 2 . 12 . 1 ) .

Geochemical. The anomalously high concentrations in Cretaceous - Tertiary boundary sediments of iridium, and other elements scarce in the Earth's crust but abundant in asteroids, launched the Alvarez hypothesis that an end-Cretaceous impact caused the mass extinction . Since 1979 this anomaly has been found at over 75 localities world-wide (Fig . 1) in deep-sea, shallow-marine, and continen­ tal palaeoenvironments, usually in a distinctive clay layer that coincides (within stratigraphic uncer­ tainty limits) with the extinction event (W. Alvarez 1986; L . W . Alvarez 1987) . Excursions in oxygen and carbon isotopes near the boundary also suggest a low-productivity episode that may have lasted 1 . 0 million years or more, accompanied by detectable but unexceptional temperature oscillations (Zachos et al. 1989) . The direction of the stable isotopic fluctuations is appropriate to impact hypotheses, but the duration seems too long and the temperature changes too mild (but see above discussions on uncertainties in impact models and limits in stratigraphic resolution) . None of the other four major mass extinctions of the Phanerozoic has such strong geochemical

2 The Evolutionary Process and the Fossil Record

168

.. . . .

. . ...

.

•

Fig. 1

Global distribution of iridium anomalies in Cretaceous - Tertiary sediments . (After L.W. Alvarez 1987.)

anomalies known from so many localities, although far less effort has been devoted to the search (Jablonski 1986a; Donovan 1987a; Raup 1987) . Slight end-Ordovician iridium enrichments seem to be terrestrial in origin; the end-Triassic results are negative so far; the reported end-Permian anomaly, at the largest mass extinction of them all, has not been repeated by other laboratories, and the boundary clays seem volcanic in origin; the Late Devonian (Frasnian -Famennian) anomaly occurs in an unusual stromatolitic deposit and has not been replicated in other boundary sections . Among lesser extinction events, iridium anomalies are geographically widespread near the Eocene - Oligocene extinction boundary, along with a series of microtektite horizons whose impact origin is virtually uncontested (Hut et al. 1987) . An iridium anomaly was recently discovered (L . W. Alvarez 1987) for the small Middle Miocene extinction that forms the most recent peak in periodicity analyses, although the global extent of the iridium is as yet unknown . The Cenomanian - Turonian boundary has excess iridium, but other impact signatures are lacking and a terrestrial origin may be involved . An anomaly at the Middle - Upper Jurassic bound­ ary - where no extinction event occurs but is predicted by periodicity models - occurs (like the Frasnian - Famennian example) in stromatolitic sediments, raising the spectre of biological or dia-

genetic concentration . Age uncertainty of an iridium anomaly in a 2 - 3 mm iron-rich crust at an uncon­ formity in the Southern Alps overlaps with another weak or 'missing' (i . e . predicted by periodicity models) extinction peak in the Bajocian (Rocchia et al. 1986) . An iridium anomaly, with other cosmic debris, is recorded from Late Pliocene sediments in the Southern Ocean, coinciding in time but not in space with a regional extinction event in the North Atlantic . The situation is further complicated by an anomaly near the base of the Cambrian, at a level lacking mass extinction and well after the beginning of the Cambrian radiation of skeletonized organisms (Donovan 1987b) . The degree to which all of these iridium anomalies denote impacts is still debated (Hallam 1987; Officer et al. 1987; Section 2 . 12 . 1 ) . Iridium enrichments may extend for metres around the Cretaceous ­ Tertiary boundary in some key sections; the signifi­ cance of these new observations is unclear, with interpretations ranging from diagenetic mobili­ zation from an impact-fallout layer to prolonged deposition from volcanic aerosols . An aerosol from the Hawaiian volcano Kilauea was highly enriched in iridium, apparently derived from the deep mantle; however, other elements in the aerosol do not mimic the extra-terrestrial abundances in end­ Cretaceous boundary sequences (W. Alvarez 1986) so that, again, the significance of these data is uncer-

2 . 1 2 Mass Extinction : Processes

169

tain. Boundary clay compositions do not always correspond to extra-terrestrial elemental abun­ dances and isotope ratios, and yield conflicting evi­ dence regarding the nature of the hypothesized bolide . It is not clear whether post-impact diagenetic overprint or multiple impacts by bolides of different compositions (expected in cometary bombard­ ment?) can account for such inconsistencies . New analytical techniques (1 . W. Alvarez 1987) will per­ mit much more extensive stratigraphic coverage, both at extinction boundaries and at quiet times in between, and thus greatly improve understanding of the global iridium flux and potential nonextra­ terrestrial enrichment mechanisms . Fig.

Mineralogical. Potential independent ' evidence for impact comes from shock-metamorphosed quartz and other sedimentary particles . Like iridium, quartz grains with at least two and up to nine intersecting sets of shock lamellae have been found in Cretaceous- Tertiary boundary sequences throughout the world, in both marine and conti­ nental settings (Fig. 2) (Bohor et al . 1987a; Izett 1987) . Such multiple lamellae are known only in particles from nuclear testing sites and impact craters . Shock-metamorphosed minerals do form near certain explosive volcanic eruptions (Hallam 1987), but the multiple lamellae and the world-wide distribution of the relatively large grains (0 . 1 0 . 2 mm in North Pacific and New Zealand sedi­ ments, up to 0 . 6 mm in North America) are difficult to reconcile with volcanic activity (W . Alvarez 1986; Bohor et al. 1987a) . The search for shock­ metamorphosed minerals at other extinction events has been negative so far, except for an intriguing preliminary report near the Triassic -Jurassic boundary in Austria (Badjukov et al. 1987) .

Sedimentological. Microtektites (glassy droplets formed by bolide impacts) are almost undoubtedly present at three horizons near the Eocene ­ Oligocene boundary (Hut et al. 1987) . A similar origin has been suggested for spherules of dis­ ordered potassium-feldspar (sanidine), glauconite, goethite, and magnetite found world-wide in Cretaceous -Tertiary sequences (W . Alvarez 1986), but recent evidence suggests an authigenic, non­ impact origin for at least some spherules (Hallam 1987; Izett 1987) . Microspherules of varying com­ position occur in Permo-Triassic boundary sedi­ ments in Sichuan, China (Gao et al. 1987); their significance is uncertain in the light of the seemingly volcanic origin of the boundary clays in China.

2 Shocked quartz grain from Cretaceous -Tertiary boundary clay in a non-marine section at Brownie Butte, Garfield County, Montana. Scanning electron micrograph, width of field 0 . 14 mm. (Courtesy of B.F. Bohor.)

More work is needed in separating spherules of different origins before interpretations are possible (Bohor et al. 1987b) . Soot particles are abundant in Danish and New Zealand Cretaceous - Tertiary boundary clays (W . Alvarez 1986; 1 . W . Alvarez 1987) . If these clays rep­ resent only one year of deposition, as postulated by most impact models, the carbon flux would have been 10 3 - 104 above background levels, suggesting extensive wildfires triggered by the heat of impact or propagated among the remains of forests killed by the hypothesized post-impact cold interval . How­ ever, the uniqueness of such soot occurrences is uncertain, and the high flux depends on the duration of clay layer deposition, which is still debated (Hallam 1987) .

Cratering. Major impacts should leave craters at least an order of magnitude larger than the bolide itself. Age uncertainties are troublesome and the data are extremely sparse, but the association be­ tween extinction events over the past 250 million years and the 26 well dated craters of 5 km or more in diameter may be statistically significant (re­ viewed by Shoemaker and Wolfe 1986, who are sceptical) . Simulations by Trefil & Raup (1987) suggest that this cratering record comprises about one-third periodic impacts (presumably comet showers) and two-thirds random collisions with asteroids . Shoemaker & Wolfe (1986) reach a similar conclusion by different means . Questions emerge about the best-studied extinc­ tion event, however . The only well dated craters of appropriate size near the Cretaceous- Tertiary

170

2 The Evolutionary Process and the Fossil Record

boundary are in the U . S . S .R. (Shoemaker & Wolfe 1986), but the size and density of shocked quartz grains suggests an impact in North America (Bohor et al. 1987a; Izett 1987) . Further, the shocked quartz suggests an impact in sedimentary rocks, i . e . a con­ tinental or shallow-water setting, whereas mag­ netite and other spherules suggest altered basalt, and thus an oceanic impact (although impact deri­ vation of the spherules is now questioned, as noted above) . These contradictions are perhaps resolvable with an end-Cretaceous comet shower and the consequent multiple impact, but the problem of impact site(s) remains (Hallam 1987) . A volcanic interpretation is no more satisfactory in this regard .

Conclusion and prospects Although no one indicator is definitive, at present the diverse physical and chemical evidence at the Cretaceous -Tertiary boundary is most readily ex­ plained by a bolide impact . Volcanism is the chief rival, but as W. Alvarez (1986) argued, evidently only quiet basaltic eruptions yield iridium aerosols and melt microspherules, whereas violent siliceous eruptions are needed to produce shocked minerals . Neither kind of eruption will produce all of the observed impact signatures, nor can either account for the world-wide distribution of shocked quartz, iridium and other geochemical anomalies . The periodic mantle plume hypothesis might yield both explosive and non-explosive volcanism on a global scale (Hallam 1987; Section 2 . 12 . 1), but this model awaits evaluation . The palaeontological data are generally consistent with, but provide little con­ clusive support for, impact-driven extinction mech­ anisms . As many authors have noted, marine re­ gression at this and other extinction events obscures biological and physico-chemical signals and may even play a role in extinction (Section 2 . 12 . 1 ) . The most definitive evidence for or against extra­ terrestrial factors in mass extinctions (apart from the discovery of the hypothesized solar companion, Nemesis) will come with an assessment of the strength of temporal association between Phanero­ zoic mass extinctions and physico-chemical signa­ tures of bolide impacts . This work is under way, and it is impressive that the three or four most recent extinction peaks recognized in global data sets and/or local strati graphic sections (Middle Miocene, Eocene - Oligocene, Cretaceous- Tertiary, and Cenomanian - Turonian) bear at least some impact indicators . The weak but significant cluster­ ing of crater ages at extinction events over the past

250 million years should prompt analyses around other boundaries, with ongoing refinement of hypotheses . Assessment of negative evidence re­ mains a problem, however, so that impact hypoth­ eses can be remarkably elastic and difficult to falsify : absence of craters, shocked quartz and even iridium anomalies are consistent with impact on now­ subducted ocean, basaltic impact site, and cometary rather than meteorite impact, respectively. Addition­ ally, not all major craters, microtektite horizons, or iridium anomalies coincide with extinction events . Better understanding of the potential effects of im­ pacts, and of the distribution of potential impact signatures through the stratigraphic record, should lead to the framing of more refined hypotheses regarding the role of extra-terrestrial factors in the evolution of life on Earth .

References Alvarez, L.W. 1987. Mass extinctions caused by large bolide impacts . Physics Today 40, 24 -33. Alvarez, W., 1986 . Toward a theory of impact crises . Eos 67, 649, 653 -655, 658. Badjukov, D . D . , Lobitzer, H. & Nazarov, M.A. 1987. Quartz grains with planar features in the Triassic-Jurassic boundary sediments from Northern Limestone Alps, Austria. Lunar and Planetary Science 18 , 38 -39. Bohor, B . F . , Modreski, P.]. & Foord, E.E. 1987a . Shocked quartz in the Cretaceous -Tertiary boundary clays: evidence for a global distribution. Science 236, 705 - 709 (see also 666 - 668) . Bohor, B . F . , Triplehorn, D . M . , Nichols, D.J. & Millard, H.T., ]r. 1987b. Dinosaurs, spherules, and the 'magic' layer: a new K-T boundary clay site in Wyoming. Geology 15, 896 - 899 . Donovan, S.K. 1987a. Iridium anomalous no longer? Nature 326, 331 -332 . Donovan, S.K. 1987b . Confusion at the boundary. Nature 329, 288 . Gao Zhengang, Xu Daoyi, Zhang Qinwen & Sun Yiyin 1987. Discovery and study of microspherules at the Permian­ Triassic boundary of the Shangsi section, Guangyuan, Sichuan. Geological Review 33, 203-211 (in Chinese with English abstract) . Hallam, A. 1987. End-Cretaceous mass extinction event: argument for terrestrial causation . Science 238, 1237-1242. Hut, P., Alvarez, W., Elder, W . P . , Hansen, T., Kauffman, E . G . , Keller, G . , Shoemaker, E . M . & Weissman, P.R. 1987. Comet showers as a cause of mass extinctions. Nature 329, 118- 126. Izett, G.A. 1987. Authigenic 'spherules' in K-T boundary sediments at Caravaca, Spain and Raton Basin, Colorado and New Mexico, may not be impact derived . Bulletin of the Geological Society of America 99, 78- 86. ]ablonski, D. 1986a . Causes and consequences of mass extinctions : a comparative approach . In: D.K. Elliott (ed . )

2 . 1 2 Mass Extinction: Processes Dynamics of extinction, pp . 183-229 . Wiley, New York. Jablonski, D. 1986b . Evolutionary consequences of mass extinctions . In: D.M. Raup and D. Jablonski (eds) Patterns and processes in the history of life, pp . 313-329 . Springer­ Verlag, Berlin . Officer, C B . , Hallam, A . , Drake, C L . & Devine, J . D . 1987. Late Cretaceous and paroxysmal Cretaceousrrertiary extinctions . Nature 326, 143 - 149 . Prinn, R.G. & Fegley, B . , Jr. 1987. Bolide impacts, acid rain, and biospheric traumas at the Cretaceous-Tertiary boundary. Earth and Planetary Science Letters 83, 1 - 15 . Raup, D.M. 1987. Mass extinction: a commentary. Palaeon­ tology 30, 1 - 13 . Rocchia, R., Boclet, D . , Bonte, P . , Castellarin, A. & Jehanno, C 1986 . An iridium anomaly in the Middle - Lower Jurassic of the Venetian region, northern Italy . Journal of Geophysical Research 91, E259 - E262. Shoemaker, E . M . 1984. Large body impacts through geologic time. In: H.D. Holland & A.F. Trendall (eds) Patterns of change in Earth evolution, pp. 15 -40. Springer-Verlag, Berlin. Shoemaker, E . M . & Wolfe, R. 1986 . Mass extinctions, crater ages, and comet showers. In: R.S. Smoluchowski, J.N. Bahcall & M . 5 . Matthews (eds) The galaxy and the solar system, pp. 338 -386. University of Arizona Press, Tucson. Trefil, J . 5 . & Raup, D.M. 1987. Numerical simulations and the problem of periodicity in the cratering record. Earth and Planetary Science Letters 82, 159 - 1 64. Wolfe, J.A. 1987. Late Cretaceous- Cenozoic history of de­ ciduousness and the terminal Cretaceous event. Paleo­ biology 13, 215-226. Zachos, rC & Arthur, M.A. 1986. Paleoceanography of the Cretaceousrrertiary boundary event: inferences from stable isotopic and other data. Paleoceanography 1, 5-26. Zachos, J.C, Arthur, M.A. & Dean, W.E. 1989 . Geochemical evidence for suppression of pelagic productivity at the Cretaceous/Tertiary boundary . Nature 337, 61 -64.

2 . 12 . 3 Periodicity J . J . S E P K O S KI , Jr

Introduction Periodicity of extinction is a hypothesis that ex­ tinction events (both mass extinctions and their less severe analogues) have occurred at regularly spaced intervals through geological time . It is an empirical claim based upon statistical analyses of the fossil record which indicate that maxima in extinction intensity, recognized in both biostratigraphic studies and taxonomic data compilations, are de­ cidedly non-random with respect to time and seem to fit a regular, periodic time series. This hypothesis was introduced by Fischer & Arthur (1977) for

171

patterns of diversity in open-ocean pelagic com­ munities and later supported by Raup & Sepkoski (1984), who claimed a 26 million year periodicity in extinction of global marine families . The hypothesis has since proved very controversial largely as a result of association with suggestions of catas­ trophic, extra-terrestrial forcing agents .

Meaning of periodicity A perfectly periodic time series has regularly spaced events separated by invariant waiting times (Fig. 1 ) . I n most o f the debate about periodicity, this pattern has been contrasted with a 'random', or Poisson, time series . A Poisson series can arise when events are independent of one another and determined by a large number of unrelated factors . A classic example is coin flipping, in which the outcome (heads or tails) of each trial results from a multitude of in­ dependent forces . The lower time series in Fig. 1 was generated by flipping a pair of coins and re­ cording when both came up heads . The frequency of events (one in four trials) is the same as in the upper, periodic series, but the appearance is very different . The lower series is composed of loose clusters of events with irregular gaps in between; waiting times approach an exponential distribution with the median waiting time shorter than the average frequency . The relevance of these considerations to the study of extinction events is that traditionally each event has been analysed in isolation from others and independent causal hypotheses have been formu­ lated . Implicit in this is the assumption that extinc­ tion events must be randomly spaced in time . Observation of regular spacing, however, implies some organizing principle to extinction events, either some set of factors that governs waiting times so that they appear invariant, or some single ulti­ mate forcing agent that has clock-like properties . Periodicity can also imply that the proximate agent of any one extinction event is the same for all, although this is not a necessary implication if the chain of causation is complex. The association of periodicity with catas trophism comes from these last considerations . In particu­ lar, it has been suggested that: (1) the claimed 26 million year periodicity of extinction events is too long to have been produced by any known terres­ trial process with periodic behaviour, leaving some astronomical clock as the likely forcing agent; and (2) the association of the Cretaceous - Tertiary mass extinction with evidence of a large extra-terrestrial

2 The Evolutionary Process and the Fossil Record

172 Pe r i od i c :

I

R a n do m :

11

I I I I

II

I II

I

I

I

I

I I

Fig. 1 The contrast between a periodic and a random time series . Both series have the same frequency of 'events', but the random series is characterized by irregular clusters of events with variable gaps between them. (After Sepkoski 1986 . )

impact (Section 2 . 12.2) suggests (as a hypothesis to be tested) that other events in the periodic series might have been similarly caused . Note that these arguments suggest only a possible association, and other, terrestrial mechanisms are still conceivable (Section 2 . 12 . 1 ) .

Evidence for periodicity The hypotheses of periodicity put forward by Fischer & Arthur (1977) and Raup & Sepkoski (1984) were based upon compilations of diversity data and ex­ tinction times for taxa in the marine fossil record . Fischer & Arthur were concerned with recurrent fluctuations in the diversity of globigerinid species, ammonoid genera, and large pelagic predators through the Mesozoic and Cenozoic . They argued, without rigorous statistical testing, that these fluc­ tuations were cyclic with a 32 million year waiting time . Using family-level data for the entire marine ecosystem, updated time-scales, and a variety of statistical tests, Raup & Sepkoski corroborated the Fischer -Arthur hypothesis but concluded that the period length was closer to 26 million years . Their statistical tests (which included parametric Fourier and autocorrelation analyses and non-parametric randomization analysis) all indicated a significant non-randomness in the distribution of extinction events and a good, but not perfect, fit to a periodic series. Raup & Sepkoski's (1984) treatment and testing of familial extinctions were somewhat complex and have led to some confusion . Their analysis was limited to families in the Late Permian through Neogene, where stratigraphic stages are shorter and more accurately dated than in the preceding Palaeozoic. To enhance resolution, only families with extinctions known to the stage level were used

and taxa of soft-bodied and lightly sclerotized ani­ mals, or of very uncertain taxonomic position, were rejected . These manipulations left a data set of 567 extinct families ranging over 39 stratigraphic stages. Extinction intensity was measured by percent ex­ tinction, the number of extinctions in a stage div­ ided by diversity . This metric (statistical measure) scales extinction to the number of families at risk in any stage but does not incorporate estimates of stage duration, which have limited accuracy . Percent extinction for families exhibits very low values over the Cenozoic, leaving peaks of extinction difficult to discern; Raup & Sepkoski therefore used only the diversity of families extinct before the Recent in the denominator of the metric, inflating its values in the Cenozoic . The time series constructed by this treatment (Fig. 2) contains 'peaks', or local maxima, that vary considerably in height. Raup & Sepkoski recognized that some of these (e . g . the Guadalupian, Rhaetian, and Maastrichtian) correspond to well documented mass extinctions, but that some of the lower peaks might be spurious . Nevertheless, they chose to ana­ lyse all peaks rather than a selected subset, in order to avoid possible subjective bias . Unfortunately, they referred to all 12 peaks as 'mass extinctions' . A randomization test for periodicity was favoured by Raup & Sepkoski (1984) because it permitted fitting a wide band of period lengths and was not sensitive to unequal spacing of data (imposed by the stratigraphic time-scale) or to variation in mag­ nitudes of extinction peaks (which were presumed to fluctuate freely) . The test (which is akin to boot­ strap procedures) involved fitting periodicities to the observed extinction peaks and then comparing the goodness of fits to randomized (i . e . shuffled) versions of the data . The peaks were treated as if they all fell at the ends of stages; this, however, was merely a formalization, and equivalent results would have obtained if the peaks were consistently placed at the middles or beginnings . The shuffling procedure converted the data into what was essen­ tially a random walk with the only constraint being that peaks must be separated by at least two stages. The randomization test showed that periodicity fits the observed data better than 99 . 99% of random walks at 26 million years, even though the fit to the peaks (especially the smaller peaks) was not perfect (Fig . 2) . On this basis, Raup & Sepkoski concluded that there was a 26 million year periodicity to 'mass extinctions' through the Mesozoic and Cenozoic Eras . No periodicity was found in the Palaeozoic, however. Rampino & Stothers (1984) corroborated

2 . 1 2 Mass Extinction: Processes

173

70 50

30 20 c

.Q U c

xQ) Q)

�

10

c

Q) 0-

S

1 00

1 50

200

250

50

Geological t i m e ( 1 0 6 yrs ) Fig. 2 Raup & Sepkoski' s (1984) time series for familial extinction from the Late Permian to Neogene, computed from their highly culled sample of marine animal families . The 12 'peaks' of extinction are highlighted by stippling, and the fit to the 26 million year periodicity is indicated by the vertical lines (note the lack of fit in the Middle Jurassic and Early Cretaceous) . Geological systems and stages (small boxes with initial letters of names) are indicated along the bottom of the graph (note the variable stage lengths, especially through the Cretaceous and Tertiary) . The y axis is logarithmic. Abbreviations along the time axis: D Dzhulfian; S Scythian, A Anisian, L Ladinian, C Carnian, N Norian, R Rhaetian; H Hettangian, S Sinemurian, P Pliensbachian, T Toarcian, B Bajocian + Aalenian, B Bathonian, C Callovian, o Oxfordian, K Kimmeridgian, T Tithonian; B Berriasian, V Valangian, H Hauterivian, B Barremian, A Aptian, A Albian, C Cenomanian, T Turonian + Coniacian, S Santonian, C Campanian, M Masstrichtian, PI Danian, P2 Upper Palaeocene, El Lower Eocene, E2 Middle Eocene, E3 Upper Eocene, 0 1 Lower Oligocene, O2 Upper Oligocene, MI Lower Miocene, M2 Middle Miocene, M3 Upper Miocene, P Pliocene . =

=

=

=

=

=

=

=

=

=

=

=

=

=

this result, even after eliminating the three smallest peaks in the time series . A similar period-fitting technique applied to the nine remaining peaks gave a 26 million year period . However, a regression­ based technique resulted in a 30 million year period, which they favoured on other grounds . Subsequent analyses performed by Raup & Sepkoski were designed to counter criticisms of their data manipulation and statistical procedures, and to explore the correspondence between global taxonomic data and information from biostrati­ graphic studies . Sepkoski & Raup (1986) re analysed the familial data using all extinctions (other than those of soft-bodied animals tied to Konservat ­ Lagerstatten; Section 3 . 1 1 ) and employing total di­ versity in the metrics . Fig . 3 illustrates the time

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

series for percent extinction in this analysis . Three other metrics of extinction were also computed and an attempt was made to assess which extinction peaks could be considered statistically significant . Sepkoski & Raup determined only eight o f their previous 12 peaks to be significant and found that the heights of these peaks were generally lower than in the highly culled data set . They argued, however, that seven of the peaks corresponded to extinction events recognized by palaeontologists working at the species level with material collected from outcrops and cores . This indicated to the authors that global familial data could be trusted to reflect important extinction patterns among species in the fossil record . Sepkoski & Raup (1986) found that the random-

2 The Evolutionary Process and the Fossil Record

174

40 r--r----�--r_--_r--��

26-Ma Periodicity

30

c

.9 U c

xQ)

C :::

Q)

Tith

Q) Cl..

Fig. 3 Percent extinction for all marine animal families based on a data set with 970 extinctions . Peaks that are 'significantly' higher than background are labelled . (Guad Guadalupian, uNori Upper Norian or Rhaetian, Plie Pliensbachian, Tith Tithonian, Ceno Cenomanian, Maas Maastrichtian, uEoc Upper Eocene, mMio Middle Miocene, other abbreviations as in Fig. 2) . Error bars indicate one estimated standard error on either side of the data points. The y axis is a square root scale. (After Sepkoski & Raup 1986, by permission of the AAAS . ) =

=

=

=

=

=

=

=

250

200

1 50

1 00

Geological t i m e (1 0 6 yrs)

ization test applied to the eight extinction peaks still indicated a significant periodicity at 26 million years, with the standard error judged to be about ± 1 million years when imprecisions in the time­ scale were accommodated . In a companion paper, Raup & Sepkoski (1986) showed that the level of statistical significance of the randomization test varied somewhat if different ages were assigned to the less precisely dated events (end-Permian and end-Triassic) . Still, they concluded that most fits of the 26 million year periodicity were significant at or above the 95% level, even after adjustment for the problem of multiple tests (i . e . testing many fre­ quencies in the 1 2 - 60 million year band) . Raup & Sepkoski (1986) and Sepkoski (1986) also conducted analyses at the generic level, using a new compilation for global marine animals . This was done to increase sample size and to obtain a better approximation of species patterns . Higher taxa tend to damp the signal of species extinction since all species within a polytypic taxon must disappear for the taxon to register an extinction event . The new data set contained nearly 10 000 genera in the inter­ val from Upper Permian to Recent . It also incorpor­ ated a refined stratigraphic time-scale with 51 intervals (in contrast to the previous 39 -43 stages) . Fig . 4 illustrates one of four time series for generic extinction . As expected, the eight peaks of extinction

50

o

are more prominent than in the familial data . The peak in the Middle Miocene seems to be confirmed and an extinction event is suggested in the Aptian, which previously appeared as a gap in the periodic sequence (Fig . 3) . A gap still exists in the Middle Jurassic despite two questionable peaks (Lower Bajocian and Oxfordian) . These two peaks, as well as the Carnian peak (to the left of the Upper Norian peak in Fig . 4), fluctuate erratically with different metrics of extinction, suggesting that they are not robust features of the data . Raup & Sepkoski (1986) performed the random­ ization test on these data and concluded that they contained the 26 million year periodicity of extinc­ tion . Sepkoski (1986) also performed autocorrelation analyses (i . e . correlating a time series with itself at a given time lag, which assesses amplitude as well as wavelength) and obtained statistically significant results consistent with a 26 million year periodicity . Finally, Fox (1987) performed an elaborate series of Fourier analyses on the generic data and also found a significant 26 million year periodicity . This was true even when he split the time series into two parts: both halves displayed a periodicity with the same wavelength and, very importantly, nearly the same phase . None of these analyses of the generic data showed decisive evidence for a periodicity prior to the Permian, however, although Sepkoski

175

2 . 1 2 Mass Extinction: Processes I

Guad

50

26-Ma Periodicity

Fig. 4

Percent extinction for marine genera. (After Sepkoski 1986 . ) The data illustrated comprise 9773 genera, of which 5594 are extinct. Note that peaks of extinction (labelled as in Fig. 3) are better defined than in the total familial data, and a seemingly periodic Aptian ( Apti') peak appears in the Early Cretaceous . No clearly periodic peak is present in the Middle ]urassic, although low, possibly spurious peaks appear over the Lower Bajocian and Oxfordian; a peak also appears over the Carnian within the Late Triassic. '

c o

"-6

40 Maas

uNori

c

xQJ 30

C QJ

::'

QJ 0...

20

250

(1986) did suggest a possible longer (greater than 30 million years) periodicity in the late Palaeozoic .

Critiques of periodicity The hypothesis of periodicity in extinction engen­ dered immediate attention from scientists as well as the popular press . Not surprisingly, this led to intense scrutiny of both the data and the statistical analyses . The result has been a complex series of critical discussions with various responses by Raup & Sepkoski (see Sepkoski 1989), which can only be briefly summarized here .

Data. The validity of compilations of taxonomic data has been questioned by several authors . Hoffman (1985) argued that familial data are very noisy and that different treatments, including ap­ plication of alternative time-scales, results in dif­ ferent, seemingly random patterns of extinction peaks . This claim was countered by Sepkoski & Raup's (1986) demonstration of consistency of eight extinction peaks under four different metrics and by Sepkoski' s (1986) argument that even Hoffman's composite data display strong periodicity . The pres­ ence of the same periodic extinction peaks in the much larger generic data would also seem to indi­ cate signal rather than noise . Stigler & Wagner (1987), however, argued that periodicity even in the generic data could be an artifact of imperfect sampling of the fossil record . Failure to sample taxa in their last stage of existence will smear the record of extinction backward in time . This will tend to swamp some minor extinction

200

1 50

1 00

50

o

G e o l ogical t i m e ( 1 0 6 yrs)

peaks between major maxima and cause the time series to appear more regular than expected for a Poisson distribution . The counterargument to this claim (Sepkoski & Raup 1986) is simply that detailed biostratigraphic investigations corroborate most of the extinction peaks evident in the generic data, and do not indicate many smaller extinction events in other stages (although some major extinction events may be composites of tightly clustered steps) . Patterson & Smith (1987) questioned the accuracy of any taxonomic compilation that contains para­ phyletic taxa (see Section 5 . 3) . They claimed that three-quarters of the families of echinoderms and vertebrates used by Raup & Sepkoski were para­ phyletic, monotypic, and/or misdated. When a cor­ rected monophyletic component (equivalent to 10% of Raup & Sepkoski's total data set) was examined, no periodicity was evident . Sepkoski (1987) re­ sponded that paraphyly in itself should not be a problem since family extinctions simply represent a sample of species extinctions . He further noted that the monophyletic taxa in Patterson & Smith's analysis failed to show some well documented extinction events (e . g . the Maastrichtian mass extinction) and suggested that this might be due to small sample size, idiosyncracies in the echinoderm and ver­ tebrate records, or biases inherent in the cladistic culling. Inaccuracies in the estimated ages of stratigraphic intervals used in the data sets pose numerous prob­ lems . As noted above, Hoffman (1985) argued that use of different time-scales causes differences in extinction peaks . Shoemaker & Wolfe (in Smoluchowski et al. 1986) assessed the estimated

176

2 The Evolutionary Process and the Fossil Record

ages of Raup & Sepkoski's (1984) 12 extinction peaks and concluded that only three (the Cenomanian, Maastrichtian, and Upper Eocene) were reliable; this was too small a sample to support periodicity . Raup & Sepkoski (1986), however, showed that their randomization test did give significant results for the last four, best-dated extinction events (including the Middle Miocene event which Shoemaker & Wolfe rejected on the basis of familial data, but which Raup and Sepkoski accepted on the evidence of the generic data) . Stigler & Wagner (1987) questioned the strength of this test, arguing that the 26 million year period­ icity might be embedded in the time-scale . This is not surprising, however, since some stratigraphic boundaries are placed at points of major turnover (e . g . the Palaeozoic - Mesozoic and Mesozoic ­ Cenozoic boundaries) . Potential coupling of the stratigraphical and biological records was recog­ nized by Raup & Sepkoski (1984, 1986), who shuffled the time-scale in their tests in order to avoid this problem . It should be noted that the 51-interval time-scale used in the generic data, with longer stages subdivided and shorter stages amalgamated, does not display any embedded 26 million year periodicity .

Statistical analyses. Many technical aspects of the statistical tests conducted by Raup & Sepkoski have been questioned. Hoffman & Ghiold (1985) claimed that the analyses did not properly test for a random walk . They argued that the familial data displayed a mean frequency of one peak in every four stages, which is indistinguishable from the expectation of a random walk. But these authors (and likewise Noma & Glass 1987) failed to recognize that Raup & Sepkoski' s randomization procedure in essence converted the extinction data into random walks (although perhaps with less variance than proper, as pointed out by Quinn (1987» . Also, Sepkoski's (1986) auto correlation analysis with the refined time-scale showed a peak every fifth interval in the generic data, which is not consistent with a random walk . Noma & Glass (1987) used turning points in the familial data to argue that the hypothesis of ran­ domness could not be rejected. However, their test was very sensitive to variance in stage durations (which range from 1 million years for the Coniacian to 1 5 . 5 million years for the Albian), and it is unclear whether Noma & Glass demonstrated anything more than this variance . They also argued that there were flaws in the selection of 'significant' extinction

peaks by Raup & Sepkoski (1986) (as well as Sepkoski & Raup 1986) . This argument is valid, and at best Sepkoski & Raup merely eliminated demon­ strably insignificant peaks from their familial analyses. However, other evidence presented by Sepkoski & Raup suggests that the remaining eight peaks were not insignificant since : (1) the same peaks appeared even more prominently in the gen­ eric time series (Fig. 4); and (2) most of the peaks correspond to independently identified events in biostratigraphic analyses . Kitchell & Pena (1984) re analysed the familial data assuming equal durations of stages and apply­ ing a series of autoregressive models (i . e . regression equations in which values in each time interval are predicted from values in preceding intervals) . They rejected a simple model with periodic impulses but found adequate fits with a model incorporating five-stage memory, which they concluded dem­ onstrated only pseudoperiodicity in the data. How­ ever, the rejected simple periodic model imposed a regular amplitude as well as wavelength, and re­ quired equal numbers of stages between extinction peaks . (The number of stages between Raup & Sepkoski's periodic peaks varied from two to six; see Fig . 2) . Again, Sepkoski's (1986) autocorrelation analysis of the generic data suggested that a simple periodic impulse model could provide a statistically significant fit when the stratigraphic intervals were adjusted to be more equal in length . Quinn (1987) criticized Raup & Sepkoski's (1984) randomization test for ignoring the auto correlation in the data (although Stigler & Wagner (1987) did not consider this to be a problem) . Quinn failed to note that Raup and Sepkoski had recognized this problem and used only randomizations that had the same number of peaks as observed in the data . Quinn offered an alternative test that compared waiting times between peaks to the expectation of random events (a broken-stick distribution) . This test, he claimed, failed to demonstrate any evidence of periodicity in either the familial or the generic data . Unfortunately, he used an arbitrary definition of 'mass extinction' (either all stages with extinction intensities in the upper quartile of the data, or all peaks exceeding the mean intensity after log- linear adjustment for temporal trends) . His test appears to be sensitive to the number of points selected and could reject a moderately noisy sine curve if the number of points exceeded the number of cycles . Running Quinn' s test for different numbers of cycles or points would have presented difficulty in assessing the significance level for multiple tests .

2 . 1 2 Mass Extinction : Processes Quinn (1987) complained that Raup & Sepkoski (1984) did not calculate the joint significance level for the 49 independent tests that were conducted in assessing all periodicities between 12 and 60 million years (although Raup & Sepkoski did attempt to tackle this, albeit incorrectly) . Quinn claimed the joint significance level was only 39%, given a sig­ nificance of 99% for the fit of the 26 million year period . Tremaine (in Smoluchowski et al. 1986) cal­ culated the joint significance level to be 95 .4%, using a recomputed significance of 99 . 74% for the 26 million year period . Tremaine went on to argue, however, that random simulations run over the 1 2 6 0 million year band indicated a joint significance level of less than 90% for the 12 peaks of Raup & Sepkoski (1984) and less than 50% for the eight peaks of Sepkoski & Raup (1986) . But these results may have been sensitive to his assumption that variance in fit was directly proportional to period length in his tests . Raup & Sepkoski (1986) used Tremaine's procedure without this assumption and obtained joint significance levels greater than 95% . All of these tests and arguments have used a Poisson model of randomness as a basis of compari­ son . Lutz (1987) argued that this is not the only alternative in testing for periodicity . He tested Raup & Sepkoski's (1984) familial time series against models for Poisson distributions, 'noisy' period­ icities, and constrained episodicities (i . e . y distri­ butions in which the standard deviation in waiting times is less than the mean waiting time) . He found that the Poisson model could be rejected at the 95% significance level, but he could not distinguish be­ tween fits of noisy periodicities and of episodicities with variances less than 30% of mean waiting time (although it is not clear how sensitive these results are to selection of events and to errors in the time­ scale) . Lutz (1987) concluded that an exogenous forcing agent with clock-like behaviour was not necessary to explain the data . Stanley (1987) proffered a similar argument on qualitative grounds . He suggested that extinction events eliminate particularly vulnerable taxa and that there is a lag time after each event during which few vulnerable taxa are available for extinction . Thus, palaeontologically recognizable perturbations should be spaced more widely than expected from a Poisson distribution . The counter to this argument is that recovery times observed for most extinction events in the Mesozoic and Cenozoic are only one or two stages, which is within the lag time built into Raup & Sepkoski's randomization procedure .

177

It cannot be claimed that any of these arguments and counterarguments is decisive, and it is doubtful whether new, more accurate data could settle the matter (although more precise data would certainly promote better understanding of extinction in the fossil record) . A definitive settlement will be reached only if a clear agent of periodic extinction is discovered .

Possible causes o f periodicity Both terrestrial and extra-terrestrial mechanisms have been suggested as ultimate causes of period­ icity in extinction. The terrestrial mechanisms in­ volve hypothetical quasiperiodic processes in the deep Earth that lead to episodes of intense volcan­ ism . The extra-terrestrial mechanisms involve a var­ iety of observed and hypothesized astronomical clocks that might induce periodic cometary bom­ bardments of the Earth . Evidence that extra-terrestrial impacts might be important in periodic extinction come from two sets of observations (see also Section 2 . 12.2) : 1 Materials presumed to b e of impact origin (excess iridium, microtektites, and/or shocked mineral grains) are associated with several periodic extinction events, including the Cenomanian, Maastrichtian, Upper Eocene, and Middle Miocene . 2 Ages of terrestrial craters seem to exhibit a weak periodicity, involving 25 - 50% of impacts, that has a phase and period length (variously estimated at 27 - 32 million years) that are roughly congruent with the extinction periodicity (see Shoemaker & Wolfe in Smoluchowski et al. 1986) . The periodic impactors are presumed to be comets derived from the Oort Cloud at the outer fringes of the Solar System. It has been hypothesized that a gravitational perturbation from a body as small as four times Jupiter'S mass could induce a comet shower that would bring up to 109 comets into the inner Solar System; about 25 of these on average would strike the Earth over a 1 million year interval. Four mechanisms, all of which are flawed, have been suggested to produce such comet showers periodically (reviewed by Sepkoski & Raup 1986; Shoemaker & Wolfe in Smoluchowski et al. 1986) : 1 A dim binary companion to the Sun, dubbed 'Nemesis' . This small star is hypothesized to have a highly eccentric orbit with a mean revolution time of 26- 28 million years . At aphelion, it would pass through the Oort Cloud and induce a comet shower. However, a distant companion has never been ob-

178

2 The Evolutionary Process and the Fossil Record

served, and simulations indicate it would be un­ stable and easily stripped from its orbit by passing field stars and molecular clouds . 2 An unobserved tenth planet, usually called 'Planet X' . If it had a slightly eccentric orbit inclined to the plane of the Solar System, orbital precession could bring the perihelion into the solar plane twice every 52 - 56 million years, at which time the planet would scatter comets from the inner edge of the Oort Cloud . However, a tenth planet has never been observed, and it is not clear whether it would have sufficient mass to scatter enough comets to leave a recognizable periodic signature on Earth . 3 Oscillation of the Solar System perpendicular to the Galactic plane . This well known behaviour moves the Solar System every 31 - 33 million years through the dense plane of the Galaxy, where gravi­ tational encounters with molecular clouds might perturb the Oort Cloud . However, the oscillation is out of phase with the extinction periodicity, and it has been argued that the mass of the Galaxy is not sufficiently concentrated in the plane to affect any distinct periodicity over a 270 million year interval . 4 Quasiperiodic transit of the Solar System through the spiral arms of the Galaxy . During its galactic orbit, the Solar System passes through either two or four arms, where concentrated mass may perturb the Oort Cloud . However, the intervals between transits are either about 60 or 125 million years, which is much longer than the observed periodicity of extinction . Alternative hypotheses that deep-Earth processes could induce periodic extinction are based on two lines of evidence (see also Section 2 . 12 . 1 ) : (1) there is an arguable periodicity of around 30 million years in the frequency of reversals of the Earth's magnetic field, suggesting some kind of regularity in deep-Earth dynamics (Loper et al. 1988); and (2) several periodic extinction events are associated with immense volcanic deposits (e . g . the Siberian traps, Deccan traps, and Columbia River basalts), which were produced during major episodes of basaltic volcanism . Such episodes could release large quantities of particulates, sulphates, and carbon dioxide into the atmosphere, perturbing climate and inducing extinction . Loper et al. (1988) argued that major volcanic episodes would be quasiperiodic if they were caused by variation in the thickness of the thermal layer at the base of the mantle . Thickening of this layer through time could lead to dynamical ins ta-

bilities that would spawn mantle plumes and cause widespread basaltic volcanism . Release of the plumes would draw material from the thermal layer, re-establishing stability and thus limiting the dur­ ation of the volcanic episode . This hypothesis of terrestrial forcing challenges, but does not negate, a role for extra-terrestrial im­ pacts in producing the observed distribution of extinction events : coincidental impact during a vol­ canic episode could greatly amplify a biotic crisis . Both sets of hypotheses are consistent with the implication from periodicity that most Mesozoic ­ Cenozoic extinction events share a common ultimate cause . But, as Lutz (1987) noted, the deep-Earth mechanism is not strictly clocklike but would oper­ ate by constraining waiting times between f'vents to generate the non-random distribution that is seen in the fossil record of extinction .

References Fischer, A . G . & Arthur, M.A. 1977. Secular variations in the pelagic realm. In: H . E . Cook & P. Enos (eds) Deep-water carbonate environments . Special Publication of the Society of Economic Paleontologists and Mineralogists No . 25, pp . 19-50. Fox, W.T. 1987. Harmonic analysis of periodic extinctions. Paleobiology 13, 257- 271 . Hoffman, A. 1985 . Patterns of family extinction: dependence on definition and geologic time scale . Nature 315, 659 662 . Hoffman, A. & Ghiold, J. 1985 . Randomness in the pattern of 'mass extinctions' and 'waves of originations' . Geological Magazine 122, 1 - 4. Kitchell, J.A. & Pena, D . 1984 . Periodicity of extinctions in the geologic past: deterministic versus stochastic expla­ nations . Science 226, 689 - 692. Loper, D . E . , McCartney, K. & Buzyna, G. 1988 . A model of correlated episodicity in magnetic-field reversals, climate, and mass extinctions. Journal of Geology 96, 1 - 16 . Lutz, T . M . 1987. Limitations t o the statistical analysis of episodic and periodic models of geologic time series . Geology 15, 1 1 1 5 - 1 117. Noma, E . & Glass, A . L . 1987. Mass extinction pattern : result of chance . Geological Magazine 124, 319 - 322. Patterson, C. & Smith, A . B . 1987. Is the periodicity of extinc­ tions a taxonomic artefact? Nature 330, 248-25l . Quinn, J . F . 1987. On the statistical detection of cycles in extinctions in the marine fossil record . Paleobiology 13, 465 -478. Rampino, M.R. & Stothers, R.B. 1984. Terrestrial mass extinc­ tions, cometary impacts and the Sun's motion perpendicu­ lar to the galactic plane . Nature 308, 709 - 712. Raup, D.M. & Sepkoski, J.J., Jr. 1984. Periodicity of extinctions in the geologic past. Proceedings of the National Academy of Sciences, U. 5.A. 81, 801 - 805. Raup, D.M. & Sepkoski, J.J., Jr. 1986 . Periodic extinction of

2 . 13 Mass Extinction: Events families and genera. Science 23 1 , 833- 836. Sepkoski, J .J . , Jr. 1986. Global bioevents and the question of periodicity. In: O . Walliser (ed . ) Global bio-events, pp . 47-6 1 . Springer-Verlag, Berlin. Sepkoski, J.J., Jr. 1987. Is the periodicity of extinction a taxonomic artefact? Response . Nature 330, 251 -252 . Sepkoski, J.J., Jr. 1989 . Periodicity in extinction and the problem of catastrophism in the history of life . Journal of the Geological Society of London 146, 7 - 19. Sepkoski, J .J . , Jr. & Raup, D . M . 1986. Periodicity in marine

179

extinction events . In : D.K. Elliott (ed . ) Dynamics of extinction, pp. 3 - 36 . Wiley, New York. Smoluchowski, R . 5 . , BahcalI, J . N . & Matthews, M . S . (eds) 1986. The galaxy and the solar system . University of Arizona Press, Tucson . Stanley, S.M. 1987. Extinction . Scientific American Books, New York. Stigler, S . M . & Wagner, M.J. 1987. A substantial bias in non­ parametric tests for periodicity in geophysical data. Science 238, 940- 945 .

2 . 13 Mass Extinction: Events

2 . 1 3 . 1 Vendian M . A . S . McMENAMIN

Introduction The earliest known, reasonably well documented mass extinction is of Vendian age, and seems to have occurred in the middle part of the Vendian, about 650 Ma. The severity and timing of this ex­ tinction is somewhat obscured by the difficulty of obtaining precise dates for Vendian sediments . Also, some losses of Vendian diversity appear to be the continuation of declines that began before the be­ ginning of the Vendian, such as the loss of many different types of stromatolites .

Micro-organisms Stromatolites reached a peak in diversity (nearly 100 recognized taxa) in the Late Riphean (c. 850 Ma) . Following this acme, stromatolites underwent a precipitous decline (see also Section 1 .5) starting in the second half of the Late Riphean and con­ tinuing through the Vendian . Stromatolite diversity bottomed out at less than 30 taxa by the beginning of the Cambrian . Although this decline does not necessarily represent the extinction of any of the individual microbial species that participated in the formation of stromatolites, it does indicate that the conditions became much less favourable for many formerly successful types of benthic microbial communities . For example, well formed specimens of the conical Proterozoic stromatolite Conophyton

are unknown after the Vendian. The advent of burrowing and grazing metazoans, and disturbance to microbial mats as a result of their activities, has been hypothesized as the factor responsible for the decline of stromatolites . Individual taxa of benthic microbial organisms (Section 1 .2), represented by delicate unicells and filamentous chains of cells preserved in chert, seem to have been largely unaffected by extinction during the Vendian, although it is difficult to recognize taxonomic turnover in floras consisting primarily of morphologically simple coccoidal and filamentous microbes . This problem is further compounded by the fact that fossilized benthic microbiotas are rare after the beginning of the Cambrian; apparently the conditions necessary for fossilization of microbes in chert became much less common after the end of the Vendian . A different situation exists with acritarchs, a het­ erogeneous group of organic-walled microfossils recovered from sediment by acid maceration . By comparison with modem dinoflagellate cysts, most acritarchs are thought to represent the resting stages of planktic, eukaryotic marine algae (Section 1 . 7. 2) . Both within-flora and total taxonomic diversity of these planktic microfossils underwent a severe de­ cline during the Middle to Late Vendian, which Vidal & Knoll (1982) regarded as indicative of major extinctions in the eukaryotic phytoplankton . Diag­ nostic acritarch taxa such as Trachysphaeridium laufeldi and the distinctively striate Kildinella lopho­ striata (Vidal & Knoll 1982) disappeared by the Middle Vendian . These distinctive Late Riphean - Early Vendian acritarchs were succeeded by a depauperate flora typified by Bavlinella faveolata (an acritarch that

180

2 The Evolutionary Process and the Fossil Record

resembles the existing colonies of spherical cyano­ bacteria called chroococcaleans) and the ribbon­ shaped vendotaenid algae . The sediments containing this depauperate flora also have curious­ ly large amounts of organic matter (sapropel) derived from the burial of acritarchs and other organic-walled objects . The re-radiation of the plank­ ton from this low-diversity interlude was slow . Acritarch diversity in most stratigraphic sections did not recover to Early Vendian levels until well into the Lower Cambrian, when very spiny forms such as Skiagia became abundant (but see Zang & Walter 1989) .

A

B

Metazoans The soft-bodied fossils of the Ediacaran fauna are generally thought to be metazoans (Section 1 .3) . Frondose or leaf-like Ediacaran forms such as Charnia and Charniodiscus are known throughout the world in sediments of Vendian age . Some of these organisms attained sizes of up to one metre in length . The second half of the Vendian (the Kotlin Horizon) is marked by local extinction on the Russian Platform of many of these large, distinctive soft-bodied creatures . Possibly coincident with the decline in phytoplankton diversity, Late Vendian metazoan faunas of the Russian Platform were re­ duced to rare problematic forms of medusoids and small trace fossils (Fedonkin 1987; Sections 1 . 3, 1 . 5) . The Ediacaran fauna seems to have died off by ' the end of the Vendi an (the top of the Rovno Horizon of the Siberian Platform), although a few of these soft-bodied forms may have survived into the Early Cambrian. Seilacher (1984) argued that the end of the Vendian witnessed a mass extinction of the soft-bodied Ediacaran forms, and that these extinc­ tions were real and were not an artifact of preser­ vation . It must be noted, however, that the intensity of burrowing increased greatly in the terminal Vendian. The trace fossils at this time became more complicated, deeper and larger, indicating an in­ crease in the dimensions of infaunal animals . This development may have reduced the potential for preservation of soft-bodied animals . The Late Vendian increase in burrowing intensity was accompanied by an explosion in the diversity of trace fossils . Numerous new ichnotaxa appeared that have ranges continuing through most or all of Phanerozoic time . Of the dozens of new ichnogenera that first appeared in the Vendian, only six became extinct by the end . Of these, Neonoxites, and Palaeo­ pascichnus were horizontal grazing or very shallow

c

2 mm

f------i

Three ichnogenera known from the late Vendian . A, Harlaniella. B, Neonoxites . C, Palaeopascichnus . All three are shallow deposit feeding or grazing, horizontal bedding plane trace fossils .

Fig. l

deposit feeding traces (Fig. 1 ) . If there was indeed a mass extinction at the end of the Vendi an, it was overshadowed by the metazoan diversification occurring at this time . The study of Vendian extinctions is hampered by a paucity of well preserved macrofossils . Neverthe­ less, the disappearance of acritarchs suggest that the Middle Vendian was marked by a mass extinc­ tion event that rivalled in magnitude the better known mass extinctions occurring later in the Phanerozoic . This Middle Vendian acritarch extinc­ tion event was linked to the Varangian glaciation by Vidal & Knoll (1982), who invoked climatic cool­ ing as a causal mechanism. More evidence is needed to clarify the timing, severity, and possible climatic control of these extinction events . Of particular in­ terest is the unresolved question of whether global metazoan mass extinctions occurred in the Vendian, and whether or not they were coincident with the phytoplankton extinctions .

References Fedonkin, M.A. 1987. The non-skeletal fauna of the Vendian and its place in the evolution of metazoans . Nauka, Moscow (in Russian).

181

2 . 13 Mass Extinction: Events Seilacher, A. 1984. Late Precambrian and Early Cambrian Metazoa: preservational or real extinctions? In: H.D. Holland & A.F. Trendall (eds) Patterns of change in Earth evolution, pp. 159 - 168. Springer-Verlag, Berlin. Vidal, G. & Knoll, A.H. 1982. Radiations and extinctions of plankton in the late Proterozoic and early Cambrian. Nature 297, 57-60. Zang, W.L. & Waiter, M . R. 1989. Late Proterozoic plankton from the Amadeus Basin in central Australia. Nature 337, 642 - 645 .

2 . 1 3 . 2 End-Ordovician P . J . BRENCHLEY

Brachiopods. Thirteen families of brachiopods be­ came extinct at or near the Ordovician- Silurian boundary . Of the 27 families which crossed the boundary, nine showed a marked decline in abun­ dance (Sheehan 1982) . Amongst the rich brachiopod faunas of the Ashgill of northwest Europe, 25% of genera disappeared at the top of the Rawtheyan and another 40% at the top of the Hirnantian (Fig. 1). Graptolites. The diversity o f graptolite species de­ creased from a high point in the Late Caradoc to a nadir in the Climacograptus extraordinarius and Glyp­ tograptus persculptus zones, when the total world graptolite fauna consisted of only a few genera . Primitive echinoderms. The diversity of cystoid, edrio­

Introduction About 22% of all families became extinct in the Late Ordovician, which makes this one of the largest episodes of mass extinction (Raup & Sepkoski 1982) . Although there were some extinctions throughout the Ashgill, the main phase of extinction was in the Late Ashgill . The Late Ordovician extinctions can­ not be related to a single stratigraphic level, but occurred in at least two steps. One phase coincided with the start of a major regression at the end of the Rawtheyan (the penultimate stage in the Ashgill) and a second phase coincided with a transgression .at the end of the Hirnantian (the last Ashgill stage), about 1 -2 million years later (Brenchley 1984) . There may in addition have been some extinctions throughout Hirnantian times . The two major phases of extinction have been best documented from clastic sequences in Europe . Upper Ordovician extinctions of comparable magnitude are known from carbonate sequences in North America but have not been clearly differentiated into two phases .

Extinction patterns

asteroid, and cyclocystoid families declined sharply in the Late Ashgill . The sharpest drop in num­ bers of cystoid genera in the families Diploporita and Dichoporita was at the Rawtheyan - Hirnantian boundary, when the rich and varied Rawtheyan fauna with 26 genera was reduced to a small but distinctive Hirnantian fauna with only eight. Most of the latter fauna apparently disappeared at the end of the Hirnantian .

1 00

\ \ \ \ \ \ \ \ \ \ \ \

80 rti ;;.. ;; �

60

�

40

'"

c:

IJ) u

0..

�

20

O�

----- -

Trilobites. Only about 14 o f the 38 families o f Middle Ashgillian trilobites survived into the Silurian. The main decline in diversity occurred at the end of the Rawtheyan when about 15 families disappeared. Generic diversity which had been fairly constant throughout the Ashgill declined by 45- 75% at the top of the Rawtheyan (Fig . 1) and species diversity was possibly reduced even more .

/

/

/

/

Brach iopods ( 1 57 gene r a ) Trilobites ( 1 1 2 gene ra )

�____�____�____�____�____� c c c c c c '" Cl Cl Cl Cl Cl >>C C (j) GJ c Cl C .c Cl -0 � 0 -0 :: � :; Cl � 0.. Cl U I .c � �

__ __

The Late Ordovician extinctions involved signifi­ cant changes to a number of groups :

/

/

/ /

.�

Fig. 1 Percentage survival of genera from stage to stage in the Upper Ordovician, e . g . 98% of Cautleyan trilobite genera survived into the Rawtheyan and 25% of Rawtheyan genera survived into the Hirnantian. The graphs are based on genera found in sequences with good stratigraphical control mainly in Europe. (After Brenchley 1984, by permission of John Wiley & Sons Ltd . )

182

2 The Evolutionary Process and the Fossil Record

Conodonts. The mainly clastic Hirnantian se­ quences of Europe have yielded very few species of conodonts, even where collecting has concentrated on the more promising limestone horizons. In the carbonate sequences of North America di­ verse conodont faunas declined a little in the Gamachian and disappeared almost completely at the Ordovician - Silurian boundary. Chitinozoa, acritarchs, and ostracodes. All three groups show major decreases in diversity and changes in taxonomic composition at or near the Ordovician ­ Silurian boundary .

Corals. The best data for the Late Ordovician show that c. 50 of the 70 tabulate and heliolitoid gen­ era became extinct in Late Ordovician times . It is not clear whether this was an end-Rawtheyan or Hirnantian extinction . Following the first wave of extinction at the end of the Rawtheyan there was a residual fauna, domi­ nated by brachiopods, which is usually referred to as the Hirnantia fauna . This fauna is unusually cosmopolitan and appears to have ranged from cir­ cumpolar to sub-tropical latitudes, though it was not well developed in the carbonate environments of tropical regions. The Hirnantia fauna is commonly considered to have been a relatively cool water fauna . The second wave of extinction at the top of the Hirnantian (top Gamachian in Canada) was rela­ tively modest in the clastic sequences of Europe . Several elements of the Hirnantia fauna disappeared at this level, and coral and ostracode faunas may have been heavily depleted . Coral-stromatoporoid reefs which occur at the top of the Hirnantian are rare or absent in the lower levels of the succeeding Silurian . In North America the diversity of brachiopods, trilobites, conodonts, acritarchs, and ostracodes greatly diminished at the end of the Ordovician (Lesperance 1 985 ), but because the detailed strati­ graphy is uncertain the extinctions could be Early or Late Hirnantian . Environmental changes In most shelf sequences there is a change of facies at the Rawtheyan - Hirnantian boundary, reflecting the start of the regression which reached its maxi­ mum in the Middle or Upper Hirnantian . The re­ gression partially drained many clastic shelves leaving a variety of shallow-marine sandy deposits .

A major part of the world's carbonate platforms became exposed with widespread development of karst surfaces and disconformities . At the top of the Hirnantian there is generally a sharp change in facies indicating a rapid trans­ gression . In many clastic sequences the shallow­ marine rocks of the Upper Hirnantian are overlain by black graptolitic shales . In carbonate regions there is a progressive return to more offshore carbonate facies . I t has been estimated that the regression in­ volved a fall in sea-level of 50- 100 m (Fig . 2) . In several Hirnantian sequences there is some evi­ dence of fluctuations of sea-level (two to four re­ gressions) but the pattern is not clear on a global scale .

Causes The cause or causes of the extinctions are debatable . The stepped nature of the extinctions makes an extra-terrestrial cause, such as meteorite impact, unlikely . Furthermore no iridium anomaly was discovered in detailed investigations of the Ordovician- Silurian stratotype at Dob' s Linn or in the carbonate sequence of Anticosti Island . The very precise correlation between the disappearance of faunas in many sections and the first evidence of regression makes it likely that the extinctions were related to contemporaneous environmental changes such as the following : 1

Sea-level changes. The fall in sea-level during the Hirnantian would have drastically reduced the size of continental shelves and platforms and hence the habitable area for shelf benthos. Many very exten­ sive platforms (N. America, Baltica and the Russian Platform) were covered by shallow seas during most of the Ordovician so a sea-level fall of tens of metres would have had a profound effect . The main argument against a major role for sea­ level change in causing extinctions is that the faunal changes were concentrated at the Rawtheyan ­ Hirnantian boundary while the regression appears to have continued throughout the early part of the Hirnantian . The second phase of extinction at the top of the Hirnantian coincides with a rise in sea­ level, and consequently a potential increase in hab­ itable area. However, following the transgression, black shales were deposited on many clastic shelves, indicating widespread anoxic or dysaerobic con­ ditions hostile to benthic faunas (see also Section 2.12.1).

183

2 . 13 Mass Extinction: Events z « et:: =:l

Vl

...l

Bathy m e t r i c cu rve Rhuddanian

�

« « �

0 ", c-

'" E (; Qj

"0

cL

�W w«

3-

o b: t--

� -'

-' w

e: .�

1:

::J

�I l "

0 "-

,',I

'" 0 "-

3

2

1

I " I I I

E (; Qj

1

'"""' 0 20 %

�

B

& ' 8 0%oPDB

Post K·T evo l u t i o n

-D

4

:J 01) :J

'" 13

S u rvivors

"0

cL


Cl.

::!

Z

u Z o e o �

Cl>

�

Cl>

�I Cl> 01 �

Cl> �

Cl>

C E R ATO PS I A I I .------.. ANKYLOSAU R I A .------,

ORN I T HOPODA

lJJ ­ « u L.U o

I�E�a_rl�Y

M I N E RA L I ZAT I O N

�

La --,

__ __ __ __ __ __ __ __ __ __ __ __ __ __

Fig. 1 The relationship between decay and mineralization in the preservation of soft-bodied and lightly skeletonized organisms . Reduced decay and early diagenetic mineralization are required for preservation of volatile tissues such as muscle. The preservational field of each tissue type extends from the bottom left-hand corner of the box to the boundary fence of the next most refractory type. (After Allison 1988 . )

Fig. 2 A , Three-dimensional preservation of mantle and appendage musculature in squid from the Jurassic Oxford Clay of Wiltshire, U . K . ; Bristol City Museum, Cb7661 . B, Close-up of muscle fibres in A. C, Flattened polychaete worm from the Upper Carboniferous Mazon Creek biota of Illinois, U.s.A. (A, B from Allison 1988b; reproduced with permission from the Lethaia Foundation. )

for over 1 1 km. Freshy killed organisms subjected to tumbling were hardly damaged (Fig. 3A), while car-

215

casses which had been allowed to decompose for several weeks were disarticulated and fragmented (Fig. 3B) . A sealed glass jar filled with seawater and carcasses of Palaemon was used as a control. The carcasses became buoyed up to the surface with decomposition gases and gradually disarticulated to produce a carpet of skeletal fragments upon the floor of the jar. Thus freshly-killed organisms could tolerate turbulent transport without fragmenting, while at the opposite extreme, carcasses were dis­ articulated when buoyed up by decay gases, even in the absence of currents . It is therefore primarily decay and not transport which determines the degree of fragmentation and disarticulation in soft bodied and lightly skeletonized taxa . Completeness or preservation is therefore no indicator of duration or nature of transport. This interaction between decay and hydrody­ namic processes has produced some difficult taxo­ nomic problems. The most common instance of this form of distortion is provided by fossil plants. A living plant will produce a number of different preservable structures such as pollen, seeds, fruit, and leaves. Upon death, the stem of the plant is commonly fragmented and separated from its root system. Thus, plant fossils are rarely encountered as whole entities . As a result of this bias the remains of most fossil plants are given form names (Section 5 . 1 . 3) . Animal remains too may be subject to this bias. An unusual example is provided by the large Middle Cambrian predator Anomalocaris, from the celebrated Burgess Shale of British Columbia (Section 3 . 1 1 .2) (Whittington and Briggs 1985) . This animal was one of the largest predators of its time, although for many years it was only known from disarticulated elements. The limbs were ori­ ginally identified as arthropod bodies and named Anomalocaris canadensis, while the mouth parts were thought to be a medusoid coelenterate (Peytoia nathorsti) . An incomplete body of the animal was named Lagania cambria and classified as a holo­ thurian . These 'animals' are in fact all part of the same organism. When Anomalocaris died and began to decompose, the mouth parts, body, and append­ ages were separated and deposited according to the hydrodynamic properties of each particular element. The recognition of this decay-induced distortion of fossil taxonomy was only achieved by the discovery of a number of rare complete individuals . The pres­ ervation of complete animals required deposition prior to decay-induced fragmentation. Conversely, the occurrence of disarticulated skeletal elements indicates a period of decay prior to final burial .

216

3 Taphonomy reduction, iron reduction, sulphate reduction, or methanogenesis) used by microbes in the decom­ position process. Sedimentary pyrite, for example, is produced as a by-product of bacterial sulphate reduction (Section 3 . 8 . 3), and manganese carbonates may be produced during manganese reduction (Section 3 . 8 . 2) . Similarly, the fractionation of carbon isotopes during bacterial decay and their incorpor­ ation into the lattice of carbonate minerals is diag­ nostic of specific decay reactions (Coleman 1985) The rarest and most spectacular characterization of decay processes is the preservation of fungi and bacteria in fossil organisms (Allison 1988a) . When bacteria die they undergo autolysis, whereby en­ zymes and other cell contents begin to corrode and eventually destroy the cell wall . Such a process takes hours or days. Thus the mineralization of microbes implies extremely rapid diagenetic growth . Further work on these microbe - carcase associations is required in order to fully understand their significance . References

Fig. 3 Carcasses of Nephrops . A, Freshly-killed individual after tumbling in rotating barrel . Note that although carcase is decapitated, delicate structures such as the appendages have survived . B, Individual tumbled after 26 weeks of decay : a, rostrum; b, c, segments of chelae nearest to coxae; d, pincer; e, mandible; and f, segment of chela attached to pincer.

Characterization of decay

Decay in the fossil record can be characterized on three levels : (1) the identification of information loss and decomposition structures ir, particular fossil organisms; (2) the recognition of particular minerals and geochemical markers associated with particular decay regimes; and (3) the preservation of fossil microbes involved in the decomposition process . The most basic characterization of decay, that of level of preservation in macro-organisms (e . g . per­ mineralized muscle, tissue impressions), merely documents extent of decay prior to mineralization (Figs 1, 2) . A more detailed characterization relates specific geochemical markers to particular decay pathways (i . e . aerobic decay, nitrate reduction, manganese

Allison, P.A. 1986 . Soft-bodied fossils : the role of decay in fragmentation during transport. Geology 14, 979 - 98l . Allison, P.A. 1988a . Konservat-Lagerstatten : cause and classifi­ cation. Paleobiology 14, 331 - 344 . Allison, P.A. 1988b . Phosphatized soft-bodied squids from the Jurassic Oxford Clay. Lethaia 2 1, 403 -410. Coleman, M.L. 1985 . Geochemistry of diagenetic non-silicate minerals: kinetic considerations. Philosophical Transactions of the Royal Society of London A315, 39 - 56 . Glob, P . V . 1969 . The bog people. Faber, London . Jones, G . F . 1969 . The benthic macrofauna of the mainland shelf of Southern California. Allan Hancock Monographs in Marine Biology 4, 1 -219. Jorgenson, B . B . 1982 . Ecology of the bacteria of the sulphur cycle with special reference to anoxic- oxic interface en­ vironments . Philosophical Transactions of the Royal Society of London B298, 543 - 56l . Jorgenson, B . B . 1983 . Processes at the sediment -water inter­ face. In: B. Bolin & R.B. Cook (eds) The major biochemical eye/es and their interactions, pp. 477-561 . John Wiley & Sons, Chichester. Stout, J . D . , Goh, K.M. & Rafter, T.A. 1981 . Chemistry and turnover of naturally occurring resistant organic com­ pounds in soil. Soil Biochemistry 5, 1 - 73 . Whittington, H.B. & Briggs, D . E . G . 1985 . The largest Cambrian animal Anomalocaris, Burgess Shale, British Columbia. Philosophical Transactions of the Royal Society of London B309, 569 - 609 . Williams, A.M. 1963. Enzyme inhibition by phenolic com­ pounds. In: J . B . Pridham (ed . ) The enzyme chemistry of phenolic compounds, pp. 57-85. Pergamon Press, Oxford .

3 . 2 The Record of Organic Components and the Nature of Source Rocks P . FARRIMOND & G . EGLINTON

of the algal material is changed during passage through the gut of the grazing organism; various organic components are preferentially assimilated and modified during digestion, and other lipids may be contributed from tissues of the grazer. An example of this 'editing' process is the observed increase in certain sterols, such as cholesterol (Fig. 1), in faecal pellets of zooplankton fed on phytoplankton (Harvey et al . 1987) . Microbial ac­ tivity, proceeding both in the gut of the feeder and, later, within the faecal pellets, also plays a role in modifying the molecular composition of the organic matter in its descent to the sea floor. Upon arrival at the sediment, organic matter is further modified by a variety of processes acting during early burial. It is during this early diagenesis that biological compounds and debris are incorpor­ ated into insoluble sedimentary organic matter. In addition to the free lipids, the organic matter entering the sedimentary regime comprises bio­ polymers such as carbohydrates, proteins, cutins, and lignins, all of which are available for consump­ tion and modification by benthic macro- and micro­ organisms . There is evidence that a variable fraction of carbohydrates and proteins is initially converted to individual sugars and amino acids by enzymatic microbial attack prior to the use of the resulting monomers by microbes as a source of energy and to form new cell material . The remainder, not utilized in this way, can undergo polycondensation to form geopolymers; these complex, high molecular weight materials may incorporate fulvic and humic acids . This heteropolymeric debris has been termed 'protokerogen' - the precursor of kerogen. With further sediment burial, increasing condensation and insolubilization accompanies the slow dia­ genetic conversion to kerogen, which constitutes the bulk of the organic matter in ancient sediments . Biolipids may be incorporated into kerogen in a similar way, or may be preserved in the sediment with only minor modification . Diagenetic reactions at various stages of burial appear to convert some lipids to hydrocarbons (Fig. 1) through the loss of functional groups via dehydration, hydrogenation,

Preservation and diagenesis

Organic molecules are abundant constituents of many sediments and sedimentary rocks . These components have been referred to as 'chemical fossils' in recognition of their biological origin, but the terms 'biological marker' or simply 'biomarker' are more commonly used . Macro- and micro fossils are readily apparent in rocks, but the identification of chemical fossils requires sophisticated techniques of sample work-up and analysis; nevertheless, they too preserve a remarkably detailed record of past biological activity. 'Biological markers' are defined as organic compounds present in sediments (or petroleums) which possess chemical structures un­ ambiguously related to present day biologically­ occurring organic molecules (Fig. 1 ) . Obviously, the possible sources of biomarkers in geological samples are almost limitless, comprising all organisms in the palaeoenvironment of deposition - aquatic, land, and air . Consequently, the molecular record is in­ variably complex. Furthermore, numerous chemical reactions, both biologically and non-biologically mediated, proceed within the water column and then during sedimentation and burial of the organic debris; these serve to modify and diversify the record of organic components still further. Only a relatively small proportion of the organic matter produced within, or supplied to, ocean sur­ face waters ever reaches the underlying sediments; the vast proportion of this material is recycled (much of it 'remineralized' to carbon dioxide) within the water column, particularly in the euphotic zone . Many processes act to modify the organic flux, including photo-oxidation, microbial activity, and predation by grazing organisms. Of the very small fraction of the original marine organic material which arrives at the sediment, a large proportion is generally transported in the form of faecal pellets released by zooplankton or organisms higher in the food chain. Such faecal pellet transfer is relatively rapid, allowing marine organic matter produced in the euphotic zone largely to escape photo-oxidative degradation . However, the molecular composition

217

3 Taphonomy

218

BIOLlPID

GEOLl PID

HO C h o lesterol

S a ( H ) -C h o l esta n e

Po r p h yr i n C h l o ro p h y l l a

OH

OH

OH Bacte r i o h o pa n etetro I

C35 H o p a n e

and decarboxylation. Such hydrocarbons cannot be readily incorporated into geopolymers by poly­ condensation reactions . However, a proportion may become trapped in the kerogen structure . The remaining free hydrocarbons and other related compounds comprise only a small proportion of the organic matter in a sediment (typically less than 5%), although they have a high information content. These 'chemical fossils' have been introduced into the sediment from their source organisms with only relatively minor changes to their molecular structure . It is only through knowledge of the reactions proceeding during sedimentation of organic matter through the water column, and during its sub­ sequent burial in the sedimentary record, that

Fig. 1 Three biologically widespread molecules and their geologically occurring products after diagenesis . Note that, in each case, structural specificity is maintained - the geolipids are thus 'chemical fossils', having an unambiguous link with their precursor biolipids .

geologically-occurring lipids may be used as a source of information . Certain lipid classes, notably the steroids (Mackenzie et al. 1982), are becoming well understood in this respect, although other classes await study. A knowledge of precursor ­ product relationships allows the use of sedimentary organic components as indicators of biological sources of organic matter, depositional environment or conditions, climatic variations, and organic matter maturity. Biological marker compounds and their uses

Biological marker compounds have a wide variety of structures, all specifically indicative of a biological origin. The degree of specificity in the structure

3 . 2 Organic Components and Source Rocks may enable inferences to be made as to the precursor molecules, and hence the ultimate origin in a par­ ticular family, class, or even genus of organism (see also Section 2 . 1 ) . Of course, detailed chemotaxo­ nomic information for modern organisms is the essential basis for successful correlation with such biological sources . Furthermore, when applying biomarkers as source indicators in ancient sedi­ ments it is necessary to make the major assumption that ancestor organisms possessed similar molecular compositions to their modem descendants . How­ ever, there are often good biosynthetic grounds for such assumptions . Obviously, it is desirable for links to be established between specific fossils (macro and micro) and the molecular record. For example, what does the brown or black material comprising a leaf fossil really consist of? Is there a molecular record of the original lignin, cutin, or wax? Similar questions apply to other macrofossils (e .g. fish re­ mains) and microfossils . Unfortunately such work is only in the early stages . Nevertheless, despite these constraints, a considerable number of indica­ tive compounds, and, indeed, classes of compound, are generally accepted as reflecting certain biological inputs, as discussed below. These and other bio­ logical markers are reviewed by Brassell et al. (1978) and Philp (1985) . Straight-chain alkanes (n-alkanes) and their func­ tionalized equivalents (n-alcohols (alkanols), n-fatty acids (alkanoic acids), and n-alkanones) are common constituents of the majority of organisms (e . g . leaf waxes of higher plants, membrane lipids of algae, etc . ) . In addition, the distributions of carbon chain lengths of these compounds are informative as to the origin of organic matter in a sediment. In gen­ eral, short- (C 1 S - C 1 9) and medium-chain (C20 - C24) compounds reflect algal and/or bacterial sources, whilst long-chain compounds (C27-C33) typify a higher plant contribution. A class of organic compounds known as hopa­ noids are ubiquitous constituents of sediments . Several biological precursors of the geological hop­ anoids have been identified - almost all are bac­ terial in origin (Fig. 1 ) . More specific biological marker compounds have also been proposed . For example, certain long-chain acyclic isoprenoids are common constituents of archaebacteria; further­ more, some compounds appear to be restricted to methanogens (Brassell et al. 1981 ) . Other widely accepted biological marker compounds include 18<x(H)-0Ieanane (higher plants), 4-methylsteroids (especially dinosterol; dinoflagellates), long-chain alkenones (prymnesiophyte algae), and botryo-

219

coccane or botryococcenes (only observed in the fresh- or brackish-water alga Botryococcus braunii) . An appraisal of the biological sources of the sedi­ mentary organic matter, and the relative importance of specific contributions, aids the reconstruction of the environment of deposition of the sediment. For example, freshwater and marine sediments may usually be distinguished by their molecular signa­ tures, owing to the contribution of organic matter from different organisms in the two environments . Furthermore, in the marine realm, the abundance of terrestrial organic matter is related to proximity to land and the importance of fluvial and/or aeolian transport of land-plant debris . In petroleums, the molecular composition is the best (if not the only) source of information regarding the environmental setting of its source rock. In addition to providing clues to the broad depo­ sitional setting of a sediment, the molecular record is instructive with regard to the environmental conditions prevailing at the time of deposition . Of prime concern here is the oxicity of the water column . Didyk et al. (1978), in an extension of the work by Powell & McKirdy (1973), proposed the ratio of two related organic compounds, pristane and phytane, as an indicator of oxygen levels at the site of deposition . Whilst the basic rationale behind this indicator is sound - namely two different reaction pathways from the same precursor (the phytol side chain of chlorophyll; Fig. 1), the one followed being dependent upon the oxygen level of the environment - the effects of differences in organic matter sources and maturity complicate its use . However, when used in conjunction with other evidence, such as porphyrin content, this ratio can be a useful indicator of the degree of oxygenation. Sediments deposited in hypersaline environ­ ments are frequently characterized by distinctive distributions of biomarkers (ten Haven et al . 1988) . These unusual molecular signatures presumably reflect a contribution of organic matter from salinity­ tolerant organisms, coupled with the presence of highly reducing conditions of deposition . During sediment burial and organic matter matu­ ration, biological marker distributions are modified through chemical reactions . Whilst early diagenesis is characterized mainly by reactions involving the loss of functional groups, during late diagenesis and catagenesis the biomarker reactions are domi­ nated by isomerization and degradation processes (Mackenzie et al . 1980) . Each reaction proceeds over a specific range of maturity, dependent upon time, temperature, and to a lesser extent, pressure (Tissot

220

3 Taphonomy

& Welte 1984) . Consequently, determination of the extent of such reactions in a sediment (typically by molecular product- precursor ratios) allows the assessment of maturation stage - critical in oil generation studies . A further application of the molecular components of sedimentary organic matter lies in the recon­ struction of palaeoclimatic fluctuations . Recent progress in this area includes the recognition of a molecular 'palaeothermometer' in a group of organic compounds called alkenones (Brass ell et al. 1986) . A simple molecular parameter is now available which can be used to illustrate past fluctuations in sea­ surface temperatures, as prymnesiophyte algae modify their molecular composition in response to long-term temperature changes . This approach is currently being employed to record glacial or interglacial cycles in deep-sea sediment cores, and compares well with classical oxygen isotope measurements on foraminifera . Biological marker compounds may also record marine productivity changes, and variations in aeolian transport of terrestrial organic debris.

The nature of source rocks

Exactly what constitutes a hydrocarbon source rock has long been a matter of debate, although advances in petroleum geology and geochemistry have re­ sulted in the general acceptance of a broad defi­ nition. Brooks et al . (1987) define a source rock as 'a volume of rock that has generated or is generating and expelling hydrocarbons in sufficient quantities to form commercial oil and gas accumulations' . A potential source rock is a volume of rock which has the capacity to generate commercial hydrocarbon accumulations, but is of insufficient maturity. Most source rocks are fine-grained, typically dark­ coloured shales or marls . However, the organic matter within a sediment must meet minimum re­ quirements for organic richness and quality or type in order for the rock to be considered a source bed . Most potential source rocks contain between 0 . 8 and 2 % organic carbon; an approximate limit of 0 .4% is commonly accepted as the lowest organic carbon content for hydrocarbon generation and expulsion to occur. Of course, there is no general upper limit of organic richness, and many of the best source beds contain upwards of 5 - 10% organic carbon. The kerogen in a source rock may contain par­ ticulate organic matter from a variety of sources in fact, the nature of the hydrocarbons generated is

strongly dependent upon the kerogen composition . Most kerogens are mixtures of two types of organic matter: terrigenous higher plant debris and aquatic (marine or lacustrine) lower plant material. Micro­ scopic analysis of source rocks reveals that most of the sedimentary organic matter is amorphous, with only a minor part comprising recognizable biologi­ cal debris. Sediments containing large quantities of yellow-brown amorphous organic matter of algal and/or bacterial origin (i . e . types I or 11; Tissot & Welte 1984) will produce petroleum given sufficient maturation. In contrast, sediments containing type III kerogens, comprising abundant particulate land plant debris, will liberate mainly gas . There are two main prerequisites for the ac­ cumulation of significant quantities of organic matter in sediments : production of organic matter, and its subsequent preservation . Both are controlled by many variable factors (Fig. 2) .

Production of organic matter. Source rocks may be deposited in marine or lacustrine environments . Owing to their greater importance, only marine source rocks will be discussed here, although many of the factors controlling organic matter accumu­ lation apply in both environments . Marine primary productivity typically supplies the bulk of organic matter to marine source rocks, although processes within the water column utilize much of the organic debris before it can reach the sediment. Surface productivity is largely controlled by water temperature, light intensity, and the avail­ ability of nutrients . The latter may be influenced by sea-level (with the flooding of coastal areas during periods of high sea-level introducing terrigenous nutrients), water column overturn (resulting from storm activity or improved deep circulation), and up welling of nutrient-rich water. Upwelling is, in turn, controlled by the action of prevailing winds and the Earth's Coriolis forces, and by the distri­ bution of land masses . Present-day upwelling areas overlie many of the most organic-rich sediments in the oceans. Terrigenous higher plant debris, which may also be a significant constituent of hydrocarbon source rocks, may be introduced into the marine environment by flooding of coastal areas during transgression, or by aeolian or fluvial transport.

Preservation of organic matter. The accumulation of organic debris in sediments depends to a large extent upon the inhibition of chemical oxidation and biochemical degradation processes during transport, deposition, and early burial . These pro-

221

3 . 2 Organic Components and Source Rocks

Terrest rial influx of o rga n i c matter

Water tem pe r ­ atu re

co l u m n stratificati o n

Balance o f evaporation vs. prec i p i tation

Fig. 2 Flow diagram showing various interrelated factors which influence the production of organic matter in the biosphere, and its subsequent preservation in the geosphere . These factors may all exert some control upon the accumulation of organic matter in marine sediments .

cesses, in turn, depend upon sediment particle size, sedimentation rate, mode of transport of organic matter to the sediment, and water column oxicity. Thus, organic-rich sediments are typically fine­ grained, and are favoured by relatively high sedi­ mentation rates, resulting in rapid burial. Rapid transit of organic debris through the water column, either through faecal pellet transport or sediment re deposition (turbidity currents, etc . ), also favours organic matter preservation. However, the oxicity of the water column, particularly at the sediment surface where residence time of organic matter is generally high, has long been recognized as the major control on the preservation of organic carbon in sediments . Under oxygen-depleted conditions, aerobic bacterial activity is absent, and degradation of organic matter is limited to the action of the less efficient anaerobic bacteria (see also Section 3 . 1 ) . Furthermore, the grazing o f macro-organisms on the sediment surface ceases in low-oxygen con­ ditions; consequently, there is no bioturbation to promote the access of oxygen and aerobic bacterial degradation within the upper sediments . The resulting sediments are usually finely laminated,

and typically contain relatively large amounts of organic matter. The oxygen content at any point in the water column is controlled by oxygen demand (which is controlled by organic matter degradation), oxygen supply, and oxygen solubility (which is greatly reduced in warmer or more saline water) . Oxygen supply in the marine environment is largely a func­ tion of deep-water circulation, although oxygen is supplied to surface waters by exchange with the atmosphere and photosynthetic production. Demaison and Moore (1980) discussed several models for the deposition of oil source beds - all involving highly oxygen-depleted conditions . One such model is that of a restricted and/or stratified basin. Oxygen-deficient conditions may develop in sedimentary basins where physical barriers tend to inhibit water circulation, particu­ larly in basins with a positive water balance (i. e . river inflow exceeding evaporation) . The present­ day Black Sea is a much-cited example of an anoxic silled basin with organic-rich sediments . Depo­ sition of potential source beds is also favoured in permanently stratified lakes (e . g . Lake Tanganyika) .

222

3 Taphonomy

Density stratification in basins may be induced by the influx of dense, oxygen-poor, saline water (formed in shelf areas where evaporation is high), or by the influx of low density freshwater (in areas of high precipitation) . Such stratification in the water column inhibits circulation, and hence oxygen replenishment is poor. The second type of oxygen-deficient environment where organic-rich sediments are characteristic is that of an expanded mid-water oxygen-minimum layer. The best developed of these form in response to coastal upwelling of nutrient-rich waters in areas where oxygen supply cannot match demand as or­ ganic matter degrades in the water column (e . g . Peru Upwelling) . Alternatively, oxygen-minimum layers may develop in areas where productivity is normal, but oxygen supply is poor due to isolation from a source of well oxygenated water. In either case, organic-rich sediments may be deposited where the oxygen-minimum layer impinges upon a continental slope or shelf. Open ocean oxygen minima, covering wide areas of the oceans, may have been important during specific times in the past - the so-called 'oceanic anoxic events' . These relatively short periods of geological time were characterized by widespread accumula­ tion of organic-rich sediments . The best known examples occur in the Cretaceous (Aptian ­ Albian, Cenomanian -Turonian, and Coniacian ­ Santonian), although another well defined oceanic anoxic event occurs in the Toarcian Gurassic) . Organic-rich sediments from these intervals com­ prise a large proportion of the world's potential and actual source rocks . References Brassell, S . c . , Eglinton, G . , Maxwell, J.R. & Philp, R.P. 1978. Natural background of alkanes in the aquatic environ­ ment . In: O. Hutzinger, L H . van Lelyveld & B . C .J . Zoeteman (eds) Aquatic pollutants : transformation and bio-

logical effects, pp . 69 - 86. Pergamon Press, Oxford . Brassell, S . c . , Eglinton, G . , Maxwell, J.R. Thomson, L D . & Wardroper, A.M.K.' 1981 . Specific acyclic isoprenoids as biological markers of methanogenic bacteria in marine sediments. Nature 290, 693 - 696. Brassell, S . c . , Eglinton, G . , Marlowe, L T . , Pflaumann, U. & Sarnthein, M. 1986. Molecular stratigraphy: a new tool for climatic assessment . Nature 320, 129 - 133. Brooks, J., Cornford, C. & Archer, R. 1987. The role of hydro­ carbon source rocks in petroleum exploration . In: J. Brooks & A.J. Fleet (eds) Marine petroleum source rocks, pp. 17-46. Special Publication of the Geological Society of London No . 26. Blackwell Scientific Publications, Oxford. Demaison, G.J. & Moore, G.T. 1980 . Anoxic environments and oil source bed genesis . Organic Geochemistry 2, 9 - 31 . Didyk, B . M . , Simoneit, B . R T. , Brassell, S . c . & Eglinton, G . 1978. Organic geochemical indicators o f palaeoenviron­ mental conditions of sedimentation. Nature 272, 216-222. Harvey, H.R., Eglinton, G., O'Hara, S . C . M . & Corner, E . D . 5 . 1987. Biotransformation and assimilation of dietary lipids by Calanus feeding on a dinoflagellate . Geochimica et Cosmochimica Acta 51, 3031 - 3040 . Haven, H . L . ten, de Leeuw, J.W. & Sinninghe Damste, J . S . 1988. Application o f biological markers i n the recognition of palaeo-hypersaline environments . In: A.J. Fleet, K . Kelts & M.R. Talbot (eds) Lacustrine petroleum source rocks . Special Publication of the Geological Society of London, No . 40, pp . 123 - 140 . Blackwell Scientific Publications, Oxford. Mackenzie, A.5., Patience, R.L., Maxwell, J.R., Vandenbroucke, M. & Durand, B. 1980. Molecular para­ meters of maturation in the Toarcian shales, Paris Basin, France - L Changes in the configurations of acyclic isoprenoid alkanes, steranes and triterpanes. Geochimica et Cosmochimica Acta 44, 1709 - 1 721 . Mackenzie, A . 5 . , Brassell, S . c . , Eglinton, G. & Maxwell, J.R. 1982. Chemical fossils - the geological fate of steroids. Science 217, 491 - 504. Philp, R.P. 1985 . Biological markers in fossil fuel production. Mass Spectrometry Reviews 4, 1 - 54. Powell, T.G. & McKirdy, D.M. 1973. Relationship between ratio of pristane to phytane, crude oil composition, and geological environment in Australia. Nature, Physical Sciences 243, 37-39. Tissot, B . P . & Welte, D.H. 1984. Petroleum formation and occurrence, 2nd Edn . Springer-Verlag, Berlin.

3 . 3 Destructive Taphonomic Processes and Skeletal Durability C . E . B RE T T

Destructive processes

evidenced by decay experiments using controls in cages that exclude larger organisms . Scavenging and burrowing processes are precluded in anaero­ bic environments, thus favouring articulated pres­ ervation. Physical agents, such as current and wave turbu­ lence, also produce disarticulation in skeletons which have undergone some decay. It is frequently assumed that the transport of carcasses over any distance will result in their disarticulation. How­ ever, if organisms are transported just prior to death, or immediately thereafter, this may not be the case (Allison 1986; Section 3 . 1 ) . Conversely, once connective tissues have decayed, even very minor currents (less than 5 cmls) may be effective in producing complete disarticulation . Interlocking structures of skeletons inhibit disar­ ticulation. For example, the interlocking hinge-teeth of certain brachiopods (such as terebratulids) may prevent disarticulation of the valves for extended periods of time . The tightly crenulated sutures of some pelmatozoans and echinoids appear to be similarly resistant. Thus, most multielement skeletons can only be preserved as articulated remains if they are buried extraordinarily rapidly (hours to a few days) . Anoxic environments promote articulated preservation, as does an absence of turbulence . However, these fac­ tors are not sufficient in themselves to explain this mode of preservation . Tightly sutured skeletons (e . g . the tests of echinoids, and crinoid stems), on the other hand, may withstand much longer periods of exposure in marine environments .

Durability refers to the relative resistance of skel­ etons to breakdown and destruction by physical, chemical, and biotic agents . The processes of skel­ etal destruction can be subdivided into five cat­ egories which follow one another, more or less sequentially, as remains of organisms are exposed in different environments (Seilacher 1973; Muller 1979; Brett & Baird 1986) : (1) disarticulation; (2) frag­ mentation; (3) abrasion; (4) bioerosion; and (5) corrosion and dissolution. Depending on the physi­ cal characteristics of the sedimentary environment, one or more of these processes may be more active . 1

Disarticulation is the disintegration of multiple­ element skeletons along pre-existing joints or articulations . There is a paucity of hard data on disarticulation rates, although this has been partly alleviated by several observational and experimental studies (Schafer 1972; Allison 1986, 1988; Meyer & Meyer 1986; Plotnick 1986) . Disarticulation may occur even prior to death in the case of moulting, which yields recognizable exuviae in many arthro­ pods. In most cases, disarticulation proceeds very rapidly after the death of an organism, and may in­ volve biochemical breakdown of tissues by enzymes present in the body of the organism itself. Bacterial decay (see also Section 3. 1) of ligaments and con­ nective tissues proceeds at a variable rate depending upon the nature of the tissues, as well as the local environment of decay. Aerobic decay of tissues proceeds rapidly in most cases; e . g . , the ligaments binding echinoderm ossicles are broken down within a matter of hours to a few days after death . Hinge ligaments composed of conchiolin in bivalves are evidently more resistant, and can remain intact for periods of months, despite fragmentation of the shells. Anoxia obviously inhibits bacterial decay. None the less, recent experiments indicate that anaerobic bacteria destroy ligaments and connective tissues within a matter of a few weeks to months . Biotic agents, including scavengers and infaunal burrowers, may greatly accelerate disarticulation as

2

Fragmentation of skeletons results both from physical impact of objects and from biotic agents such as predators and scavengers . Some fragmen­ tation may occur prior to death, such as that pro­ duced by attempted predation (see also Section 4. 13) . Distinct fragments or patterns of breakage may be recognizable in certain instances, e . g . the curved fractures produced by peeling of gastropod apertures by crabs. However, more commonly, pre223

224

3 Taphonomy

dation damage is indistinguishable from physical breakage . Shells tend to cleave along pre-existing lines of weakness such as growth lines, or ornamentation such as ribbing, and yield consistent types of frag­ ments (Fig. 1 ) . Resistance to fragmentation relates to several aspects of skeletal morphology and com­ position, including thickness and curvature of shells, microarchitecture, and percentage of organic matrix. In general, nacreous (pearly) skeletal fabric in mollusc shells are most resistant to breakage by impact, whereas foliated shells are more fragile . Bacterial decomposition of organic matrix greatly weakens shells, and makes them much more sus­ ceptible to fragmentation by other agencies; hence, for example, the high organic content of the shells of certain nuculid bivalves has probably resulted in their under-representation in the foss�l record . Sur­ ficial exposure time is also critical; microborings of endolithic algae and fungi greatly weaken shell structure and facilitate breakage . Delicate skeletons of corals, bryozoans, graptolites, and other fossils are particularly prone to fragmentation, even in slightly agitated waters . Hence, they form key taphonomic indicators of changes in current energy among facies . A high degree of fragmentation sug­ gests persistent breakage and reworking, perhaps within normal wave base . Extraordinary events, such as storms, may also generate currents or waves that impinge on otherwise quiet environments and cause intermittent fragmentation.

3 Abrasion, or physical grinding and polishing, results in the rounding of skeletal elements and loss of surficial details (Fig. lE). The extent of abrasion in any given type of skeleton is related to environ­ mental energy, exposure time, and particle size of the abrasive agent. In general, the rate of abrasion increases with increasing grain size: clay-sized grains do not significantly abrade skeletons; sand­ and gravel-sized material is probably the most effective agent. Semiquantitative measurements of abrasion rates have been obtained by tumbling shells artificially (Fig. 2; Chave 1964) . Two factors strongly influence the relative resistance of skel­ etons : size relative to the grain size of the sediment, and microarchitecture . Not surprisingly, small bi­ valve shells are fragmented and abraded much more readily than large ones. Furthermore, dense skeletal microstructures, such as crossed-lamellar structure in molluscs, are relatively hard and resistant to abrasion. Gastropods with dense shells may survive over one thousand hours of continuous tumbling.

B Fig. 1 A, Abrasion and breakage may produce fragments diagnostic of different depositional environments. Stably anchored shells are abraded from the top down (anchor faceting) . Pounding of shells by surf produces fractures that follow medial and concentric lines of weakness in the shell. Rolling and gliding of the shell will abrade the outer edge (glide faceting) . B, Roll fragments are produced by shells tumbling in an abrasive medium which preferentially destroys thinner parts of the shell, leaving thickened umbonal parts intact: spiriferid brachiopods from the Lower Devonian Oriskany Sandstone of Maryland. (After Seilacher

1973 . )

Moderately porous and/or organic-rich shells dis­ play intermediate durabilities, while very porous skeletons, such as those of bryozoans and algae, abrade very rapidly and will be selectively removed from the fossil record of high energy environments . However, porous particles such a s echinoderm os­ sicles may be cushioned against abrasion by their low density . Thus, it is commonly assumed that echinoderm ossicles can be strongly abraded only if they have undergone some early diagenetic permineralization .

4 Bioerosion, commonly associated with recogniz­ able trace fossils such as the borings of clionid sponges and various endolithic algae, proceeds at

225

3 . 3 Skeletal Durability

1 00�-------r--r---1

Fig. 2 Durability of invertebrate skeletal material in a tumbling barrel filled with chert pebbles. Numbers following each skeleton name give the initial size range in centimetres . (After Chave 1964; reproduced with the permission of John Wiley & Sons, Inc . )

o

Cor a l l i n e �AlgQe

very high rates in most shallow-marine environ­ ments . Rates from 16% to over 20% weight loss per year, as a result of algal and sponge boring, have been observed for modern marine mollusc shells. It is not clear whether such rates pertained in the Palaeozoic when clionid sponges were much less abundant. As with abrasion, shell thickness, organic content, and perhaps density may influence the relative resistance of skeletal material to destruction by bioerosion. 5

1 00

1

Corrosion and dissolution of skeletons result from chemical instability of skeletal minerals in seawater or in sediment pore-waters . Dissolution may begin at the sediment- water interface and continue to considerable depths within the sediment. Biotur­ bation of sediments commonly promotes dissol­ ution by the inmixing of fresh seawater and by oxidation of sulphides to produce weak acids within sediment pore-waters . A general ordering of the stability of minerals is as follows : phosphate > silica > echinoderm calcite > other skeletal calcite > aragonite . In addition, skeletal materials containing a high proportion of organic matter, such as nacreous shell, are relatively more resistant to dissolution than those with pure carbonate mineralogies, a trend which runs counter to destruction by abrasion or fragmentation. This differential stability results in biases in the records of different groups: e . g . , calcitic brachiopod shells may be extremely well preserved where aragonitic molluscs occur as highly compacted internal ­ external moulds. In practice, the effects of mechanical abrasion, most bioerosion, and corrosion are difficult to dis­ tinguish in fossils . Hence, Brett and Baird (1986) suggested the use of the term corrasion to indicate the general state of wear in shells resulting from any

1 000

T I M E I N H O U RS ( l o g sca l e )

combination of these processes . Corrasion provides a general index of exposure time to various agencies of wear on the sea floor.

Skeletal durability

Destructive processes of disarticulation, fragmen­ tation, and corrasion are readily evident in the fossil record. These processes affected different skeletal types in different ways . Most marine skeletonized organisms can be assigned to one of five skeletal architectural categories : massive, arborescent, uni­ valved, bivalved, or multielement. Table 1 provides a summary of biostratinomic processes, such as fragmentation and disarticulation, with respect to their influence upon each of the five skeletal types . I n general, massive skeletons are the least subject to breakage and are most resistant to mechanical destruction. However, because they remain on the sea floor for prolonged spans of time, such massive skeletons often display the effects of corrasion to a greater extent than other skeletons. Arborescent skeletons are probably the most sensitive indicators of fragmentation; an absence of breakage in such skeletons is an excellent indicator of minimal dis­ turbance of the sedimentary environment. Most bivalved skeletons become disarticulated relatively rapidly after death, although those with tough con­ chiolin ligaments may remain articulated for exten­ sive periods . Finally, multielement skeletons provide the best indicators of rapid burial, as they disarticu­ late extremely rapidly in the absence of sediment cover. Taken together, various skeletal types and their varied sensitivities to destructive agents may provide excellent indicators of sedimentary pro­ cesses, and can be used to define taphonomic facies (Section 3 . 9) .

3 Taphonomy

226

Potential utility of various invertebrate skeletal types as qualitative indicators of physical environmental parameters . In each case the types of evidence useful for inferring a given condition (e . g . high energy) are listed as symbols, defined at the bottom of the table . (From Brett & Baird 1986.)

Table 1

Current/wave transport of skeletons

Skeletal type

Azimuthal (compass-bearing) orientation

Burial rate

Environmental energy

Convex up/down

Low

Single unit Massive

High

Slow, reworked

++

++

(do)

(cor)

Very rapid

++

Encrusting

(cor) Ramose, robust Ramose, fragile Univalved shell

Multiple unit Bivalved shell, thick Bivalved shell, thin Multielement, tightly sutured Multielement, loosely articulated

+

+

+

++

(la)

(fr)

(fr)

(cor)

++

+

+

(la)

(fr)

(fr)

++

+

+

+

(la, d)

(do)

(cor)

(fr)

+

+

+

+

+

+

(la)

(do)

(fr)

(fr)

(fr, cor)

(da)

+

++

++

+

(la)

(do)

(da, fr)

(fr)

+

+

+

+

+

(la)

(da)

(da, fr)

(da, cor)

(da)

+

++

++

(la)

(da)

(da)

Utility as indicator of given condition: - not generally usable; + usable indicator; Type of indicator: cor = degree of corrosion; do = disorientation (overturning); fr articulation; la = long axis lineation; d = direction of apex .

References Allison, P.A. 1986 . Soft bodied animals in the fossil record: the role of decay in fragmentation during transport. Geology 14, 979 - 981 . Allison, P.A. 1988 . The role of anoxia in the decay and mineralization of proteinaceous macro-fossils . Paleo­ biology 14, 139 - 1 54 . Brett, c . E . & Baird, G . c . 1986. Comparative taphonomy : a key to paleoenvironmental interpretation based on fossil preservation. Palaios 1, 207- 227. Chave, K.E. 1964. Skeletal durability and preservation. In: J . Imbrie & N.D. Newell (eds) Approaches to paleoecology, pp. 377- 387. Wiley, New York. Meyer, D . L . & Meyer, K.B. 1986. Biostratinomy of Recent

++ =

very important indicator. fragmentation (or lack of); da

=

. . . disarticulation!

crinoids (Echinodermata) at Lizard Island, Great Barrier Reef, Australia. Palaios 1, 294- 302. Miiller, A.H. 1979 . Fossilization (taphonomy) . In: R.A. Robison & C . Teichert (eds) Treatise on invertebrate paleontology. A, fossilization (taphonomy), biogeography, and biostratigraphy, pp. A2 -A78 . Geological Society of America, Boulder, and Kansas University Press, Lawrence . Plotnick, R.E. 1986. Taphonomy of a modern shrimp : impli­ cations for the arthropod fossil record. Palaios 1, 286-293. Schafer, W. 1972. Ecology and paleoecology of marine environ­ ments . University of Chicago Press, Chicago . Seilacher, A. 1973. Biostratinomy: the sedimentology of bio­ logically standardized particles . In : R. Ginsburg (ed . ) Evolving concepts in sedimentology, pp. 159 - 1 77. Johns Hopkins University Studies in Geology No. 21 .

3 .4 Transport - Hydrodynamics

and unpredictable, chiefly because of the huge di­ versity of forms involved . The s h ell properties of greatest influence are : (1) the kind and degree of symmetry; (2) the degree of elongation; (3) the degree of shell curvature (brachiopods, bivalve mol­ luscs, ostracodes) or the apical angle (gastropods); (4) the character and distribution of ornament and the presence of teeth or processes along the hinge (brachiopods, bivalve molluscs); (5) the mean mass per unit shell area; and (6) the distribution of mass. Aside from fluid properties, the other factors con­ trolling behaviour are : (7) the agent transporting the shell (river, tidal stream, waves, turbidity cur­ rent); (8) the force exerted by the agent; (9) the nature of the bed on which the shell alights or over which it moves; and (10) the character and distribution of any other particles, either already deposited or moving with the shell . The ultimate response of the shell is to assume a characteristic attitude and orientation on the sedimentary surface; these properties, when summed over a sample of shells, constitute a biofabric (Kidwell et al. 1986; Section 3 . 5), which may be diagnostic of current direction and/or agency. Attitude, whether concave­ up or convex-up, is especially important in the analysis of transported brachiopod, bivalve mollusc, and ostracode valves . Introducing the pointing di­ rection afforded by an apex or umbone, shell orien­ tation may be measured with respect to either the axis of symmetry of the shell (gastropods, ortho­ cones, belemnite guards, crinoid columnals) or some convenient feature such as the line of elongation, the hinge, or a straight edge (brachiopods, bivalve molluscs, ostracodes) .

3 . 4 . 1 Shells J . R . L . ALLEN

Introduction

A consideration of the following as sedimentary particles exemplifies the range of behaviour of shelly hard parts: the shells of brachiopods, bi­ valve, gastropod and cephalopod molluscs (includ­ ing those with internal hard parts), ostracodes, and articulated crinoid columnals . All but the crinoids typically have hard, calcareous coverings marked by a low mass per unit surface area . The brachiopods are protected by two normally unequal but bilater­ ally symmetrical opposed valves, which may separ­ ate after death on the decay of muscle tissue . Equal but asymmetrical (about the umbone) valves typify the bivalve molluscs; separation depends on the decay of the ligament. In the brachiopods, and particularly in the bivalve molluscs, there may be teeth and other processes projecting from the hinge . Typically, gastropods have spiral shells with a wide range of apical angle and external ornament, which approximate to axial symmetry. External shells in the cephalopods are chambered and vary from straight (axially symmetrical) to more or less tightly coiled (bilaterally symmetrical) . The internal shell of the coleoid cephalopods varies from straight and axially symmetrical (e . g . belemnite guards) to flattened with bilateral symmetry (cuttlefish) . Ostracode valves are equal, but not symmetrical normal to the hinge . Articulated crinoid columnals are axially symmetrical and virtually cylindrical. Little is understood of the hydrodynamic behaviour of these hard parts, so abundant and ecologically important in modern shallow-water environments and the fossil record . Field studies are few (Nagle 1967; Salazar-Jimenez et al . 1982) and what laboratory experimental work exists (Kelling & Williams 1967; Brenchley & Newa1l 1970; Futterer 1978; AlIen 1984) seldom faithfully reflects natural conditions. There is a particular paucity of data on the behaviour of shells en masse. The hydrodynamic behaviour of shells is complex

Settling

Shells will eventually settle to the bed after having been either carried from shallow- to deep-water by turbidity currents or swept up into the water column by storm waves on a shelf. Laboratory experiments give some insight into the settling of bivalve mollusc valves . A terminal settling velocity i s reached when the upward drag acting on the sinking shell equals the downward-acting immersed weight. Valves of all studied species eventually fall concave-up (Fig. lA) .

227

228

3 Taphonomy

The centre of mass of the shell then lies below the centre of fluid force, there being no turning couple. Released convex-up, a turning couple at once ap­ pears because, in this attitude, the centre of action of the prevailing fluid forces underlies the centre of particle mass (Fig . lE). Valves with a length similar to the height sink steadily on a helical path, the shell spinning once about a vertical axis for each turn of the trajectory (Fig . IC) . The sense of the trajectory, either clockwise or anti-clockwise, varies with the species and whether the valve is on the left or the right. Valves with a length more than about 1 . 6 times the height settle unsteadily, the shell dis­ playing a regular oscillation (pitching), amongst other motions, while settling either spirally or irregularly (Fig. ID) . The drag coefficient of sinking mollusc valves is invariably substantially larger than for dynamically equivalent smooth spheres (i. e . those with the same

t

b

-

� A

S t a b l e fa l l

B

U o ,, , b / ,

\�) �

"

,

" ' I

�_

.

j)

",

, -- - - ....... ,

Some understanding of the complex process of transport in one-way currents has come from field observations and laboratory experiments, but much remains unknown, particularly concerning shells in bulk. In the case of dispersed bivalve molluscs, entrain­ ment depends on the orientation and particularly the attitude of the shell, and on the roughness of the

R I V E R S , T I DAL STREAM S

�

U n ste dY fa l l

(1 -bla )

'

(1 - bla ) � o . 4 _____ _ __

(,.'

.

F

Turritella

L

. .

.

.

"

•

:.

Cerastoderma

�

__

H

K

E

• •

0 .4

' '� __�

__ __ __ __ __ __

WAVE ( O S C I L LATO RY) C U R R E NTS

,

�

/' I

... 1

Steady fa l l

__ __ __ __ __ __ __

Transport in one-way currents (rivers, tidal streams)

_ < : � w� "..::.� . _ . . . . . . . . .. .. . . . . . . . � � (�::��: {� . �

�.

S ETTL I N G

balance of inertial and viscous forces) . Valves that settle unsteadily differ most from spherical particles, affording drag coefficients up to three times greater. Thus the 'quartz equivalents' (the size of a quartz grain or pebble with the same terminal settling velocity) of mollusc valves are much smaller than the valves themselves .

; :

: .. ! .: .. . : : :

. . .:

:: :

Mytilus

G

0

,

B e l e m n ite g u a r d s

Mytilus

S t r o n g c u rrents

Fig. 1 Schematic summary of the behaviour and idealized biofabrics of representative shells (bivalve molluscs (Cerastoderma, Mytilus); gastropods (Turritella); belemnite guards) when settling in water, and when transported and deposited from one-way and oscillatory currents .

3 . 4 Transport - Hydrodynamics bed relative to the scale of the valves . For planar beds of particles much smaller than the valves, convex-up shells require a larger fluid force for entrainment than concave-up ones, the force for convex-up entrainment varying from a few to many times greater, depending on shell shape and mass per unit area. Hence the drag coefficient of a convex­ up valve is smaller than for the same valve when concave-up; consequently a given valve is most streamlined when convex-up (Fig. lE). Once entrained, convex-up valves take a variety of orien­ tations depending on shell shape and the promi­ nence and distribution of teeth and other processes along the hinge, which can act like a storm anchor (Fig. IF, G) . Convex-up valves tend, without change of attitude, to glide over relatively immobile planar beds, but on mobile, sandy ones they may speedily become partly buried and thus halted. Concave-up valves entrained on relatively fine-grained planar beds also maintain their attitude while gliding over the bed but become tilted downcurrent. Overturn­ ing into the more resistant and stable convex-up position occurs only where the moving valve en­ counters a substantial obstacle on the bed. An ex­ ception is the stout-shelled Mytilus edulis, valves of which at once turn over when entrained from the concave-up attitude . As natural beds abound in obstacles, and concave-up valves are the least resist­ ant to entrainment, it is not surprising that the convex-up attitude is the norm for shells on river beds and beneath tidal currents . Bivalve mollusc shells appear to undergo frequent changes in attitude as they travel over ripples and dunes, which are bedforms much larger than them­ selves. A valve that is transported convex-up over the upstream side of the bedform is liable to over­ turn on being propelled into the sluggish wake to leeward, with the result that the shell could slide concave-up into and be buried in the trough . Because of their narrow conical form, high-spired gastropods become oriented with the apex upcur­ rent (Fig. IH) . Low-spired and coarsely ornamented forms assume a more random orientation. Cylindri­ cal shells (tentaculitids, orthocones, belemnite guards, articulated crinoid columnals) develop a variety of orientations beneath a current, depending on flow and bed conditions (Fig. 11, J) . Particles of this form tend to roll over the bed, and so develop a flow-transverse biofabric. The fabric changes in­ creasingly towards a flow-parallel one as the shells become rotated into the current direction on meeting obstacles, and as the amount of rotation increases with growing current strength.

229

Fig. 2 Mainly vertically packed and tightly nested shells of Macoma balthica forming a beach deposit in a laboratory wave tank.

Transport in oscillatory (wave) currents

Dispersed shells on smooth beds affected by wave swash and backwash behave much as in one-way flows . Wave action on concentrated bivalve shells forming beaches commonly results in a distinctive biofabric, the valves packing mainly vertically in nests and rosettes (Fig . 2) . In wave-affected shallows, however, where genuinely oscillatory currents exist, field and laboratory experiments point to a different mode of behaviour. The shells either glide (convex­ up if brachiopod or bivalve) or roll over the bed and become orientated so that the long dimension is in most cases parallel with the wave crests (Fig. IK, L) . The combination of oscillatory with steady (e . g . tidal) currents creates more complicated patterns which are as yet little understood. Biofabrics due to organic activity

Some instances of a concave-up attitude assumed by disarticulated bivalve and brachiopod shells found in shallow-marine deposits are with little doubt a consequence of the reworking of the shelly sediment by scavenging organisms, but it is not known how exactly the biofabric arises . Shells dis­ turbed by organisms should possess a random orientation, in contrast to concave-up shells that have settled on the bed in the presence of a current strong enough to swing the particles . References AlIen, J . R. L . 1984 . Experiments on the settling, overturning

230

3 Taphonomy

and entrainment of bivalve shells and related models. Sedimentology 31 , 227-250. Brenchley, P.J. & Newall, G . 1970 . Flume experiments on the orientation and transport of models and shells. Palaeogeography, Palaeoclimatology, Palaeoecology 7,

palaeoclimatology (e .g. Spicer 1981, 1989; Ferguson 1985; Spicer & Greer 1986; Spicer & Wolfe 1987) . Attention here is focused on potential megafossils of terrestrial plants .

1 85 - 220 .

Futterer, E. 1978. Untersuchiingen iiber die Sink- und Transport-geschwindigkeit biogener Hartteile. Neues Jahrbuch fUr Geologie und Paliiontologie, Abhandlungen 155, 318-359 .

Kelling, G . K . & Williams, P.F. 1967. Flume studies of the reorientation of pebbles and shells. Journal of Geology 75, 243- 267.

Kidwell, S . M . , Fiirsich, F.T. & Aigner, T . 1986 . Conceptual framework for the analysis and classification of fossil concentrations . Palaios 1, 228- 238 . Nagle, J . S . 1967. Wave and current orientation of shells . Journal of Sedimentary Petrology 37, 1124 - 1 138. Salazar-Jimenez, A., Frey, R.W. & Howard, J.D. 1982. Concavity orientations of bivalve shells in estuarine and nearshore shelf sediments, Georgia . Journal of Sedimentary Petrology 52, 565 - 586 .

3 . 4 . 2 Plant Material R . A . S PI C E R

Introduction

Allochthonous plant fossil assemblages usually re­ present variously degraded fragmented parts of dif­ ferent individuals and species that lived at varying distances from their ultimate site of deposition and burial. Individual plants are composed of, and pro­ duce, an indeterminate number of organs. Whole plants are almost never found in the fossil record, so palaeobotanical systematics has to handle isolated organs (Spicer & Thomas 1986; Section 5 . 1 .3) that have greatly differing potentials for transport, deposition, and preservation. The interaction of a detached plant organ (or organ fragment) with a fluid medium is governed by its density in relation to that of the fluid medium, together with its shape, size, and surface character­ istics . The transportability of a plant part is largely a function of its terminal fall velocity. Many plant parts are flexible planar objects containing air spaces (e . g . leaves) and their hydrodynamic properties are difficult to model theoretically . Empirical ap­ proaches have proved more successful. Leaves have received most taphonomic attention because of their abundance and utility in biostratigraphy and

Organ dispersal by wind

Aerial transport determines what organ sample a river or lake, for example, receives and therefore 'sees' of the surrounding vegetation . Factors affect­ ing fall velocity in still air include :

Leaf weight. Weight per unit area at abscision is the most critical intrinsic property of a leaf that affects 'flight' and ground dispersal (Spicer 1981 ; Ferguson 1985) . Evergreen taxa typically are heavier and have higher settling velocities .

Leaf shape. Leaf shape has a n effect on fall velocity but shapes with major axes of markedly different length (long and narrow) tend to rotate about the longer axis; such behaviour slightly increases fall time and therefore the chance of greater dispersion from the source (Ferguson 1985) . Leaf size. Although not obviously correlated with fall rate, leaf size affects movement through the branch and trunk space within a forest. Large leaves tend to encounter static obstacles more frequently than small leaves, and any such event either traps the leaf directly, or affects its fall rate . Ferguson (1985) noted a weak positive correlation between leaf size and weight per unit area. Such a correlation would tend to favour the transport of smaller leaves. However, while this may be true for a tree crown as a whole, 'sun' leaves at the top of a tree tend to be smaller but have a higher weight per unit area. Long-distance dispersal of these leaves (and result­ ing preservational bias) is a function of their ex­ posure to high wind energies and their initial height from the ground (Spicer 1981 ) .

Petiole effects. The petiole rarely exceeds 20% of total leaf weight, and even large petioles have negli­ gible effect on fall rate .

Dispersion resulting from air fall. Aerial dispersal of leaves away from a source follows a negative ex­ ponential model (Rau 1976; Spicer 1981) . Rau, in a study of litter deposition in an open lake, used the following equation: Zx = Zr exp

(-k[r-x]),

3 . 4 Transport

Hydrodynamics

where x distance from the lake centre, Zx deposition occurring at distance x, r distance from the lake centre to the shoreline, Zr deposition 1 at the shoreline, and k = - ( r - x ) - l In(Zx Zr- ) . Under some circumstances estimates o f ancient litter productivity may be obtained from the fossil record . =

Water absorption by leaves is governed by the thickness of cuticle and of epicuticular wax, abun­ dance of stomata and/or hydathodes, lamina or petiole damage, water temperature and chemistry, and to a lesser extent by leaf anatomy (Spicer 1981; Ferguson 1985) . Floating times range from several hours to several weeks (Fig. 1); thin chartaceous leaves tend to sink earlier than thick coriaceous leaves (Spicer 1981; Ferguson 1985) . Intact com­ pound leaves float longer than their individual leaflets . Dispersed fruits and seeds (diaspores) exhibit a greater range of floating times than do leaves (e . g . Collinson 1983) . Floating times d o not appear t o be directly related to diaspore size or to the habit of the parent plant (Collinson 1983) . Wood, and in particular logs, can remain afloat for several years and potentially therefore the only hindrance to log dispersal downstream from the growing site is water (channel) depth and in stream obstacles .

=

=

=

Post-descent dispersion over the ground. Leaves blown along the ground are distributed laterally by a com­ bination of saltation and rolling. In laboratory ex­ periments the greatest dispersion was found with circular shapes that tended to roll (Spicer 1981) . Dispersion is little affected by leaf size but, as with fall velocities, weight per unit area did prove impor­ tant with light leaves travelling the furthest. Dry curled leaves are easily transported, but wet leaves tend to stick together and suffer minimal ground dispersion. Field experiments (Ferguson 1985) show that most woodland leaves are never disseminated very far from the parent trees (although dispersion is greater in open sites) and that, barring flooding or volcanic activity, even the tallest temperate trees must be growing within 50 m of a body of water to stand any chance of becoming fossilized .

Transport in the water column . Progressive plant tissue saturation follows an 's' shaped curve and does not cease until long after the object has sunk (Greer, unpublished data) . Progressive post-sinking water absorption continues to affect the submerged density of the object, and therefore its behaviour during transport in the water column, until full saturation is reached . When and where the object eventually settles is determined mostly by sub­ merged density and shape, two factors important in determining settling velocity and entrainment be­ haviour. In an aquatic environment plant debris is degraded by biological and mechanical agents, both of which affect the hydrodynamic properties but which produce characteristic degradation patterns seen in fossil material (Spicer 1981, 1989; Ferguson 1985) .

Water transport

Floating. Immediately upon landing on water, plant material absorbs water and soluble substances begin to be leached out. Initially a dry leaf will float and may remain buoyed up by surface tension for sev­ eral weeks, provided that only the bottom surface of the leaf is wetted and the water surface is calm (Spicer 1981) . Plant material could be transported long distances this way, but such conditions are likely to pertain only in slow-flowing rivers pro­ tected from wind (i . e . subcanopy streams), situ­ ations in which long-distance transport is unlikely to occur. 1 00 00 c

Fig. 1 Floating times for freshly abscised leaves of coriaceous evergreen Rhododendron, chartaceous deciduous Fagus sylvatica (abscised dry and brown), and Alnus glutinosa (abscised moist and green) . The leaves were placed in mildly agitated water at room temperature. (Data from Spicer 1981 . )

;;

o ;;:: '" Cl)

� .!!! C Q)

�

0...

.- -

,

75 50

"

,

- - - - Rhododendron s p . "

. . Fagus sylvatica

"

_ .

"

25

o

2

231

4

6

"

"

"-

"-

8

'-

- . - . - Alnus glu tinosa

.....

10

Days

12

14

16

- - -

18

20

232

3 Taphonomy

Settling (fall) velocity in water. In spite of their irregular two-dimensional shape, angiosperm leaves exhibit within-taxon uniformity of settling velocities, as do the more prismatic shapes of coni­ fer needles (Spicer & Greer 1986) . Even irregularly shaped fern pinnules and moss leafy shoots have settling velocities that fall within narrow, moder­ ately well defined, limits . Statistically there is no significant difference between the settling velocities of different broad-leaved taxa (including Ginkgo and fern pinnules), but significant differences do exist between conifer needles and broad-leaved taxa, and between individual conifer taxa (Greer, unpublished data) . In general, conifer needles have a higher settling velocity (e . g . 3 . 03 cm/s for Picea pungens at full saturation) than angiosperm leaves (e . g . 1 . 5 cm/s for Fagus sylvatica at full saturation) . Individual leaves of other broad-lamina taxa such as Ginkgo biloba, however, exhibit fall velocities as high as 6 . 7 crn/s when petiole and lamina configur­ ation produce a hydrodynamically efficient shape that results in a stable gliding fall. Hydraulic sorting, primarily related to settling velocities, has been observed in both low and high energy fluviolacus­ trine delta systems and modelled in relation to spatial and temporal pattern in the source vegetation (Spicer 1981 ; Spicer & Wolfe 1987) . Entrainment. For any given flow, particles concen­ trated near the stream bed are mostly those with the greatest settling velocity. The heaviest particles are transported as bedload and are only in suspension for brief periods of time . As current flow wanes, the lighter fractions progressively settle out. Con­ versely, increases in current flow progressively en­ train material. Flow rate in natural streams and rivers is rarely constant and plant debris is likely to undergo several cycles of deposition and entrain­ ment before permanent burial takes place . Leaf aspect and orientation to fluid flow influence entrainment. Curved leaves, or planar particles in­ clined with their raised edge facing into the flow, are entrained at lower flow rates than those lying flat on the stream bed or inclined with their raised edge pointing downstream . Bed roughness, including bedforms, affects plant particle entrainment (Spicer & Greer 1986) . If bed­ forms (e . g . ripples) are large enough for the plant particles to settle between, the particles are protected from entrainment and often buried rapidly by bed­ form migration . Larger particles pass through the system. Thus, if ripples are noted in a fossil deposit, and only conifer needles are preserved, it cannot be

assumed that angiosperms were not present (even in large numbers) within the source vegetation: they may have been deposited elsewhere because they were too large to be trapped between the ripples . References Collinson, M . E . 1983. Accumulations of fruits and seeds in three small sedimentary environments in southern England and their palaeoecological implications. Annals of Botany 52, 583 - 592. Ferguson, D.K. 1985 . The origin of leaf assemblages new light on an old problem. Review of Paleobotany and Palynology 46, 1 17- 144. Rau, G . H . 1976 . Dispersal of terrestrial plant litter into a subalpine lake . Oikos 27, 153- 160. Spicer, R.A. 1981 . The sorting and deposition of allochthonous plant material in a modern environment at Silwood Lake, Silwood Park, Berkshire, England . US Geological Survey Professional Paper No. 1 143. Spicer, R.A. 1989 . The formation and interpretation of plant fossil assemblages . Advances in Botanical Research 1 6 ,

95- 19 1 .

Spicer, R . A . & Greer, A . G . 1986. Plant taphonomy i n fluvial and lacustrine systems. In: T.W. Broadhead (ed . ) Land plants . pp. 27-44. University of Tennessee Department of Geological Sciences Studies in Geology No. 15. Spicer, R.A. & Thomas, B . A . (eds) 1986. Systematic and taxo­ nomic approaches in palaeobotany. Systematics Association Special Volume 31 . Oxford University Press, Oxford. Spicer, R.A. & Wolfe, ].A. 1987. Plant taphonomy of late Holocene deposits in Trinity (Clair Engle) Lake, northern California. Paleobiology 13, 227-245.

3 . 4 . 3 B ones A . K . BEHRENSMEYER

Introduction

After death, vertebrate skeletons interact with bio­ logical, physical, and chemical processes at or near the Earth's surface . These processes determine whether the bones are destroyed (i. e . recycled) or fossilized. Transport is one of the important pro­ cesses that can affect bones prior to fossilization. Both physical and biological mechanisms of trans­ port may alter life associations of organisms by carrying bones away from the original environment and by mixing taxa from different habitats and time

3 . 4 Transport periods . Such processes also cause abrasion and other types of damage, as well as sorting and differ­ ential preservation of body parts (see also Section 3.3). Biological transport

Biological mechanisms of transport include pred­ ators and scavengers that derive some benefit from collecting bones. One notable modern bone collec­ tor is the African hyaena (Crocuta crocuta); other mammals such as canids, felids, humans, elephants, porcupines, and pack rats also transport bones or parts of carcasses (Shipman 1981) . Predatory birds carry off carcasses and leave accumulations of bones of small vertebrates as regurgitated pellets or debris below a favoured perch . Various small mammalian carnivores leave concentrations of bones in their faeces . Harvester ants (Messor barbarus) also collect bones of small vertebrates and transport them un­ derground (Shipman 1981) . Trampling causes bones to move outward from disintegrating carcasses (Hill 1979) . Fossil accumulations in preserved burrows and cave deposits attest to the bone-transporting activi­ ties of ancient species . It is likely that the fossil record includes examples of biological bone­ transporting processes for which there are no mod­ ern analogues . The Phanerozoic history of such processes and their effect on the vertebrate fossil record is not yet known . Physical transport

Physical processes causing bone transport include water currents and wave action, wind, and gravity. Unfossilized bones are relatively light, with high surface area to volume ratios and irregular shapes, all of which make them readily transportable by moving water. Although bone mineral (hydroxy­ apatite) has a density of about 3 . 2, bones themselves have densities varying from less than 1 .0 (i. e . they float) to 1 . 7 (Behren smeyer 1975) . This is because pore spaces and organic components make up a significant percentage of a fresh bone . Pores may retain air or other gases and keep bones relatively buoyant for hours to days (Behrensmeyer 1975) . Weathered bones that have lost their organic ma­ terial and become cracked are less buoyant. Currents of 20- 30 cm/s can move bones of small to medium­ size vertebrates (e .g. rodent to sheep) but stronger currents are required for bones of larger animals (e . g . cow, elephant) . Teeth have densities ap-

Hydrodynamics

233

proaching 2.0 and almost always require stronger currents for transport than do bones, regardless of the size of the animal. Experiments in natural rivers with flood velocities of 1 .0 m/s demonstrate that bones can be transported a kilometre or more in a single year. The hydrodynamic behaviour of bones can be predicted to some extent by considering them as sedimentary particles and calculating their 'quartz­ equivalents' . This is the size of a quartz grain with a settling velocity equal to that of the bone, and it can be calculated based on measurements of actual set­ tling velocities of bones in water (Behrensmeyer 1975; Shipman 1981) . Bones and teeth of approxi­ mately equal sizes can have very different quartz­ equivalents (Fig. 1 ) . Those with smaller equivalent quartz grains are generally more transportable, although shape and orientation to a current can cause exceptions to this rule . Scapulae are small and light relative to other bones in a skeleton, but their shape is also streamlined so that they are less easily moved than an equivalent quartz sphere . In fossil deposits, the difference in grain size of matrix sediment and bone has been used to assess transport history (Shipman 1981) . If bones are pre­ served with grains of near-equal quartz diameters, this is interpreted as an indication that the bones were transported and hydraulically sorted. In con­ trast, if bones are associated with sediment of much finer quartz-equivalents, then minimal transport is inferred . Since the relationship between transport and quartz-equivalents can be influenced by in­ dividual bone shapes, considerable caution is necessary in such interpretations . Moreover, the grain sizes that are available for transport, rather than hydraulic sorting, can control which quartz­ equivalents are associated with bones at the time of burial.

Sorting. Differing hydrodynamic behaviour of bones in a single skeleton results in sorting (separation of body parts according to transport rates) and win­ nowing (removal of the lighter elements, leaving a 'lag' of the heavier, less transportable bones) . Experi­ ments in flumes and natural rivers have demon­ strated that there are three distinct transport groups ('Voorhies Groups') for medium to large mammals (in order of decreasing mobility) : Group I vertebrae, ribs, sternum; Group 11 - limb parts; Group III - skulls, mandibles, teeth (Voorhies 1969; Behrensmeyer 1975; Shipman 1981) . Bones from a single point source (i. e . a skeleton) show progressive sorting with continued current action. Distinct pat-

3 Taphonomy

234 26.9

mm

8.6 mm

14.8

mm

2.6 mm

3.1

mm

S h eep m o l a r Ast raga l u s m ed i u m artiodactyl

( Oamaliscus)

Verteb ral cent r u m g i a n t forest h o g

( Hylochoerus)

H o rse m o l a r

terns of sorting in a fossil deposit thus indicate the interaction of currents with a localized bone source . Input from multiple sources of bones along natural rivers or beaches obscures patterns of sorting for individual carcasses . If there are many point sources of bones, all body parts (from different individuals) can co-occur in deposits formed by water currents (Hanson 1980) . Bones of small vertebrates also can be moved by wind action on beaches, dune fields, and ephemeral river beds. Dust-devils are effective mechanisms for transporting and scattering small bones in arid en­ vironments . Gravity assists in bone movements on steep slopes, as in caves and sinkholes . The low density of bones helps to keep them near the surface and exposed to slope wash and mass move­ ments of sediment.

Orientation . The orientations of individual bones indicate the influence of hydraulic processes on the bone assemblage (see also Section 3.4. 1 ) . In strong currents, elongate bones generally orient parallel to the flow direction, with the heavier end upstream . In shallow water or in weak currents, such bones may orient perpendicular to the current and roll downstream around their long axis (Voorhies 1969; Behrensmeyer 1975; Shipman 1981) . Determining flow direction from bone orientations must take both of these patterns into account. Large bones may act as 'traps' for smaller ones, which accumulate against the upstream side or in the downstream zone of flow separation (shadow) . The orientation of the larger bones may influence those of associated smaller ones.

D e r m a l s c u te

(Crocodylus)

Fig. 1 The hydraulic equivalents of examples of Recent bones, as determined by their settling velocities and calculations of the diameter of a quartz sphere that would settle at the same velocity. Bones and quartz grains are drawn to the correct relative sizes. (After Behrensmeyer 1975; reproduced with permission from the Museum of Comparative Zoology, Harvard University. )

Concentration . Dense concentrations of fossil bones in river deposits are often attributed to hydraulic processes. However, experiments in modern rivers indicate that fresh bones are generally too light to form permanent patches or 'bone bars' unless there is a point source nearby and/or an obstruction that causes transport to cease abruptly . Repeated win­ nowing and reworking can concentrate denser elements (e.g. teeth) as part of the gravel 'lag' deposit in fluvial sediments . Previously fossilized bones with higher densities may be incorporated into such lags, mixing remains from periods of 100 10 000 or more years to create a time-averaged sample of the original vertebrate communities (Behrensmeyer 1984) . Abrasion . Hydraulic transport can abrade and break bones, but experiments in tumbling machines and in natural rivers indicate that it takes considerable bone - sediment interaction to cause significant damage to fresh bones and teeth (see also Section 3 . 1 ) . Weathered bones are more vulnerable to abra­ sion and breakage in transport situations . Fresh experimental bones can travel over 3 km in a sand­ and gravel-bed river with only minor loss of surface detail due to abrasion. Exposure to poorly sorted sand for up to 35 h in a tumbling machine is necess­ ary to produce gross changes in fresh bones. Fossil dinosaur and crocodile teeth subjected to the equivalent of 360 -480 km of transport in a tumbling machine with coarse sand showed negligible dam­ age to enamel surfaces (Argast et al. 1987) . Thus, bones and teeth can experience considerable trans-

3 . 5 Fossil Concentrations

235

port without showing significant abrasion. Con­ versely, stationary bones may be 'sand-blasted' by water or wind-borne sediment and heavily abraded without significant transport. Thus, it is very dif­ ficult to judge the transport history of a bone from its appearance .

thus contain evidence of vertebrate palaeoecology for areas and time periods that often are not directly comparable with samples of vertebrate communi­ ties from modern ecosystems.

Burial. The same physical processes that transport bones also bury them. In channels and on beaches, bones are buried and exhumed many times prior to destruction or fossilization as they move along with bedload sediment. They may be overtaken by mov­ ing bedforms (ripples, sand waves), and scour on their downstream sides also promotes burial. Per­ manent burial happens when the bone is removed from the active zone of sediment transport; this can occur when the channel abandons its course or when unusual flood conditions alter its bottom morphology . The taphonomic history of bones should be ana­ lysed prior to ecological interpretations of species represented in a vertebrate fossil assemblage be­ cause they are readily transported by both physical and biological processes . The tendency for bones to be buried and reworked repeatedly in fluvial and shoreline environments also implies that trans­ ported remains may represent a substantial amount of time-averaging. Transported bone assemblages

Argast, 5 . , Farlow, J . O . , Gabet, R.M. & Brinkman, D . L . 1987. Transport-induced abrasion of fossil reptilian teeth : impli­ cations for the existence of Tertiary dinosaurs in the Hell Creek Formation, Montana . Geology 15 , 927-930. Behrensmeyer, A.K. 1975 . The taphonomy and palaeoecology of Plio-Pleistocene vertebrate assemblages east of Lake Rudolf, Kenya . Bulletin of the Museum of Comparative Zoology 146, 473- 578 . Behrensmeyer, A.K. 1984. Taphonomy and the fossil record . American Scientist 72, 558 - 566. Hanson, C.B. 1980 . Fluvial taphonomic processes: models and experiments . In: A. Behrensmeyer & A. Hill (eds) Fossils in the making, pp. 156 - 1 8 1 . University of Chicago Press, Chicago. Hill, A. 1979 . Disarticulation and scattering of mammal skel­ etons . Paleobiology 5, 261 -274 . Shipman, P. 1981 . Life history of a fossil. Harvard University Press, Cambridge. Voorhies, M.R. 1969 . Taphonomy and population dynamics of an early Pliocene vertebrate fauna, Knox County, Nebraska. University of Wyoming, Contributions to Geology

References

1,

1 - 69.

3 . 5 Fossil Concentrations and Life and Death Assemblages F . T . FURSICH

Fossil concentrations

A fossil concentration is defined as any relatively dense accumulation of fossils, irrespective of taxo­ nomic compOSition, state of preservation, or degree of post mortem modification. Fossil concentrations are nearly exclusively accumulations of hard parts. They are therefore regarded here as synonymous with skeletal concentrations of fossil organisms (Kidwell et al. 1986) . As the size of the biogenic hard parts is not restricted, this definition includes dinosaur bone beds as well as coral reefs, shell beds of bivalves and those of ostracodes (see also Section 3.4) .

Fossil concentrations also include several types of Fossil-Lagerstatten, especially those formed by rapid burial or by condensation (Section 3 . 6) .

Descriptive classification. Fossil concentrations can be described in several ways, stressing either taxo­ nomic composition, bioclastic fabric (degree of packing), geometry, or the internal structure of the deposit (Fig . 1 ) . Each of these aspects carries some genetic significance . The taxonomic composition largely depends on the ecology of the component taxa and, to a lesser degree, on the hydrodynamics of their accumulation . The biofabric, that is the three­ dimensional arrangement of skeletal elements in

236

3 Taphonomy

TAXO N O M I C COMPOS I T I O N

G E O M ETRY

PAC K I N G _

m o n otyp i c

-: matrix s u p po rted

po lyty p i c

b i oclast - s u p p o rted

-

-

stringer pave m e n t pod clump lens wedge bed

I NT E R N A L STRUCTU R E S I MPLE C O M P L EX

CJ -

CJ -

Fig. 1 Major features used in the descriptive classification of fossil concentrations and their genetic significance. Shaded box predominant; white box rare. (After Kidwell et al. 1986 . )

1-----+--4L .--t--JJ-----I--.------I ECOLOGY

HYDRODY­ NAM I CS

TOPOG RA P H Y

the matrix, includes skeletal orientation, degree of packing, and sorting by size and shape . The bio­ fabric is strongly influenced by hydrodynamic con­ ditions, whilst ecology and compaction may be additional factors . The geometry of a fossil concen­ tration depends on the pre-existing topography of the depositional surface (e .g. burrow fills), the ecol­ ogy of the hard part producers (e .g. clumps of mussels) and other organisms (e .g. shell-lined bur­ rows), and on the hydrodynamic conditions which, at the time of hard part concentration, produce a topography (e . g . by the migration of ripples, exca­ vation of scours, etc . ) . The internal structure provides information on the ecological and hydrodynamic history of the deposit. Simple (i . e . internally homo­ geneous or, at the most, with unidirectional trends) and complex skeletal concentrations can be dis­ tinguished . In the latter, features such as grain size, degree of articulation, and orientation vary in a complicated pattern .

Genetic classification . Fossil concentrations can also be classified genetically, based on the main concen­ trating processes . The formation of concentrations is governed by the interplay of net rate of sedi­ mentation, net rate of production of biogenic hard parts, and to a lesser extent diagenetic processes . Ac­ cordingly, biogenic, sedimentological, and diagen­ etic concentrations can be distinguished (Fig. 2) . Biogenic concentrations result from the gregarious behaviour of organisms with hard parts (e .g. mussel beds) or of organisms which concentrate skeletal elements (e . g . during the feeding process) . Sedimen­ tological concentrations are produced by hydraulic processes which may represent short-term events (e . g . storms) or long-term processes (e . g . back­ ground current and wave action) . Examples include shelly storm lags or condensed shell beds . Diagenetic concentrations are the result of post-burial physical

ECO LOG I CAL & H Y D RODYNAM I C H I STORY

=

=

. .

0· · · · .

. . . . ·.·...

·a.Q '

.

.

.

.

'

.

B i ogen i C

Sed i m entological

D i agenetic

Fig. 2 Genetic types of fossil concentrations based on biogenic, sedimentological, and diagenetic processes . White area of the triangle represents concentrations of mixed origin. The longer the time-span involved in the formation of a concentration, the more likely it will be mixed in origin. (After Kidwell et al. 1986 . )

or chemical processes, including compaction as well as selective pressure solution of matrix in bioclastic limestones . These processes act in most fossiliferous sediments, but are rarely as significant as biogenic or sedimentological processes . Most fossil concentrations are formed by more than one process. For example, a storm-reworked mussel bed is of mixed biogenic and se dim en to­ logical origin; a strongly compacted layer of bivalves killed by drastic changes of salinity represents a diagenetically enhanced biogenic concentration.

237

3 . 5 Fossil Concentrations Of particular importance for the formation of fossil concentrations is the combination of low net rates of sedimentation with high net rates of bio­ genic hard part production. Zero net rates of sedi­ mentation (i. e . omission) and negative net rates (corresponding to erosion) result in different types of fossil concentrations which frequently exhibit sharp lower or upper contacts (Kidwell 1986) . When subsequently undisturbed, such concentrations and their bed contacts can be interpreted very pre­ cisely. In reality, however, the contacts are com­ monly modified by burrowing organisms and/or diagenesis .

Geological and palaeontological significance. Fossil concentrations are a useful tool in basin analysis, furnishing information on bathymetry, rate of sedi­ mentation, hydrodynamic regime, and environmen­ tal gradients . The prevalence of particular types of concentration, such as those produced by storms, and their frequency through time allow inferences about basin configuration and evolution. Along onshore - offshore gradients for example, sedimen­ tological concentrations which dominate in shallow, nearshore environments are gradually replaced by biological concentrations in deeper shelf areas . Among sedimentological concentrations those exhibiting wave influence are most prominent in very shallow water, those of storm origin in shallow to intermediate shelf depths, whilst in lower shelf regions sediment starvation and condensation are the governing factors . Biostratinomic features of the skeletal elements such as biofabric, articulation, sorting, fragmen­ tation, abrasion, bioerosion, and encrustation pro­ vide additional data on residence time on the sea floor, wave versus current influence, degree of re­ working, and sediment starvation (see also Section 3.3). The palaeontological significance of fossil concen­ trations varies greatly depending on their genesis (see below) .

Life and death assemblages

Definitions. The term assemblage has several mean­ ings (Fagerstrom 1964) . According to some authors, assemblages consist of organisms derived from more than one community (i . e . they exhibit signs of mixing) . In another, broader definition adopted here, the term refers to any group of organisms

from a geographical locality. A life assemblage accord­ ingly is defined as any group of living organisms from a geographical locality . It may represent a whole community or only parts of it. A death assem­ blage ( thanatocoenosis) consists of the preserved elements of a life assemblage after its death and decay. As a rule, soft-bodied organisms are no longer represented in a thanatocoenosis . The species diversity of a thanatocoenosis is therefore much lower than that of the life assemblage . The trophic composition of a living community is commonly not adequately reflected in the thanatocoenosis . The term taphocoenosis ( allochthonous thanatocoenosis of some authors) refers to hard parts of organ­ isms which became embedded together after having been subject to biostratinomic processes such as sorting, admixture of shells from adjacent habitats or from stratigraphically older units, mixture of skeletal elements resulting from time-averaging, or selective destruction by physical, chemical, or bio­ logical agents . A fossil assemblage differs from a taphocoenosis in that post-burial diagenetic pro­ cesses have been operating which led to lithification or partial destruction of the hard parts . In cases where biostratinomic and diagenetic pro­ cesses did not play a significant role, the fossil assemblage will be roughly identical to the thanato­ coenosis . The terms taphocoenosis and thanatoco­ enosis are often used as synonyms . In the definition given here they do, however, characterize different stages of the transition of a life assemblage to a fossil assemblage (Fig. 3) . =

=

Time averaging. The quality of a taphocoenosis (and thus often of the resulting fossil assemblage) largely depends on the time factor. Rapid in situ burial of a life assemblage (e .g. during storms) may produce a taphocoenosis which faithfully records species composition and age structure of the organisms with hard parts . Under normal circumstances, how­ ever, taphocoenoses represent time-averaged relics of life assemblages . Time-averaging refers t o the mixing o f skeletal elements of non-contemporaneous populations or communities . This telescoping of biological data representing tens, hundreds, or even thousands of years into a single geological time plane drastically alters the quality of the data. Short-term fluctuations in species composition and in the morphological range of species, reflecting variations in salinity, oxygen, or other environmental factors, cannot be recognized from time-averaged assemblages . The occasional dominance of opportunistic species in a

3 Taphonomy

238 L I F E AS S E M B LAG E

THANATO C O E N O S I S

TA P H O C O E N O S I S

C o m m u n i ty

I I



�

(!)

::

;:

'"

C§'"

o

�

o '0 (!) u c (!)

� >

�

Cl.

c (!)

u

'" .D

'0

(!) u c (!)

2

Cl.

.

�(!)

3

C

(!) Cl.

c

o

�

� >

�

5 4 3 2

Cl.

(!)

::

c (!) Cl.

o

5

5

4

4

3

(!) u c (!)

� >

�

2

::

Cl.

c (!)

::

32

(!) Cl.

(!) Cl.

o

�"L

Pseudopathology

The study of palaeopathology requires a detailed knowledge of both pathology and the processes of fossilization . One of the major problems is in iden­ tifying specific pathologies and distinguishing them from pseudopathologies . The action of boring bi­ valves can simulate dental abscesses; burial in an acidic environment can result in the surface erosion of bone, which can mimic periostitis or even osteo­ myelitis . Even the excavation of fossil bones can produce pseudopathologies : depressed fracturing

Fig. 4 Distribution through time of disease and trauma in Quaternary elephants . (From Bricknell 1987.)

of skull bones may result simply from pressure on the overlying sediments, while digging implements can produce pseudohunting injuries . Erosion of the bone surface can often be produced by rootlets . Saprophytic fungi attack bone, and they are known from the time of the first bone preserved in the fossil record . In general, however, a close exami­ nation of the details of surface structure and also in cross-section (as a thin section under the light microscope) will determine the authenticity or otherwise of the supposed pathologies .

385

4 . 1 6 Trophic Structure References Baker, J . & Brothwell, D. 1980 . Animal diseases in archaeology. Academic Press, London and New York. Bricknell, I. 1987. Palaeopathology of Pleistocene proboscid­ eans in Britain. Modern Geology 11, 295 -309 . Brothwell, D. & Sanderson, J.T. 1964. Diseases in antiquity. Thomas Publications, London . Halstead, L . B . 1974. Vertebrate hard tissues . Wykeham Publications, London .

Halstead, L.B. & Middleton, J. 1972. Bare bones: an exploration in art and science. Oliver & Boyd, Edinburgh. Moodie, R. 1917. General considerations of the evidence for pathological conditions found among fossil animals. Science 4 3 , 425 -452. Tarlo, L . B . 1959 . Stretosaurus gen . nov . , a giant pliosaur from the Kimeridge Clay . Palaeontology 2, 39 -55. Wells, C . 1964. Bones, bodies and disease. Thames & Hudson, London.

4 . 16 Trophic Structure J . A . CRAME

One o f the most exciting breakthroughs in com­ munity palaeoecology in recent years was the dis­ covery that fossil assemblages can be classified according to the feeding characteristics of their con­ stituent species . This resulted in an entirely new way of comparing and contrasting palaeocommuni­ ties . The trophic structure of a community can be defined as the cumulative feeding habits of its com­ ponent species . These feeding habits are in turn based on two fundamentally different food chains: a grazing one centred on green plants and a detritus one centred on dead organic matter. Both these chains are terminated by predators . It is important to emphasize at the outset the distinction between feeding habit and trophic level. Whereas the former relates to what an organism eats, the latter refers to its position in the steps of energy transfer (Scott 1976) . Each species occupies a specific position (or positions) in a food web (Fig. 1 ) . Marine biologists had, o f course, been classifying feeding mechanisms for many years . However, their schemes were based largely on features such as food particle size, and little attention was paid to precisely what was eaten or where . These were just the sorts of details that were of interest to the community palaeoecologist, and, when added to existing schemes, produced a number of basic feed­ ing groups (or trophic categories) . Several simplified classification schemes for benthic marine inver­ tebrates are now in existence (Table 1) (Walker & Bambach 1974) .

When fossil communities were analysed using these new schemes, it became apparent that the vast majority of species fell into just three basic categories : suspension feeders, detritus feeders, and predators (Table 2) . In fact, so striking was this regular tripartite division that it was suggested that these categories could form the end members of a triangular (or ternary) diagram and be the basis of a rapid system for classifying palaeocommunities Table 1 A simplified classification of marine invertebrate trophic groups . (Adapted from data given by Walker & Bambach 1974.)

Group

Feeding habits

1

Suspension feeders

Higher level (epifaunal or infaunal)

2

Suspension feeders

Lower level (epifaunal or infaunal)

3

Deposit feeders

Sediment-water interface (epifaunal)

4

Deposit feeders

Shallow, in-sediment (infaunal)

5

Deposit feeders

Deep, in-sediment (infaunal)

6

Browsers

(Epifaunal)

7

Predators

(Passive or active; epifaunal or infaunal)

8

Scavengers

(Epifaunal or infaunal)

9

Parasites

386

/ }

4 Palaeoecology

P R E DATO R S

CONSUMfRS s " ,pe",o feed ers

F i l te r feed ers

B mw,",.,

1

PRO D U C E RS Table 2

RECUP ERA TORS Det r i t u s fe ed e rs Depos i t feed ers

B ro w s e r s

Scave ngers

�

DET R I T U S

The relationship between feeding habit and trophic level . Within each of the basic food chains ( a consumer one based on green plants and a recuperator one based on dead organic matter), species of various feeding habits can be arranged into trophic levels at each stage of energy transfer. Heavy arrows energy-flow pathways; light arrows donors of organic detritus . (From Scott 1976.) Fig. 1

=

=

The principal marine invertebrate trophic groups (Data from Walker & Bambach 1974, and Scott 1976 . )

Group

Examples

Food and feeding method

1

Suspension feeders

Food - small particles such as phytoplankton and zooplankton; dissolved and colloidal organic molecules; resuspended organic detritus. Feed b y - flagellae, ciliated lophophores, ctenidia and tentacles .

Sponges, anthozoans, hydrozoans, stromatoporoids, bryozoans, brachiopods, many bivalves, some gastropods, some annelids and crustaceans, pelmatozoans and graptoloids .

2

Deposit feeders*

Swallow or scrape particulate organic detritus, living and dead smaller members of benthic flora and fauna, and organic-rich grains.

Some crustaceans, echinoids, ophiuroids, bivalves, gastropods and annelids; scaphopods, holothurians .

3

Predators

Either active search and seizure (involving swallowing whole, biting and chewing or external digestion) or passive (waiting for prey to pass) techniques.

Larger anthozoans, cephalopods, many gastropods, some annelids and crustaceans, asteroids, some echinoids and ophiuroids.

* In palaeontological studies, deposit feeders are usually included in the more general category of 'detritus feeders' . This also includes scavengers, which eat larger particles and dead organisms upon and within the sediment (e . g . some gastropods), and most browsers (or herbivores) . The latter are first level consumers that scrape, rasp, or chew live algae and other plants ( e . g . Amphineura and some gastropods) .

(Fig . 2; Scott 1976, 1978) . In practice, it was found that precise habitat requirements needed to be recorded too, and so a second 'substrate-niche' tri­ angle is usually depicted alongside the 'feeding­ habit' one (Fig . 2) . Substrate-niche names are usually appended as prefixes to the feeding-habit ones, giving a community a title such as 'vagrant­ epifaunal, detritus-suspension feeding' . One of the obvious applications of this method of describing trophic structure is in differentiating contemporaneous communities within particular environmental settings . For example, both Cretaceous and Cenozoic communities occurring along onshore - offshore gradients plot in distinctive fields within the ternary diagrams (Fig. 3; Scott 1978) . Although the database is still comparatively small, it is also possible to trace the trophic groups associated with certain habitats through time .

For example, in lower shore face and nearshore communities between the Early Palaeozoic and the Cretaceous, there was a marked shift from epifaunally- to infaunally-dominated detritus­ suspension feeding types (Scott 1976) . There are, of course, other ways of depicting the trophic structure of palaeocommunities . In his very detailed analysis of macrobenthic assemblages from the Korytnica Clays (Middle Miocene, central Poland), Hoffman (1977) used a series of tables to illustrate 'trophic- substrate - mobility niches' . These are simple two-dimensional diagrams which plot food location (infauna and epifauna, subdivided into mobile and sessile) against feeding category (suspension and deposit feeders, predators etc . , subdivided into various positions i n the water column and sediment) . Each of these diagrams is supported by a histogram showing the distribution

4 . 1 6 Trophic Structure

387 SUBSTRA TE NI CHES

FEEDING HA B I TS

VAG D ET

SUSP S u spension

Detritus­ suspension

50�--�----.---�--� 50 Vagrant­ i n fa u n a l

Vagrant­ e p i fa u n a l Predator

A------+--� 80 I n fa u n a l

E p i fa u n a l

PRED

50

D ET

E PS U S

INSUS

50

Descriptive trophic structure . SUSP suspension feeders, DET detritus feeders, PRED predators, VAGDET vagrant detritus feeders, EPSUS epifaunal suspension feeders, INSUS infaunal suspension feeders . Within the feeding habit triangle, a suspension-feeding community consists of more than 80% suspension-feeding species. A detritus-suspension community would be one composed of 50 -80% suspension feeders, 10-50% detritus feeders, and less than 10% predators. Within the substrate niche triangle, an infaunal community comprises less than 20% vagrant organisms and no more than 50% epifauna . A vagrant-infaunal community comprises 20 - 80°;;, vagrant animals and more infauna than epifauna . (From Scott 1976.) Fig. 2

=

=

=

=

=

=

FEEDINGS HA B I TS

S U B S TRA TE NICHES

SUSP

VAG D ET

Oyste r b a n k Middle shelf

I nner shelf

Oyster b a n k

B a y centre DH

PRED

EPSUS

I NS U S

Fig. 3 Trophic structure o f some Cenozoic shallow-marine environments. Note that the only significant overlap occurs between Bay centre and Outer bay -inlet communities. Abbreviations as for Fig. 2. (From Scott 1978 . )

4 Palaeoecology

388 B

A

Lowe r level s u s p e n s i o n feed e r s I n sed i m e n t parasites Scave ngers In sed i m e n t d e p o s i t fee d e r s

Perce n t of total

Percent of total

b i ovo l u m e

b i ovo l u m e

o

50 Lowe r l e v e l s u s p e n s i o n feed ers In sed i m e n t p red ato rs Scavengers In sed i m e n t d e p o s i t feed ers B rowsers Para s i t e s Sed i m e n t-wate r i n te rface p redators H i g h e r l evel s u s p e n s i o n feed e rs

B rowsers Paras ites

o

c

Perce n t of total b i ovo l u m e

o

50

50

B row; e r s Lowe r level s u s p e n s i o n feed e r s Scave ngers Para s i t e s I n sed i m e n t p a ras ites I n sed i m e n t d e p o s i t feed ers Sed i m e n t-wate r i n terface p redators H i g h e r l evel s u s p e n s i o n feede rs

a

Trophic structure of three communities from the Middle Miocene Korytnica Clays, central Poland . Their frequent association is thought to reflect an ecological succession in a small shallow basin from a pioneer stage on barren muddy bottoms (Corbula community), through an intermediate stage (Corbula- scaphopod community), to a mature (climax) stage marked by the development of extensive seagrass stands (Turboella - Loripes community) . An index of trophic uniformity within each community is derived from the dispersion of total biovolume amongst the various trophic groups (upper histograms) . The index (or Nesis) value (6 2) is the total biovolume of the assemblage divided by the number of trophic groups (see Hoffman 1977, p. 244). The high value in the Corbula community (A) is due to the dominance by a single species (the shallow-burrowing, suspension­ feeding bivalve Corbula gibba) . Much lower values in the Corbula - scaphopod (8) and Turboella - Loripes communities (C) can be linked to the wider dispersion of biovolume amongst the various trophic troups . The trophic web reconstruction for the Corbula community (a) is a very simple one based largely on a short suspension-feeder food chain . However, in the succeeding Corbula - scaphopod community (b) the web is appreciably more complex, and comprises two distinct subwebs (the suspension-feeding and deposit-feeding ones) . Finally, in the climax Turboella - Loripes community (c) there are at least three equally important subwebs. Note that separate subwebs can be terminated by the same predator (top row of boxes) . epr epifaunal predators, ipr infaunal predators, sc scavengers, par parasites, br browsers, df deposit feeders, sf suspension feeders . (From Hoffman 1977, 1979.) Fig. 4

=

=

=

=

=

=

=

of biovolumes (representing biomass) amongst the various trophic categories (Fig. 4) . In fact, the pres­ ervation of these assemblages was so good that trophic webs could be constructed for each com­ munity . Although this entailed estimating unpre­ served components of the ecosystem, and involved a considerable amount of simplification, the resulting models proved to be valuable aids in the study of community structure and function (Fig. 4) .

One potentially serious drawback to the tech­ niques outlined so far is the fidelity of the fossil record (Section 3 . 12) . Estimates of the proportion of a biocoenosis unlikely to be preserved (i. e . the soft­ bodied component) range from 50 to 75% . Deposit­ feeding groups of annelids and athropods are particularly likely to be missing from a fossil as­ semblage . There may, however, be ways of partially offsetting this problem. Firstly, there may be direct

389

4 . 1 6 Trophic Structure evidence of deposit feeders in the form of either faecal pellets or bioturbation . It may also be possible to calculate the proportion of individuals at various trophic levels in an assemblage and compare these ratios with those in a modern community. Using estimates of the efficiency of energy transfer be­ tween trophic levels of 10-20%, it may be clear, for example, that the ratio of carnivores to primary consumers is too high . In that case, soft-bodied organisms must have been an important component of the ecosystem (Stanton et al . 1981) . A trophic web based solely on numerical abun­ dances arguably gives a poor picture of the original community. Ideally, trophic analyses should be based on estimates of energy flow from one level to another and this would appear to�be beyond the scope of the palaeontologist. Nevertheless, in a pioneering study of molluscan assemblages from the Middle Eocene Stone City Formation of south­ east Texas, Stanton et al. (1981) suggested several ways in which abundance data could be refined to make them more representative of the passage of energy through a community . Such techniques, which are particularly amenable to certain predatory gastropod taxa, include the production of survivor­ ship curves (Fig. 5) . From these it is apparent that Polinices aratus, the numerically dominant predator, was subject to very high juvenile mortality . In at least five other predatory taxa (such as the fascio­ lariid Latirus moorei; Fig. 5), it was found that a much greater proportion of their populations at­ tained adulthood; these were thus inferred to have been more important components of the Stone City Formation foodweb (Stanton et al . 1981, Table 2) . Another useful measure of a species' significance in a trophic web may be its total biomass, for in predators this should be related to the biomass of prey consumed. The number of prey required to support a predator population is proportional to the amount of energy input necessary to produce and maintain the predator population (Stanton et al. 1981; see their Table 2 where biovolumes are used to calculate the biomass formed by secondary pro­ duction) . In the Stone City Formation, biomass values again showed that predatory gastropods, such as Latirus moorei, two species of turrids, and Retusa kellogii (opisthobranch), were much more prominent than Polinices aratus (the naticid) . The influence of the turrids is particularly important as they fed on non-preservable prey (primarily polychaetes) . If further turrids have similar size ­ frequency distributions, it has been estimated that the proportion of soft-bodied prey in the food web

B

A

1 00 ...-----, 1 00 .......-----. 50

"�

bJ)

.� >

10

C

5

:::>

50 � 10

"2:

>

:::>

if>

.5

if>

a;

:::

a; CL

1 5

5

20 40

60

80 1 00

Perce n t m a x i m u m age

20 40

60

80 1 00

Perce n t maxi m u m age

Fig. 5 Survivorship curves . A, for Polinices aratus (naticid) . B , for Latirus moorei (fasciolariid) . Note that size values were converted to age values using the logarithmic relationship that exists between the two . (From Stanton et al. 1981 . )

may have been as high as 50% . A more serious problem faced by the community palaeoecologist is that many benthic organisms defy simple trophic classification. Take, for example, those types that can readily interchange between deposit- and suspension-feeding strategies . These include representatives of several polychaete famil­ ies, a number of ophiuroids and irregular echinoids, and a significant proportion of the bivalve super­ family Tellinacea (species of Scrobicularia, Macoma, Tellina, etc. ) . By switching between these two feeding modes, organisms may well be able to over­ come periods of food shortage . This in turn would permit colonization of unpredictable habitats, such as estuaries or shallow temperate seas . Some recent discoveries from the deep seas have had a profound effect on our understanding of trophic structure . Not the least of these is the start­ ling revelation that many abyssal bivalves (perhaps 25 - 30% of the total fauna) are predators! Using a raptorial inhalant siphon, representatives of at least four families within the subclass Anomalodesmata (Parilimyidae, Verticordiidae, Poromyidae and Cuspidariidae) actively seek out and capture prey (Fig. 6) (Morton 1987) . The giant tube-worm and bivalve communities recently discovered around certain sea floor volcanic vents (such as in the Galapagos rift) also display some very unusual feeding traits . Forms such as the pogonophoran Riftia pachyptila, the vesicomyid bivalve Calytogena

390

4 Palaeoecology

Sulphide-oxidizing symbiosis in the Lucinacea . In addition to suspension-feeding, many species within this bivalve superfamily obtain nutrients from endosymbiotic bacteria . In common with other groups demonstrating this phenomenon, lucinaceans live in deep burrows; it is thought that these (or complementary tube-like structures) are essential for the accumulation of both dissolved oxygen (from surface waters) and hydrogen sulphide (from the enclosing anoxic muds) . Besides having a hypotrophied gut, lucinaceans typically display prominent ctenidia (vertical shading) which are packed full of bacteriocytes. The postero­ dorsal margins of the gills are typically fused with the muscular mantle edge and it is thought that this arrangement may facilitate the pumping of sulphide-rich waters over the bacteriocytes (especially via the exhalant siphon). Anterior foot in black . (From Reid & Brand 1986.) Fig . 7

In predatory bivalves active prey capture is achieved using an inhalant siphon which can be rapidly extended (through hydraulic pressure changes within the mantle cavity) . The siphon is retrieved (with the prey enclosed) by the action of pallial retractor muscles. Here the poromyid Poromyn grmllllntn uses a large hood at the end of the siphon to ensnare a tiny crustacean, the principal diet item of predatory bivalves . (From Morton 1987.) Fig. 6

magnifica (giant white clam), and the mytilid Bathymodiolus thermophilus are now known to be gutless . They obtain nutrients by means of endo­ symbiotic sulphide-oxidizing bacteria contained in the gill regions . These act to detoxify the sulphide­ rich volcanic waters, producing a series of carbo­ hydrates and amino acids that can then be utilized by the organisms as a food source . Similar gutless bivalves, together with forms possessing hyper­ trophied alimentary systems, have recently been shown to characterize other environments with extraordinary energy sources (such as the anoxic muds associated with marine grass beds and the effluent from fish farms and pulp mills) . Among these are shallow-water solemyid, lucinid, and thyasirid bivalves, all of which possess rich supplies of sulphide-oxidizing bacteria (Fig . 7; Reid & Brand 1986) . No longer can bivalves such as these be simply classified into either deposit- or suspension­ feeding categories . Finally, a note o f caution should b e expressed about interpreting the role of predators within palaeocommunities; in many instances it is virtually impossible to pinpoint their exact prey. Some Recent species of whelks, for example, feed on rep­ resentatives of up to eight separate phyla, and the Nassariidae (also frequently interpreted as car­ rion feeders) include at least one deposit-feeder (Ilyanassa obsoleta) . Similarly, the Cymatiidae con­ tains an algal-grazer (Apollon natator) and several

members of the Cancellariidae are probably parasitic (Taylor 1981). Clearly, the maintenance of trophic structural analysis as a viable technique in palaeoecology is going to involve the very close collaboration of palaeontologists and biologists . References Hoffman, A. 1977. Synecology of macrobenthic assemblages of the Korytnica Clays (Middle Miocene; Holy Cross Mountains, Poland) . Actn Geologica Polonica 27, 227-280 . Hoffman, A. 1979 . A consideration upon macrobenthic assemblages of the Korytnica Clays (Middle Miocene; Holy Cross Mountains, central Poland) . Acta Geologica Polonica 29, 345 - 352. Morton, B . 1987. Siphon structure and prey capture as a guide to affinities in the abyssal septibranch Anomalodesmata (Bivalvia) . Sarsia 72, 49 -69. Reid, R . G . B . & Brand, D.G. 1986. Sulfide-oxidizing symbiosis in Lucinaceans : implications for bivalve evolution. The Veliger 29, 3-24. Scott, R.W. 1976 . Trophic classification of benthic com­ munities. In : R.W. Scott & R.R. West (eds) Structure and

4. 1 7 Evolution of Communities classification of paleocommullities, pp. 29 �66. Oowden, Hutchinson & Ross, Stroudsburg, Pennsylvania. Scott, KW. 1978. Approaches to trophic analysis of paleo­ communities . Lethaia 11, 1 � 14. Stanton, R.J . , Powell, E.N. & Nelson, P.e. 1981 . The role of carnivorous gastropods in the trophic analysis of a fossil community. Malacologia 20, 451 �469 .

391

Taylor, J . O . 1981 . The evolution of predators in the Late Cretaceous and their ecological significance . In: P.L Forey (ed .) The evolving biosphere, pp. 229 �240 . British Museum (Natural History), London & Cambridge University Press, Cambridge . Walker, K.R. & Bambach, R.K. 1974. Feeding by benthic invertebrates: classification and terminology for paleo­ ecological analysis . Lethaia 7, 67�78.

4 . 17 Evolution of Communities A . J . BOUCOT

It has long been known that fossils do not occur in a random manner. Particular mixtures of taxa, with particular relative abundances, characterize every time interval and environment. These 'mixtures' may be termed communities, employing biological parlance, although some prefer the term association or assemblage. Some palaeontologists refuse to use the term community, preferring assemblage, be­ cause of the absence of soft-bodied organisms . But it is obvious that virtually all descriptions of both modern and fossil communities deal with only a small part of the total biota present, i . e . the term community commonly approximates to the definition of a guild. Thus, we have rodent communities, coral communities, brachiopod communities, trilobite communities, planktic foraminiferal communities, benthic foraminiferal communities, lichen com­ munities, larger carnivore communities, tree communities, etc. Palaeontologists and biostratigraphers of the nineteenth century did not pay much attention to what are now termed fossil communities . This probably reflects their overwhelming concern with dating and correlation of beds in which emphasis was placed on the taxa common to different collec­ tions . The study of communities emphasizes instead the taxonomic differences between collections . Only in the latter part of this century has extensive interest developed in communities, particularly because of their potential for providing a better understanding of past environments . A community may be defined as a recurring associ­

ation of taxa, in which relative taxonomic abundances remain more or less fixed. For example, brackish water

oysters remained dominant in their community from the later Mesozoic to the present, just as the brachiopod Pentamerus remained dominant in its community from the later Early Silurian to the earlier Late Silurian, and the shells in a lower dominance, later Cenozoic Pecten community retained similar relative abundance through time . The term biofacies is used in various ways (Section 4. 18) : some workers employ it when referring to what are essentially biogeographical units, such as a 'Gondwanic biofacies' ; others employ it for very broad environmental units within a biogeographical unit, such as 'deep-water biofacies' or 'black shale biofacies'; still others use it for individual com­ munity types, such as 'pentamerid biofacies', 'stringocephalid biofacies', or 'brackish water oyster biofacies' . Community evolution deals with the Darwinian species-level evolution shown by the genera present in specific community types . These narrowly de­ fined 'community types' may be referred to as com­ munity groups (Boucot 1978, 1981, 1983, 1986, 1987; Fig . 1); the term was devised to describe a com­ munity type undergoing species-level evolution among its constituent genera in evolutionary time - particularly the less abundant, commonly more endemic, and more stenotopic genera . Those who study modern communities commonly name them after the dominant, abundant taxa - those which evolve very little - whereas the changes in communities chiefly affect the uncommon genera and their rapidly evolving species. It makes sense, therefore, to name community groups after the abundant, slowly evolving genera, and the com-

392

4 Palaeoecology H igh G e n e ra Samples

�g

A

E

B

C

�

34 33

-

30

29

26 25

-

24 22

21 20 19

16

15

14

13

12

7

6

5

4 3

2

-

!

-

I

·

3·

I

1 21

-J!l_ 2.I! -- ; li

18

9 8

tI

2

17

11 10

-

r!

31 [-

28 27

Time

E F G H I

•

I I:1 I 1I I �

I

-

I .

•

61

I 51

41

4 . 3 •

31

2 ·

31

I •

7 I 6 ' 5 •

• •

I II 1 41

31

23

D

_ 4. 3 •

32

low

:>

A b u n d a nce

2 I I

21

•

I

1 1

21

11

I1

1 Diagrammatic representation of the fossil record of a community group through an interval of time measured in a few fives or tens of millions of years . The genera (lettered) and their described species (numbered) are assigned line widths based on their relative abundance as individual specimens. Note that species of the rarer genera tend to be both more rapidly evolving, and also have a much poorer fossil record correlating with their numerical abundance (i. e . are rarer) as individual specimens . The ancestor - descendant relations are presumably phyletic in nature . Sampling gaps indicated on the figure may be due to non-deposition, erosion, collecting failure (commonly caused by covered intervals), or presence of another community group because of local or regional environmental shifts. (From Boucot 1983 . ) Fig.

munities after the less common, rapidly evolving genera and their time-sequences of species . An excellent example o f community evolution was provided by Ziegler (1966), whose late Lower Silurian time sequence of species of the brachiopod Eocoelia has been well tested in eastern North America, the U.K., and Scandinavia (Boucot 1975) for nearshore, subtidal marine, level-bottom, mod­ erate turbulence, possibly turbid conditions . The

Eocoelia community, as now construed, commonly comprises more than 90% Eocoelia shells, chiefly disarticulated in a muddy matrix, and with a high population density (so-called 'pearly beds') . [Com­ munity succession, defined as the presence of one taxon making it possible for the same area to be subsequently colonized by a second taxon, is not community evolution . ] Community evolution also deals with the initial,

393

4. 1 7 Evolution of Communities quantum evolution of distinctly different, narrowly defined biofacies types from either newly evolved higher taxa or new mixtures of previously existing taxa. This phase of community evolution sees the first appearance of new community groups, and is followed by subsequent species-level changes within the initial genera, particularly, the less abundant ones. The community groups presem- within any one major subdivision (level-bottom environment, reef complex of communities, sponge forest, pe1matozoan thicket, etc . , as well as comparable non­ marine units) have concurrent times of appearance and disappearance from the base of the Cambrian to the present. These concurrent times within the level-bottom environments mark the boundaries of ecological-evolutionary units (see below) . The level-bottom community groups of the mar­ ine environment tend to be not only dominant in terms of area occupied and overall numbers of specimens per time interval, but also in strati­ graphical duration. The fossil record consists of a fixed number of time intervals, each one of which contains a relatively homogeneous biota . Boucot (1983, 1987) termed these units ecologic-evolutionary units . There are 12 such major units from the Cambrian to the present (Fig. 2) . Within marine benthic environments the level­ bottom community groups have the full time range of the appropriate ecological-evolutionary units, but the non-level-bottom community groups (reef complex communities, pelmatozoan thicket com­ munities, sponge forest communities, bryozoan thicket communities, etc.) commonly appear in time significantly later than the level-bottom groups . They do, however, commonly share the same extinction time . The reasons for relative fixity of community groups are poorly understood . They may involve a significant measure of both coevolution at one trophic level or another, and of stabilizing selection in so far as the taxa present are concerned . Biologists have not yet uncovered any very effective means of measuring levels of coevolution in modern com­ munities, although they suspect that there are major differences in levels of coevolution (viz . the coral reef community complex vs . level-bottom com­ munities, and tropical rainforests vs . grasslands) . The boundaries of ecological- evolutionary units commonly mark major extinctions followed by adaptive radiations (Boucot 1983, 1987) . Ecological ­ evolutionary subunit boundaries, such as those between the Silurian-Devonian and Mississippian-

L E V E L BOTTOM COM M UN I T I E S EcoloQicalEvol u t i o n a r y U n i ts

Time I n t e rv a l s

MOjor

E d iocarion

Lt:

D

M - U t:

m

LO

IV

M-UO

V

VB

2/3 L S

�LS-�UO � UO

vm

C+ P

IX X

XI

D

MOjor

*

Adoptive R adiations

* Major Community Reorl;lonizotions

AI

Cl

AI

Cl

A3

C3

A3

C3

Major

*

Dispersal Event

El

I

VI

*

E x t i nc t ions

L

El

E5

E5

E2

J +K

Cenozoic

C2

C3

E2

C4

01

C4

El

Tr

M-UTr

A2

A2

C2 C4

E3

A2 A2

E4

C2

C2

C4

Phanerozoic ecological- evolutionary units . E major extinctions (numerals indicate rank, 1 highest), A major adaptive radiations (numerals indicate rank, 1 highest), C major community reorganizations (1 highest), D major dispersal event. (From Boucot 1987.) Fig. 2

=

=

=

=

Pennsylvanian, are similarly marked by minor extinctions followed by minor adaptive radiations . One consequence o f community evolution at the species level is that the number of species (and genera) does not change significantly within a specific community group in any specific ecological- evolutionary unit, i . e . diversity is not in a continual state of flux. There have been many suggestions made about the factor(s) involved in extinctions (Section 2 . 1 2), but few about those controlling adaptive radiations . Major, time-concurrent, adaptive radiations within the same portion of the ecosystem affect varied, taxonomically unrelated organisms and many dif­ ferent community groups . The level of randomness involved is unknown, as is whether the presence of potentially empty niches is most important (most adaptive radiations tend to follow major extinctions) . However, the absence of the reef com­ plex of communities during many ecological ­ evolutionary units (Middle - Upper Cambrian,

4 Palaeoecology

394

-

�

�

A

Jf

Eco l o g l c­ evo l u t i o n a ry unit or subunit boundaries

'-'- - - - - B

Fig. 3 A, CIa doge ne tic pattern characteristic of the organisms belonging to individual ecological -evolutionary units. Note that cladogenesis is restricted to that brief moment in time when new community groups first appear . Cladogenesis here refers to metacIadogenesis, i.e. quantum evolution mediated phenomena. Biogeographically mediated diacIadogenesis can, of course, occur anywhere within an ecological- evolutionary unit. Metacladogenesis refers to major cladogenetic events resulting in new families and higher taxa (Boucot 1978), whereas diacladogenesis refers to minor cladogenetic events giving rise merely to new genera and species, such as the post-Miocene species occurring on either side of the Isthmus of Panama. B, Cladogenetic pattern of the standard, hypothetical, random through time type, which ignores the constraints imposed by what we know about community evolution. Note that this view permits cladogenesis to occur at any time within an ecological- evolutionary unit, and is also consistent with important changes in species-level diversity within any ecological - evolutionary unit as contrasted with the conclusion outlined in A. Such random, within-unit changes in diversity do not occur . It is only by 'superimposing' family trees derived from ecologically unrelated, major parts of the global ecosystem (such as level-bottom, reef complex of communities, pelmatozoan thickets, etc . ) that one can simulate the unnatural random cladogenetic pattern . (From Boucot 1986 . )

Lower Ordovician, Famennian half of Upper Devonian, Mississippian, Lower Triassic) suggests that the empty niche possibility is incapable of explaining all the facts . Another consequence o f community evolution is that the commonly presented, hypothetical, random type of family tree (Fig. 3B) does not agree with the

more espaliered tree (Fig. 3A) indicated by species­ level evolution within community groups . This is a consequence of the fact that the community types (community groups) within each ecological ­ evolutionary unit remain relatively constant in their generic content. The evolutionary changes consist of phyletic-anagenetic changes within each genus (particularly the more stenotopic, more endemic genera, that also tend to be far less common as individual specimens) . There is not a constantly changing overall species- or genus-level diversity within either individual community groups or within major portions of the ecosystem, such as the level-bottom, during any one ecological ­ evolutionary unit. An apparently ever changing overall diversity may, however, be observed stat­ istically if disparate portions of the ecosystem are grouped together uncritically, (such as the level­ bottom, reef complex of communities, sponge forests, bryozoan thickets, pelmatozoan thickets) which commonly have different origination times within any particular ecological -evolutionary unit. References Boucot, A.J. 1975 . Evolution and extinction rate controls . Elsevier, Amsterdam. Boucot, A.J. 1978 . Community evolution and rates of clado­ genesis . Evolutionary Biology 11, 545 - 655 . Boucot, A.J. 1981 . Principles of benthic marine paleoecology. Academic Press, New York. Boucot, A.J. 1983. Does evolution take place in an ecological vacuum? 11 . Journal of Paleontology 57, 1 - 30. Boucot, A.J. 1986 . Ecostratigraphic criteria for evaluating the magnitude, character and duration of bioevents . In: O . H . Walliser (ed . ) Lecture notes i n Earth sciences, N o . 8 , pp. 25 -45. Springer-Verlag, Berlin. Boucot, A.J. 1987. Phanerozoic extinctions: how similar are they to each other? III Journadas de Paleolltologia, Leioa (Vizcaya), Palaeontology and evolution: extinction events, 50 - 82. Ziegler, A.M. 1966 . The Silurian brachiopod Eocoelia hemisphaerica (J. de C. Sowerby) and related species . Palaeontology 9 , 523 - 543.

4 . 18 Biofacies P . J . BRENCHLEY

mentally significant, but it has no value in correlation . The 'Posidonia' Shales (Section 3 . 1 1 . 6) are bitumi­ nous, laminated shales of Toarcian (Jurassic) age, widely developed throughout Germany. They are characterized by the exceptional preservation of a variety of fossils, but particularly by an abundance of the bivalve 'Posidonia' ( Bositria) . The onset of deposition of bituminous shales was apparently nearly synchronous over a wide areas and was re­ lated to the Toarcian transgression. However, depo­ sition of the shales persisted longer in basin areas than on more positive regions, so the upper bound­ ary is diachronous . The Toarcian sequence can be effectively zoned and correlated using the common ammonites, while the 'Posidonia' Shales represent a distinctive unit in that succession reflecting particu­ lar environmental conditions which determined a characteristic biota . The Wilsonia Shales (cf. Sphaerirhynchia (Wilsonia) wilsoni) are a subdivision of a generally monotonous sequence of black shales and thin laminated siltstones found at the shelf to basin transition in the Ludlow (Silurian) of the Welsh borderland . The slight vertical changes in the rather sparse fauna are more obvious in the field than any changes in lith­ ology, and it has been found useful to recognize a sequence of Cyrtoceras Mudstones, Wilsonia Shales, and Orthonota Mudstones . The Wilsonia Shales pass shelfwards into mudstones with a deep-shelf fauna and basinwards into shales containing mainly graptolites . Here we have an example of a mappable field unit comparable to a formation, but more effectively recognized on the basis of its fauna . The use of fossil names to characterize formal

Definition

The term biofacies refers to 'the total biological characteristics of a body of rock' (Moore 1949) but has been used in two rather different ways: in a stratigraphic sense to refer to 'a body of rock which is characterized by its fossil content which dis­ tinguishes it from adjacent bodies of rock', and in an ecological sense to refer to 'a biota or association of fossils which characterize a region or body of rock' .

=

Biofacies in stratigraphy

When the term 'biofacies' is used in a stratigraphic sense, the emphasis is on a geographically or verti­ cally restricted body of rock which is distinct be­ cause of its fossil content (Fig. 1 ) . The stratigraphic value of biofacies is illustrated by the widespread use of fossil names to characterize particular rock units, e . g . Pen tamer us Beds, 'Posidonia' Shales (Posidonienschiefer), and Wilsonia Shales. These three examples can be used to illustrate the role of biofacies and show how biofacies differ from biozones . The Pentamerus Beds are a varied sequence of mudstones with interbedded calcareous sandstones containing abundant Pentamerus oblongus devel­ oped in the Lower Silurian of the Welsh borderland . Pentamerus is part of an ecologically controlled community and occurred diachronously across the shelf during the Lower Silurian transgression . Thus the Pentamerus biofacies occurs locally wherever the depth, substrate, and other combination of ecologi­ cal factors were suitable . Its presence is environ-

Sea- level

Fig. 1 Distribution of five biofacies developed during a transgression. Note that biofacies 1 and 2 coincide with lithofacies, but that biofacies 3, 4, and 5 are developed within a single mud lithofacies .

D WAmid

5 4 3 2

MUd Silt Sand

395

4 Palaeoecology

396

lithostratigraphic units is now deemed invalid and most of the 'biofacies' have either been redefined as formations based on their total litho logical and bio­ logical characteristics, or have become obsolete . Particularly distinctive biofacies will, however, probably persist in the literature as informal units . Ecological use of biofacies

The term 'biofacies' in an ecological and environ­ mental sense is usually used to express the lateral or vertical variation in biota in relation to differences in environment. For example, trilobite biofacies in the Cambrian of the western U . S . A . have been de­ scribed in terms of their environmental position, i . e . inner shelf, outer shelf and slope . Biofacies have also been applied in an even more general sense to express broad lateral changes in biota according to environment. Thus, House (1975) described the marine Devonian of Europe in terms of biofacies regimes which represent the characteristic Devonian faunas of the near shore, shelf, and basinal regions . In a more detailed biofacies analysis of Ordovician rocks in the Upper Mississippi Valley, U . 5 . A . , factor analysis was used to define seven faunas distributed over 65 000 km2 of outcrop within a single member (the Mifflin Member of the Platteville Formation) . Each of these bivalve/brachiopod faunas differed and their geographical occurrence could be mapped to show biofacies distributions in an extensive, shallow epeiric sea (Bretsky et al. 1977) .

M)N::;jj

t8 § �

One of the valuable facets of the term 'biofacies' is that it can be applied to faunal associations iden­ tified with very different degrees of taxonomic pre­ cision. In some environmental studies it may be important to define the associations by the assem­ blages of species, but in other situations a very broad characterization of the fauna, such as 'coral! brachiopod association', may be appropriate . Terms commonly used in palaeoecology which are closely related to biofacies are: faunal association, community, palaeocommunity, and benthic assemblage. The first three of these terms attempt to express a distinct association of taxa which probably lived together. The regional distribution of faunal as­ sociations or communities on the sea floor or in a body of rock can be referred to as a 'biofacies' . For example, Bretsky (1969) described three com­ munities, the Sowerbyella- Onniella community, the Orthorhynchula-Ambonychia community, and the Zygospira-Hebertella community, from Upper Ordovician rocks of the Central Appalachians . The communities were recognized as distinct associ­ ations of taxa, and the geographical distribution of such communities can be shown on a biofacies map (Fig. 2) . The term benthic assemblage has been used by Boucot (1975) to identify communities which lived in the same position relative to the shoreline . Benthic assemblages are therefore approximately depth related, although temperature, substrate, and other ecological controls may be important in deter-

S a n dston e S i ltsto n e shale L i m e st o n e

Fig. 2 Biofacies map of Upper Ordovician community distribution in the Central Appalachian region, U.S.A. 1 Sowerbyella- Onniella community, 2 Orthorhynchula- Ambonychia community, 3 Zygospira- Hebertella community, 4 offshore faunas . (After Bretsky 1969 . ) =

=

=

=

4 . 1 8 Biofacies mining their location . Boucot identified six benthic assemblages in his treatment of Silurian -Devonian communities . According to Boucot, nearshore and inner shelf faunas can be referred to benthic assem­ blages 1 and 2, mid-shelf faunas to benthic assem­ blages 3 and 4, and outer shelf and upper slope faunas to benthic assemblages 5 and 6 (Fig. 3) . Faunas of quite different ages can be assigned to the same benthic assemblage because they have the same range of fossil groups though the taxonomic composition is different in detail. Such a similarity of faunas was recognized in nearshore carbonate communities of Ordovician and Devonian age and was characterized as congruent communities . They could equally well have been characterized as eco­ logical stable biofacies . This usage has been applied to the persistent community found in dysaerobic sediments of Middle Devonian - Early Permian age, which has been referred to as the dysaerobic biofacies (Section 4 . 19.4) .

397

Recent biofacies

The concept of biofacies draws strength from studies of modern benthic faunas . Petersen (1915), using grab samples in Danish waters, showed that there were areas of the sea floor characterized by particu­ lar associations of bivalve, echinoid, and polychaete worm species . The distribution of these level­ bottom communities is approximately related to distance offshore and depth of water, and hence it is possible to map out approximately shore-parallel biofacies . In general, ecological zonation of faunas on level-bottom shelves is relatively simple because: (1) there is an absence of dense flora, and hence the communities are principally influenced by physico­ chemical aspects of the environment and biotic interactions amongst members of the community; (2) level-bottoms lack micro-landscape and are environmentally relatively homogeneous, so the faunal associations occupy a single habitat rather

Q u iet water

·: : �:·:·::"·: m '; .;:. : : : : : . ::0.:.:.:.:; : :°:°

I n te r m ed iate R o u g h water

Relationship between Silurian epifaunal benthic assemblages and communities . Three possible applications of the term 'biofacies' are shown. A, Term refers to rocks containing a particular community. B, Term refers to all nearshore communities . C, Term refers to rocks containing a group of communities with some taxonomic similarities . (After Boucot 1975 . ) Fig. 3

398

4 Palaeoecology

than a mosaic of micro-habitats; (3) substrate type approximately reflects the hydrodynamic conditions of the area and is often roughly correlated with several physico-chemical parameters, such as tur­ bidity, mobility of substrate, oxygen, and organic content; (4) a large proportion of level-bottom dwelling animals are suspension feeders and the trophic structure is relatively simple (Thorson 1971 ) . I t has been claimed b y Thorson that level-bottom communities which occupy approximately the same position relative to shore (e .g. belong to the same benthic assemblage) have a similar appearance wherever they are found in the world . These belts of similar communities, with different species but sharing some genera were termed parallel

communities . Communities which inhabit rocky shorelines and to a lesser extent carbonate-producing regions have a much more complex and patchy distribution . Although there is a general shore-parallel distri­ bution of communities in carbonate regions, local heterogeneity of the sea floor, particularly where reefs are present, produces a mosaic pattern of biofacies on a local scale . Biofacies distribution

Most marine biofacies are broadly related to water depth and sediment substrate, but other physico­ chemical parameters, such as availability of oxygen, salinity, or substrate mobility may modify any sim­ ple distribution pattern . Biofacies which are pre­ served in adjacent positions in vertical sequence are believed to have generally occupied adjacent pos­ itions on the sea floor and therefore behave in a similar manner to lithofacies, according to Walther's Law . At any one time, biofacies distribution, re­ flecting the community distribution of that time, will be somewhat different in different faunal prov­ inces, particularly if they belong to different climatic belts . In general, biofacies have a greater biomass and are more diverse in tropical than arctic regions, and tropical carbonate shelves are probably par­ titioned into more biofacies than are clastic shelves in cold er regions . In any one faunal province, biofacies generally show distinct changes in taxonomic diversity and biomass in any onshore to offshore transect . In modem environments, biofacies generally decline in biomass towards the deeper parts of the shelf and then rapidly down the continental slope . Diver­ sity is usually low in any one nearshore environ­ ment, but because of the large number of nearshore

environments the total diversity of the nearshore region can be high . Diversity is relatively high across the continental shelves and can remain moderately high to the base of the slope, to depths as great as 4000 m. The pattern of onshore - offshore change in different for different parts of geological history, as will be described below. Biofacies distribution in the Phanerozoic

Cambrian biofacies . On Cambrian clastic shelves bio­ facies containing trilobites, inarticulate brachio­ pods, and molluscs are rather poorly differentiated and no clear onshore - offshore trends have been defined . The very diverse, mainly soft-bodied, Middle Cambrian, Burgess Shale fauna of British Columbia (Section 3 . 1 1 . 2) is evidence that Cambrian faunas were far richer than is suggested by the shelly assemblages, and that ecological zonation may have been more refined than is recorded by the commonly preserved fossils . In carbonate shelf and slope environments of the western U . S . A . a clear differentiation has been rec­ ognized between trilobite faunas living on the plat­ form and shelf edge (Hungaia fauna) and a trilobite fauna which occupied deeper-water sites (Hedinaspis fauna) . Ordovician and Silurian biofacies. Following the ex­ tinction of a substantial part of the Cambrian fauna, there was a major radiation in Early Ordovician times which produced a varied benthic fauna of suspension-feeding animals with skeletons . These suspension-feeding faunas became partitioned into a number of biofacies, approximately related to water depth, in the early part of the Ordovician. Progressively during the Early and Middle Ordovician there was colonization of the deeper parts of the shelves (Sepkoski & Sheehan 1983) . This established a pattern where about five or six communities (biofacies) occurred in any transect across a level-bottom clastic shelf in the later Ordovician and Silurian (see Fig. 3) . The communi­ ties were generally brachiopod-dominated, but bryozoa, corals, and crinoids were also important elements, and trilobites and bivalves were often associated with the fauna (McKerrow 1978) . Nearshore biofacies typically had a low species diversity and a very variable biomass which could be high when environmental conditions were favourable . The faunas were characterized by ar­ ticulate brachiopods, particularly large orthids,

4 . 1 8 Biofacies bivalves, and sometimes inarticulate brachiopods such as Lingula. Along a traverse outwards across the shelf there was a general increase in taxonomic diversity towards the shelf margin, though there could be local areas of particularly high diversity in mid-shelf regions where carbonate build-ups de­ veloped . Abundance of fossils and total biomass was generally high into mid-shelf regions, but de­ creased towards the outer shelf and declined very rapidly down the slopes . The exact patterns of di­ versity and biomass depended on the depth of the shelf- slope break and the influence of other limiting factors, such as oxygen. Typical mid-shelf biofades in the Silurian have diverse brachiopod faunas with pentamerids, atrypids, strophomenids, and orthids, with variable numbers of associated bryozoa, tabulate and rugose corals, crinoids, and

399

trilobites . In biofacies of greater shelf depths, tabu­ late corals, bryozoa, and crinoids are rarer, trilobites are relatively more important, and the brachiopod fauna, though still diverse, is commonly composed of relatively small-shelled forms . At the shelf edge the diversity and abundance of the brachiopod fauna declines sharply, and slope faunas are usually sparse with some trilobites and molluscs (and rela­ tively more pelagic elements) with increasing depth . This pattern of biofacies applies to clastic shelves in temperate regions . Diversity and biomass were generally lower in colder regions where there were fewer communities and the shelf was partitioned into fewer biofacies . In contrast, carbonate fades of tropical regions had very diverse faunas and a large biomass . Carbonate environments on broad epeiric platforms were often very extensive in the Lower

A (10

1 00 '" c .2 75 Cl..

L.U

50 25

c

'" 60

.2 40 c

20

B

Mean d iversity

A, Diagram to show the relative abundance of epifaunal and infaunal bivalves according to type of substrate in the Corallian. B, Diversity indices (using rarefaction method at 60 individual levels) of the macroinvertebrates in different substrates . 1 condensed limestone, 2 condensed ferruginous sediments, 3 patch reefs, 4 clays, 5 silts, 6 fine sand, 7 medium- coarse sand, 8 oolites, 9 fine-grained limestone, 10 coarse-grained limestone. Note diversity change between clays (4) and coarse sand (7) . (After Flirsich 1976 . ) Fig. 4

24

20 16

=

=

12 -

=

=

=

=

=

=

8

=

=

4

400

4 Palaeoecology

Palaeozoic, giving rise to regionally developed bio­ facies . However, in areas where reefs were developed, the biofacies could be very patchy.

Upper Palaeozoic biojacies. The broad patterns of diversity and abundance established in the Lower Palaeozoic were probably continued into the Upper Palaeozoic. The largest changes in the appearance of biofacies arose from the increasing importance of terebratulid and productid brachiopods . There was a significant change in Devonian biofacies caused by the Late Frasnian extinction (Section 2 . 13 . 3) of the Pentameridea, Atrypoidea, and Orthacea and the subsequent enrichment of the rhynchonellid and productid faunas (McKerrow 1978) . Mesozoic biojacies. The end-Permian mass extinction (Section 2 . 1 3 .4) destroyed the major part of the Palaeozoic benthic associations and new faunal as­ sociations established themselves in the Triassic. Mesozoic biofacies are generally dominated by bi­ valves with variable proportions of gastropods, brachiopods, echinoids, crinoids, and corals . The partitioning of level-bottom shelves into belts oc­ cupied by distinct benthic assemblages has not been clearly demonstrated in the Mesozoic, although there are obvious changes in faunal com­ position with increasing water depth . There is gen­ erally a correlation between taxonomic composition and substrate but, because some bivalves are eury­ topic, this correlation is not always pronounced. A detailed study of Corallian (Jurassic) palaeoecology (Fiirsich 1976) recognized 17 faunal associations oc­ cupying a range of environmental situations from nearshore to open shelf. Some of these associations showed a close correlation with substrate, others little or no such correlation. In general, deposit­ feeding bivalves had a preference for fine silts or argillaceous silts, but avoided clays; epifaunal sus­ pension feeders dominated condensed facies and patch reefs but avoided soft clay substrates; and infaunal suspension feeders were particularly com­ mon in clay substrates (Fig. 4) . Mean diversities of the fauna were high in condensed facies and clays, and generally showed a marked decrease in diver­ sity with increasing grain size (Fig. 4) . This diver­ sity- grain size relationship was interpreted to reflect an increasing environmental instability in

more energetic nearshore environments relative to the more stable offshore regions . Tertiary to Recent. The end-Cretaceous extinction (Section 2 . 1 3 . 6) modified the Mesozoic biota by the loss of ammonites, belemnites, inoceramid and rudist bivalves, and several groups of gastropods . In addition, the abundance of brachiopods was reduced and many echinoid taxa disappeared . Several groups which were often present i n the Cretaceous, but in subordinate proportions, diver­ sified in the Early Tertiary; predatory gastropods, Neogastropoda, polychaete worms, heterodont bi­ valves, the Veneracea, and Tellinacea all probably diversified at this time . Reef-building corals and associated algae also apparently diversified in the Eocene . Typical Tertiary biofacies are dominated by bivalves and gastropods, and from Miocene times had a taxonomic composition, at generic level, similar to biofacies found in the Recent. References Boucot, A. 1975 . Evolution and extinction rate controls. Elsevier, Amsterdam, Oxford. Bretsky, P.W. 1969 . Central Appalachian Late Ordovician communities . Bulletin of the Geological Society of America 80 , 193-212. Bretsky, P.W., Bretsky, S5. & Schaefer, P.J. 1977. Molluscan and brachiopod dominated biofacies in the Platteville Formation (Middle Ordovician), Upper Mississippi Valley. Bulletin of the Geological Society of Denmark 26, 1 1 5 - 132. Fiirsich, F.T. 1976 . Fauna - substrate relationships in the Corallian of England and Normandy. Lethaia 9, 343 - 356. House, M.R. 1 975. Faunas and time in the marine Devonian. Proceedings of the Yorkshire Geological Society 40, 459 -490 . McKerrow, W5. (ed . ) 1978. The ecology offossils . M.LT. Press, Cambridge, Ma. Moore, R.C 1949 . Meaning of facies . Memoirs of the Geological Society of America 38, 1 - 34 . Petersen, C C .J . 1 9 1 5 . O n the animal communities o f the sea bottom in the Skagerrak, the Christiana Fjord and the Danish waters. Report of the Danish Biological Station 23, 3-28. Sepkoski, J .J . , Jr. & Sheehan, P.M. 1983. Diversification, faunal change and community replacement during the Ordovician radiations . In: M.J5. Tevesz & P.L. McCall (eds) Biotic interactions in Recent and fossil benthic communities, pp. 673 - 717. Plenum, New York. Thorson, G. 1971 . Life in the sea . Weidenfeld and Nicholson, London.

4 . 19 Fossils as Environmental Indicators

4 . 19 . 1 Climate from Plants

Leaf margins. In modern vegetation the ratio of non­ entire (toothed) to entire (smooth) margined leaves correlates strongly with mean annual temperature (MAT) (Wolfe 1979; Fig. 2) . Generally, in the North­ ern Hemisphere a change of 3% in this ratio corre­ sponds to a change in MAT of 1°C. In the Southern hemisphere, with a higher proportion of evergreen taxa, a 4% change corresponds to 1°C . Because major tooth types had evolved by the Cenomanian, and because Cenomanian leaf margin ratios cor­ relate with palaeolatitude, this technique seems applicable from the early Late Cretaceous to the present. A minimum of 20 leaf species are required at any one locality to make this technique reliable, and taphonomic factors have to be taken into consideration.

R. A . SPICER

Vegetational physiognomy

Vegetation, unlike marine organisms, is directly exposed to the atmosphere . The physiognomy (structure and composition) of environmentally equilibrated (climax) vegetation is in large part con­ trolled by, and therefore reflects, climate (Wolfe 1979) . Interpretations of climate based on veg­ etational physiognomy, foliar physiognomy, or wood anatomy are more reliable for pre-Neogene studies than taxon-dependent climate signals (those used in Nearest Living Relative - NLR methods) . Fundamental vegetational types can be recognized in modern vegetation and, provided water is not limiting, correlate with temperature regimes (Fig. 1 ) . These vegetational types can be recognized with some confidence back to late Cretaceous (Cenomanian) times .

Leaf size. This is related strongly to temperature, humidity/water availability, and light levels . Large leaves occur in humid understories, and size decreases with decreasing temperature or precipi­ tation. Size classes are used to characterize veg­ etational types and to construct leaf size indices (which are used to characterize overall leaf size parameters for a given vegetational type) . In fossil assemblages leaf size suffers strong taphonomic bias.

Features o f leaves useful in determining palaeoc1imate

Angiosperm vegetative organs exhibit considerable morphological diversity and flexibility with respect to climate . The following features are those of angiosperms except where indicated:

Drip tips. Highly attenuated leaf apices occur most frequently in evergreen leaves in humid environ­ ments, and are particularly common in the under­ storey of multistratal rain forests . Drip tips may

30 �

� 20 ������ ::>

(J) 0.

E � 15 2 10 '" '" ::>

c '"

Correlation of fundamental vegetational types with temperature regimes . Fig. 1

�

:�::�--�==�����§I� Mean a n n u a l range of te m p e ratu re (OC)

401

402

4 � N o rt h e r n

-

' -

�-

-

'' -"

- -

''

--

,..

..

-

..

..

" "

-

:.�

_ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

'"

- - -

---

S o u t h e r n h e m i s p h e r-e

-

..

::;. 3:: ro ro ' CTO 3 '" '" 3:: '

h e m i s p h e re

... ... .. ... .... - - - - - - ------- - -

"

'" � - --- -"

Palaeoecology

---

-

- _ __ _ _ _ _

_ _ __ _ _ _ _ _ _ _ _ _

::;. � � 3 0 QJ

-

I

.. ... ... ...

E 10 c '" c '" (l)

�

O ,r-----,r----r--.--� 1 00

80

60

40

20

o

Perce n t e n t i re- m rt r g i n e d s p e c i e s

Correlation of ratio of non-entire (toothed) to entire (smooth) margined leaves with mean annual temperature .

Fig. 2

enhance drainage of surface water from the leaf and thus retard the growth of epiphytes .

2).

Leaf texture. Leathery (coriaceous) leaves typically are evergreen and predominate in megathermal and mesothermal vegetation (see Fig. Thin (char­ taceous) leaves are typically deciduous and are most common in micro thermal climax or successional mesothermal vegetation .

Leaf shape. Stream-side vegetation contains a high proportion of narrow (stenophyllous) leaves. Lobed or compound leaves (also associated with decidu­ ousness) occur with greatest frequency in suc­ cessional vegetation or under storey communities, and therefore warn of bias in the climate signal. Thick cylindrical leaves in any plant group are evi­ dence of aridity, growth in saline water, or an inefficient vascular system .

Leaf cu tides. In all terrestrial plant groups thick cuticles with numerous trichomes (hairs) are characteristic of plants adapted to desiccating con­ ditions (drought or salinity) . Sunken stomata, par­ ticularly if overarched by papillae, and low stomatal density are also indicative of water stress . Con­ versely, thin, smooth cuticles suggest water-rich conditions . Wood anatomy

Manoxylic (parenchymatous) wood (e . g . modern relict cycads) is frost-sensitive, while pycnoxylic (mostly composed of secondary xylem) wood (coni­ fers and angiosperms) is usually frost-resistant.

Tree rings. In situations where climatic conditions

vary frequently, pycnoxylic wood produces rings as a consequence of variations in growth rate . Rings may be produced on an annual basis where tem­ perature, light, or water availability fluctuates on a yearly cycle, or less regularly in environments with more erratic variations in growth conditions (e .g. sporadic droughts) . Annual rings consist of early (spring) wood with large cell lumina and thin cell walls that grade into late (summer) wood, in which the lumina are smaller and the walls thicker. Wide rings generally reflect benign conditions, but ring width is also a function of position within the tree (position within the trunk, or trunk versus branch) (Creber & Chaloner 1985) . High early wood - late wood ratios indicate a high rate of spring and summer growth followed by rapid onset of dormancy . At high latitudes this may be controlled by light rather than temperature . Inter-annual variations in ring width are de­ scribed using a statistic known as mean sensitivity:

XI

ms

1

=

-

n

1

12(XIXI+1+ 1 - XIXI) I , Xt+ 1

,, 1 = ,, - 1 L... 1= 1

+

where is a ring width, is the width of the adjacent younger ring, and n is the number of rings so measured in a sequence . Woods with a mean sensitivity of < 0.3 are termed complacent and in­ dicate uniform growing conditions from year to year. Sensitive woods (ms > 0.3) are typical of trees growing at the edge of their range and/or in vari­ able environments . Intra-annual variation is marked by false rings . Unlike true rings these often do not form complete circles as seen in cross-section, and may be gra­ dational to normal (usually early) wood on both concave and convex sides . False rings reflect tem­ porary trauma within the growing season caused by waterlogging of roots, low temperatures, or severe insect attack, for example . Frost rings are characterized by cell wall disruption due to freezing of cell contents . Pycnoxylic wood is known from the Late Devonian to the present. In angiosperm woods average vessel diameter divided by the frequency of vessels per mm2 in cross-section estimates the susceptibility of a tree to air embolism (formation of air bubbles and damage to conductive elements) caused by transpirational stress or freezing . As with leaves, the reliability of the climate signal in wood increases with sample size and taphonomic understanding .

4 . 1 9 Fossils as Environmental Indicators NLR methods Plant reproductive organs have little inherent cli­ matic signal but climate may be deduced from extra­ polation of the tolerances of their NLRs . Following Axelrod & Bailey (1969), four steps are required in NLR analysis : 1

NLR o f all taxa i n a n assemblage should be identified to modem genus level. 2 NLR determinations should also be attempted at species level (because generic tolerances are too broad) . 3 The average MAT and average mean annual range of temperature (MAR) are estimated based on habit 'preferences' of modem NLRs . 4 The effective temperature (average temperature at the beginning and end of a period free from frost or chill) and equability of the palaeoclimate are calculated using the average MAT and MAR.

References Axelrod, DJ. & Bailey, H.P. 1969. Paleotemperature analysis of Tertiary floras. Palaeogeography, Palaeoclimatology, Palaeoecology 6, 163- 195. Creber, G.T. & Chaloner, W.G. 1985 . Tree growth in the Mesozoic and early Tertiary and the reconstruction of palaeoclimates . Palaeogeography, Palaeoclimatology, Palaeoecology 52, 35-60. Wolfe, ].A. 1979 . Temperature parameters of humid to mesic forests of eastern Asia and their relation to forests of other areas of the Northern Hemisphere and Australasia, US Geological Survey Professional Paper No . 1 106.

4.19.2 Temperature from Oxygen Isotope Ratios T . F . ANDERSON

Introduction

Oxygen isotope ratios e80 : 160) of well preserved marine calcareous fossils are indicative of the tem­ perature of ancient ocean waters . This approach is based on the fact that the difference in 180 : 160

403

ratios between calcium carbonate and the water from which it precipitates is a function of tempera­ ture . Oxygen isotope ratios are expressed in the 0 notation :

Units are per mil or parts per thousand . The standard material for carbonates is PDB, a late Cretaceous belemnite from the Pee Dee Formation of South Carolina; for water, the standard is SMOW, i . e . standard mean ocean water (see Anderson & Arthur 1983) . Oxygen isotope palaeotemperatures for calcite can be calculated from: TOC

=

16.0 - 4 . 14� + 0 . 13� 2,

(2)

where � 0180 calcite (vs . PDB) - 0180 water (vs . SMOW) (Anderson & Arthur 1983) . Thus, 0180 of calcite increases as temperature decreases . Palaeo­ temperature estimates can be made with an uncer­ tainty of ± O . soC, because 0180 values are measured to a precision of 0 . 1 per mil . Factors other than analytical precision control the uncertainty in isotopic palaeotemperatures : 1 The manner in which isotopic fractionation be­ tween biogenic calcium carbonate and water varies with temperature must be known . Equation (2) applies to inorganic precipitation of pure calcite at isotopic equilibrium and to a number of low­ magnesium calcite fossil groups including bivalves, belemnites, brachiopods, and planktic foramini­ fera. Slightly different equations apply to preserved aragonite and high-magnesium calcite shells (Anderson & Arthur 1983) . In addition, physio­ logical effects during shell secretion in some organisms result in departures from equilibrium fractionation; notable examples are corals and echinoids . 2 I t i s necessary to estimate the 0180 o f the water in which the shell grew. In the hydrologic cycle, evap­ oration preferentially removes H2160 from water, while precipitation and runoff returns H2160. Local variation in the hydrologic balance of ocean waters of normal salinity can produce small variations in 0180. (The range for modem seawater is 2.S per mil . ) This effect i s normally ignored in estimating isotopic palaeotemperatures because hydrologic data on ancient ocean water is lacking. Also, because H2160 is preferentially stored in polar icecaps and conti­ nental ice sheets, oceans are enriched in 180 during glacial epochs relative to nonglacial epochs. For example, the growth and decay of continental ice =

404

4 Palaeoecology

sheets during the Late Quaternary produced excur­ sions of at least 1 per mil between glacial and inter­ glacial oceans . The effect of Palaeozoic glaciations on the 0180 of contemporaneous seawater was probably similar. 3 Reliable isotopic palaeotemperatures can be obtained only from those fossils that have been preserved from diagenetic alteration . Cemented or partially recrystallized fossils will generally give er­ roneous palaeotemperatures, because secondary carbonates reflect the temperature and isotopic composition of diagenetic solutions . Isotopic palaeotemperatures from the Cenozoic and Late Cretaceous

The most continuous record of marine temperature variations for the past 100 million years has been constructed from isotopic analyses of well preserved foraminifera in deep-sea sediments . Diagenetic alteration of foraminiferal tests is minor and rela­ tively easy to determine microscopically. In ad­ dition, the effects of continents on the temperature and 0180 of ocean water in the pelagic realm is minimal. The Quaternary oxygen isotope record of fora­ minifera shows oscillations with periods of about 10 5 years between 0180 maxima during glacials and 0180 minima during interglacials (see Savin 1977, fig . 8; Anderson & Arthur 1983) . Although the direction of these isotope shifts is qualitatively compatible with temperature changes, it is now generally accepted that the amplitude of Quaternary 0180 oscillation reflects changes in continental ice volumes more than changes in seawater temperatures .

0

';;0

Marine temperatures for the Tertiary and Late Cretaceous have been estimated from isotopic data on Deep Sea Drilling Project cores. Composite oxygen isotope records for planktic and benthic foraminifera from subtropical sites in the North Pacific illustrate the major features of palaeoclimatic changes over the past 130 million years (Fig. 1 ) . The planktic record reflects temperature and 0180 vari­ ations in low-latitude surface waters; the benthic record reflects conditions at the high-latitude source regions of deep-water masses. The data suggest general cooling in the Pacific over the past 100 million years . Temperatures of subtropical surface waters were evidently warmer in the Albianl Cenomanian and the Eocene than in intervening times . Temperature trends in deep waters are cor­ related with those of surface waters from the Middle Cretaceous through the Early Tertiary. However, bottom waters during this interval (especially during the Cretaceous) were considerably warmer than at present. In other words, the latitudinal con­ trast in ocean temperatures had increased during the Tertiary, resulting principally from apparent cooling at high latitudes. Abrupt positive shifts in the Tertiary benthic 0180 trend probably reflect the initiation (Eocene - Oligocene) and rapid expansion (Middle Miocene) of the Antarctic icecap, as well as a decrease in high latitude surface temperatures . Palaeotemperature trends from shallow-marine bivalves from northwest Europe (Fig. 2) are similar to those for Pacific low-latitude surface waters (Fig. 1), suggesting that global palaeodimatic changes were not obscured by the influence of continents on the temperature and 0180 of nearshore seawater. In contrast to the deep-sea record, the isotopic data from bivalves suggest that the shallow

-3

30

-2

25

-1

20

PC

0 10

'"

15

+1 10

Benthic foraminifera

+2

(l)

� a, �

+3 +4 +5

0

20

40

60

80 " Age (1 0 yrs)

1 00

1 20

E

(l)

�

u 0

5 5 0 0 1 40

Fig. 1 Oxygen isotope record of foraminifera from the central North Pacific for the past 130 million years . 6180 scale is shown on the left vertical axis; the corresponding temperature scales for modern and ice-free oceans (6 1 80 0 and -1 per mil respectively) are shown on the right vertical axis . (Reproduced, with permission, from the Annual Review of Earth and Planetary Sciences, Vol . 5 © 1977 by Annual Reviews Inc. ) =

405

4 . 1 9 Fossils as Environmental Indicators Alb

Bivalves

Oxygen isotope record of shallow-marine bivalves from northwest Europe for the past 100 million years . /)180 and tempera­ hIre scales are the same as Fig . 1 . (Data from Buchardt 1977 and Burdett, ]. & Arthur, M. pers . comm. 1987.) Fig. 2

35 40 re 30 35 25 30 20 25 E > 0 Z

� I

4 zones

�gwickii

Low

8 zones

...J

«

­ u c U! ::>

>­ u c U! ::>

0-

0-

�

�

u...

1

u...

0.0

0 .0

Facto r l o a d i n g s

Facto r l oad i n gs

1

B

1 .0

0.0

1 .0

Facto r l oa d i n g s

1

D

U!

F

U!

E f.=

E f.=

63 L- 1 .0

-L��__�m

__ __ __

0.0 Factor l o ad i n g s

1 .0

���____�

63 � - 1 .0

__ __ __

0.0 Facto r l o a d i n g s

1 .0

63 L- 1 .0

���L___-W

__ __

0.0

1 .0

Facto r l o a d i ng s

Fig. 1

The approach that yielded these results combines simulation modelling with a statistical analysis dependent on computer solution. The research question requires that the expected distributions of temporal covariation among clades generated by a random process be known . Because there is no analytical solution to the problem, a random branching process was used to generate 45 000 simulated monophyletic clades, where the differences between each clade's history are due to the random elements of the branching algorithm. Each evolutionary 'trial' of 90 such clades, allowed to evolve for 63 time steps (where 90 and 63 were chosen to match the empirical data of number of taxa and stratigraphic stages, respectively), was then subjected to Q-mode factor analysis (to match the method of analysis of the empirical data) . The frequency distribution of these 500 factor analyses are shown in A, C, and E which represent Factors I, II, and Ill, respectively. The stippled areas of B, D, and F represent the corresponding patterns not significantly different from expectations of a random branching process. (After Kitchell and MacLeod 1988 . )

coupled logistic) the exploration of behaviour involves only kinetics . Kinetics are more inherently intuitive than dynamics, which incorporates feedback. A more ambitious undertaking of theory develop­ ment and simulation exploration involving feed­ back, nonlinearity, and complexity, is the work of DeAngelis et al. (1985) on potential coevolutionary dynamics, a series of studies motivated by (but not confined to) palaeontological questions . What this work has gained is a new intuition to replace the old expectation of linear escalation . In addition, it

has shown the salient features of nonlinear dynam­ ics (Fig. 2) : how the behaviour of the individual parts are qualitatively different from the behaviour of the whole; and the influence of evolutionary change on itself, where 'playing the game changes the rules' . Computer-intensive statistical inference

Science is argument focused on the differential credibility of competing hypotheses . Palaeontology, a historical science, must make argument of process

6 . 1 Computer Applications in Palaeontology (where the interest generally lies) from evidence of pattern (where the information generally lies) . Fortunately, hypotheses of process contain predic­ tions of pattern, and so there can be effective argu­ ment provided by historical pattern. Statistics similarly deals with an end product (namely, some observed set of data) and makes arguments, among others, regarding what factors are, and to what extent, causally responsible . The power of computing is currently changing the field of statistics . In general, the computer has allowed even classical statistical methods to be applied to what would once have been unmanage­ ably large data sets . Palaeontology has benefited from this increased capability; the compilation and analyses of large databases have changed the tenor of arguments, for example, on patterns of diversifi­ cation (Section 2 . 7), extinction (Section 2 . 12 . 3 ), rates of phenotypic evolution, and taxonomic turnover (Section 2 . 1 1 ) . Palaeontology, however, has been hampered by the limits of classical statistics : the need to make a priori assumptions about the form of the probability distributions that are sampled by the data, and the restriction to measures whose theoretical properties are simple enough to have analytical proofs . These limits have been trans­ cended recently by computer recursion techniques that replace analytical solutions with enormous numbers (105 - 10 9 ) of computations . Boots trapping represents such a computer­ intensive method, described as the 'substitution of raw computing power for theoretical analysis' (Efron & Gong 1983) . Using the traditional approach, one would hypothesize a process (or model) and deduce (or simulate) its behaviour, to compare these outcomes with empirical data . The boots trapping approach is logically different. Bootstrapping derives its power from the assumption that the empirical sample provides an informative 'glimpse' of the real or underlying process. This empirical sample is resampled with replacement a large number of times, with the statistic(s) of interest calculated for each boots trapped sample, in order to construct the bootstrapped probability distribution, against which the empirical sample is compared. The boots trap is especially useful in cases where the probability distribution is unknown, or if the data violate certain (particularly parametric) distri­ butional assumptions . A large number of palaeonto­ logical cases fall into these categories . The boots trap method has been applied i n palae­ ontology to problems that include estimating confi­ dence limits around phylogenies, assessing patterns

495

of extinction probability and the shape of clade diversity histories, and the significance of differ­ ences in rates of evolution . A problematic feature of much palaeontological data for such methods is that the data are often ordered by (geological) time . The original bootstrap method was designed for data that are identical and independently distributed; time series do not satisfy this criterion. A method applicable to palaeontological (time series) data sampled at intervals that may or may not be constant is now available . In particular, the method recog­ nizes the necessity of coupling the magnitude of evolutionary change with the magnitude of the time interval over which that change is measured (Kitchell et al. 1987) (Section 2 . 1 1 ) . The method also works with two types of time series: those in which a change in the time series is recognized on the basis of independent criteria, and those in which a segment of the time series is identified as excep­ tional simply on the basis of that change (post hoc recognition) . Such computer-intensive methods of statistical inference will undoubtedly play an in­ creasing role in fields such as palaeontology that rely little on laws, axioms, and deductions to gain understanding. Sensitivity of initial conditions

Palaeontologists have used computer simulation methods to generate samplers of patterns produced by a variety of random processes, because much of the evidence in palaeontology since the nineteen­ seventies is pattern data. Mathematicians and statis­ ticians had already shown that random processes are capable of producing orderly pattern . Many of the properties of random processes were known by analytical solution. However, the ability to display these randomly-produced patterns graphically and by simulation did most to convince palaeontologists of the fallacy of the expectation that orderly patterns required deterministic explanations . It was shown that palaeontologically significant patterns, such as some trends and the topology of branching patterns, could be produced by random models (see review by Raup 1977) . The purpose of this work was both to enlarge the intuitive understanding of palaeonto­ logists so that they would not incorrectly equate pattern with non-randomness, and to better identify non-randomly produced patterns . The opposite side of this coin, namely that com­ pletely deterministic processes lacking randomness can nevertheless produce random patterns, required the computer for its development . Until recently

6 Infrastructure of Palaeobiology

496 X �

Q) C -'


..c

I

"

..c

'" '" Cl,) C -'


2 lJ

-

-

--

-

-- -

----

- -- - - -

- - -

-

- -- - -

-

�1 1 .0 -

- - -

-

-- -

- - -

��-' � _1 0 . 5 )0.6

-

0.9 1 . 0 '---'-_.....I..._-'-_L..--'-_-'--_� 2 0 0 3 0 0 400 5 0 0 6 0 0 700 800 900 G rowt h st rategy d a y o f s w i tc h i n g t o r e p rod u ct i o n , X,

o O r--r--'--'-I-"--"--�

0.1 0.2

=

=

0.3

� 0.8

..c

2 lJ

Fig. 2 This example of computer­ intensive analysis involves simulation modelling of a set of simultaneous algorithms that represent the opposing interests of a species interaction . The axes represent trade-off energetic options of a prey species (vertical axis morphological allocation option; horizontal axis repro­ ductive strategy option) . The contours depict the fitness con­ sequences (in terms of total expected reproduction of the prey over its lifetime) for all possible combinations . Both the magnitude and positions of the contours change dramatically as predation intensity is increased, from zero in panel A to increasingly high values towards panel D, despite there being no change in the predatory strategy of prey selection. (From DeAngelis et al . 1985 . )

0 . 9 f--

1 . 0 '---'-_.....I..._-'-_L..--'-_-'--_� 200 300 400 500 600 700 800 900 G rowth st rategy d a y of s w i tc h i n g to r e p ro d u c t i o n , X,

1 . 0 '---'-_-'--_�_L-_J_-'-_.....I...� 2(J0 300 400 500 600 700 8 00 900 G rowth s t rategy d ay of s w i tc h i n g t o r e p rod u c t i o n . X l

within all the sciences, complex patterns were con­ sidered to be the consequence of complex causes . It has now been shown, however, that apparently random behaviour can derive from even simple deterministic processes. A small difference in initial conditions, for example, can lead to unexpectedly divergent behaviours . The term 'chaos' has been applied to such patterns and processes, to dis­ tinguish them from randomness. In chaos, the dis­ order is ordered . Such ordering is apparent in the detail of the patterns, a detail made increasingly evident by computer techniques and images . In palaeontology, it was shown that the most simple model of diversification, and the one being applied to empirical analyses of taxonomic diversity,

had chaotic behaviour . Using computer simulation runs to map the surprising array of behaviours and their abrupt and ordered thresholds, Carr & Kitchell (1980) showed that the 'coupled logistic' model of Sepkoski (1979) could produce not only logistic patterns of diversity change with time but also extremely complex and chaotic patterns of diversity change . In this latter case, the oscillations are driven internally, without external perturbation. Whereas earlier work, by warning that a high degree of order can be generated by purely random pro­ cesses, had tried to dispel the palaeontologist's bias that randomness implies a random pattern, Kitchell & Carr (1985) warned against the bias that deter­ minism implies an ordered pattern . They showed

6 . 1 Computer Applications in Palaeontology that even a completely deterministic and remarkably simple process can produce patterns of bewildering complexity. The understanding of chaotic behav­ iours is now being pursued in a number of cognate fields within biology, physics, and chemistry, promising to revolutionize our collective under­ standing of a class of complex phenomena, until recently unknown .

Phylogenetic inference The methodology of inferring phylogenetic (evol­ utionary) relationships among organisms has become both increasingly explicit and empirical (Section 5 . 2) . Phylogenies are constructed from data on the distribution of characters (the empirical component, such as that resulting from morpho­ metric studies), according to some criterion made operational by a computing algorithm (the explicit component) . These criteria and associated algo­ rithms used to form phylogenetic hypotheses rely either on parsimony methods, maximum likelihood methods, or compatability methods; reviews that examine the fundamental assumptions of each method were given by Felsenstein (1983) . These methods are derived from a class of prob­ lems in mathematics and statistics that focus on maximizing or minimizing some aspect of the data. In such optimality methods, the assertion is not that the historical process of evolution is optimal. Rather, optimization methods are used to choose among all tree topologies generated by an algorithm for a given set of data. Parsimony methods, for example, evaluate phylogenetic hypotheses on the basis of number of homoplasies (convergences and parallel­ isms); the 'best' genealogy is the one of minimum homoplasy. Because the criteria for evaluating phy­ logenies are unique to the method, comparing methods in terms of finding the 'true' genealogy is not possible . Instead, types of parsimony, maximum likelihood, and compatability algorithms can be compared with one another in terms of a practical goal (efficiency in computer time) and a method­ ological goal (minimizing tree 'length' or the required independent origins of each character) . Although small data sets may be analysed by hand (using the 'brute force' method of generating all possible cladograms; there are 15 possible for four taxa), large data sets require computer-assisted analyses (there are more than two million clado­ grams for only nine taxa, and more than 1020 clado­ grams for 20 taxa) . Even the latter is too much for computer analysis . This raises an interesting situ-

497

ation; there is no exact solution to the problem of finding the minimum tree for even moderate-sized data sets . This problem may not be soluble : among mathematicians, there is agreement that NP- (not polynomial)-complete optimization problems (such as these) cannot be solved given current approaches and algorithms . Within palaeontology, phylogenetic approaches principally make use of morphological character data. An interesting discussion was pro­ vided by Gauthier et al. (1988) who showed, using both palaeontological and neontological character data, the importance of palaeontological data. Strato­ cladistic methodology may also prove useful as a means of integrating both character data and strati­ graphic data in an analysis of phylogeny, where a total parsimony debt (summed from morphology and stratigraphy) serves as the minimization criterion. A problem in need of redressing is that most palaeontological analyses of taxonomic data sets (e . g . patterns of diversity change, extinction, rates of evolution) have made use of data currently avail­ able . Much of these data do not reflect the meth­ odology discussed above . As a recognized con­ sequence, monophyletic and non-monophyletic groups are not distinguished from one another. This presents a problem of interpretation since 'monophyletic groups have a unique history that exists and is to be discovered, whereas paraphyletic groups may start off with a unique history, but their boundaries are adjusted a posteriori and they are in part a human invention' (Benton 1988) .

Computer-aided vision systems The most severe restriction on palaeontology today is the lack of adequate databases to test hypotheses of interest. It is likely that major advances in the future will be made in the rapid acquisition of morphological and character-state data from auto­ matic vision systems . Although the systems described below have not yet been widely used in palaeontology and are still in stages of development, the future of advanced computer techniques in palaeontology will undoubtedly move in these directions . With laser disc technology, it is now possible (and currently in use in some research laboratories) to store all known species' images (e . g . holotypes) and their descriptions, and to make use of them with a dichotomously driven, interactive algorithm to resolve the identification of an unknown species. This technology permits exact comparisons on the

498

6 Infrastructure of Palaeobiology

screen. Access is also virtually instantaneous, with more than 50 000 analog images currently capable of being stored per disc and with the ability to access more than one disc at a time . The system utilizes answers provided by the user to a computer-driven key to select the most likely known species . It then automatically compares the unknown image with these selected known species, making comparative diagnostic measurements. Because of the interactive nature of the algorithm, the user maintains control of the final decision. Programs designed for palaeontological appli­ cations that make use of artificial intelligence pro­ gramming have also begun to be developed (e . g . Riedel in press) . These programs explicitly attempt to deal with objects that are naturally variable (organ­ isms), and may be made even more variable by preservational processes, yet are members of a single category (the species) . These systems use character­ state descriptions entered by the user and work within a hierarchy of character-states necessary for discrimination between possible species . As above, the final result is a 'narrow as possible' reporting of species that have these characters . Algorithms associated with image analysis sys­ tems are also now available (and being developed) for converting data from serial sections of any fossil (whether actually sectioned or not) to three­ dimensional models of that fossil, thereby allowing the user in many instances to bypass the building of physical models . The reconstructed three­ dimensional form can also be viewed from all per­ spectives by rotation and movement simulation algorithms .

Acquiring morphometric data by image analysis Palaeontologists, of necessity, rely on morphological data to make evolutionary inference . A recurring problem in the sciences is that the theories of a field may occasionally far exceed the capacity of that field to acquire and analyse data necessary for evaluating those theories . Such a situation occurred in palaeon­ tology, e . g . with the proposal of punctuated equilib­ rium and its associated prediction of morphological stasis . The imperative quantitative data on morpho­ logical change within and between species, and over time and geography, were not copiously avail­ able . Much of the problem stemmed from the diffi­ culties of acqumng quantitative data on morphology in a rapid and accurate manner. Widespread interest within numerous fields in the study of biological shape and its transformation

has resulted in a series of important advances . In terms of technique, advances in computer tech­ nology have made possible increasingly powerful image analysis systems that combine image acqui­ sition and image processing capabilities with pattern recognition analyses. Such image analysis or optical pattern recognition systems have made the acquisition of quantitative data on morphology rapid, accurate, and affordable . The field o f morphometries has been redefined recently as 'the analysis of biological homology as well as geometric change' (Bookstein et al. 1985) . Morphometrics is relevant to questions of phylogen­ etics, ontogenetic trajectories and their evolutionary potential for heterochrony, patterns of anagenesis and cladogenesis, ecophenotypy, and morphologi­ cal integration. Such analyses are particularly informative when they combine hypotheses of phylogenetic descent with hypotheses of morpho­ logical (character) transformation. Reviews of methodology and examples of the application of outline methods and landmark methods were given by Lohmann (1983) and Reyment (1985), respectively. The approach rec­ ommended by Bookstein et al. (1985) focuses more on the dynamics of change in shape . Analyses begin with a study of the major dimensions of morpho­ logical variation in time and space that characterize each species . Analytical procedures determine which parameters contribute most to intraspecific characterization and to interspecific discrimination within respective geographical and temporal con­ texts. A recent application of outline and landmark methods was given by Stanley & Yang (1987) who assessed the rates of morphological evolution in separate lineages of Neogene bivalves . Schweitzer et al . (1986) used the same basic techniques to evaluate the relative contribution of development (heterochrony) and structural regulation in two closely related species .

Prospects Palaeontology today is actively engaged in computer­ aided research programs . The evolution of the interaction between palaeontology and computer technology is following much the same path as that of the evolution of the human brain, as we currently understand it. The computer has not simply resulted in an increase in the speed, efficiency, and size of the problems we analyse . It has introduced novelty or true innovation. It is well recognized that the biological and evolutionary sciences deal with a much £reater de£ree of comnlexi tv in th eir svstpms

6 . 2 Practical Techniques of study than do the physical sciences. Computer techniques are beginning to open up the field of study of complex systems and, through vision sys­ tems, to relieve the human investigator of some of the effort in amassing empirical data .

References Benton, M.J. 1988 . Mass extinction in the fossil record of reptiles: paraphyly, patchiness and periodicity (?) . In: G.P. Larwood (ed . ) Extinction and survival in the fossil record, pp . 269 -294. Systematics Association Special Volume, No. 34. Bookstein, F . , Chernoff, B . , Elder, R . , Humphries, J . , Smith, G. & Strauss, R. 1985 . Morphometries in evolutionary biology. The Academy of Natural Sciences of Philadelphia, Philadelphia. Carr, T.R. & Kitchen, J.A. 1980. Dynamics of taxonomic diversity. Paleobiology 6, 427-443 . DeAngelis, D . L . , Kitchen, J.A. & Post, W.M. 1985 . The influence of naticid predation on evolutionary strategies of bivalve prey: conclusions from a model. American Naturalist 126, 817-842 . Efron, B. & Gong, G. 1983 . A leisurely look at the bootstrap, the jackknife, and cross-validation. American Statistician 37, 36-48. Felsenstein, J. 1983 . Parsimony in systematics: biological and statistical issues . Annual Review of Ecology and Systematics 14 , 313-333. Gauthier, J., Kluge, A . & Rowe, T. 1988 . Amniote phylogeny and the importance of fossils. Cladistics 4, 105 - 209 . Gleick, J. 1987. Chaos: making a new science. Viking Penguin,

499

New York. Judson, S. 1980. The search for solutions . Holt, Rinehart & Winston, New York. Kitchen, J.A. & Carr, T.R. 1985 . Nonequilibrium model of diversification : faunal turnover dynamics. In : J.W. Valentine (ed . ) Phanerozoic diversity patterns: profiles in macroevolution, pp . 277- 309 . Princeton University Press, Princeton. Kitchen, rA. & MacLeod, N . 1988 . Macroevolutionary inter­ pretations of symmetry and synchroneity in the fossil record. Science 240, 1 190- 1993 . Kitchen, J . A . , Estabrook, G. & MacLeod, N. 1987. Testing for equality of rates of evolution . Paleobiology 13, 272-285. Lohmann, G.P. 1983. Eigenshape analysis of microfossils : a general morphometric procedure for describing changes in shape . Mathematical Geology 1 5 , 659 - 672. Raup, D.M. 1977. Stochastic models in evolutionary paleon­ tology . In: A. Hanam (ed. ) Patterns of evolution, pp. 59 - 78 . Elsevier, New York. Reyment, R.A. 1985 . Multivariate morphometrics and ana­ lysis of shape . Mathematical Geology 17, 591 - 609 . Riedel, W.R. 1989. Identify: a Prolog program to help identify variable things. Computers and Geosciences (in press) . Schweitzer, P . N . , Kaesler, R.L. & Lohmann, G.P. 1986 . Onto­ geny and heterochrony in the ostracode Cavellina Coryen from Lower Permian rocks in Kansas. Paleobiology 12, 290-301 . Sepkoski, J.J., Jr. 1979 . A kinetic model of Phanerozoic taxo­ nomic diversity. 11. Early Phanerozoic families and mul­ tiple equilibria . Paleobiology 5, 222 - 251 . Stanley, S.M. & Yang, X. 1987. Approximate evolutionary stasis for bivalve morphology over millions of years : a multivariate, multilineage study . Paleobiology 13, 1 13 - 139.

6 . 2 Practical Techniques 6 . 2 . 1 Preparation of Macrofossils P . J . W H Y B R O W & w. L I N D S A Y

Mechanical methods A rock is invariably physically weakened by the presence of fossils, usually because the chemical constituents of fossils differ from those of the en­ closing matrix. For at least three centuries, palaeon­ tologists have exploited this difference by using percussion methods, normally a hammer and a chisel, to expose and to collect fossil material. Fol­ lowing the introduction of electricity into museums and universities in the nineteenth century, power tools were developed that 'automated' the basic

manual techniques . Today, three mechanical tech­ niques are widely used in palaeontology labora­ tories: percussive, grinding, and abrasive (Rixon 1976) .

Percussive and grinding techniques. Percussive electric or pneumatic engraving pens (Fig. 1) are hand-held and equipped with a tungsten carbide tip . Invari­ ably the tip supplied by the manufacturer is too coarse for most preparations and has to be substi­ tuted by tungsten carbide rod welded onto the oscillating shank of the pen . The fitting of the rod also enables a choice of either chisel or pointed tips to be fashioned . Before commencing preparation not only should the concealed morphology of the fossil be imagined (by reference to published in­ formation concerning similar fossils) but also the petrology of the matrix must be investigated (in case acid techniques can be better utilized) . If

500

6 Infrastructure of Palaeobiology state in a vibrating pressure vessel, through a nozzle of small diameter. Various hardnesses of powder can be used, ranging from sodium bicarbonate to the cast iron shot used in large industrial machines . Similarly, various diameters of nozzle can be selec­ ted . The abrasive action depends on particle size and the amount of gas pressure used . Exposed parts of a fossil can be protected by a coating of rubber latex from any polishing effect of the powder, and a box with a dust extraction system protects the oper­ ator from possibly hazardous particulates . A binocu­ lar microscope is essential for this work to see the degree or variability of abrasion of the rock.

Fig. 1 A hand-held, pneumatic engraving pen used to remove rock matrix. In the foreground is a heavy duty pneumatic chiseL

the rock cover is excessive, it can be removed by grinding . Diamond or carborundum wheels and burrs used in dentistry are ideal; for larger blocks, parallel grooves are cut using a pneumatic diamond saw and the thin rock wedges then removed by percussive methods . All preparations should be carried out at high magnifications using a binocular microscope so that the fossil -rock interface can be easily seen; a cold-light, fibre optic light source is invaluable for this (especially a system with contrasting colour filters) . The position of the per­ cussion point should ideally be at right angles to the plane of the fossil surface being exposed . The degree of force required to chip or flake away the rock and leave an unmarked specimen comes about by trial and (infrequently) error. Extreme care must be taken when microbedding planes pass through and around a fossil as flakes may contain part of it. Extensive preparation gradually weakens the structural integrity of a fossil but the percussive force used normally remains constant. Therefore, the specimen must be supported firstly by a shock absorbing cushion (such as a sandbag) and secondly by embedding in a water soluble polyethylene glycol wax of high molecular weight. For supporting delicate areas of a vertebrate skull, this wax is essen­ tial and can itself be strengthened while in its fluid state by the addition of surgical gauze (Whybrow 1982) .

Abrasive techniques. 'Airbrasive' or 'sand-blast' machines are quick and effective aids for removing rock that is softer than the fossil . An inert gas (compressed air, nitrogen, or carbon dioxide) propels an abrasive powder, which is kept in a fluid

Chemical methods

Rocks and the fossils they enclose do not always respond well to mechanical techniques . The hard­ ness of an ironstone or some limestone matrices may prohibit mechanical preparation, while the complexity or abundance of fossil remains may defy methods reliant on manual dexterity . As with mechanical methods, chemical methods aim to remove the matrix without damaging the specimen. However, in both cases, there are occasions when the information required can only be obtained by destroying the fossil and retaining the natural impression left in the rock. Chemicals used in fossil preparation are chosen for their ability to disrupt or dissolve the rock matrix, but they must achieve this without causing the same effect on the fossil. Such differentiation is determined by the chemistry of both rock and fossil. Furthermore, the long-term conservation of the fossil in a collection, with all the hazards associated with handling, must be considered .

Chemical disruption . Water, sometimes in conjunc­ tion with a detergent, readily breaks down some soft shales and muds . The clay minerals swell as the strongly polar water impregnates their structure . Detergents and other surfactants assist the process by reducing surface tension at the clay-water inter­ face . A similarly disruptive effect occurs in the presence of hydrogen peroxide (H202) . Solutions of H2 02 are unstable and deteriorate giving off oxygen . In the presence of alkalis, rough surfaces, and metals, the process is accelerated . In rock matrices the oxygen bubbles released within the pores disrupt the sediment and weaken the matrix (see also Section 6.2.2).

Sequestrants and chelating agents. Polyphosphates, such as sodium hexametaphosphate (NaP03)6, act

501

6 . 2 Practical Techniques as water softeners, sequestering the calcium, mag­ nesium, and iron salts present. Clayey and muddy sediments are broken down in solutions of poly­ phosphates. In a manner similar to that of water softeners, ch elating agents form stable complexes of metalic ions (such as calcium and magnesium) in rock forming minerals . Ethylene diaminetetracetic acid and its sodium salts in solution can corrode rock matrices, but it will also attack fossil material and careful control is therefore required .

Acids. Acids are extensively used in chemical methods of preparation (Lindsay in Crowther & Collins 1987) . Hydrochloric acid was used in the late nineteenth century to dissolve limestone con­ taining carbonized graptolites. Subsequently hydro­ fluoric, nitric, formic, acetic, and thioglycollic acids have been used in both vertebrate and invertebrate palaeontology. Hydrofluoric and nitric acids are employed for the maceration of sediment samples containing fossil pollen (Section 6 . 2 . 2) and pose particular problems of safety. The development of vertebrate material using aqueous solutions of acetic acid was first carried out in the nineteen-forties and followed from earlier techniques devised at the British Museum (Natural History) (Rixon 1976) . Acetic acid is the most commonly used acid for this work and is readily controlled and reasonably safe at low concen­ trations . Used in solutions of 1 - 10%, the reaction between the acid and calcium carbonate in the matrix occurs more readily than that between the acid and phosphates in fossilized bone (Fig. 2) . The differential rate of dissolution is controlled by vary­ ing the length of immersion time and the acid concentration . The time of exposure to acid at each step of the process may vary from a few hours to several days, and the development of a specimen may take years to complete . Bone that undergoes prolonged exposure to acid will be significantly affected; for this reason the dissolution of the matrix is interrupted regularly to wash, dry, and lacquer any newly exposed bone .

Consolidants and adhesives Consolidation (hardening) of a specimen must be carried out during preparation in order to conserve it for subsequent study. A number of adhesives and consolidants are used; they should be reversible in the long term as further work on a specimen may be required . In mechanical preparation the surface of the fossil is coated with a consolidant as the rock is removed in order to prevent fractures caused by

Fig. 2 The partially exposed, post cranial skeleton of the Jurassic dinosaur Scelidosaurus harrisoni during preparation with acetic acid.

I I I I I I

Anterior skull and jaw elements of the Lower Cretaceous dinosaur Baryonyx walkeri after mechanical and chemical preparation. Scale in cm . Fig. 3

any excessive vibration (Fig. 3) . Polyvinyl butyral resin, dissolved in a variety of solvents, has now replaced polyvinyl acetyl resins and serves as an adhesive when dissolved in ethyl acetate . Poly­ methyl-methacrylate, also dissolved in ethyl acetate, is a useful adhesive but shrinks markedly on drying and should never be used as a consolidant. Supplied as a powder monomer with a liquid polymer cata­ lyst, polymethyl-methacrylate effectively seals wide cracks . Cynoacrylate adhesives are effective for the fast repair of small pieces of fossil, but their long­ term stability is at present poorly understood and they are practically insoluble when set. Chemical methods of preparation require adhesives and consolidants that protect the fossil from chemical attack as well as supporting and strengthening it.

6 Infrastructure of Palaeobiology

502

Polybutyl-methacrylate is used as an acid resistant con solid ant and can withstand long periods of immersion in acids . Polymethyl-methacrylate as an adhesive is similarly resistant to attack by organic acids; the cynoacrylates also seem to be unaffected. In all methods of preparation, which by necessity expose the fossil to risk, good records must be kept (Rixon 1976) . Photographs, drawings, and written descriptions are essential and must be prepared as the specimen passes through various stages of treat­ ment. Their value can only be appreciated when a dismembered fossil needs to be reassembled .

References Crowther, P.R. & Collins, C]. (eds) 1987. The conservation of geological material. Geological Curator 4, 375 -474. Rixon, A . E . 1976. Fossil animal remains: their preparation and conservation . The Athlone Press, London . Whybrow, P.J. 1982 . Preparation of the cranium of the holo­ type of Archaeopteryx lithographica from the collections of the British Museum (Natural History) . Neues Jahrbuch fUr Mineralogie, Geologie und Paliiontologie Mh H3, 184- 192.

6 . 2 . 2 Extraction of Microfossils R . J . A L D RI D G E

Extraction techniques have been developed princi­ pally to recover microscopic fossils from rock samples, but may also be adopted for larger speci­ mens . A variety of chemical and mechanical pro­ cedures for rock disaggregation are employed, dependent upon the composition of the rock and of the fossils sought. Residues from these processes are often large, and some concentration of the micro­ fossil specimens may be required . Many of the chemicals used in dissolving samples and in concen­ trating residues are highly hazardous or toxic and the safety aspects of all techniques should be fully investigated before they are applied . Full attention must be given to hazard warnings given by the suppliers of chemicals .

Releasing microfossils from rocks

Calcareous rocks. Limestones, dolomites, and cal­ careous clastic rocks can be broken down with dilute organic acids (e .g. acetic acid, CH3COOH;

formic acid, (HCOOH) to release microfossils com­ posed of calcium phosphate (conodont elements, fish remains) or with resistant organic walls (sco­ lecodonts, chitinozoans, palynomorphs) (Fig. 1 ) . Some workers crush the samples into 1 - 3 cm chips, but this is only necessary for very impure lime­ stones . Standard procedure is to place the sample in a polythene bucket or beaker which is then filled with warm, 10- 15% acetic acid; formic acid acts more rapidly and may be used at higher concentra­ tions, but is more corrosive and hazardous . Phos­ phatic material may be attacked by acetic acid in the absence of calcium acetate to buffer the sol­ ution, so powdered calcium carbonate should be added to samples with low lime content. Alter­ natively, samples may be buffered by using a sol­ ution comprising 7% concentrated acetic acid, 63% water, and 30% of filtered liquid remaining after digestion of previous samples . Hydrochloric acid (HCl) dissolves phosphate, but may be used at a concentration of about 10% to recover organic-walled microfossils and siliceous (e . g . radiolarians) or silicified material. When buf­ fered by calcium acetate, HCl can be used to extract phosphatic, siliceous, and organic specimens from a single sample, but there is always a risk of damage to the phosphate, especially when all the limestone is allowed to dissolve . When effervescence fades or ceases, the sample is sieved; the mesh sizes of the sieves employed are dictated by the sizes of the microfossils sought. For conodont elements, an upper sieve of 1 mm mesh and a lower of 75 [tm are adequate, but chitinozoans and palynomorphs require much finer bottom sieves, down to 5 [tm. Undissolved rock remaining on the upper sieve is placed in new acid solution, while the sieved residue is dried and retained for concentration and picking. There is no easy technique for recovering cal­ careous microfossils from calcareous rocks . Soft limestones and marls may be treated in a similar way to soft shales, but for hard limestones and chalks only crude mechanical methods are available . Normally, these involve pounding the moistened sample with a pestle in a mortar, followed by wash­ ing and concentration . An intermediate step is sometimes inserted in which the pulverized sample is washed into a container and placed in an ultra­ sonic cleaner for a period of two minutes to two hours . Delicate microfossils will not survive these techniques and are best studied in thin section. The procedure may be successful, though, for calcareous nannofossils such as coccoliths .

6 . 2 Practical Techniques

. : . .. . . . . D i ssolve calcareo u s s a m p l e s i n 1 0-1 5 % acet i c acid

503 S o a k p re-d r i e d s h a l e o r clay � s a m p l e s i n petro l e u m e t h e r o r s i m i l a r solvent fo r at least one hour

1 \

n ��'�

Ret u rn u n d i s so lved rock to b u cket w i t h fresh aCid

o_� � L

1

III

1 /Wet s l eve

_o_�

C ov e r s a m p l e I m med i at e l y w i t h hot wate r and l eave fo r at l east D i saggregated res i d u e o n e h o u r

I.�. . �. . .� . .� . �. . � . ?'. . �. I

Fig. 1 Techniques for extracting and concentrating microfossils from limestones, calcareous clastic rocks, and soft or partly indurated mudrocks .

I

�

.0, 0 0,0 0.0

. . . .. . . . . . .

D ry fi n e res i d u e a n d sepa rate i n to l i g h t a n d heavy fract i o n s u s i n g b ro m o fo r m , tetra b r o m oet h a n e , o r sod i u m p o l ytu n gstate s o l u t i o n ; was h res i d u es i n app rop r i ate solvent a n d d ry

---, r•

E

er �

I \ \ \

Where necessary , fu rt h e r c o n c e n t rate � res i d u e s u s i n g e l ect r o m a g n e t i c � separat i o n o r a d d i t i o n a l h eavy l i q u i d s Hand-pick m i c rofoss i l s under m i c roscope

Argillaceous rocks. Soft or partly indurated clays and shales may be disaggregated by a number of tech­ niques . A relatively gentle procedure involves the use of petroleum ether, paraffin, or similar solvent on thoroughly pre-dried samples (Fig. 1 ) . All of these solvents are highly flammable, and due regard must be given to fire risks . The rock is soaked in solvent for at least one hour; the solvent is then poured off and the rock immediately inundated with hot (not boiling) water. The clay is reduced to an uncohesive, muddy slurry, which then can be wet-sieved as appropriate . Black shales and other mudrocks that do not respond to this treatment may disaggregate on immersion in a 10- 15% solution of hydrogen peroxide (H202) in water (see also Section 6 . 2 . 1 ) . The reaction involves the oxidation of organic matter, which may also be accomplished by other oxidizing agents, such as sodium hypochlorite (NaClO) . Hard clays may also disintegrate when boiled in water with a dispersing agent. Those commonly used include a few grams of sodium carbonate (Na2C03) or 20% sodium hydroxide (NaOH) . Some

samples respond to boiling in the detergent Quat­ ernary '0', with a 20% solution added to boiling water containing the sample . A combination of techniques may be applied, perhaps involving treat­ ment with buffered acetic or formic acids for samples containing some calcium carbonate . Mechanical disaggregation may sometimes be achieved by alter­ nate freezing and thawing of samples soaked in water, or by boiling the rock in sodium thiosulphate (Na2S203.5H20), which will crack the shale apart as it crystallizes when allowed to dry.

Sandstones. For most microfossils there is no tech­ nique for extraction from sandstones or siltstones, unless the rock is poorly-cemented, when mechan­ ical methods may be successful, or calcareous, when acids may be employed . For organic-walled micro­ fossils, palynological techniques (below) may be tried, but palynomorphs are not normally well preserved in coarse clastic rocks . Cherts. Phosphatic microfossils, such as conodont elements, can be recovered from cherts and other

504

6 Infrastructure of Palaeobiology

siliceous rocks using dilute hydrofluoric acid (HF) . The sample is crushed into 1 - 5 cm fragments, any carbonate removed with acetic acid, and the frag­ ments placed in 5 - 10% HF in an acid-resistant plastic container in a fume cupboard . After 24 hours the HF is decanted off and neutralized with calcium hydroxide; the residue is first washed with dilute HCl, then several times with water before being sieved and reprocessed as necessary . The technique works through fluoridization of the apatite of the conodont elements and is accompanied by some fracturing and distortion of specimens . Hydrofluoric acid is extremely dangerous and must be used in properly designed fume cupboards with the handler wearing full protective clothing.

Concentration techniques Residues from disaggregation procedures can be concentrated into light and heavy fractions by using various heavy liquids . Bromoform (CHBr3, specific gravity 2 . 89) and tetrabromoethane (C2H2Br4, specific gravity 2 . 96) are commonly used to produce a heavy concentrate containing phosphatic micro­ fossils, but these chemicals are severely toxic. A safer alternative involves the use of water-soluble sodium polytungstate (3Na2W04.9W03. H20), which can be made up at any required specific gravity, but is best at 2 . 75 or slightly higher to avoid problems of high viscosity and crystal precipitation . Light or buoyant microfossils, such as hollow foraminiferans, radiolarians, and chitinozoans, may be removed in a light concentrate by adjusting the specific gravity of the sodium polytungstate accord­ ingly. Electromagnetic separation is useful in deal­ ing with large residues containing iron oxides or iron-rich dolomite grains .

(nitric acid) prior to crushing to 1 -2 mm fragments . Any carbonate in the rock must be completely re­ moved using warm 10% HCl, followed by thorough washing in distilled water. Silica and silicates are dissolved using HF . Cold, concentrated HF is poured onto the sample in a polypropylene beaker and stirred daily with a teflon rod until all the rock has disaggregated . The reaction may be speeded up by warming the containers in a water bath . After digestion the sample is washed with warm water and fluoride precipitates are removed by treatment with warm 40 -50% HCl, followed by at least four washes in warm water. Ten per cent HCl is added to the last washing to discourage flocculation . Mineral particles may be separated from the organic residue by centrifuging in zinc bromide solution (specific gravity 2 . 0); if examination reveals the presence of pyrite, 10% HN03 may be added to the organic fraction for ten minutes to remove it. Unwanted, undecomposed, or partially decomposed organic material can be removed by careful oxidation (although experience is needed to avoid destruction of microfossils during this process) . Concentrated HN03 is a commonly used oxidant. Fine organic debris may be removed by alkali treatment with 5% potassium hydroxide (KOH) . After processing, the remaining organic-rich resi­ due is sieved, using appropriate mesh sizes for the palynomorphs present. Generally a 53 !-tm sieve is employed to retain chitinozoans and large paly­ nomorphs, while a fine sieve of 5 - 7 !-tm is necessary for the smallest specimens . The fossils may be further concentrated prior to sieving by swirling in a large watch glass . The palynological concentrate, or a representative fraction of it, is finally strew­ mounted onto slides, using glycerine jelly for tem­ porary mounts and Canada balsam or a plastic mounting medium for permanent mounts .

Palynological techniques Procedures for the recovery and concentration of palynological microfossils are complex, with the steps tailored to the nature of the sample being processed . A full account was given by Phipps & Playford (1984), who emphasized the dangers of HF, zinc bromide (ZnBr2), and other chemicals used.

Palynological processing should only be undertaken in a purpose-built laboratory with efficient fume-cup­ boards, full protective clothing, and neutralization and disposal facilities available. All equipment must be kept absolutely clean to avoid contamination . Rock samples should be thoroughly cleaned by scrubbing and, if necessary, etching in HCl or HN03

References Austin, R.L. (ed . ) 1987. Conodonts: investigative techniques and applications . Ellis Horwood, Chichester. Brasier, M.D. 1980 . Microfossils . George Allen & Unwin, London. Phipps, D . & Playford, G . 1984. Laboratory techniques for extraction of palynomorphs from sediments . Papers, Department of Geology, University of Queensland 11, 1 -23 .

6 . 2 Practical Techniques

6 . 2 . 3 Photography D . J . SIVETER

Introduction The photography of fossils involves a wide range of techniques, materials, and object sizes . Large fossils, in excess of about 15 cm in length, fall within the range of normal cameras with standard lenses; specimens up to about 2-3 mm long are best photo­ graphed using the scanning electron microscope (SEM). The middle ground between normal and SEM photography (Section 6.2.4) is generally known as macrophotography, and covers a magnification range on the negative from about x 0 . 2 to x 20 or more . Macrophotography in incident light, for which there are numerous systems available, is the type of photography used for most macroinvertebrates . The Leitz 'Aristophot' system (Whittington in Kummel & Raup 1965) was first used for the macropho­ tography of fossils in the nineteen-fifties, and has since been widely adopted (Fig. IH) . It was modified in various ways before production was discontinued in the early nineteen-eighties . In its image range the quality of photographs produced by this appar­ atus is excellent . The comparable Nikon 'Multi­ phot' system gives similar results and is still (1991) marketed. In the last decade Wild-Leitz (now Leica) have introduced a quite different system for the macrophotographic range, the photomacroscop . The most useful source on the photography of fossils is Kummel & Raup (1965); many of the techniques described therein have not been superseded.

Preparation: cleaning and coating Prior to photography any extraneous sediment should be removed from the surface of the fossil (Section 6 . 2 . 1 ) . If the specimen is embedded in matrix, particular effort should be concentrated on cleaning its margins . This obviates the need for any retouching of or cutting round the fossil outline to delete non-organic material on the final print. The handling of testaceous specimens should be mini­ mized, and they should be cleaned with an organic solvent (such as acetone) to remove any surface grease marks . When photographing most fossils, particularly those that are of variable or light shade, better

505

results are obtained if the specimen is first coated. A matt, uniformly dark surface is applied to the fossil which is then lightly dusted with a whitening agent for contrast. Fountain pen ink and particularly black photographic opaque have been used as darkening agents; these should ideally be applied to impart a dark grey (not black) colour. The former can be removed in large part with a mixture of ammonia and hydrogen peroxide solutions, and the latter with warm soapy water. The cleaning of darkening agents from natural mould specimens (especially in medium to coarse clastics) is very difficult or impossible, as they are fully absorbed into porous sediments; this is particularly so where Indian ink has been used . The excellent opaque produced by Phillips and Jacobs (Philadelphia) is now discontinued . Practitioners should experiment with alternatives; poster paint has, for example, been successfully used . Various inks and carbon powder (soot) have been used to darken latex and silicone rubbers . A whitening agent sympathetically applied on the darkened surface considerably enhances the contours and surface sculpture of the fossil, as it falls more densely on those areas of greater relief, which are thus highlighted (Fig. 1 ) . It also provides an even, glare-free reflecting surface for photogra­ phy and results in prints of a similar tone - which are desirable when making plates for publication. Ammonium chloride, magnesium oxide, and anti­ mony oxide have all been used for whitening. Ammonium chloride and antimony oxide are heated in a glass bulb and the resulting sublimate cloud directed onto the fossil (Teichert 1948; Marsh & Marsh 1975) . Magnesium oxide is produced by burning magnesium ribbon and the fossil is held over the smoke . Ammonium chloride should be washed off immediately after use as it combines with water vapour in the air to form hydrochloric acid capable of etching the fossil; its deliquescence also renders it impracticable for use in areas or on days of high humidity, as the sublimate quickly becomes coarse-grained after coating. Nonetheless, many authors favour the use of ammonium chloride as control on the application of magnesium oxide is not very precise . All coating should be done in a fume cupboard, but the draught should not be so strong as to affect the flow direction of the whitening agent. After coating and prior to photography a check should be made under a binocular microscope for hairs or other artifacts . The implications for future conservation of the specimen should be considered before employing these techniques .

506

6 Infrastructure of Palaeobiology

Fig. 1 A and B taken using Nikon 'Multiphot', C -G using Leitz 'Aristophot' . A taken with the Nikon 'Makro-Nikkor' 12 cm lens, B with Nikon 'Makro-Nikkor' 6.5 cm lens, C - G with the Leitz 'Summar' 12 cm lens . A - C, E, G photographed on Kodak 'Panatomic X' film, D and F on Ilford 'Pan-F film. All specimens are coated with ammonium chloride on top of matt black opaque. A-G, Silurian trilobites . A, Cranidium, odontopleurid, Ireland; dorsal view, x 4. B, Glabellar sculpture, proetid, Ireland; dorsal view, x 22 . C, Thoracic pleural facet, calymenid, Gotland; lateral view, x 10. D, Stereo-pair, complete specimen, calymenid, West Midlands, U.K. ; dorsal view, x 2. E, Glabellar sculpture, calymenid, Welsh Borderland; dorsal view, x 10. F, Eye, phacopid, Ireland; oblique view, x 7. G, Complete specimen, calymenid, Gotland; oblique view x 2.5. H, Leitz 'Aristophot' with anglepoise and ring light illumination, laboratory jack, and tilt-table for taking stereo-pairs.

6 . 2 Practical Techniques Macrophotographic equipment and methods The photographic film should be fine grained (50 ASA or less) and have good resolving properties so that when enlarged it suffers minimal loss of definition; Ilford 'Pan-F' and Kodak 'Panatomic X' are both suitable . Fossil size on the final print depends on the negative magnification multiplied by that selected on the enlarger. Macrophotography of fossils is for most purposes adequately and econ­ omically performed with the use of 35 mm format, with final prints of up to X 30 to X 40 being satisfactorily obtained . Recourse to larger format apparatus and film (e . g . 9 X 12 cm) is preferable only where excessive enlargement is demanded, or where a wider field of view is required at a given magnification . The photographic stand should be sturdy and capable of absorbing vibrations . The camera body is not one of the more critical pieces of equipment but the action of the shutter should be smooth if this is to be used to control exposure, and those with a reflex mirror lock-up facility that negates the vibrations from this source are most useful . Leica 'M' cameras have been used on the 'Aristophot' in combination with a separate reflex mirror unit that also incorporates a focusing magnifier and focusing screen . Nikon 'F' cameras for use with the 'Multiphot' house the reflex mirror and focusing system within the camera body . At high magnifications requiring long bellows exten­ sions, where the slightest vibration is ruinous, it is best to control the exposure by means of the lens shutter rather than the camera shutter . When using the 'Aristophot' in the 35 mm format, the correct exposure time is best assessed empirically with the use of test films and records of film speed, lens type, aperture setting, lighting, and magnification . Through the lens metering (TTL) is available in this format with the 'Multiphot', utilizing in particular the Nikon F3 camera . However, with over-long exposures (in excess of about 1 second the readings from any type of metering system will be inadequate due to reciprocity failure, and extra time must be allowed, depending on the film type. Much macro­ photography of fossils falls within the 1 - 15 second exposure time . The focusing screen on the camera should be of the finely ground glass or clear glass type and focus­ ing done at full aperture . The specimen - lens and lens - film (bellows length) distances combine to determine magnification on the negative, and at any given magnification these distances will vary according to the focal length of the lens employed .

507

Manufacturers' handbooks normally contain graphs plotting magnification against distance for each lens . Sometimes it is desirable to produce negatives at set, whole number magnifications; this requires retention of the camera and lens in the appropriate positions and focusing by moving the specimen vertically, either by means of a heavy duty labora­ tory jack or a rack and pinion operated 'lift' . The specimen can be mounted by plasticine onto the jack or 'lift', the surface of which should be painted matt black to provide a contrasting background to the whitened fossil . Photographs other than those of surface sculpture should not be focused on the upper surface of the fossil but more towards the median plane of the specimen to take into account depth of field . Lighting comprises two basic components . A directional light source, by convention shining from the northwest, is beamed at the fossil at an angle (normally low) suitable for emphasizing its relief. The shadows thus produced are partially filled in and the specimen lit overall by means of soft, dif­ fuse, even illumination . One of the several ways of achieving the desired effect is to use an anglepoise lamp with a frosted bulb, the light strength of which is controlled by a dimmer switch, together with a fluorescent ring light (about 30 cm diameter and 60 watts) capped by a reflector (Fig. IH) . Any extraneous light should be prevented from entering the lens . It is important to ensure that any lens used for enlarging small objects gives, in addition to sharp resolution at the plane of focus, good imaging throughout the depth of the specimen . Increased depth of field is achieved by reducing the size of the lens aperture, but beyond a certain limit (which can be empirically determined for each lens) the effect of diffraction gives progressively poorer resolution and makes it pointless to stop down further . Lenses optically corrected for the macrophotographic range, for use with the ' Aristophot' or 'Multiphot', come in several focal lengths from about 12 mm to 120 mm. Lens selection depends on the desired scale of repro­ duction, those with shorter focal lengths being used for greater magnifications . The original 'Aristophot' lenses, the Leitz 'Milar' and particularly the 'Summar' range, and also the later, compatible first generation 'Photar' range, give excellent results; Nikon have consistently produced four macro lenses with high resolving power for use with the 'Multi­ phot' . The latest, more restricted generation of 'Photar' lenses reproduce over the X 1 to X 16 range and are combined with the Leica 'R' system of

508

6 Infrastructure of Palaeobiology

cameras and bellows, featuring through-the-Iens metering, for use on a copying stand . Other Leica 'R' macro lenses for use with this system enable reproduction from infinity to X 3 . The 'M400' photomacroscop o f Wild-Leitz i s a fully integrated unit featuring 35 mm to 9 x 12 cm format, automatic exposure control, and a 1 : 5 macrozoom objective, focusing being accomplished via a binocular system. With the optional use of three additional objectives and using the 35 mm format it covers the macro range X 1 to X 20 . A I : 6 'Apozoom' objective has recently been introduced for this set-up . Wild-Leitz also offer a similar auto­ matic system in the macro range on their 'M420' zoom macroscop . It is debatable whether the macro­ zoom lenses used on these systems can out-perform the individually computed Leitz or Nikon macro lenses, although the photomacroscop would seem to win over the 'Aristophot' and 'Multiphot' in terms of convenience of operation, combined with the relative lack of experience required to obtain reasonable results.

Special techniques More specialist techniques are sometimes em­ ployed in the macrophotography of fossils . Stereo­ photography involves photographing the specimen in two slightly different attitudes differing by an angle of rotation of 7- 10° (Fig . 10) . The resultant two photographs give a three-dimensional image when optically fused by means of a steroscope (Evitt 1949) . Immersion of a specimen in a liquid such as alcohol, water, glycerin, or xylene is under­ taken particularly if the fossil is of low relief, and where the distinction between the fossil and the surrounding matrix needs enhancement, and also to make clearer internal structures (Rasetti in Kummel & Raup 1965) . Photographs taken in ultraviolet radiation at low inclinations also bring out features of low relief (e . g . Whittington 1985) . Lastly, X-ray photography with the use of long exposures has been successfully employed on pyrit­ ized material (StUrmer et al. 1980) . A combination of the above techniques is possible, as with stereo and X-ray photography .

Processing and printing A fine grained developer should be used for the film, to maintain detail . The enlarger should have a good quality lens and hold the film perfectly flat. Resin coated paper has advantages over tra-

ditional fibre-based paper in speed of development, fixing, and washing, and the fact that glazing is un­ necessary - it can be simply air-dried if required . Multigrade paper (either fibre-based o r resin coated) is convenient to use and enables very fine contrast control on the finished print by utilizing enlarger filters graduated to half a grade of monograde papers; it also makes redundant the potentially wasteful practice of having five boxes of different grade paper open simultaneously . Glossy paper pro­ vides a wider range of contrast and tone, and more detail than matt paper . Optimum use of space and prints of matching tone with parallel edges are necessary for an aesthetically pleasing plate (Fig . 1 ) .

References Evitt, W.R. 1949 . Stereophotography as a tool of the paleonto­ logist. Journal of Paleontology 23, 566- 570. Kummel, B. & Raup, D . 1965 . Handbook of paleontological techniques . W.H. Freeman and Co . , New York. Marsh, R . e . & Marsh, L . F . 1975 . New techniques for coating paleontological specimens prior to photography . Journal of Paleontology 49, 565 -566 . StUrmer, W., Schaarschmidt, F. & Mittmeyer, H.-G . 1980. Versteinertes Leben in Rontgenlicht. Kleine Senchenberg­ Reihe No. 11, Verlag Waldermar Kramer, Frankfurt am Main. Teichert, e. 1948 . A simple device for coating fossils with ammonium chloride. Journal of Paleontology 22, 102- 104. Whittington, H . B . 1985 . The Burgess Shale. Yale University Press, New Haven.

6 . 2 . 4 Electron Microscopy

D . CLAUGHER & P . 0 . TAYLOR

Both the transmission (TEM) and scanning (SEM) electron microscopes have wide-ranging appli­ cations in palaeobiological research, including studies of skeletal microstructure and growth, func­ tional morphology, and taphonomy .

Transmission electron microscopy The TEM produces an image by passing a beam of electrons through a specimen which must be very thin (90 - 250 nm) and must fit onto a 3 . 5 mm dia­ meter microscope grid . Methods for investigating fossils using the TEM were developed in the early days of carbon replication . This technique involved

6 . 2 Practical Techniques coating a specimen with carbon, dissolving the specimen, and examining the carbon replica of the specimen surface in the microscope . Although much useful information could be gained using carbon replicas, the technique was relatively unpopular because of limitations on specimen orientation in the microscope, and the delicate nature of the rep­ lica. Before the advent of SEM, however, small speci­ mens, such as coccoliths and diatoms, and fragments of larger specimens were routinely examined in this way. Fossil plant and animal tissue is generally miner­ alized and unsuitable for direct study with the TEM. However, unmineralized tissue may be prepared for TEM examination by releasing it from the matrix using acids or other solvents . The released tissue is thoroughly washed in distilled water to remove any remaining acids or solvents, and is then dehydrated through a graded series of acetone solutions . After two changes in pure acetone, it is embedded in an epoxy resin. Sections are cut with a glass or diamond knife on an ultramicrotome, then mounted on grids, dried, and examined in the TEM (see Glauert 1974) . Using this method Urbanek & Towe (1974) were able to produce some excellent micrographs of unstained graptolite tissue, and the palaeobotanical literature contains many similar examples .

Scanning electron microscopy The introduction of the SEM in 1968 gave palaeon­ tologists an instrument of such versatility that 20 years later new techniques for investigation are still being developed . The SEM produces an image by bombarding the surface of a specimen held in a high vacuum with a stream of electrons . This pro­ vokes the generation of X-rays, secondary electrons, and backscattered electrons, which may be collected and processed to form a visual image of the speci­ men on a cathode ray tube (see Goldstein et al. 1981 ) . The method is non-destructive, and some microscopes can accommodate specimens up to 10 cm in diameter. Stereoscopic images can be prepared with SEM . Two photographs are taken at a separation of 8° and, when examined using a stereo viewer, these may give much additional information on the spatial arrangement of the specimen. Good examples of this application can be found in issues of A Stereo-Atlas of Ostracod Shells (British Micro­ palaeontological Society, London) . A disadvantage of the early SEMs was that all

509

material to be examined had first to be coated with a thin layer of conducting metal such as gold, plati­ num, or aluminium. Many museum curators are unwilling to commit type or other valuable speci­ mens to this treatment, despite the fact that some coatings can be subsequently removed (e . g . gold by treatment with cyanide) . A device known as CFAS (charge free anticontamination system) is now available which allows uncoated specimens to be examined (Taylor 1986) . The microscope chamber is pumped to a poorer vacuum than the gun and column, and a back scattered electron detector is used in place of the normal secondary electron detector to collect the signal. Specimens do not have to be glued or permanently attached to a stub, but are simply held on a metal plate with plasticine or a similar substance which does not contaminate the inside of the microscope . Clear micrographs of un­ coated specimens can be obtained using CFAS (compare Fig . lE, F) . The coating of valuable specimens may also be avoided by preparing replicas (Fig. lC) for examin­ ation in the SEM. Hill (1986) investigated various replicating materials and concluded that cellulose acetate (which must be used with care on delicate material) gave the best results, whereas the more commonly used latex rubber gave poor results . The method of attachment specimens to stubs is of paramount importance, especially if the specimen is later to be recovered for examination of the reverse side . Double-sided adhesive tape is commonly used because it is convenient and permits specimen re­ moval using an organic solvent . However, this is not a recommended procedure; the volatile compo­ nents of the adhesive tape evaporate in the micros­ cope and deposit in the form of carbon on the inside of the column and apertures, giving rise to poor image resolution. A simple and inexpensive method for attaching microfossils (e . g . foraminifera, pollen, and spores) and small fragments of macro­ fossils is as follows: (1) cut dried processed film into small squares and glue it to stubs with the emulsion side of the film uppermost; (2) moisten a small area of the film with water using a fine paintbrush to soften the gelatin; and (3) manipulate the specimens onto this area and leave them to dry (after examin­ ation, removal or re orientation can be achieved using a wet paintbrush) . Permanent attachment of material to stubs should be made with epoxy resin (not the quick setting varieties, which may not set as hard as normal types) . Only small quantities of epoxy should be used for small specimens, and special care should

510

6 Infrastructure of Palaeobiology

Fig. 1 Scanning electron micrographs illustrating some of the diverse palaeobiological applications of the SEM . A, Umbilical view of the benthic foraminifer Pseudorotalia yabei (Ishizaki), a species from the Miocene of Borneo potentially useful in stratigraphy, x 33. (Micrograph courtesy of Dr r E . p . Whittaker . ) B, Proximal end of a rodent femur from a British Pleistocene cave site showing evidence of digestion by a predator, x 10. (Micrograph courtesy of Dr P.J. Andrews. ) C, Dow Corning silicon rubber replica of in situ spores of the fern Qasimia schyfsmae (Lemoigne) from the Permian of Saudi Arabia, x 1000. (Micrograph courtesy of Dr CR. Hill) . D, fractured shell of the British Jurassic bivalve DeItoideum delta (Smith) showing prismatic microstructure with endolith borings, x 335 . E,F, part of a colony of the bryozoan Metrarabdotos moniliferum (Milne Edwards) from the Pliocene of U.K. depicted as a conventional secondary electron image of the gold-coated specimen (E) and a backscattered electron image of the uncoated specimen (F) prepared using CFAS, x 13.

511

6 . 2 Practical Techniques b e taken with porous material which tends t o absorb the adhesive and in some cases obscures surface detail . For very small fossils (e . g . diatoms), which are not practicable to mount individually, the fol­ lowing method is advocated: (1) abrade a clean stub with very fine wet and dry emery paper, wash thoroughly in an ultrasonic bath and dry; (2) rub epoxy resin into the abraded surface using a cocktail stick and remove the excess adhesive with a lint­ free tissue such as 'vellin' to leave the epoxy only in the very fine grooves; and (3) onto this surface place the specimens which will adhere permanently . Coccoliths are among the most difficult fossils to prepare, but dry material can be treated as above if the stub is very finely abraded and the excess epoxy wiped off very thoroughly. The most successful method for mounting coccoliths is simply to abrade a stub with very fine emery paper, wash, dry, and then pipette a suspension of material onto the stub; dry and coat before examining. Many fossils in the SEM accumulate charge which degrades image quality. Charging may be related to the composition of the specimen, poor attachment to the stub, or inadequate coating. The use of CFAS, a backscattered electron detector, or reduction of the accelerating voltage may eliminate charging but often does so at the cost of poorer resolution . One of the most promising developments to help overcome the charging problem is a method of collecting and pro

� :0

(j) Q.

300 250

E � E

200

::;:

1 00

:0

E x ro

1 50 ......

50 1 000

0.1 D u ra t i o n o f h e at i n g

( 1 0(, y r s )

A Karweil type diagram from which maximum temperature can be determined from known vitrinite reflectivity and estimated duration of heatinglburial. (From Bostick et al. 1979 . )

Fig. l

Colours are estimated on a visual scale (Fig . 2) by reference to a set of spore/pollen standards (most of which have been produced by commercial labora­ tories and are therefore not widely available because of cost and confidentiality) . The range of colour in spores is continuous and the scale boundaries are imposed arbitrarily. The colours are also difficult to describe in words so that, without recourse to standards, these scales can only be crudely applied . They are also frequently non-linear when compared to both depth of burial and other maturity indicators; brown and darker colours become unpredictable in their occurrence, rendering the scales of limited value at higher temperatures . The influence of time is important since changes in colour are not instantaneous . Thus a time - temperature cross plot like that used for vitrinite can be employed (Fig . 3) to estimate maximum temperature . The correlation of colour against vitrinite reflectivity in different depositional basins does not give a constant relationship since these materials behave differently kinetically. Differing geological histories result in different durations of thermal input and each basin has a somewhat different correlation .

Fluorescence. The walls of spores and pollen (in common with plant cuticle, acritarchs, dinoflagellate cysts, and certain types of A . a . M . ) fluoresce in the visible spectrum when excited with ultraviolet light. Fluorescence colour (Fig. 2) varies both with organic matter composition and thermal maturity. The generation of these colours requires a sophisti­ cated microscope with an incident ultraviolet light

6 . 2 Practical Techniques S po re co l o u r

SCI

Co l o u r l e s s-pa l e ye l l ow

1

Pa l e ye l l owl e m o n ye l l ow

2

L e m o n ye l l ow

3

G o l d e n ye l l ow

Rv (SCI)

5

O ra n g e

6

O ra n g e b rown

7

D a r k b row n

8

RCH

1 .5 2.2

0.4

9 10

/\

400

0.5

G re e n-ye l l

0.5 0.7

Ye l l ow

0.6 2.4

0.7

1 .0

Ye l l owo ra n g e

0.8 2.6

0.9

3.0

1 .0

3.5 4.0

1 .5

700

F /'

�

�

'"

u Vl

Q.) Cl.

E 2 E '" E '"

� x

'" o

"0 �

u

�

o Cl. Vl

60 40 20

1 .2

2.8 1 .4

D a r k b rownblack

S p o r e fl u o re s c e n c e co l o u r & s pectra B l u e-g reen

0.3

4

Ye l l ow o ra n g e

B lack

TA l

513

2.0

/ /

o ra n ge-red

B rown No further fl u o re s c e n c e

A comparison of the Spore Colour Index (SCI) and Thermal Alteration Index (TAl) scales with verbal colour description . Vitrinite Reflectivity (Rv), Chitinozoan Reflectivity (Rch), spore fluorescence colours, and spectra are also included . The latter are measured over the range 400 - 700 nm and normalized to the same relative intensity. (From Otterjahn et al. 1974; Fisher et al. 1980; Smith 1983 . )

D u ra t i o n of h e at i n g

( 1 06 yrs)

A Karweil type diagram from which maximum palaeotemperature can be determined from a known spore colour (SCI only) with an estimated duration of heatinglburial. (From Cooper 1978 . )

Fig. 3

Fig. 2

source and dichroic beam splitters . The colours are difficult to estimate (in comparison to spore colours in white light) as they are pastel shades and rather faint. Colour can also be quantified with a photometer/monochrometer that generates a curve relating intensity -wavelength (nm), for which the maximum peak height, width, and position change with maturity (Fig. 2) . Quantitative fluorescence measurements are complicated by additional factors, such as intensity fading, microscope correc­ tions, and uncertainty over absolute standards . The Table

technique is therefore only employed in specialist laboratories .

Conodonts Conodonts have proved a popular group for deter­ mination of rank in Palaeozoic rocks due' to the widespread adoption of a single colour scale and the availability of standards . The conodont alteration index (CAI) is an eight-point scale (Fig. 4) that covers the temperature range < 50° to > 700°C . It is thus applicable to the widest range of maturity, including schists, although above CAI 6 difficulties occur in the event of hydrothermal alteration . The essential difference between conodonts and other microfos­ sils used in organic maturation studies is that cono­ donts are composed of a phosphatic mineral, with only trace amounts of organic matter. The initial colour changes (1 -5) result from maturation of this

1 Use of fossils as indicators of organic maturity. Microfossil group

Reflectivity

Spores Pollen Acritarchs Dinoflagellate cysts Chitinozoans Conodonts Vitrinite Note :

*

=

minor application;

Colour

Fluorescence

Geological range

***

**

***

**

Silurian - Recent Devonian - Recent Precambrian - sub-Recent Triassic- Recent Ordovician - Devonian Cambrian-Triassic Silurian - Recent

*

*

*

*

*

***

***

***

=

significant use .

6 Infrastructure of Palaeobiology

514

material, whilst above CAI 6 the mineral itself re­ crystallizes with oxidation of the organic matter and becomes clear. The phosphatic composition generally restricts their recovery to carbonate rocks and low rank shales, whilst the trace amount of organic matter limits the colour changes so only two CAI points are available within the oil window . Like all colour scales it is a series of points imposed on a continuous colour series and difficult to express in words, so standards are required for serious work.

Acritarchs Acritarchs, like pollen and spores, undergo changes in wall colour with increasing maturity. They have not been as widely used because the colour changes are more subtle and difficult to determine on the thinner walled tests, whilst thicker walled forms are frequently pigmented with significant natural colour. Consequently, where their geological ranges overlap, pollen and spores are used in preference to acritarchs . In the Lower Palaeozoic, where spores and vitrinite are largely absent, acritarch colour (and fluorescence) become important (although conodont colour is also available for this interval) . An acritarch colour alteration index has been produced (Fig . 4), with a five point scale based on colour changes in simple leiospheres . CAI

Te m p . r a n g e (QC)

1

600

4 S

-

C o l o u r l e s s or c rystal c l e a r

Chitinozoans Polished sections through chitinozoan walls show reflectivity properties similar to the vitrinite maceral (a calibration is given in Fig . 2) . Usage is still only at an initial stage but chitinozoans should provide early Palaeozoic researchers with a quantitative scale as precise as that of vitrinite . Chitinozoans have the advantage of being large and thus easily measured in comparison to acritarch walls; they can also be recovered from every type of sediment within which they occur, unlike conodonts . Prob­ lems with low numbers in polished whole-rock preparations can be solved by using polished thin sections .

Other microfossil indicators of thermal maturity Kerogen components, such as dinoflagellate cysts, plant cuticle, and most types of A . O . M . , generally show colour and fluorescence changes with increas­ ing rank in a similar way to spores and pollen, although changes relate differently to temperature . For various reasons they have not become estab­ lished as routine thermal maturity indicators but can be used if required, and related approximately to the major points of existing scales . Situations where they are used include A . O .M . rich or distal

Acritarch co l o u r

Hyd roca r b o n g e n e ration

T ra n s l u ce n tl i ght ye l l ow

I m matu re

L i g h t ye l l owp a l e ye l l ow Pa l e ye l l owo range O ra n ged a r k b rown

-- - - - -

O i l & wet gas

1- - - - - - - Black

-

D ry gas

--------

Correlation of conodont colour and acritarch colour with verbal colour descriptions, temperature ranges, and the main zones of hydrocarbon generation . (From Legal! et al. 198 1 ; Rejebian et al . 1987, by permission of the Geological Society of America . )

Fig. 4

6 . 3 Museology oxic marine kerogen facies which may either lack a terrestrial input, with no spores, pollen, or vitrinite, or have had it diagenetically modified and/or diluted .

References Bostick, N . H . , Cashman, S . M . , McCulloh, T.H. & Wad dell, CT. 1979 . Gradients of vitrinite reflectance and present temperature in the Los Angeles and Ventura Basins, Cali­ fornia . In: D.F. Oltz (ed . ) Low temperature metamorphism of kerogen and clay minerals, pp . 65 -96. Society of Economic Paleontologists and Mineralogists (Pacific Section), Los Angeles, Ca. Cooper, B.5. 1978 . Estimation of the maximum temperatures attained in sedimentary rocks . In: G . D . Hobson (ed . ) Developments in petroleum geology Vo!. 1 . Applied Science Publishers, London . Epstein, A . G . , Epstein, J.B. & Harris, L.D. 1977. Conodont colour alternation - an index to organic metamorphism . U.5. Geological Survey, Professional Paper, No . 995 . Fisher, M.J., Barnard, P . C & Cooper, B . 5 . 1980 . Organic maturation and hydrocarbon generation in the Mesozoic sediments of the Sverdrup Basin, Arctic Canada. Pro­ ceedings of the Fourth International Palynological Conference, Lucknow (1 976 - 77) 2, 581 -588.

515

Heroux, Y., Chagnon, A . & Bertrand, R. 1979 . Compilation and correlation of major thermal maturation indicators. Bulletin of the American Association of Petroleum Geologists 63, 2128-2144. Legall, F . D . , Barnes, CR. & MacQueen, R.W. 198 1 . Thermal maturation, burial history and hotspot development, Paleozoic strata of southern Ontario- Quebec, from conodont and acritarch colour alteration studies. Bulletin of Canadian Petroleum Geology 29, 492 -539 . Otterjahn, K . , Teichmuller, M. & Wolf, M. 1974. Spectral fluorescence measurements of sporinite in reflected light, a microscopical method for the determination of rank in low rank coals . Fortschrifte in der Geologie von Rheinland und Wesifalen 24, 1 -36. Rejebian, V.A., Harris, A.G. & Heubner, J.5. 1987. Conodont colour and textural alteration: an index to regional meta­ morphism, contact metamorphism, and hydrothermal al­ teration . Bulletin of the Geological Society of America 99, 471 -479 . Smith, P . M . R . 1983 . Spectral correlation o f spore colouration standards . In: J. Brooks (ed . ) Petroleum geochemistry and exploration of Europe, pp. 289 -294. Special Publication of the Geological Society of London, No. 12. Staplin, F . L . , Dow, W . G . , Milner, CW.D., O' Connor, 0 . 1 . , Pocock, S.A.J . , van Gijzel, P . , Welte, D.H. & Yukier, M.A. 1982. How to assess maturation and paleotemperatures. Society of Economic Paleontologists and Mineralogists, Short Course No . 7.

6 . 3 Museology

6 . 3 . 1 Collection Care and Status Material P . R . CROWTHER

logical ordering of specimens enables material to be found as required, often without recourse to a man­ ual index or computerized documentation system (Section 6 . 3 . 2) .

Storage environment Introduction The fundamental aim of good fossil storage is to ensure the long-term survival of specimens, thus guaranteeing their future availability for study and display . The clean, ordered storage of specimens in a controlled environment is the physical basis of a good collection (Brunton et al. 1985; Rickards in Bassett 1979) . The ability to view and handle fossils easily, the use of appropriate containers, and the

Storage areas should be as free as possible from fluctuations in temperature and relative humidity (r. h . ) . Extremes and rapid changes of r . h . are the most common cause of damage to fossil material in museums. The vulnerability to oxidation of pyri­ tized fossils and pyrite-bearing matrices ('Pyrite Disease') increases to unacceptable levels when r . h . rises above about 55% . Neutralization o f affected material (Cornish in Crowther & Collins 1987) is no protection against future damage, which can

516

6 Infrastructure of Palaeobiology

only be prevented by keeping r . h . down. On the other hand, subfossil bone and some shale matrices shrink and crack when r.h. falls below about 45% . Rapid fluctuations of r . h . causes some clays and shales to swell and shrink alternately, leading to deterioration and loss. Monitoring and control of r . h . is therefore essential in geology storage areas, to keep conditions at 50 ± 5% r . h . This is achiev­ able either through full air conditioning or, more economically, through the use of portable de­ humidification and humidification equipment . Conditioned silica gel can maintain small, sealed volumes (storage boxes or display cases) at whatever r . h . is required. Temperature variation alone has little detrimental effect on fossil material, but because temperature is so intimately associated with r.h. (a fall in tempera­ ture causes r.h. to rise, and vice versa), its stabiliz­ ation is essential for r.h. control . A combination of high r . h . and high temperature accelerates the hydrolysis of hemicellulose to acetic acid in the wood of oak or birch ply cabinets; this may attack calcareous fossils and matrices (the so-called 'Bynes Disease') and such woods are best avoided for cabinet construction. Airborne dust is a particular and obvious menace to collections . It makes material difficult to examine and its removal is both time consuming and poten­ tially damaging to fragile specimens . Dust proofing can be incorporated at several levels within a quality storage system: individual storage trays can be made deep enough to support acetate tops; storage drawers and boxes should have tightly fitting lids; and the mobile bays in a compactable racking system can be edged with seals which mesh together when picking aisles are fully closed.

Storage furniture The ordered physical storage of fossils in a controlled environment cannot be realized cheaply . The specialized storage requirements of 'difficult' cat­ egories of specimens dictate particular solutions, e . g . large vertebrates (Brunton et al. 1985; Gentry in Bassett 1979) . Inside more generalized storage units, specimens should sit in paper-lined card trays (made of acid-free materials) to prevent abrasion and mixing . Storage unit design should be flexible regarding the use of drawers or shelves, and in the variety of drawer or shelf depths . They should in­ corporate good dust seals . Wooden cabinets are preferable (but not oak or birch ply for the reason given above) since they buffer against changes in

r.h. and cushion vibration. Mobile, rail-mounted, compactable racking systems make the most effec­ tive use of limited space, but they require strong floors and inevitably subject their contents to more vibration.

Status material Article n(g) of the 1985 International Code of Zoological Nomenclature (see Section 5 . 1 . 1) states that name-bearing types (holotypes, syntypes, lec­ totypes, and neotypes) are international standards of reference and are held in trust for science by those responsible for their safe keeping. Insti­ tutional responsibility in this regard is set out in the Code's Recommendation nG as follows : Every institution in which name-bearing types are deposited should : 1 Ensure that all are clearly marked so that they will be unmistakably recognized as name-bearing types . 2 Take all necessary steps for their safe pres­ ervation . 3 Make them accessible for study. 4 Publish lists of name-bearing types in its possession or custody. 5 So far as possible, communicate information concerning name-bearing types when requested . Failure to heed this code of practice hinders the progress of science and puts type material at risk. Any museum holding fossil type material should have a geologist on its permanent establishment; any university department or museum with types but no designated curator should deposit them elsewhere (Owen 1964) . It is the responsibility of the name giver to ensure that types go to an appro­ priate repository, and it follows that editors must insist on authors carrying out this duty as a con­ dition of publication . Indeed, taxonomic practice would be greatly enhanced if all status material (type, figured, and referred specimens) had to be registered in an appropriate institution as a condition of publication . The question of how best to store status material has provoked some disagreement (Brunton et al. 1985, p . C25) . Arguments that favour separating status material from the main collections include : meeting the ICZN and ICBN requirements regard­ ing type specimen care; convenience of access; increasing its physical security in better quality storage by improving protection from theft and damage from fire, flood, etc . ; and ease of evacuation in emergencies . Disadvantages of isolating such

6 . 3 Museology material include the risk that users might overlook its existence when working through the main col­ lection; and that it might be totally destroyed in the event of a disaster affecting that one particular part of the store . Whether type material should remain in its country of origin is another vexed question, although the ability of the home country to properly house and curate such material must be a primary consideration . Practices differ from nation to nation : Canada has legislation to ensure that types erected on Canadian material taken abroad for study are eventually repatriated; the Palaeontological Museum, Oslo, functions in effect as the National Museum for Norway and preferred policy is for all Norwegian primary type material to be held there (Bruton in Bassett 1979) ; the British Museum (Natural History) regards its holdings as inter­ national in scope, and considers it essential that related collections from different parts of the world are kept together, because palaeontology is a com­ parative science (see comment by Ball in Bassett 1979, p. 147) . Publication of a museum' s holdings of status specimens should be given high priority, since the dissemination of such information to the world at large is one of the important responsibilities that goes with being a type repository . Many of the larger museums can utilize an in-house publication for this purpose (Section 6.4), while smaller repositories can still fulfil their obligations to the wider scientific community via specialist journals such as the Geological Curator.

References Bassett, M . G . 1975 . Bibliography and index of catalogues of type, figured, and cited fossils in museums in Britain. Palaeontology 18, 753- 773 . Bassett, M . G . (ed . ) 1979 . Curation of palaeontological collec­ tions . Special Papers in Palaeontology, No . 22. Brunton, C H . C , Besterman, T . P . & Cooper, J.A. 1985 . Guidelines for the curation of geological materials . Miscellaneous Paper of the Geological Society, No. 17. Crowther, P.R. & Collins, CJ. (eds . ) 1987. The conservation of geological materiaL Geological Curator 4, 375 -474. Owen, D . E . 1964. Care of type specimens . Museums Journal 63, 288-291 Torrens, H . 5 . 1974. Palaeontological type specimens. Newsletter of the Geological Curators ' Group 1, 32 -35 .

517

6 . 3 . 2 Collection Management and Documentation Systems P . R . C ROWTHER

Introduction All science depends on being able to 'repeat the experiment', to check the data on which conclusions drawn by others were based . In whatever way the fields of palaeobiology, biostratigraphy, taxonomy, evolutionary studies, etc. are delineated, each relies to a greater or lesser extent on our interpretation of the fossil record . It follows that fossil collections and their associated data represent the primary material evidence that underpins the intellectual structure of these elements of Earth science . The survival and availability of such collections is crucial to the ad­ vancement of knowledge, so that past results can be checked and new observational and analytical tech­ niques can be applied . Without museum collections, palaeobiology could not exist. The management of museum collections concerns the accessioning, control, cataloguing, use, and dis­ posal of specimens . The accelerating awareness of the importance of collections management has been triggered by: pressures on museums to demonstrate accountability for their collections; modern security and audit requirements; and the higher standards of inventory control expected by governing author­ ities (Roberts 1988) . Effective documentation is the key to collections management and is essential if the legitimate aspirations of museum users are to be met.

Information storage and retrieval A fossil without certain basic information (locality, stratigraphy, collector, etc.) is of little scientific value, however visually attractive it may be . Conversely, the most unprepossessing fossil fragment can con­ tinue to provide answers to new questions if it was effectively documented at the outset. A precise re­ cord of where, when, and how such a fossil was collected, and by whom, guarantees its future utility . All serious collectors have an obligation to science to ensure the long-term survival of their fossil material - which may represent an irreplaceable resource from a temporary exposure, and was per­ haps collected at great public expense from a remote part of the globe .

518

6 Infrastructure of Palaeobiology

The principle of being able to repeat an obser­ vation is as axiomatic to the large numbers of speci­ mens associated with biometrics, population dynamics or phylogenetics as it is to the holotype concept in taxonomy . Today's collectorlresearcher can ensure the continued availability of primary source material - be it a unique type specimen or the thousands of measured specimens in a statistical study - by : 1 Allocating a unique identifying number to each specimen at the earliest opportunity. 2 Securely recording certain essential data about the specimen (locality, stratigraphy, collector name, collection date) and keying such a record to the specimen via its unique identifier. 3 Greatly improving the specimen' s chances of sur­ vival and long-term availability for future study by depositing it in an appropriately staffed and funded museum . This procedure ensures that the specimen's latent scientific potential is protected . The additional benefits that follow from producing specimen labels, classified catalogues, indexes (by donor, locality, age, etc . ) are many, and certainly make a collection more accessible to the user. But they can all be created later as required, manually or by computer, if the essential collection data has been properly recorded. Museum collections are assemblages of facts in the form of specimens, specimen-related data (Light et al. 1986), and (increasingly) site-related data (Raup et al. 1987; Crowther & Wimbledon 1988) . Many of these facts are available nowhere else; in the museum they remain available for re­ examination, reinterpretation, and restructuring, over and over again (Waters ton in Bassett 1979) . The physical well-being of collections (Section 6 . 3 . 1 ) and the dissemination o f information relating to those collections are both fundamental to the role of museums in Earth science today . An effective documentation system is the key to fulfilling such a role . The theory of how best to permanently link a specimen with its essential data is well established in museology, using a unique number and a secure register of sequentially ordered data entries . Unfor­ tunately, museum performance in this area has often left much to be desired, either through curatorial incompetence or (more usually) through under­ staffing and the tug of conflicting priorities . The efficient retrieval of specimen data i n a form capable of satisfying the needs of all museum users has proved an intractable problem. The traditional, manual approach is to maintain alongside the main

register as many running card indexes (by taxon, locality, age, donor, etc . ) as staff time allows. Any collection is to some extent 'self indexing' through the classified storage strategy adopted - by taxo­ nomic group, stratigraphical division, geographical location, or (more likely) by a combination of these . But in reality many museums are unable to keep pace even with basic registration of specimens, and it is very rarely possible to resource a fully effective manual system.

Computerized documentation systems Computing techniques are having a major impact on both the scale and type of problems being at­ tacked within contemporary palaeobiology (Sec­ tion 6 . 1 ) . Museums were quick to appreciate the potential of computer-based information techno­ logy for the sorting and selective retrieval of specimen-related data. Any conceivable index can be generated from a single input of specimen data, and interactive retrieval can be used to interrogate the database directly. Some museums with access to mainframe computers, either in-house or through computer bureaux, now have more than 15 years of experience to draw upon . The more recent devel­ opment of the desk-top microcomputer, with its increasingly more powerful data-processing abili­ ties and data storage capacity, has opened up the same advantages to a much wider spectrum of potential users . The sophisticated inputting, sorting, batch, and interactive retrieval routines that characterized the mainframe software packages of the nineteen-seventies can now be duplicated on a micro, while the storage capacity of hard discs enables typically lengthy museum specimen records to be held in sufficient numbers to cater for large collections . The availability of powerful relational database packages for microcomputers opens up exciting possibilities for the interactive interrogation of large complex files on low-cost hardware, in a way that would have seemed impossible just a few years ago . The capacity of the newest optical storage media makes it likely that within a very short time storage will cease to be any kind of limiting factor, even for the very largest collections . The effectiveness of computerized information retrieval has had additional benefits on the way museums deal with specimen-related data. The in­ formation must be structured in a standard form before inputting, and the terminology applied to

6 . 3 Museology different data categories must be rigidly controlled if sorting procedures are to produce useful output. This inflicts higher standards of data recording on museums than has traditionally been the case . Taken to its logical extreme, the adoption of a single data standard by museums, combined with an agreed thesaurus for terminology control, opens up the exciting prospect of combining museum databases and of their remote interrogation by users . How­ ever, there is as yet little international agreement about the structuring of museum data, and the question of terminology control is at an even more rudimentary stage . The V.K. is probably as far advanced as any other country in this regard, with the Museum Data Standard of the Museum Docu­ mentation Association (MDA) now in widespread use by museums, whether they employ the MDA's manual recording cards and/or supporting software packages or choose to develop in-house applications of commercial database packages . Full computerization o f specimen records entails a massive short-term commitment of data preparation time, since it obviously involves keying in all the manual records accumulated during a museum's history. Crucially, it also entails structuring the data and terminology to conform with agreed standards - and rigorously checking the data input. This is beyond the staffing resources of most museums, and computerization is commonly restricted initially to upgrading inventories; detailed computer catalogu­ ing is often limited to new material entering the museum . Nevertheless, the automatic scanning of manual records using developments of the 'optical character readers' already available open up the exciting possibility of direct input of typed or even handwritten records to a computer database, thereby drastically reducing data preparation time . At a time when museums are coming under in­ creasing pressure to make their reserve collections more accessible, new technology has an important role to play. As the efficient management of large taxonomic collections in the public domain becomes increasingly expensive, those responsible for such collections must become more adept at justifying their unique role to those who ultimately pay the cost through taxes or entrance charges . A database compiled for basic collections management pur­ poses can be made available to the general visitor via interactive terminals, after only minor modifications to strip out sensitive information (donor address, insurance value, storage location, etc . ) . Linked to a video disc (which are already capable of holding 50 000 images), such a system could provide instant

519

visual access on demand to a collection, yet involve no physical risk to the specimens themselves .

References Bassett, M . G . (ed . ) 1979 . Curation of palaeontological collec­ tions. Special Papers in Palaeontology, No . 22. Brunton, C . H . C . , Besterman, T.P. & Cooper, J.A. 1985 . Guidelines for the curation of geological materials . Miscellaneous Paper of the Geological Society, No . 17. Crowther, P.R. & Wimbledon, W.A. (eds) 1988 . The use and conservation of palaeontological sites. Special Papers in Palaeontology, No. 40. Light, R. B . , Roberts, D.A. & Stewart, J.D. (eds) 1986 . Museum documentation systems . Butterworths, London. Raup, D . M . , Black, c . c . , Blackstone, S., Dole H., Grogan, S . , Larsen . , P . , Jenkins, F., Pojeta, J . , Robinson, P . , Roybal, c . , Schopf, J.W., Stehli, F . G . & Wolberg, O . 1987. Palaeontological collecting. National Academy Press, Washington, DC. Roberts, D.A. (ed . ) 1988 . Collections management for museums. Museum Documentation Association, Cambridge, U.K.

6 . 3 . 3 Exhibit Strategies R . S . MILES

Introduction The purpose of mounting exhibits is normally to communicate information, so this section looks at some of the principles behind successful distance communication . By distance communication we imply the existence of a gap, either in time or space, between the sender and receiver of a message . This mode of communication applies typically to exhibits, whether comprising single posters or en­ tire museums . Classroom teaching, lectures and demonstrations, on the other hand, involve face-to­ face communication, in which the sender is there in person. It is important to distinguish between these two modes of communication, because if the sender is not there to answer questions, an effort must be made, at the stage of designing the communication, to ensure that it is intelligible to its intended audi­ ence . Good communication is selected for a purpose, and has a sound logical structure . Successful com­ municators know their audience, and attend both to the content ('what to say') and the form ('how to say it') of their communications .

520

6 Infrastructure of Palaeobiology

The audience The organizers of exhibits inevitably form a mental image of their audience . Ideally this is based on hard data, e . g . the audience's vocabulary, under­ standing, interest in the subject, and commitment to its study. Care must be taken to avoid creating a false image, either through professional self-interest or limited experience of the audience . Where suf­ ficient data are not available, survey techniques similar to those found in market research can be used . Questionnaire design and sampling methods have been described by Loomis (1987) and Miles et al. (1988) . Accurate knowledge of the audience helps the communicator to connect his or her message with the viewer's world, and use language that is matched to the viewer's requirements . It is also helpful to know about the audience' s misunder­ standings (or 'alternative conceptions'), for these may need to be removed before the desired in­ formation can be imparted . For example, an intuit­ ively held Lamarckian view of evolution might block understanding of Darwin's theory of natural selection.

Selection and structure of content What to communicate, where to start, and how to continue? - in other words, the selection and order­ ing of the content - are basic questions in organiz­ ing any exhibit . Generally, there is more to say about a subject than the space or other resources allow, or the viewer's stamina permits, and there has to be some selection . The basis of this selec­ tion is a clear statement of purpose . For a group of exhibits this statement takes the form, at the broadest level, of a series of aims . But a more detailed statement of purpose is required for individual exhibits, and this is best provided by listing the teaching points, i . e . the facts, concepts, relationships, procedures, and so on that need to be communi­ cated . Teaching points generally divide into key concepts and ancillary points . Thus some are in­ cluded simply in order to define other concepts or to remove misconceptions, others to ensure the positive transfer of knowledge . Teaching points also help to promote clear communication among those responsible for exhibits, and provide a basis for judging the success of exhibits as pieces of communication (below) . The ordering or sequencing of content is done with the help of a strong central theme, to give a good flow of ideas and a framework that unifies the

facts, theories, and so on that are spelled out in the teaching points . It is important to tell only one story at a time; to organize things so that the audience knows what is going on (e . g . where they are going and how long it will take to get there); and to make the status of each message clear (e . g . is it the main conclusion or a supporting argument, is it a question or an instruction?) . A common way of ordering ideas is to place them in a logical sequence, e . g . concept A is dealt with before concept B because concept A must be under­ stood before concept B can be understood . However, it is often unwise to argue from first principles in exhibitions for the lay public, because of the need to attract and keep the viewer's interest and connect the message to his or her familiar world . Thus, if no particular sequence of concepts can be chosen on grounds of logical relations, it might still be better to deal with concept A before concept B, because on psychological grounds it is easier for the viewer to understand concept A before concept B (Fig. 1 ) . T o help the audience know what i s going o n it should be told, in an introductory exhibit, what the exhibits are about and how they are organized . In large exhibitions it may be necessary to repeat such information in different places. In addition to con­ ceptual orientation, it may also be necessary to provide topographic orientation, i . e . signposts, maps, and exhibit numbers . The aim is to indicate the correct route through the exhibits, and such orientation devices must be designed to make sense to viewers who have no prior understanding of the content and arrangement of the exhibits (Miles et al. 1988) .

Selection of media Communications media are the physical means of transporting messages from the sender to the re­ ceivers . Some media are normally used in the static mode, e . g . three-dimensional objects, graphics, and text; others are used in the dynamic mode and undergo a change of state during operation, e . g . audiovisuals and interactive computers . Selecting the appropriate medium for a particular message is important, yet never easy . There are few rules to assist the selection procedure, and the assessment of setting-up and maintenance costs is likely to weigh as heavily as educational advantage . If an exhibit is to communicate change over time or movement (e . g . continental drift), it is useful to use a dynamic medium, possibly a film or working model. But the basic exhibit media still remain

6 . 3 Museology 1.

D i v e rs i ty of l i fe

I I I

-------

2.

521

C h a r l e s D a rw i n

3 . Domestic selection 4.

TNt .. .. ....1..-.1. """" et I..... � Hy �"" � " f1y .mn � """. ,..... ... ,.... rI¥>t

S p e c i e s as taxon o m i c g r o u p s

• ..., ... tho �� ....

5 . S p e c i e s as b re ed i n g g r o u p s

I

6.

Rep rod u ct ive pote n t i a l

7.

E n v i ro n m e n t

8.

I n h e ritance

I I I

i::'=I.:.

1r· ..t.bea> ....

+

11""'''''.1 ?

,
o-....... .. "" .--.. I6Ioo