Origen Herviboria

Evolution of Herbivory in Terrestrial Vertebrates
Perspectives from the fossil record
Although herbivory probably first appeared over 300 million

years ago, it only became established as a common feeding
strategy during Late Permian times. Subsequently, herbivory
evolved in numerous lineages of terrestrial vertebrates, and the
acquisition of this mode of feeding was frequently associated
with considerable evolutionary diversification in those lineages.
This book represents the first comprehensive overview of the
evolution of herbivory in land-dwelling amniote tetrapods in
recent years. In The Evolution of Herbivory in Terrestrial Vertebrates
leading experts review the structural adaptations for, and the
evolutionary history of, feeding on plants in the major groups
of land-dwelling vertebrates, especially dinosaurs and ungulate
mammals. As such it will be the definitive reference source on
this topic for evolutionary biologists and vertebrate
paleontologists.
h a n s - d i e t e r s u e s is Vice-President for Collections and

Research at the Royal Ontario Museum in Toronto, and
Professor of Zoology at the University of Toronto. He is
interested in the evolution of late Paleozoic and Mesozoic
tetrapods, especially dinosaurs and their relatives, and patterns
of ecosystem change through time. Professor Sues has co-edited
two other volumes, Evolution of Terrestrial Ecosystems Through Time
(1992) and In the Shadow of the Dinosaurs: Early Mesozoic Tetrapods
(1994).
Evolution of
Herbivory
in Terrestrial
Vertebrates
Perspectives from the fossil record
Edited by
Hans-Dieter Sues
Royal Ontario Museum and University of Toronto
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo
Cambridge University Press

The Edinburgh Building, Cambridge CB2 2RU, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521594493
© Cambridge University Press 2000
This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.
First published 2000

This digitally printed first paperback version 2005
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication data

Evolution of herbivory in terrestrial vertebrates: perspectives from the fossil record /
edited by Hans-Dieter Sues.
p. cm.
Includes index.
ISBN 0 521 59449 9 (hardcover : alk. paper)
1. Herbivores, Fossil–Evolution. 2. Vertebrates, Fossil. I. Sues, Hans-Dieter, 1956–
QE841.E96 2000
566–dc21 00-020863
ISBN-13 978-0-521-59449-3 hardback

ISBN-10 0-521-59449-9 hardback
ISBN-13 978-0-521-02119-7 paperback

ISBN-10 0-521-02119-7 paperback
Contents
List of contributors page vii

Preface ix
1 Herbivory in terrestrial vertebrates: an introduction 1

Hans-Dieter Sues
2 Herbivory in late Paleozoic and Triassic terrestrial vertebrates 9
Robert R. Reisz and Hans-Dieter Sues
3 Prosauropod dinosaurs and iguanas: speculations on the diets of
extinct reptiles 42
Paul M. Barrett
4 The evolution of sauropod feeding mechanisms 79
Paul Upchurch and Paul M. Barrett
5 Plant-eaters and ghost lineages: dinosaurian herbivory
revisited 123
David B. Weishampel and Coralia-Maria Jianu
6 Dental constraints in the early evolution of mammalian
herbivory 144
John M. Rensberger
7 Patterns in the evolution of herbivory in large terrestrial mammals:
the Paleogene of North America 168
Christine M. Janis
8 Origin and evolution of the grazing guild in Cenozoic New
World terrestrial mammals 223
B r u c e J. M a c F a d d e n
Taxonomic index 245

Subject index 249
[v]
Contributors
Paul M. Barrett
Department of Zoology, University of Oxford, South Parks Road, Oxford
OX1 3PS, UK
Christine M. Janis
Department of Ecology and Evolutionary Biology, Brown University,
Providence, RI 02912
Coralia-Maria Jianu
Muzeul Civilizatiei Dacice si Romane Deva, B-dul 1 Decembrie Nr. 39,
2700 Deva, Romania
B r u c e J. M a c F a d d e n
Florida Museum of Natural History, University of Florida, Gainesville, FL
32611
Robert R. Reisz
Department of Biology, Erindale College, University of Toronto, 3359
Mississauga Road, Mississauga, ON L5L 1C6, Canada
John M. Rensberger
Burke Museum and Department of Geological Sciences, Box 353010, Uni-
versity of Washington, Seattle, WA 98195
Hans-Dieter Sues
Royal Ontario Museum, 100 Queen’s Park, Toronto, ON M5S 2C6, and
Department of Zoology, University of Toronto, Toronto, ON M5S 1A1,
Canada
[vii]
viii Contributors
Paul Upchurch
School of Biological Sciences, University of Bristol, Woodland Road,
Bristol BS8 1UG, UK
D a v i d B. W e i s h a m p e l
Department of Cell Biology and Anatomy, The Johns Hopkins University
School of Medicine, 725 North Wolfe Street, Baltimore, MD 21205
Preface
This book presents a collection of paleontological perspectives on the

origin and evolution of herbivory in various major taxa of terrestrial ver-
tebrates. The complex interactions between plants and their animal con-
sumers have long been the subject of much interest to evolutionary
biologists. At the same time, most studies exploring coevolutionary rela-
tionships have focused on extant organisms, and generally little atten-
tion has been paid to the historical development of plant–animal
interactions through time documented by the fossil record.
Most contributors to this volume review the nature and acquisition of
structural features of the skull and dentition suitable for feeding on high-
fiber plant material in various major lineages of herbivorous tetrapods. In
some instances, they also discuss other lines of evidence (such as isotopic
data) bearing on this issue as well as the possible impact of herbivory on
the evolutionary diversification of that group. Traditionally, paleobiolog-
ical studies have assumed a direct relationship between form and func-
tion, but current research on the functional morphology of extinct
organisms is much more mindful of the inherent theoretical and practical
difficulties in reconstructing the habits of ancient organisms. The present
volume cannot and does not provide a comprehensive account of her-
bivory in extinct vertebrates. Rather, it is intended as a review of current
research on some of the key issues for advanced students of evolutionary
biology, historical ecology, and paleobiology and, it is to be hoped, as a
stimulus for further work.
Most chapters are based on contributions presented at a symposium
on the evolution of herbivory in insects and terrestrial vertebrates held
during the Sixth North American Paleontological Convention (NAPC-96)
at the Smithsonian Institution in Washington, D.C. in June 1996. Special
[ix]
x Preface
thanks are due to the co-organizer of this symposium, Conrad C. Laban-

deira (National Museum of Natural History), for his co-operation and for
sharing my enthusiasm for this subject, and to the Organizing Commit-
tee of NAPC-96 for its interest and support.
As editor, I am indebted first to the contributors, for taking part in this
project, for (in some cases) meeting deadlines, and for cheerfully putting
up with my editorial efforts. At Cambridge University Press, I would like
to thank Robin Smith, who first encouraged me to compile this collection
of papers, and Tracey Sanderson for her continuing encouragement and
support. During the editing of this volume, I have heavily relied on the
expertise of many reviewers. I would like to acknowledge the generous
assistance and thoughtful comments provided by Richard Beerbower
(State University of New York, Binghamton), Robert L. Carroll (McGill
University), Peter Dodson (University of Pennsylvania), Robert J. Emry
(Smithsonian Institution), James O. Farlow (Indiana University–Purdue
University, Fort Wayne), Mikael Fortelius (University of Helsinki), the
late Nicholas Hotton III (Smithsonian Institution), John P. Hunter (New
York College of Osteopathic Medicine, Old Westbury), Christine M. Janis
(Brown University), Gillian King (University of Cambridge), Paul L. Koch
(University of California at Santa Cruz), David B. Norman (University of
Cambridge), Robert R. Reisz (Erindale College, University of Toronto,
Mississauga), Richard K. Stucky (Denver Museum of Natural History),
David B. Weishampel (The Johns Hopkins University), and Jeffrey A.
Wilson (University of Chicago). I thank Janice Robertson for her careful
copy-editing and especially Joan Burke (Royal Ontario Museum) for her
meticulous proof-reading.
Hans-Dieter Sues
hans-dieter sues
Herbivory in terrestrial vertebrates: an

introduction
Introduction
Understanding the ecological attributes of extinct organisms has

long been a major research topic in paleobiology, dating back to the pio-
neering work of the French paleontologist Georges Cuvier in the early
nineteenth century. Inferences concerning the ecology of an extinct
organism can be based on functional interpretation of its structure, by
analogy with present-day relatives, or from the sedimentary context and
distribution of fossil remains referable to this taxon (Wing et al. 1992).
Traditionally, functional morphology has been the most widely used of
these approaches. It basically relies on the analysis of organisms as simple
machines with functional attributes that can be inferred from the physi-
cal properties of their bodies as well as from their shape and size. Chemi-
cal analyses of hard tissues (such as extraction of preserved stable carbon
isotopes) increasingly are providing significant new data for inferring
diet in extinct animals. In recent years, researchers have developed
various procedures for linking inferences concerning function in fossils
to phylogenetic analyses, increasing confidence in the robustness of these
reconstructions (see various papers in Thomason [1995]).
Herbivory, the consumption of plant tissues, is a widespread phenom-
enon among terrestrial vertebrates. It has frequently and independently
evolved in many lineages of amniotes during the last 300 million years or
so. Some major groups of herbivorous tetrapods, such as ungulate
mammals and ornithischian dinosaurs, attained great abundance and
taxonomic diversity. Indeed, the advent of herbivory among land-
dwelling tetrapods was one of the key events in the history of life on land.
It led to the establishment of ‘modern’ continental ecosystems, with vast
[1]
2 hans-dieter sues
numbers of herbivores supporting a relatively small number of carni-

vores, during the Permian period (Olson 1966; Hotton et al. 1997; Sues and
Reisz 1998).
Feeding on plants requires many morphological and physiological
modifications to facilitate the efficient reduction and digestion of plant
tissues. Plant fodder contains less caloric energy per volume unit than do
animal foods (Southwood 1973). Furthermore, in most instances, much of
that energy is tied up in substances that are difficult to digest for verte-
brates. The contents of plant cells are enclosed by walls that are primarily
composed of cellulose, hemicellulose, and lignin. Cellulose is a polymer
of glucose, but its glucose units are linked together in such a fashion that
it cannot be readily broken down. This makes plant tissues, to varying
degrees, more difficult to digest than animal tissues, which are devoid of
resistant cell walls. Extant vertebrates lack any endogenous enzymes to
hydrolyze the compounds forming the cell walls of plant tissues, but
many micro-organisms (bacteria, protists) can produce them. Thus many
plant-eating tetrapods have entered into endosymbiotic relationships
with such organisms to facilitate cellulysis in their digestive tracts, result-
ing in the production of sugars and volatile fatty acids that can be
absorbed by the vertebrate host. Not only are the walls of the plant cells
themselves resistant to unaided digestion by vertebrates but they often
protect the digestible cellular contents such as lipids, sugars, organic
acids and proteins. Initial breakdown of the cell walls by mechanical or
chemical action is thus required.
Plant tissues are highly variable in terms of their nutritional value to
vertebrate consumers. Certain plant parts, such as fruits, seeds, and
immature vegetative tissues, contain much digestible matter that is only
protected by relatively delicate cell walls and often require only little pro-
cessing to make the enclosed nutrients available to the consumer.
However, these are seasonally available resources, and most herbivores
subsist on the tougher, more cellulose-rich vegetative structures, such as
leaves, roots, shoots and stems that form the bulk of available plant
material. The effective utilization of such fodder by the vertebrate consu-
mer requires two steps. First, the plant material must be mechanically
broken up by oral processing or by comminution in a muscular foregut or
gizzard using ingested grit and pebbles. Grinding the plant fodder more
finely clearly increases the rate at which herbivores can process it (Bjorn-
dal et al. 1990). Second, symbiotic micro-organisms in the gut must
convert the cellulose into volatile fatty acids (especially acetic, butyric,
Introduction 3
and proprionic acid) that can be readily absorbed by the vertebrate host.
Not all extant herbivorous vertebrates employ microbial endosymbionts.
Some animals such as the giant panda Ailuropoda melanoleuca (Schaller et al.
1985) apparently compensate for this lack by consuming large amounts of
plant fodder as well as maintaining relatively low activity and growth
rates (see McNab 1986).
Most studies of herbivory in present-day vertebrates have focused on
the digestive performance of plant-eating mammals, particularly ungu-
lates, due to the commercial importance of the latter (McBee 1977; Chivers
and Langer 1994). The adaptations for feeding on plants in non-mam-
malian tetrapods are much less well studied. Relatively few taxa of
present-day reptiles are obligate herbivores, and the range of structural
features of the skull and dentition for feeding on plants is limited com-
pared with that in mammals (Throckmorton 1976). However, more recent
work has demonstrated that plant-eating iguanid lizards can degrade cel-
lulose and hemicellulose almost as efficiently as herbivorous mammals
(Troyer 1984).
Herbivory in extinct vertebrates
Unequivocal evidence for trophic interactions between extinct animals

and plants is only rarely found. Therefore, inferences concerning herbi-
vory in extinct vertebrates must rely almost entirely on circumstantial
evidence. Starting with Cuvier’s pioneering work, paleontologists have
correlated broadly defined feeding categories with specific morphologi-
cal attributes, primarily of the skull and dentition, in tetrapods. Such cor-
relations are based either on analogy to present-day animals with known
dietary habits or on biomechanical assessment of the suitability of tooth
structure for processing potential food items (e.g., Rensberger 1973; Lucas
1979). In a few instances, fossilized gut contents provide a more direct line
of evidence, but it is important to rule out postmortem introduction of
material into the abdominal cavity. Furthermore, some food stuffs may
have already been digested prior to the death and fossilization of the con-
sumer.
Mammals are distinguished from most other tetrapods by extensive
oral processing of food or mastication. Thus their masticatory apparatus
provides an excellent model for correlating tooth structure with diet. The
specific relationships between gross occlusal structures and mechanical
performance can be modeled (Rensberger 1973; Lucas 1979; Lucas and
4 hans-dieter sues
Luke 1984) and, in many instances, can be tested experimentally in extant

forms (Kay and Sheine 1979; Teaford and Oyen 1989). Wear on tooth
enamel at the microscopic level can be related to diet and the masticatory
system (Walker et al. 1978; Rensberger 1986), and complex functional asso-
ciations exist between enamel microstructure, gross occlusal features,
and chewing direction (Koenigswald 1980; Rensberger and Koenigswald
1980). As teeth are the most commonly preserved part of the mammalian
skeleton, they thus become a valuable resource for paleontologists and
physical anthropologists concerned with paleoecological questions.
In non-mammalian vertebrates, the correlation between diet and den-
tition is much more difficult because oral processing of food and thus
tooth occlusion are uncommon. Furthermore, in some groups (especially
birds) with herbivorous taxa, teeth are absent altogether. However, based
on biomechanical considerations and comparison with extant forms, it is
still possible in many cases to interpret craniodental attributes in general
terms of their biomechanical suitability for feeding on a particular type of
material.
Dental features suggestive or indicative of high-fiber herbivory
include dentitions adapted for crushing and grinding, or marginal teeth
with labiolingually compressed, leaf-shaped, and cuspidate crowns suit-
able for puncturing and shredding plant fodder. Features of the skull and
mandible variously associated with feeding on plants, especially in
mammals, include short tooth rows (along with foreshortening of the
snout and mandible), elevation/depression of the jaw joint relative to the
occlusal plane for increased mechanical advantage of the adductor jaw
muscles, enlargement of the adductor chambers and temporal openings
as well as deepening of the zygomatic arches and mandibular rami for the
origin and insertion of substantial adductor jaw muscles, and jaw joints
suitable for complex mandibular motion (Maynard Smith and Savage
1959; Turnbull 1970; Rensberger 1986; Janis 1995). Finally, plant-eating
tetrapods typically have longer and/or bulkier digestive tracts, and a
longer and/or broader trunk region, than related faunivorous forms
(Schiek and Millar 1985; Dearing 1993) because part of the gut is modified
to form a reservoir housing and creating suitable pH conditions for the
endosymbiotic micro-organisms involved in the fermentative break-
down of the cellulose from the ingested plant fodder. Passage of food
through longer intestinal tracts also allows for longer periods of time for
processing of resistant materials. Thus the rib cages of many herbivorous
tetrapods are either much wider and more capacious (‘barrel-shaped’)
Introduction 5
than those of their closest faunivorous relatives, or the trunk region is

elongated to accommodate a longer digestive tract. This is reflected in
overall body proportions as well as in the structure of the vertebral
column.
In this volume, nine contributors examine the structural correlates of
herbivory in various taxa of herbivorous tetrapods and discuss them in
relation to the evolutionary diversification of these groups.
Reisz and Sues (Chapter 2) review the earliest occurrences of herbivory
in amniote tetrapods. They use the craniodental and, to a lesser extent,
postcranial features discussed above to identify probably herbivorous
taxa among late Paleozoic and Triassic amniotes. The oldest known puta-
tive herbivores, of Late Pennsylvanian age, are the Diadectidae, a group of
tetrapods very closely related to the Amniota, and the basal synapsid
Edaphosaurus. Surprisingly, feeding on plants only became common
among tetrapods during the Late Permian. At that point in time, verte-
brate herbivores first became a key component of the terrestrial food web,
leading to the establishment of the basic trophic structure of modern ter-
restrial ecosystems (Olson 1961, 1966). Phylogenetic analysis indicates that
herbivory was repeatedly and independently acquired in various lineages
of amniotes during the Late Carboniferous and Permian (Sues and Reisz
1998).
In his discussion of possible diets for prosauropod dinosaurs, Barrett
(Chapter 3) injects a cautionary note concerning the identification of her-
bivory in non-mammalian tetrapods based primarily on dental features.
Citing anecdotal evidence, he argues that present-day iguanid lizards,
which have teeth with leaf-shaped, serrated crowns, have long provided a
model for inferring herbivory in certain dinosaurs with similar teeth, but
are actually omnivorous rather than strictly herbivorous. However, few
extant vertebrates are exclusively herbivorous, and many herbivores sup-
plement their diet by the intake of animal protein.
Upchurch and Barrett (Chapter 4) review the craniodental and postcra-
nial features of the constituent clades of sauropod dinosaurs in relation to
possible feeding strategies. They marshal evidence ranging from patterns
to dental microwear to browsing height inferred from skeletal propor-
tions to demonstrate a diversity of possible feeding styles among sauro-
pods. Sauropodomorph dinosaurs (prosauropods and sauropods) are
noteworthy because they represented the first diverse radiation of mega-
herbivores (sensu Owen-Smith [1988], with body weights exceeding one
metric ton) and were capable of foraging at greater heights (i.e., more
6 hans-dieter sues
than one or two meters above the ground) than other plant-eating tetra-
pods at that time.
Weishampel and Jianu (Chapter 5) present a detailed phylogenetic
analysis of the various major taxa of dinosaurian herbivores. They
emphasize the importance of taking into consideration the unrecorded
segments of lineages (‘ghost lineages’) that can be inferred based on phy-
logenetic hypotheses in estimating overall diversity. The authors find
little evidence to support the currently popular coevolutionary scenario
linking the onset of the evolutionary diversification of flowering plants
(angiosperms) to the radiations of large ornithopod and ceratopsian dino-
saurs.
Rensberger (Chapter 6) provides an elegant analysis of the biomechan-
ical factors dictating tooth configuration in herbivorous placental
mammals from the early Cenozoic. In the two most common groups of
early Paleocene ungulates in North America, the shearing component of
mastication was greatly reduced relative to the condition in more primi-
tive mammals and compression became the dominant component. Rens-
berger demonstrates that the stresses in the tooth enamel induced by
chewing are lower in low, wide cusps than in tall, sharp cusps. Prism de-
cussation (where zones of prisms with a common orientation alternate
with zones of prisms with a different orientation) in the enamel increases
resistance to fracturing. It appears in most ungulates as body size
increases later during the Paleogene. With the acquisition of stronger
enamel, the earlier trend toward more blunt cusps is reversed, and shear-
ing crests reappear.
Based on a recent compendium of Tertiary mammals from North
America, Janis (Chapter 7) reviews the diversification of the different
types of feeding strategies (as deduced from tooth shape) in herbivorous
mammals from the Paleogene (Paleocene–Oligocene) of North America.
She relates the relative abundance of the different kinds of tooth shape to
climatic changes during the Early Tertiary inferred from other lines of
evidence. The early ‘condylarths’ were presumably omnivorous rather
than strictly herbivorous. Feeding on foliage (folivory) apparently did not
occur until the latest Paleocene. The earliest artiodactyls and perissodac-
tyls appear in the early Eocene, but they had dentitions more typical of
omnivores/frugivores, and folivory in these groups was not common
until the late Eocene. Although ungulates with relatively high-crowned
teeth are known from the Paleogene, there is no evidence of true grazers
in the fossil record until the Neogene.
Introduction 7
MacFadden (Chapter 8) presents an overview of recent work by him and

various colleagues, including the application of isotopic analyses, on the
origin and early evolution of the grazing guild among Neogene mammals
from the Americas. Grazing first developed as a major feeding strategy
during the Miocene. The acquisition of high-crowned teeth, which appears
to be an adaptation to the pervasive occurrence of highly abrasive silica par-
ticles in grasses, is interpreted as a coevolutionary response to the advent of
widespread grassland communities at that time. Grazing was indepen-
dently acquired in various North American groups (artiodactyls, perisso-
dactyls, proboscideans, rodents) and in the South American notoungulates
prior to the formation of the Panamanian land bridge. Based on MacFad-
den’s work, grasslands and grazing mammals may have originated slightly
earlier in South America than in North America. Data for stable carbon iso-
topes indicate that the earliest grasslands were formed by C£ grasses. The
change to the grasslands dominated by C¢ grasses (which comprise most
grasslands today) occurred after the late Miocene and may explain the
observed decline in grazing diversity after the late Miocene.
References
Bjorndal, K. A., Bolten, A. B., and Moore, J. E. (1990). Digestive fermentation in

herbivores: effect of food particle size. Physiol. Zool. 63:710–721.
Chivers, D. J., and Langer, P. (eds.) (1994). The Digestive System in Mammals: Food, Form and
Function. Cambridge and New York: Cambridge University Press.
Dearing, M. D. (1993). An alimentary specialization for herbivory in the tropical
whiptail lizard Cnemidophorus murinus. J. Herpet. 27:111–114.
Hotton, N. III, Olson, E.C., and Beerbower, R. (1997). Amniote origins and the discovery
of herbivory. In Amniote Origins, ed. S. S. Sumida and K. L. M. Martin, pp.
207–264. San Diego: Academic Press.
Janis, C. M. (1995). Correlations between craniodental morphology and feeding
behavior in ungulates: reciprocal illumination between living and fossil taxa. In
Functional Morphology in Vertebrate Paleontology, ed. J. J. Thomason, pp.76–98.
Cambridge and New York: Cambridge University Press.
Kay, R. F., and Sheine, W. S. (1979). On the relationship between chitin particle size and
digestibility in the primate Galago senegalensis. Amer. J. Phys. Anthrop.
50:301–308.
Koenigswald, W. von. (1980). Schmelzstruktur und Morphologie in den Molaren der
Arvicolidae (Rodentia). Abh. Senckenberg. Naturforsch. Ges. 539:1–129.
Lucas, P. W. (1979). The dental-dietary adaptations of mammals. N. Jb. Geol. Paläont., Mh.
1979:486–516.
Lucas, P. W., and Luke, D. A. (1984). Chewing it over: basic principles of food breakdown.
In Food Acquisition and Processing in Primates, ed. D. A. Chivers, B. A. Wood, and A.
Bilsborough, pp. 283–301. New York: Plenum Press.
8 hans-dieter sues
Maynard Smith, J., and Savage, R. J. G. (1959). The mechanics of mammalian jaws. School
Sci. Rev. 141:289–301.
McBee, R. H. (1977). Fermentation in the hindgut. In Microbial Ecology of the Gut, ed. R. T.
J. Clarke and T. Bauchop, pp. 185–222. London and New York: Academic Press.
McNab, B. K. (1986). The influence of food habits on the energetics of eutherian
mammals. Ecol. Monographs 56:1–19.
Olson, E. C. (1961). The food chain and the origin of mammals. In International
Colloquium on the Evolution of Lower and Non-Specialized Mammals, ed. G.
Vanderbroek, pp. 97–116. Brussels: Koninklijke Vlaamse Akademie voor
Wetenschappen, Letteren en Schone Kunsten van België.
Olson, E. C. (1966). Community evolution and the origin of mammals. Ecology
47:291–308.
Owen-Smith, R. N. (1988). Megaherbivores: The Influence of Very Large Body Size on Ecology.
Rensberger, J. M. (1973). An occlusal model for mastication and dental wear in
herbivorous mammals. J. Paleont. 47:515–528.
Rensberger, J. M. (1986). Early chewing mechanisms in mammalian herbivores.
Paleobiology 12:474–494.
Rensberger, J. M., and Koenigswald, W. von. (1980). Functional and phylogenetic
interpretation of enamel microstructure in rhinoceroses. Paleobiology 6:477–495.
Schaller, G. B., Hu J., Pan W., and Zhu J. (1985). The Giant Pandas of Wolong. Chicago:
University of Chicago Press.
Schiek, J. O., and Millar, J. S. (1985). Alimentary tract measurements as indicators of diet
in small mammals. Mammalia 49:93–104.
Southwood, T. R. E. (1973). The insect/plant relationship — an evolutionary perspective.
Symp. Roy. Soc. Lond. 6:3–30.
Sues, H.-D., and Reisz, R. R. (1998). Origin and early evolution of herbivory in terrestrial
tetrapods. Trends Ecol. Evol. 18:141–145.
Teaford, M. F., and Oyen, O. J. (1989). Differences in the rate of molar wear between
monkeys raised on different diets. J. Dental Res. 68:1513–1518.
Thomason, J. J. (ed.) (1995). Functional Morphology in Vertebrate Paleontology. Cambridge
and New York: Cambridge University Press.
Throckmorton, G. S. (1976). Oral food processing in two herbivorous lizards, Iguana
iguana (Iguanidae) and Uromastix aegyptius (Agamidae). J. Morph. 48:363–390.
Troyer, K. (1984). Structure and function of the digestive tract of a herbivorous lizard,
Iguana iguana. Physiol. Zool. 57:1–8.
Turnbull, W. D. (1970). Mammalian masticatory apparatus. Fieldiana Geol. 18:149–356.
Walker, A., Hoeck, N., and Perez, L. (1978). Microwear of mammalian teeth as an
indicator of diet. Science 201:908–910.
Wing, S. L., Sues, H.-D., Potts, R., DiMichele, W. A., and Behrensmeyer, A. K. (1992).
Evolutionary paleoecology. In Terrestrial Ecosystems through Time, ed. A. K.
Behrensmeyer et al., pp. 1–13. Chicago: University of Chicago Press.
robert r. reisz and hans-dieter sues
Herbivory in late Paleozoic and Triassic terrestrial

vertebrates
Introduction
The exploitation of land plants as a major food resource by amniote

tetrapods led to profound changes in the pattern of trophic interactions
in terrestrial ecosystems during the late Paleozoic (Olson 1961, 1966; King
1996; Hotton et al. 1997; Sues and Reisz 1998). Prior to the appearance of
various forms specialized for feeding on plants, tetrapods could access the
plentiful vegetal resources only indirectly through detritivory and con-
sumption of invertebrates that fed on plants and/or plant detritus (Olson
1961, 1966). The oldest plant-eating tetrapods are known from the Late
Pennsylvanian (Late Carboniferous) of North America and Europe.
However, herbivores did not form a major component of the known ter-
restrial tetrapod assemblages from the Early Permian. Only during the
Late Permian, some 40 million years after their first appearance, did
plant-eating tetrapods become abundant and much diversified in the
fossil record. At that time, a ‘modern’ pattern of trophic interactions was
established, with a vast standing crop of herbivores sustaining a relatively
small number of top carnivores. Communities of this type are first docu-
mented by diverse assemblages of tetrapods from the Late Permian
portion of the Beaufort Group of South Africa (Kitching 1977) and, to a
lesser extent, from more or less coeval continental strata in other regions
of the world.
Morphological and physiological correlates of herbivory
The effective utilization of high-fiber plant material by vertebrates

requires two steps. First, the plant material must be mechanically broken
[9]
10 robert r. reisz and hans-dieter sues
up by oral processing or by comminution in a muscular gizzard using

ingested grit and pebbles. Present-day herbivorous lizards do not chew
their food (Throckmorton 1976), and leaves often pass through the gas-
trointestinal tract virtually intact (Bjorndal 1979; Iverson 1982). Grinding
the plant fodder more finely clearly increases the rate at which herbivores
can process it (Bjorndal et al. 1990). Second, symbiotic micro-organisms in
the gut must convert the cellulose into volatile fatty acids (especially
acetic, butyric, and proprionic acid) that can be readily absorbed by the
vertebrate host.
In this chapter, we review the possible exploitation of high-fiber plant
material by various groups of late Paleozoic and Triassic terrestrial tetra-
pods. We discuss the earliest stages in the evolution of herbivory among
terrestrial tetrapods and the possible impact of this feeding strategy on
the evolution of continental vertebrate communities. It is important to
stress here that relatively few extant reptiles are obligate herbivores (see
also Barrett, this volume), and most feed on a variety of dietary items (see
below). By analogy, we expect similar dietary versatility among late Paleo-
zoic and early Mesozoic amniotes. However, we should note that the skel-
etal modifications observed in certain extinct forms are, among extant
vertebrates, most commonly associated with diets that include high-fiber
plant material as the primary component.
In most cases, inferences concerning the dietary habits of extinct verte-
brates rely on circumstantial evidence. Dental features suggestive or
indicative of high-fiber herbivory include dentitions adapted for crush-
ing and grinding, or marginal teeth with labiolingually compressed, leaf-
shaped, and cuspidate crowns suitable for puncturing and shredding
plant fodder. Cranial features variously associated with feeding on plants,
especially in mammals, include short tooth rows (and related foreshort-
ening of the snout and mandible), elevation or depression of the jaw joint
relative to the occlusal plane for increased mechanical advantage of the
adductor jaw muscles during biting, enlargement of the adductor cham-
bers and temporal openings as well as deepening of the zygomatic arches
and mandibular rami to accommodate powerful adductor jaw muscles,
and jaw joints suitable for complex mandibular movements (Maynard
Smith and Savage 1959; Turnbull 1970; Rensberger 1986; Janis 1995). Her-
bivorous tetrapods typically have longer and/or bulkier digestive tracts,
and a longer and/or broader trunk region, than related faunivorous forms
(Schiek and Millar 1985; Dearing 1993) because part of the gut is modified
to form a reservoir housing and creating suitable pH conditions for endo-
Herbivory in early tetrapods 11
symbiotic micro-organisms for fermentative breakdown of the cellulose

from the ingested plant fodder. Passage of food through longer intestinal
tracts allows for longer periods of time for the processing of resistant
materials. Thus the rib cages of most early herbivorous tetrapods (espe-
cially caseid synapsids) are either much wider and more capacious
(‘barrel-shaped’) than those of their closest faunivorous relatives, or,
alternatively, the trunk region is distinctly elongated (e.g., in many dicy-
nodonts) to accommodate a longer digestive tract. These features are
reflected in overall body proportions and frequently in the structure of
the vertebral column as well.
The underlying assumption of this review is that high-fiber plant
parts formed the primary component of the diet for an extinct tetrapod if
the dental and skeletal features of this animal together are structurally
suitable for processing of fiber-rich foodstuffs. Once herbivory has been
established with some degree of confidence in an individual taxon or a
group of closely related taxa, we can examine the anatomical correlates in
an explicitly phylogenetic context to trace their historical development.
Although the reasoning behind this comparative approach is admittedly
somewhat circular, it permits relating the development of certain fea-
tures of the feeding apparatus to the adoption of a herbivorous diet.
Acquisition of cellulytic endosymbionts by herbivorous

tetrapods
The aforementioned anatomical adaptations for feeding on high-fiber

plant material would be relatively ineffectual in the absence of endosym-
biotic micro-organisms for cellulysis. The actual mode of acquisition of
these endosymbionts by their vertebrate hosts is unknown, and it is
unlikely that the fossil record will ever provide direct clues for resolving
this issue. Hotton et al. (1997) suggested that the micro-organisms may
have initially been picked up by detritivorous animals foraging in plant
litter. Those that could survive in the tetrapod gut could have gradually
assumed a role in the digestive processes of the host. We tested this
hypothesis by considering the phylogenetic relationships of each major
lineage of presumed late Paleozoic and Triassic herbivorous tetrapods. We
found that, in most cases, their sister-taxa were insectivorous, and thus we
hypothesize that most of these herbivorous forms were derived from
insect-consuming precursors. Gow (1978) argued the same point for the
various groups of plant-eating reptiles and non-mammalian synapsids
considered in his brief review. In some taxa, insectivory may have even
been retained in juvenile individuals, followed by a shift toward herbivory
in adults (Gow 1978; DeMar and Bolt 1981). Insects were feeding on plants
by mid-Carboniferous times, much earlier than tetrapods (Scott and
Taylor 1983; Labandeira 1997). We consider it equally plausible that
ingested insects, especially those feeding on plant material, provided the
original source for fermentative endosymbionts. The currently accepted
scheme of tetrapod interrelationships indicates that cellulytic endosymbi-
onts were independently acquired in each lineage (Hotton et al. 1997; Sues
and Reisz 1998) because the major groups of late Paleozoic plant-eating
tetrapods are only distantly related to each other.
High-fiber herbivory in tetrapods presumably developed only after a
(possibly extended) transitional period of omnivory that included con-
sumption of plant material because various changes were required for the
efficient processing and digestion of plant fodder. The morphological and
physiological requirements for omnivory and low-fiber herbivory would
not differ significantly from those for strict faunivory (Hotton et al. 1997).
The present-day herbivorous lizard Cnemidophorus murinus is an opportu-
nistic herbivore; indeed, in captivity, this reptile prefers animal foods
over any plant material offered (Dearing 1993). Comparative studies on
turtles (Bjorndal and Bolten 1993) indicate that herbivores may show
greater digestive efficiency than omnivores only on those plant diets that
are subject to extensive fermentation; the digestive performance of omni-
vores may equal, if not exceed, that of herbivores feeding on material that
does not require a significant amount of fermentative processing.
Among extant plant-eating iguanid lizards, the juveniles of each gen-
eration acquire the requisite microbes for endosymbiotic fermentation by
consuming the droppings of adult conspecifics (Troyer 1982). Modesto
(1992) argued that this intergenerational type of endosymbiont acquisi-
tion could only occur in nest-building amniote tetrapods. Although, at
first glance, this hypothesis is attractive because it draws on the most
obvious distinction between amniotes and non-amniotes, Modesto’s rea-
soning is not compelling because amphibian hatchlings and juveniles
could have easily picked up the endosymbiotic microbes by feeding at or
near sites where adults defecated. Most adult extant amphibians are fau-
nivorous, and only one, the Indian green frog (Rana hexadactyla), is a foli-
vore (Das 1996).The apparent inability of present-day amphibians to
acquire, or at least to maintain, cellulytic endosymbionts is of consider-
able interest, but remains yet to be explained. In this context, the Permo-
Carboniferous Diadectidae present an interesting conundrum. They pre-

sumably were herbivores capable of processing high-fiber plant fodder,
but they are not considered amniotes by many recent authors based on
osteological features (Reisz 1997). As the closest known relatives of amni-
otes, diadectids may have been reproductively amniotic, but, in the
absence of preserved eggs, their skeletal features place them outside the
crown-group Amniota. Perhaps diadectids were not amniotic in repro-
ductive terms, but, being closely related to amniotes, shared with the
latter certain anatomical and/or physiological attributes that allowed
them to feed on plants (see Hotton et al. 1997).
Diversity of late Paleozoic and Triassic herbivorous

tetrapods
The known fossil record of amniote tetrapods begins in the Mid-Pennsyl-

vanian (Westphalian B [Moscovian], Joggins, Nova Scotia/Canada; about
308 million years ago [Harland et al. 1990]). At that time, the two principal
lineages of Amniota, Reptilia (including birds) and Synapsida (including
mammals), had already differentiated (Reisz 1997). The first plant-eating
amniotes are known from the Late Pennsylvanian (Stephanian B; about
303 to 295 million years ago [Harland et al. 1990]). The proximate sister-
group of Amniota, Diadectomorpha, which includes presumably herbiv-
orous forms as well (Hotton et al. 1997; Sues and Reisz 1998), also first
appears in the fossil record during the Late Pennsylvanian. Herbivory is
clearly an ancient feeding strategy among amniote tetrapods.
Late Paleozoic and early Mesozoic tetrapods are separated from their
present-day relatives by a long evolutionary history, and it is difficult to
compare the craniodental features of most of these animals directly with
those in their extant relatives. Several fairly robust hypotheses for the
interrelationships of the principal groups of late Paleozoic tetrapods have
been published in recent years (Laurin and Reisz 1995; Lee 1997; Reisz
1997), and it is now possible to compare the craniodental features of pre-
sumed herbivores with the homologous ones in their faunivorous sister-
taxa within a phylogenetic framework. This approach permits the
identification of potential anatomical features suitable for feeding on
plants. For example, the shapes of the tooth crowns in most late Paleozoic
tetrapods, including those of the sister-taxa of early herbivores, range
from acutely conical to blade-like and recurved with distinct cutting
edges. By analogy to most present-day vertebrates with such teeth, this
Figure 2.1. Skulls of (A) Limnoscelis and (B) Diadectes in lateral and palatal views. Scale
bars each equal 1 cm. (Limnoscelis modified from Williston 1911; Diadectes
modified from Case 1911 and Olson 1947.)
type of tooth crown was presumably used for immobilizing animal prey
and for tearing and slicing through flesh. Only a few Permo-Carbonifer-
ous tetrapods had teeth that deviated from this common pattern. The ear-
liest forms with dentitions suitable for processing high-fiber plant fodder
date back to the Late Carboniferous, and these animals also have cranial
and postcranial features that are consistent with herbivory. Although the
fossil record demostrates that feeding on high-fiber plant material had
evolved by Late Carboniferous times, it is likely that consumption of
other plant tissues already occurred at an earlier date (Hotton et al. 1997;
Sues and Reisz 1998).
Late Paleozoic taxa
Diadectomorpha: Diadectidae
Desmatodon, from the Late Pennsylvanian of North America, and a closely
related but more derived form, Diadectes (Figure 2.1B), from the Late Penn-
sylvanian and Early Permian of North America and Europe, were appar-
ently the earliest herbivores capable of feeding on high-fiber plant
material (Hotton et al. 1997). Although not as diverse as the older taxo-
nomic literature (e.g., Case 1911) would imply, the fossil record indicates
that diadectids formed a common element of the terrestrial tetrapod

assemblages of North America and Europe during the Late Pennsylvanian
and Early Permian. Most authors now consider Diadectomorpha the
proximate sister-taxon of Amniota (Laurin and Reisz 1995; Reisz 1997).
However, Lee and Spencer (1997) have recently revived an earlier interpre-
tation of these tetrapods as basal amniotes. We do not find the arguments
advanced by these authors persuasive, mainly because their proposal
centers on issues of taxonomic stability, rather than phylogenetic rela-
tionships. In any case, the clade Diadectomorpha would retain its posi-
tion outside the crown-group of Amniota.
The spatulate, incisiform anterior teeth of Diadectes (Figure 2.1B) are
distinctly procumbent. The lower incisiforms occlude with the upper
ones; this arrangement is suitable for seizing and cropping food items
(Case 1911), which is consistent with the vertical striations produced by
wear on the incisiform teeth (Hotton et al. 1997). The more posterior
dentary and the maxillary teeth of Diadectes are molariform, with trans-
versely expanded but anteroposteriorly short crowns suitable for crush-
ing. Longitudinally oriented striations on the typically heavily worn
molariform teeth indicate fore-and-aft motion of the mandible (Hotton et
al. 1997). Berman et al. (1998) have found evidence that the lower molari-
form teeth may have occluded against a pad of perhaps keratinized tissue
supported by the palatal shelf of the palatine and the upper molariform
teeth against a similar pad on the lingual surface of the parapet of the
dentary. Diadectes has a large, barrel-shaped trunk that could have accom-
modated a voluminous digestive tract.
Comparisons with the sister-taxa of Diadectidae, Limnoscelis (Figure
2.1A) and Tseajaia (Heaton 1980) indicate that these diadectomorphs show
the plesiomorphic character-states for the aforementioned features. In
addition, there exists a suite of other skeletal features that appear to be
related to herbivory. The temporal region of the skull of Diadectes is
deeper than in other diadectomorphs and basal amniotes, the lower jaw is
more massive, and an incipient secondary bony palate is developed. The
postcranial skeleton of Diadectes is considerably more robust than in other
diadectomorphs and basal amniotes, especially in the construction of the
trunk region. In particular, the dorsal neural arches of Diadectes are tall
and massive, with massive neural spines and additional intervertebral
articular facets. The large transverse processes supported the strongly
curved ribs forming a barrel-shaped trunk (Case 1911). The trunk is pro-
portionately shorter than in other Diadectomorpha; it comprises fewer
Figure 2.2. Skulls of (A) Ianthasaurus and (B) Edaphosaurus (with attached mandible) in
lateral views and skull and right mandibular ramus of Edaphosaurus in palatal
view. The palate of Ianthasaurus is insufficiently known for a reconstruction.
Scale bars each equal 1 cm. (Ianthasaurus modified from Modesto and Reisz
1990; Edaphosaurus redrawn from Modesto 1995.)
vertebrae, and these vertebrae are relatively short. The available skeletal
evidence indicates that diadectids were herbivores, and this feeding strat-
egy evolved within the Diadectomorpha. Berman et al. (1998) noted that
the teeth of juvenile diadectids were much less molariform and had much
more limited occlusion than those of the adults, possibly indicating an
ontogenetic shift from omnivory to herbivory (or from low-fiber to high-
fiber herbivory).
Synapsida: Edaphosauridae
Another early herbivore is the Permo-Carboniferous synapsid Edaphosau-
rus (Figures 2.2B, 2.3B) from North America (Romer and Price 1940; Reisz
1986). It is best known for its spectacular dorsal ‘sail’ supported by the
greatly elongated dorsal neural spines, which are studded with promi-
nent lateral tubercles. Unlike the condition in related, presumably fau-
nivorous, forms its skull is small relative to overall body size. The crowns
of the anterior marginal teeth of Edaphosaurus have distinct cutting edges,
which are set obliquely relative to the long axis of the tooth row, and are
suitable for cutting off pieces of plants (Modesto 1995). The cutting edges
Figure 2.3. Skeletons and body silhouettes (in black) of (A) Cotylorhynchus and
(B) Edaphosaurus. The distal portion of the tail was omitted in both restorations
for layout purposes. (Original drawings by D. M. Scott.)
of unworn teeth bear fine, oblique serrations, which subsequently

became obliterated by wear. The palate (Figure 2.2B) and the lingual sur-
faces of the mandibular rami both support massive tooth plates, each
bearing several rows of numerous teeth. Wear on these teeth indicates
that food was crushed between the opposing upper and lower tooth
plates by occlusion and by fore-and-aft motion of the mandible (Modesto
1995; Hotton et al. 1997). The latter is indicated by the considerable size
disparity between the articular surfaces of the quadrate and the articular
bone; the latter is about 50% longer anteroposteriorly than the former
(Romer and Price 1940). The large, barrel-shaped trunk of Edaphosaurus is
consistent with the presence of a voluminous digestive tract suitable for
fiber fermentation.
Modesto (1994) developed a phylogenetically constrained scenario for
the evolution of feeding habits in Edaphosauridae. The most basal taxon
known, the Late Pennsylvanian Ianthasaurus (Figure 2.2A), has a dentition
comprising slightly recurved, labiolingually compressed teeth that is
consistent with faunivorous, probably insectivorous habits (Modesto and
Reisz 1990). The more derived Early Permian Glaucosaurus retains labio-
lingually compressed, sharp teeth, but its snout is much shorter, its denti-
tion is isodont, and an antorbital buttress is developed. Modesto (1994)
argued that these features indicate a more omnivorous diet, perhaps
including harder food items. Unfortunately, Glaucosaurus is currently
known only from a single incomplete skull, and thus comparisons have to
be restricted to Ianthasaurus and other basal synapsids.
Comparisons between Edaphosaurus and Ianthasaurus, as well as with
other eupelycosaurs, indicate that the following cranial features may also
be related to the adoption of herbivory in the former: reduced ratio of
skull length to trunk length, foreshortened snout, greatly enlarged
adductor chamber as a result of changes in the shape of the squamosal and
a shift of the jaw joint to a more ventral position, and massive, deep lower
jaw. In the postcranial skeleton, the cervical vertebrae are markedly
smaller than the trunk vertebrae, possibly in relation to the relatively
small head of these forms, and the trunk region is shorter (with three
fewer vertebrae) but has a wider rib cage with more strongly curved ribs
than in related faunivorous forms. The fossil record provides compelling
evidence that herbivory evolved within the Edaphosauridae.
Synapsida: Caseidae
The Caseidae, from the Permian of North America and Europe (Olson
1968; Reisz 1986), are generally interpreted as herbivores. Traditionally
considered close relatives of Edaphosauridae (Romer and Price 1940),
more recent phylogenetic analyses (e.g., Reisz 1986) have reinterpreted
them as part of the basal clade of Synapsida, Caseasauria. Unlike the con-
dition in faunivorous basal synapsids and outgroup taxa such as Limnosce-
lis, all caseids have a very small head relative to overall body size (Figure
2.3A). The unusually broad and barrel-shaped trunk indicates the devel-
opment of an enormous digestive tract. The anterior teeth are the largest
ones, and may have been used to crop plant material (Figure 2.4B). Their
crowns are anteroposteriorly compressed, spatulate, and frequently lack
apical cusps. The more posterior teeth have labiolingually compressed
crowns that become spatulate toward the apices and bear prominent
apical cusps or denticles (Olson 1968). They closely resemble the teeth of
extant plant-eating iguanid lizards. The number of apical cusps varies
between individual taxa, ranging from three to eight (Reisz 1986).
Although the upper and lower teeth did not meet in occlusion, they could
puncture and shred plant matter. In marked contrast to the condition in
iguanid lizards, the palatal dentition is very well developed in Caseidae,
Figure 2.4. Skulls of (A) Eothyris and (B) Cotylorhynchus in lateral and palatal views.
The palate of Eothyris is only partially known, preventing a detailed
reconstruction. Scale bars each equal 1 cm. (Original drawings by D. M. Scott.)
ranging from the primitive pattern seen in Cotylorhynchus (with trans-

verse, diagonal, and longitudinal rows of palatal teeth) to the highly
derived condition seen in Casea (with a massive pavement of teeth cover-
ing most of the palate). Furthermore, the palatal dentition extends onto
the ventral surface of the parasphenoid and cultriform process in all
caseids. The large, well-ossified hyoid apparatus indicates the develop-
ment of a substantial tongue for manipulating food within the oral cavity
(Olson 1968). The tongue could have positioned and moved food against
the palatal dentition. The jaw joint is situated below the level of the lower
tooth row. The forelimbs are robust and unusually large, and the hands
terminate in large claws suitable for digging (Stovall, Price and Romer
1966). Cotylorhynchus romeri, from the Early Permian of Oklahoma, reached
a length of about 3.5 m and an estimated live weight of up to 331 kg
(Stovall et al. 1966), but C. hancocki, from slightly younger Permian strata in
Texas, was up to 25% larger in comparable linear dimensions (Reisz 1986).
The Eothyrididae, the sister-taxon of Caseidae among Caseasauria
(Reisz 1986), are characterized by a distinctly heterodont dentition com-
posed of simple conical marginal teeth (Figure 2.4A), which indicate fau-
nivorous habits. Outgroup comparisons are not possible for postcranial
features because the two known eothyridid genera (Eothyris and Oedaleops)
are represented only by cranial remains. Herbivory evolved within Case-
asauria, and even the most basal caseids already fed on plants.
Synapsida: Therapsida: Dinocephalia

Tapinocephalid dinocephalians are known only from the Late Permian of
South Africa and Russia (Kemp 1982). Forms such as Moschops from South
Africa and the closely related Ulemosaurus from Russia are large, massively
built animals, some of which reached a length of 3 m. The dentition is
rather uniform and lacks distinct canines. Each tooth crown has a blunt
apex and broad heel (Efremov 1940; Boonstra 1962), and the upper and
lower teeth intermeshed during occlusion to form a cropping mecha-
nism. The typically low and broad snout is sharply demarcated from the
remainder of the skull. The jaw joint is located further forward than in
other dinocephalians, and the snout and mandible are proportionately
shorter. The deep posterior portion of the mandibular ramus created a
relatively long moment arm for the adductor jaw muscles, which,
coupled with the short jaws, probably increased available bite force
(Kemp 1982). The rib cage is unusually large, and the transverse processes
of the dorsal vertebrae are long. The trunk probably housed a voluminous
digestive tract (Gregory 1926).
The Tapinocephalidae represent a derived clade of presumed herbi-
vores within Dinocephalia, which are typically and plesiomorphically
carnivorous.
Synapsida: Therapsida: Anomodontia

Anomodont therapsids (Figure 2.5), especially the Dicynodontia, were
the most successful group of Permian continental tetrapods in terms of
both taxonomic diversity and overall abundance (King 1990, 1996). Dicy-
nodonts ranged from the Late Permian to the Late Triassic and had a
worldwide distribution. They are most common, both in terms of
number of taxa and especially number of individuals, in the Late Permian
continental strata of the Beaufort Group in South Africa (Kitching 1977;
King 1990, 1996). The largest forms have skull lengths exceeding 50 cm.
Some Middle and Late Triassic dicynodonts may have exceeded 3 m in
length and weighed one metric ton or more. The entire skeleton is
robustly built. The skull (Figure 2.5C) is highly derived compared with
that in other therapsids (Crompton and Hotton 1967; Cluver 1971; King
1990, 1996). Most derived dicynodonts have enlarged upper caniniform
teeth but lack postcanine teeth. The anterior ends of the snout and man-
dible were probably covered by an extensive keratinous beak (rham-
photheca) in life, and the lower beak bit inside the upper when the mouth
was closed. Judging from the underlying bony ridges, it seems that the
Figure 2.5. Skulls of three representatives of Anomodontia ((A) Patranomodon,

(B) Suminia, and (C) Dicynodon) in lateral (with attached mandible) and palatal
views, arranged on a cladogram depicting a hypothesis of the
interrelationships of these taxa. Scale bars each equal 1 cm. (Patranomodon
redrawn from Rubidge and Hopson 1996, courtesy and copyright of the
Linnean Society of London; Suminia courtesy and copyright of N. Rybczynski;
Dicynodon redrawn from Cluver and King 1983, courtesy and copyright of the
South African Museum.)
upper and lower beaks formed sharp cutting edges. A few dicynodonts
possessed both teeth and beaks, in contrast to the condition in turtles and
birds. Pristerodon has dentary teeth with leaf-shaped crowns that may have
occluded against the (possibly horn-covered) palatine bones. The man-
dibular symphysis is completely fused in all dicynodonts. The very large
temporal fenestra extends posteriorly beyond the level of the occipital
condyle. The shape of the squamosal is modified so that the external
adductor jaw muscle originates from its anterolateral surface, and the
lateral and medial components of the adductor jaw musculature had a
nearly horizontal orientation. The quadrate and the articular bone both
form (in side view) convex articular surfaces; the latter is almost twice as
long as the former and is placed far anteroventrally. This unique configu-
ration of the jaw joint permitted considerable mandibular retraction
(Crompton and Hotton 1967; Cluver 1971; King, Oelofsen, and Rubidge
1989). A retractive power stroke would have facilitated cutting and slicing
of fibrous plant material between the sharp edges of the opposing halves
of the keratinous beak. Most dicynodonts lack additional structures for
chewing (such as postcanine teeth), and thus additional oral processing of
the fodder by these animals is unlikely.
In many dicynodonts, the trunk region is distinctly elongated, but it is
barrel-shaped in the large Mid- and Late Triassic forms; both patterns are
consistent with the presence of an extensive digestive tract. We disagree
with the general assumption (e.g., King 1990) that all dicynodont therap-
sids were herbivorous. Although they have been the subject of numerous
studies, dicynodonts are still rather poorly known, and the inferred pres-
ence of keratinous beaks and other cranial features are insufficient for
such generalized assumptions. For example, most extant turtles are car-
nivorous although they, like dicynodonts, have a beak instead of teeth.
We consider it likely that at least some dicynodonts were omnivorous or
perhaps even carnivorous. Certain dicynodont taxa, especially the Late
Permian Cistecephalus (Cluver 1978) and its relatives, show features of the
limbs and girdles that indicate distinctly fossorial habits. Other less spe-
cialized taxa may have also had at least partially subterranean habits.
Skeletons of the Late Permian Diictodon have been found in the terminal
chambers of complex helical burrows (Smith 1987).
More basal anomodonts differ from dicynodonts especially in the pres-
ence of well-developed marginal teeth. Patranomodon (Figure 2.5A) from
the lowest of the Permian-age strata of the Beaufort Group of South Africa
is considered the most basal anomodont (Rubidge and Hopson 1996). Its
dentition is still poorly known, but it apparently lacks any specializations

for herbivory. In the absence of such dental features, it is difficult to evalu-
ate the feeding habits of this small anomodont, but some of its cranial
features, such as the short snout, are generally found in herbivorous tetra-
pods. Patranomodon lacks the typically dicynodont jaw articulation and
has a screw-shaped articular surface between quadrate and articular,
which precludes fore-and-aft motion of the mandible. In contrast,
Suminia (Figure 2.5B) from the Late Permian of Russia has teeth with leaf-
shaped, cuspidate crowns (Rybczynski 1996). Striations on the well-devel-
oped wear facets demonstrate that jaw motion comprised mandibular
elevation and retraction. Many cranial features typical of later dicyno-
donts are already present in this small anomodont. For example, the
squamosal flares out posterolaterally, providing an area for the origin of
the external adductor jaw muscle (Crompton and Hotton 1967). As the
sister-taxon of Dicynodontia, Suminia provides evidence that many
cranial characters found in later dicynodonts are not necessarily related to
the origin of herbivory in that group. Most significant among these fea-
tures is the presence of a keratinous beak, the much enlarged, dorsolater-
ally oriented temporal fenestra, and the partial development of a
secondary bony palate. In dicynodonts, these features may represent
further refinements for feeding on plants. The fossil record provides com-
pelling evidence that herbivory evolved within the clade Anomodontia.
Reptilia: Parareptilia: Pareiasauria

The Pareiasauria included some of the largest Paleozoic herbivorous
tetrapods. These bulky reptiles range in length from about 1 m to 3.5 m
and have large, massive skulls (Bystrov 1957; Lee 1997). They appear to
have been restricted to the Late Permian and had an almost worldwide
distribution. The snout is transversely broad, and the postorbital region
is much expanded transversely (Figure 2.6). Each jaw bears a single mar-
ginal row of closely spaced, non-occluding teeth with crowns that super-
ficially resemble those of caseid synapsids in being tall, strongly flattened
labiolingually, and blade-like towards the apex of the crown. The apices
of the tooth crowns each bear six or more cusps or denticles and are lined
up along the entire length of the jaw to form a continuous shearing edge.
The number of cusps increased significantly and the tooth crowns became
fan-shaped in more derived pareiasaurs (Lee 1997). The large, broad trunk
of pareiasaurs is consistent with the presence of a voluminous digestive
tract. Although the phylogenetic relationships of Pareiasauria are still
Figure 2.6. Skull of a pareiasaur, Scutosaurus, in lateral and palatal views. Scale bar
equals 1 cm. (Redrawn from Bystrov 1957.)
somewhat controversial, the Procolophonoidea are the most likely sister-

group of this taxon among Parareptilia (Laurin and Reisz 1995), with
small faunivorous forms such as Owenetta (see Figure 2.10) presumably
representing the plesiomorphic condition for that group. Using this
hypothesis of relationships for outgroup comparisons, the following
cranial features of pareiasaurs may be related to herbivory: short tooth
rows and snout, posterodorsally expanded skull table, and reduced
palatal dentition. In the postcranial skeleton, the vertebral column is
reduced to 19 or 20 presacral vertebrae (Lee 1997) and stiffened, with each
Figure 2.7. Skull of a bolosaurid, Belebey, in lateral and palatal views. Scale bar equals
1 cm. (Modified from Ivakhnenko and Tverdokhlebova 1987 and specimens.)
dorsal vertebra being short but tall and massively built. The ribs are large
and strongly curved.
Reptilia: Bolosauridae
The Bolosauridae form a clade of still poorly known basal reptiles from
the Early Permian of North America (Bolosaurus; Watson 1954, Hotton et al.
1997) and the Late Permian of Russia and China (Belebey; Ivakhnenko and
Tverdokhlebova 1987, Li and Cheng 1995). The premaxillary and anterior
dentary teeth are somewhat procumbent (Figure 2.7). The crowns of the
maxillary teeth of Bolosaurus are bulbous labially, with a single robust, lin-
gually curved cusp. A prominent cingulum extends lingual to the base of
the cusp and delimits a shallow basin between it and the base of the labial
cusp. Similarly, the dentary teeth bear a prominent lingual cusp and a
labial cingulum. The dentary and maxillary teeth are thickly enamelled.
Striations on the distinct wear facets indicate fore-and-aft motion of the
mandible, which is consistent with a greater anteroposterior length of the

articular facet of the jaw joint (Watson 1954; Hotton et al. 1997). A narrow
infratemporal fenestra is present, and a tall coronoid process of the lower
jaw extends dorsally into the adductor chamber, directly medial to the
fenestra (Figure 2.7). In contrast to faunivorous Early Permian reptiles,
Bolosaurus and Belebey show incipient development of a secondary bony
palate. One articulated skeleton of Belebey (Paleontological Museum of
the Russian Academy of Sciences, Moscow, PIN 104B/2020) has a barrel-
shaped trunk and a high ratio of trunk to skull length, both features com-
monly found in herbivores.
Reptilia: Captorhinidae
The Captorhinidae (Figure 2.8) are Permian reptiles with an apparently
worldwide distribution. Recent phylogenetic studies (e.g., Laurin and
Reisz 1995) have placed them close to the Diapsida. Basal captorhinids are
characterized by small size and, with the exception of the some representa-
tives of the Early Permian Captorhinus (Figure 2.8A) itself, have single rows
of marginal teeth in the dentary and maxilla (Heaton 1979). These forms
have generally been interpreted as faunivorous. One reported instance of
apparent predation in Captorhinus aguti involves a smaller conspecific
(Eaton 1964). Some of the stratigraphically younger, more derived, and
larger captorhinids such as Labidosaurikos (Figure 2.8C) have certain cranio-
dental features that are consistent with feeding on plants (Dodick and
Modesto 1995). Labidosaurikos and the Late Permian Moradisaurus differ from
other captorhinid reptiles in being much larger (with the skull length of
Moradisaurus exceeding 40 cm), and having broad dentaries and maxillae
that bear multiple (6 to 11) rows of teeth. The small, rather isodont teeth are
set in parallel longitudinal rows. The tooth crowns show distinct labial and
lingual wear facets. Both the pattern of tooth wear (Dodick and Modesto
1995; Hotton et al. 1997) and the structure of the jaw joint (Ricqlès and
Taquet 1982) indicate fore-and-aft motion of the mandible, and the lower
rows of teeth fit neatly between the upper rows for effective crushing and
shredding of plant material. The postcranial skeleton of Labidosaurikos is
still unknown, and that of Moradisaurus has yet to be described. Therefore,
we restrict our comparisons to cranial features. The sister-taxon of the pre-
sumably herbivorous Moradisaurinae is the single-tooth-rowed Labidosau-
rus (Figure 2.8B). It is difficult to determine the feeding habits of the latter,
but there exists no clear evidence to suggest that it was a herbivore. Com-
parisons with this form and other more basal captorhinid taxa indicate
that the Moradisaurinae have additional skeletal features that may be
Figure 2.8. Skulls of three representatives of Captorhinidae (A, Captorhinus;

B, Labidosaurus; and C, Labidosaurikos) in lateral and palatal views, arranged
on a cladogram depicting a hypothesis of the interrelationships of these taxa.
Captorhinus includes both forms with single (C. laticeps, depicted here) and
multiple marginal tooth rows (C. aguti). The hypothesis of relationships
depicted here postulates that multiple tooth rows evolved twice within this
clade (Dodick and Modesto 1995). Scale bars each equal 1 cm. (Captorhinus
modified from Heaton 1979 and specimens; Labidosaurus original drawing
by D. M. Scott; Labidosaurikos modified from Dodick and Modesto 1995 and
specimen.)
related to feeding on plants. These include some degree of elongation of

the snout (possibly to extend the reach of the premaxillary teeth that were
used either for digging or cropping), the loss of the palatal teeth, and the
posterior extension of the dermal skull roof. The latter feature represents a
way of enlarging the adductor chamber relative to the condition in Labido-
saurus and more basal captorhinid reptiles. Another significant feature is
the structure of the jaw joint. Both Labidosaurus and Captorhinus show evi-
dence of an open jaw articulation that allowed some fore-and-aft motion of
the mandible, but this feature by itself is insufficient evidence for herbi-
vory. Mandibular fore-and-aft motion in Labidosaurikos and Moradisaurus
must have been much more extensive than in Captorhinus, possibly in direct
relation to more extensive oral processing of plant material. The fossil
record provides clear evidence that herbivory evolved within the clade Cap-
torhinidae.
Triassic taxa
This section briefly reviews the principal groups of presumed herbivores
among non-dinosaurian tetrapods from the Triassic period. During the
Late Triassic, dinosaurs became the dominant herbivores in terrestrial
vertebrate communities (Weishampel and Norman 1989; Wing and Sues
1992). (For discussions of herbivory in the various major clades of Dino-
sauria, the reader is referred to the chapters by Barrett, Upchurch and
Barrett, and Weishampel and Jianu.) The dicynodonts, discussed in the
preceding section, ranged into the Late Triassic, but their taxonomic
diversity was much reduced relative to that during the Late Permian.
With the exception of Pareiasauria, which persisted to the end of the
Permian, most other groups of late Paleozoic herbivorous tetrapods had
already disappeared from the fossil record earlier during that period.
Certain other taxa of Triassic reptiles (e.g., Stagonolepididae) and non-
mammalian synapsids (e.g., Bauriidae) have often been interpreted as
possible herbivores, but these forms have no unambiguous anatomical
features for feeding on high-fiber plant material.
Synapsida: Cynodontia: Gomphodontia

The Gomphodontia formed a diverse clade of mostly Triassic cynodonts
(Figure 2.9B). They are readily characterized by the possession of trans-
versely expanded, molariform postcanine teeth that meet in complex
tooth-to-tooth occlusion (Crompton 1972). Some derived forms, such as
the traversodont Exaeretodon from the Late Triassic of Argentina and
India, may have reached a length of up to 2 m. The skull is large and
Figure 2.9. Skulls of two representatives of Cynodontia (A, Thrinaxodon and

B, Massetognathus) in lateral and palatal views. Scale bars each equal 1 cm.
(Thrinaxodon modified from Parrington 1946 and specimens; Massetognathus
modified from Romer 1967 and specimens.)
massive, with deep zygomatic arches and often pronounced ectocranial

crests. The rostral portions of the dentaries are fused to form an extensive
mandibular symphysis. In derived traversodont taxa, the tightly packed
crowns of the molariform upper and lower postcanine teeth each bear a
transverse crest. Each molariform forms multiple shearing surfaces.
Occlusion was bilateral and involved a power stroke during which the
shearing surfaces of the lower postcanine teeth were dragged posterodor-
sally across the matching surfaces on the upper teeth (Crompton 1972).
The postcanine teeth are distinctly inset from the sides of the face, pos-
sibly indicating the development of a lateral oral vestibule bounded by
muscular cheeks (Hopson 1984). The temporal fossae are much enlarged
and delimited by prominent ectocranial crests in derived gomphodont
cynodonts, and the dentary forms a tall, often recurved coronoid process,
indicating powerful development of the temporalis musculature.
The dentitions of the Early Triassic basal gomphodonts Diademodon
and Trirachodon are distinctly heterodont, comprising conical anterior
teeth, transversely expanded molariforms, and sectorial posterior teeth
(Hopson 1971; Crompton 1972; Grine 1977). The multicuspid sectorial
teeth are very similar to the postcanine teeth of faunivorous eucynodonts
such as Cynognathus and Thrinaxodon (Figure 2.9A). The pattern of dental
microwear indicates that Diademodon retained orthal jaw motion (Grine
1977). Based on the heterodont nature of the postcanine dentition, basal
gomphodont cynodonts probably were omnivorous, and feeding on
Figure 2.10. Skulls of three representatives of Procolophonoidea (A, Owenetta;

B, Procolophon; and C, Hypsognathus) in lateral and palatal views, arranged on a
cladogram depicting a hypothesis of the interrelationships of these taxa. Scale
bars each equal 1 cm. (Owenetta original drawing by D. M. Scott; Procolophon
modified from Carroll and Lindsay 1985 and specimens; Hypsognathus modified
from Sues et al. 2000).
high-fiber plant material evolved within the clade Gomphodontia. This

heterodont type of dentition was even retained in juvenile specimens of
some traversodont taxa (H.-D. Sues, unpublished data), and may reflect
an ontogenetic shift from more omnivorous habits in juveniles to a pre-
dominantly herbivorous diet in adults.
Reptilia: Parareptilia: Procolophonidae

The Procolophonidae are small to medium-sized, superficially lizard-like
parareptiles (Figure 2.10). They had a worldwide geographic distribution
and spanned the entire Triassic period. The skull is particularly character-
ized by a distinct posterior embayment of the orbit, which probably
accommodated substantial adductor jaw muscles. This embayment is
most pronounced in the highly derived Late Triassic Hypsognathus (Figure
2.10C) and Leptopleuron (Leptopleuroninae), in which the opening is
expanded posteriorly as well as laterally. The lower jaw in these forms
also has a tall, occasionally recurved coronoid process for the insertion of
the adductor jaw muscles, and the jugal is much deeper. The jaw joint in
Hypsognathus and other derived leptopleuronines is situated well below
the level of the lower tooth row. The dentition of procolophonids is dif-
ferentiated into incisiform anterior teeth and transversely widened,
bicuspid or multicuspid posterior teeth. On unworn posterior teeth,
pointed labial and lingual cusps are linked by a sharp transverse crest,
which became progressively obliterated by wear to form a more or less flat
apical wear facet (Gow 1978; Carroll and Lindsay 1985; Sues and Baird,
1998). In most taxa, the maxillary and posterior dentary teeth interdigi-
tated in a cog-like fashion. A noteworthy postcranial feature in some pro-
colophonid taxa, such as Hypsognathus, is the transversely broad rib cage.
The Owenettidae (Barasaurus and Owenetta; Figure 2.10A) are generally
considered the sister-group of Procolophonidae (Gow 1978; Laurin and
Reisz 1995). They have conical, slightly recurved marginal teeth of fairly
uniform size, with the lower teeth biting inside the upper ones. This type
of dentition is consistent with probably insectivorous habits (Gow 1978).
In addition, the vertebrae are not as robustly constructed as those of
similar-sized procolophonids, and the trunk is much more slender than
in Hypsognathus and Procolophon. Gow (1978) argued that juveniles of Proco-
lophon (Figure 2.10B) and smaller procolophonid taxa were probably
insectivorous, and only the adults of the larger forms with transversely
expanded teeth were more or less exclusively herbivorous. The derived
leptopleuronines (e.g., Hypsognathus; Figure 2.10C) show a number of fea-
tures consistent with feeding on high-fiber plant material, such as teeth
suited for shearing, a ventrally off-set jaw joint, and powerful develop-
ment of the adductor jaw muscles. Herbivory thus evolved within the
Procolophonidae.
Reptilia: Diapsida: Archosauromorpha: Rhynchosauria

The Rhynchosauria (Figure 2.11) are a derived group of Triassic archosau-
romorph reptiles with a nearly worldwide distribution. They were par-
ticularly abundant in the Middle and Late Triassic of Gondwana.
Rhynchosaurs are characterized by the distinctive structure of the skull.
Figure 2.11. Skulls of two representatives of Rhynchosauria (A, Mesosuchus and

B, Hyperodapedon) in lateral and palatal views. Scale bars each equal 1 cm.
(Mesosuchus courtesy of D. W. Dilkes; Hyperodapedon redrawn from
Benton 1983, courtesy and copyright of the Royal Society of London.)
In the most derived taxa such as Hyperodapedon, the transverse width of

the temporal region exceeds the anteroposterior length of the skull
(Benton 1983; Figure 2.11B). The temporal fossae are greatly enlarged, pro-
viding space for powerful adductor jaw muscles. The premaxilla is beak-
like and devoid of teeth, and the rostral tip of the dentary also lacks teeth.
The maxillae bear large tooth plates with multiple longitudinal rows of
short conical teeth on either side of a median groove. The teeth are deeply
rooted and firmly attached with bone (ankylothecodont implantation;
Benton 1984). The ridge formed by the dentary teeth fits into the median
groove on the maxillary tooth plate. Jaw motion was orthal and primarily
involved shearing, with the food being cut as the lower jaws were brought
diagonally into occlusion (Benton 1983, 1984). Rhynchosaurs are generally
considered herbivorous; earlier suggestions of molluscivorous habits are
inconsistent with the structure of the jaw apparatus and the pattern of
tooth wear (Benton 1983). The more derived Rhynchosauridae are charac-
terized by the possession of a broad, barrel-shaped trunk, which indicates
the presence of a voluminous digestive tract (Benton 1983). Some taxa

such as Hyperodapedon have unusually deep, mediolaterally flattened
unguals, which may have been used for scratch-digging to recover under-
ground vegetal structures such as rhizomes and tubers (Benton 1983).
The basal rhynchosaur Howesia from the Early or early Middle Triassic
of South Africa has multiple rows of ankylothecodont teeth in the maxilla
and dentary, but the upper and lower teeth met along a broad occlusal
surface, rather than in the shearing fashion characteristic of more derived
forms (Benton 1984; Dilkes 1995). The even more primitive Mesosuchus
(Figure 2.11A) has a zigzag arrangement of marginal teeth in the dentary
and maxilla, rather than a single continuous marginal tooth row in each
jaw (Dilkes 1998).
The sister-taxa of Rhynchosauria are faunivorous with the exception
of Trilophosaurus, which exhibits rather different craniodental specializa-
tions (see below). The distinctive jaw configuration for shearing and thus
presumably herbivory evolved within Rhynchosauria because Howesia
and Mesosuchus still lack this configuration.
Reptilia: Diapsida: Archosauromorpha: Trilophosauridae

Trilophosaurus from the Late Triassic of Texas (Gregory 1945; DeMar and
Bolt 1981) has posterior teeth with relatively high and transversely much
expanded crowns. The premaxilla and anterior end of the dentary are
devoid of teeth and may have been covered by a keratinous beak in life.
With the exception of a few anterior single-cusped teeth, the dentary and
maxilla bear teeth with transversely expanded crowns, each with three
subequal cusps linked by a transverse crest on unworn teeth. The trans-
versely expanded teeth typically show extensive wear. Jaw motion was
orthal and crushing, with the upper and lower teeth interdigitating in a
cog-like fashion (Gregory 1945). The adductor jaw musculature was sub-
stantial, as indicated by the deep temporal region and extensive temporal
fossa, which is laterally enclosed by a deep, unfenestrated temporal bar.
Trilophosaurus is generally referred to Archosauromorpha (see Benton
1983), but its precise phylogenetic position within that clade is still unre-
solved. Most other archosauromorph reptiles, and Diapsida plesiomor-
phically, are faunivorous.
Summary
The preceding overviews of presumed plant-eating tetrapods from the
late Paleozoic and Triassic (with the exception of dinosaurs) establish two
basic dental patterns for herbivory. The first, found in most of the taxa
discussed above, is characterized by dentitions where occlusion between

the upper and lower teeth breaks down food material. This is the most
common pattern, appearing first in the fossil record, and shows the great-
est variation in the development of structures used for oral processing.
Rhynchosaurid reptiles show a particularly intricate mechanism for
puncture crushing and shearing. Gomphodont cynodonts developed
molariform teeth with complex cusp patterns and precise tooth-to-tooth
occlusion.
The second pattern, which first developed in caseid and basal anomo-
dont synapsids as well as in pareiasaurs, is characterized by the presence
of leaf-shaped, cuspidate marginal teeth suitable for puncturing and
tearing apart fibrous plant material, in a manner similar to that seen in
extant iguanid lizards. It probably implies less extensive oral processing
of food than the first pattern of dentition. (The beaks of dicynodont the-
rapsids may have served a similar function.) In all these forms, the ante-
rior marginal teeth presumably served to seize and cut off parts of plants
for oral processing and ingestion. Among Triassic tetrapods, leaf-shaped,
cuspidate tooth crowns are known only in prosauropod and basal orni-
thischian dinosaurs.
Compared with their faunivorous relatives, both types of Paleozoic
and Triassic herbivorous tetrapods show various cranial features suitable
for enhanced jaw action, most importantly a relative increase in the size
and thus power of the adductor jaw muscles. This change is reflected by
either enlargement of the adductor chamber of the skull in forms with
closed temporal regions (Captorhinidae, Pareiasauria) or by an increase in
the size of the temporal opening in forms with a fenestrated temporal
region (such as Synapsida, which have a single temporal opening on
either side of the skull; Figure 2.3). In the Dicynodontia and their relatives
(Figure 2.4), the external adductor jaw musculature further increased in
size through expansion onto the ventrolateral surface of the squamosal.
Although they differ considerably in their craniodental structure, dicyno-
donts and gomphodont cynodonts both employed a palinal power stroke
suitable for shearing.
Most of the presumed herbivores discussed above are characterized by
medium to large body size and capacious trunks suitable for accommo-
dating extensive digestive tracts for endosymbiotic cellulysis. This is con-
sistent with the observation that large body size facilitates more efficient
processing of high-fiber plant fodder by cellulysis (Farlow 1987; Hotton et
al. 1997). The mass-specific metabolic rate of an animal decreases with
increasing body size, whereas the ratio of gut capacity to body size
remains more or less constant (Farlow 1987). A larger herbivore should
have a lower turnover rate of its gut contents than a smaller one. Indeed,
the rate of passage of food through the gut generally decreases with larger
body size (Parra 1978), so that the ingested fodder can be exposed to
microbial fermentation for longer periods of time and the yield from fer-
mentation is higher because more cellulose is processed per unit of time.
Two basic types of body form can be distinguished among presumed
early herbivores. In most Paleozoic and early Mesozoic herbivorous tetra-
pods, the rib cage is expanded laterally and even dorsally on either side of
the vertebral column, resulting in a broad, barrel-shaped trunk. An
apparent functional correlate of this change in many forms is the stiffen-
ing of the vertebral column by means of accessory intervertebral facets,
greatly expanded zygapophyses, massive neural spines, and large facets
for rib articulation. Another common feature in these forms is a reduc-
tion in the number of trunk vertebrae relative to their faunivorous sister-
taxa. The most dramatic example of this condition is observed in the
Pareiasauria. In the evolutionary history of this group, stiffening of the
vertebral column may also have been linked to the development of
dermal armor (Lee 1996), but this suggestion is not very convincing
because the largest, bulkiest forms are the first to appear in the fossil
record and show much less extensive development of armor than the
smaller, stratigraphically younger, and more derived taxa. A second type
of body form is represented by dicynodonts, many of which have a dis-
tinctly elongate trunk region. This represents an alternative condition to
accommodate an extensive digestive tract.
Conclusions
The currently known fossil record indicates that adaptations for herbi-
vory in terrestrial tetrapods appeared only some 60 million years after the
invasion of land by vertebrates. However, once this novel feeding strategy
had developed, it evolved independently in numerous lineages of terres-
trial reptiles and non-mammalian synapsids (Figure 2.12). During the late
Paleozoic, herbivory appears to have developed four times among non-
mammalian synapsids, at least once among eureptiles, and twice among
parareptiles. During the Triassic period, it evolved in at least one addi-
tional clade of synapsids and at least two groups of non-dinosaurian
archosauromorph reptiles.
Figure 2.12. Diagram illustrating the temporal distribution and phylogenetic

relationships of late Paleozoic and early Mesozoic amniotes (combined from
numerous sources), illustrating the repeated, independent development of
herbivory. Thick lines denote lineages of or including presumed herbivores;
broken lines denote unrecorded ranges inferred from phylogenetic
hypotheses.
Although some Late Carboniferous tetrapods already show cranioden-

tal features for feeding on high-fiber plant material, herbivorous forms
did not appear in significant numbers until the Late Permian, some 40
million years after the first appearance of herbivores in the fossil record of
terrestrial tetrapods. The reasons for this delay are not yet understood.
The fossil record indicates that the structure of ‘modern’ terrestrial eco-
systems apparently developed gradually, and their basic pattern of
trophic interactions was only established during Late Permian times
when herbivorous reptiles and synapsids first became an important com-
ponent of the terrestrial food chain.
All Permo-Carboniferous and, with the exception of dinosaurs, all Tri-
assic plant-eating tetrapods were quadrupedal animals that probably
foraged within one meter above the ground. It is possible that some of the
small forms may have also foraged in the trees, but none shows obvious
specializations for a scansorial or arboreal mode of life.
The evolution of large body size in various lineages of herbivorous
tetrapods appears to be linked to the acquisition of this mode of feeding.
This trend is particularly well documented in captorhinid reptiles and in
caseid and edaphosaurid synapsids. Late Permian tetrapod communities
include, for the first time, several unrelated groups of large-bodied herbi-
vores (Pareiasauria, Dinocephalia, Dicynodontia), some with a body mass
in excess of 1000 kg. Large body size apparently evolved repeatedly
among dicynodont therapsids (King 1990).
It would appear that the development of herbivory represented a
major adaptive threshold in the evolution of terrestrial tetrapod commu-
nities. Once amniotes developed the ability to acquire and maintain colo-
nies of cellulytic endosymbionts, this happened repeatedly in many
lineages of amniotes, leading to the apparently gradual emergence of the
trophic structure characteristic of modern terrestrial ecosystems.
Acknowledgments
Special thanks are due to D. M. Scott for preparing the figures. In many
instances, her work on these illustrations included substantial reinterpre-
tation of skull reconstructions from earlier publications through direct
examination of specimens. D. W. Dilkes, S. P. Modesto, and N. Rybczynski
kindly provided us with original drawings for use in this review. We are
indebted to R. L. Carroll, N. Hotton III, and especially R. Beerbower for
providing constructive comments on the manuscript. Both authors grate-

fully acknowledge continuing financial support for their research from
operating grants from the Natural Sciences and Engineering Research
Council of Canada (NSERC).
The sequence of authors’ names is strictly alphabetical, and no senior-
ity is implied.
References
Benton, M. J. (1983). The Triassic reptile Hyperodapedon from Elgin: functional

morphology and relationships. Phil. Trans. R. Soc. Lond. B 302:605–717.
Benton, M. J. (1984). Tooth form, growth, and function in Triassic rhynchosaurs.
Palaeontology 27:737–776.
Berman, D. S, Henrici, A. C., and Sumida, S. S. (1998). Taxonomic status of the Early
Permian Helodectes paridens Cope (Diadectidae) with discussion of occlusion of
diadectid marginal dentition. Ann. Carnegie Mus. 67:181–196.
Bjorndal, K. A. (1979). Cellulose digestion and volatile fatty acid production in the green
turtle, Chelonia mydas. Comp. Biochem. Physiol. 63A:127–133.
Bjorndal, K. A., and Bolten, A. B. (1993). Digestive efficiencies in herbivorous and
omnivorous freshwater turtles on plant diets: Do herbivores have a nutritional
advantage? Physiol. Zool. 66:384–395.
Bjorndal, K. A., Bolten, A. B., and Moore, J. E. (1990). Digestive fermentation in
herbivores: effect of food particle size. Physiol. Zool. 63:710–721.
Boonstra, L. D. (1962). The dentition of the titanosuchian dinocephalians. Ann. S. Afr.
Mus. 46:57–112.
Bystrov, A. P. (1957). [Skull of pareiasaur.] Trudy Paleont. Inst. Akad. Nauk SSSR 68:3–18 (in
Russian).
Carroll, R. L., and Lindsay, W. (1985). Cranial anatomy of the primitive reptile
Procolophon. Can. J. Earth Sci. 22:1571–1587.
Case, E. C. (1911). A revision of the Cotylosauria of North America. Carnegie Inst.
Washington Publ. 145:1–122.
Cluver, M. A. (1971). The cranial morphology of the dicynodont genus Lystrosaurus. Ann.
S. Afr. Mus. 56:155–274.
Cluver, M. A. (1978). The skeleton of the mammal-like reptile Cistecephalus with evidence
for a fossorial mode of life. Ann. S. Afr. Mus. 76:213–246.
Cluver, M. A., and King, G. M. (1983). A reassessment of the relationships of Permian
Dicynodontia (Reptilia, Therapsida) and a new classification of dicynodonts.
Ann. S. Afr. Mus. 91:195–273.
Crompton, A. W. (1972). Postcanine occlusion in cynodonts and tritylodontids. Bull. Brit.
Mus. (Nat. Hist.), Geol. 21:27–71.
Crompton, A. W., and Hotton, N. III. (1967). Functional morphology of the masticatory
apparatus of two dicynodonts (Reptilia: Therapsida). Postilla 109:1–51.
Das, I. (1996). Folivory and seasonal changes in diet in Rana hexadactyla (Anura: Ranidae).
J. Zool., Lond. 238:785–794.
Dearing, M. D. (1993). An alimentary specialization for herbivory in the tropical
whiptail lizard Cnemidophorus murinus. J. Herpet. 27:111–114.
DeMar, R., and Bolt, J. R. (1981). Dentitional organization and function in a Triassic
reptile. J. Paleont. 55:967–984.
Dilkes, D. W. (1995). The rhynchosaur Howesia browni from the Lower Triassic of South
Africa. Palaeontology 38:665–685.
Dilkes, D. W. (1998). The Early Triassic rhynchosaur Mesosuchus browni and the
interrelationships of basal archosauromorph reptiles. Phil. Trans. R. Soc. Lond. B
353:501–541.
Dodick, J. T., and Modesto, S. P. (1995). The cranial anatomy of the captorhinid reptile
Labidosaurikos meachami from the Lower Permian of Oklahoma. Palaeontology
38:687–711.
Eaton, T. H., Jr. (1964). A captorhinomorph predator and its prey (Cotylosauria). Amer.
Mus. Novit. 2169:1–3.
Efremov, J. A. (1940). Ulemosaurus svijagensis Riab. — ein Dinocephale aus den
Ablagerungen des Perm der UdSSR. Nova Acta Leopold., N. F. 9(59):155–205.
Farlow, J. O. (1987). Speculations about the diet and digestive physiology of herbivorous
dinosaurs. Paleobiology 13:60–72.
Gow, C. E. (1978). The advent of herbivory in certain reptilian lineages during the
Triassic. Palaeont. Afr. 21:133–141.
Gregory, J. T. (1945). Osteology and relationships of Trilophosaurus. Univ. Texas Publ. no.
4401:273–359.
Gregory, W. K. (1926). The skeleton of Moschops capensis Broom, a dinocephalian reptile
from the Permian of South Africa. Bull. Amer. Mus. Nat. Hist. 56:179–251.
Grine, F. E. (1977). Postcanine tooth function and jaw movement in the gomphodont
cynodont Diademodon (Reptilia: Therapsida). Palaeont. Afr. 20:123–135.
Harland, W. B., Armstrong, R. L., Cox, A. V., Craig, L. E., Smith, A.G., and Smith, D. G.
(1990). A Geologic Time Scale 1989. Cambridge and New York: Cambridge
University Press.
Heaton, M. J. (1979). Cranial anatomy of primitive captorhinid reptiles from the Late
Pennsylvanian and Early Permian, Oklahoma and Texas. Oklahoma Geol. Surv.
Bull. 127:1–84.
Heaton, M. J. (1980). The Cotylosauria: a reconsideration of a group of archaic tetrapods.
In The Terrestrial Environment and the Origin of Land Vertebrates, ed. A. L. Panchen,
pp. 497–551. London and New York: Academic Press.
Hopson, J. A. (1971). Postcanine replacement in the gomphodont cynodont Diademodon.
In Early Mammals, ed. D. M. Kermack and K. A. Kermack, pp. 1–21. London:
Academic Press.
Hopson, J. A. (1984). Late Triassic traversodont cynodonts from Nova Scotia and
southern Africa. Palaeont. Afr. 25:181–201.
Hotton, N. III, Olson, E.C., and Beerbower, R. (1997). Amniote origins and the discovery
of herbivory. In Amniote Origins, ed. S. S. Sumida and K. L. M. Martin, pp.
207–264. San Diego: Academic Press.
Ivakhnenko, M. F., and Tverdokhlebova, G. I. (1987). [Revision of Permian
bolosauromorphs from eastern Europe.] Paleont. Zh. 1987(2):98–106 (in Russian).
Iverson, J. B. (1982). Adaptations to herbivory in iguanine lizards. In Iguanas of the World:
Their Behavior, Ecology and Conservation, ed. G. M. Burghardt and A. S. Rand, pp.
77–83. Park Ridge, NJ: Noyes Publications.
Janis, C. M. (1995). Correlations between craniodental morphology and feeding
behavior in ungulates: reciprocal illumination between living and fossil taxa. In
Functional Morphology in Vertebrate Paleontology, ed. J. J. Thomason, pp.76–98.

Kemp, T. S. (1982). Mammal-like Reptiles and the Origin of Mammals. London and New York:
Academic Press.
King, G. (1990). The Dicynodonts: A Study in Palaeobiology. London: Chapman and Hall.
King, G. (1996). Reptiles and Herbivory. London: Chapman and Hall.
King, G. M., Oelofsen, B. W., and Rubidge, B. S. (1989). The evolution of the dicynodont
feeding system. Zool. J. Linn. Soc. 96:185–211.
Kitching, J. W. (1977). The distribution of the Karroo vertebrate fauna. Bernard Price Inst.
Palaeont. Res., Mem. 1:1–131.
Labandeira, C. C. (1997). Insect mouthparts: Ascertaining the paleobiology of insect
feeding strategies. Annu. Rev. Ecol. Syst. 28:153–193.
Laurin, M., and Reisz, R. R. (1995). A reevaluation of early amniote phylogeny. Zool. J.
Linn. Soc. 113:165–223.
Lee, M. S. Y. (1996). Correlated progression and the origin of turtles. Nature 379:812–815.
Lee, M. S. Y. (1997). Pareiasaur phylogeny and the origin of turtles. Zool. J. Linn. Soc.
120:197–280.
Lee, M. S. Y., and Spencer, P. S. (1997). Crown clades, key characters and taxonomic
stability: When is an amniote not an amniote? In Amniote Origins, ed. S. S.
Sumida and K. L. M. Martin, pp. 61–84. San Diego: Academic Press.
Li J. and Cheng Z. (1995). [The first discovery of bolosaurs from the Upper Permian of
China.] Vert. PalAsiat. 33:17–23 (in Chinese with English summary).
Maynard Smith, J., and Savage, R. J. G. (1959). The mechanics of mammalian jaws. School
Sci. Rev. 141:289–301.
Modesto, S. P. (1992). Did herbivory foster early amniote diversification? J. Vert. Paleont. 12
(Suppl. to 3):44A.
Modesto, S. P. (1994). The Lower Permian synapsid Glaucosaurus from Texas.
Palaeontology 37:51–60.
Modesto, S. P. (1995). The skull of the herbivorous synapsid Edaphosaurus boanerges from
the Lower Permian of Texas. Palaeontology 38:213–239.
Modesto, S. P., and Reisz, R. R. (1990). A new skeleton of Ianthasaurus hardestii, a
primitive edaphosaur (Synapsida: Pelycosauria) from the Upper Pennsylvanian
of Kansas. Can. J. Earth Sci. 27:834–844.
Olson, E. C. (1947). The family Diadectidae and its bearing on the classification of
reptiles. Fieldiana, Geol. 11:1–53.
Olson, E. C. (1961). The food chain and the origin of mammals. In International
Colloquium on the Evolution of Lower and Non-Specialized Mammals, ed. G.
Vanderbroek, pp. 97–116. Brussels: Koninklijke Vlaamse Akademie voor
Wetenschappen, Letteren en Schone Kunsten van België.
Olson, E. C. (1966). Community evolution and the origin of mammals. Ecology
47:291–308.
Olson, E. C. (1968). The family Caseidae. Fieldiana, Geol. 17:225–349.
Parra, R. (1978). Comparison of foregut and hindgut fermentation in herbivores. In The
Ecology of Arboreal Folivores, ed. G. G. Montgomery, pp. 205–229. Washington, DC:
Smithsonian Institution Press.
Parrington, F. R. (1946). On the cranial anatomy of cynodonts. Proc. Zool. Soc. Lond. 116
(II):181–197.
Reisz, R. R. (1986). Pelycosauria. In Handbuch der Paläoherpetologie, Teil 17A, ed. P.
Wellnhofer, pp. 1–102. Stuttgart: Gustav Fischer Verlag.
Reisz, R. R. (1997). The origin and early evolutionary history of amniotes. Trends Ecol.
Evol. 12:218–222.
Ricqlès, A. de, and Taquet, P. (1982). La faune de vertébrés du Permien Supérieur du
Niger. I. Le captorhinomorphe Moradisaurus grandis (Reptilia, Cotylosauria) – Le
crâne. Ann. Paléont. 68:33–106.
Romer, A. S. (1967). The Chañares (Argentina) Triassic reptile fauna. III. Two new
gomphodonts, Massetognathus pascuali and M. teruggii. Breviora 264:1–25.
Romer, A. S., and Price, L. I. (1940). Review of the Pelycosauria. Geol. Soc. Amer. Spec. Pap.
28:1–538.
Rubidge, B. S., and Hopson, J. A. (1996). A primitive anomodont therapsid from the base
of the Beaufort Group (Upper Permian) of South Africa. Zool. J. Linn. Soc.
117:115–139.
Rybczynski, N. (1996). Cranial Morphology and Phylogenetic Significance of Suminia
getmanovi, a Late Permian Anomodont from Russia. M.Sc. thesis, University of
Toronto.
Schiek, J. O., and Millar, J. S. (1985). Alimentary tract measurements as indicators of diet
in small mammals. Mammalia 49:93–104.
Scott, A. C., and Taylor, T. N. (1983). Plant/animal interactions during the Upper
Carboniferous. Bot. Rev. 49:259–307.
Smith, R. M. H. (1987). Helical burrow casts of therapsid origin from the Beaufort
Group (Permian) of South Africa. Palaeogeogr., Palaeoclimat., Palaeoecol. 60:155–170.
Stovall, J. W., Price, L. I., and Romer, A. S. (1966). The postcranial skeleton of the giant
Permian pelycosaur Cotylorhynchus romeri. Bull. Mus. Comp. Zool., Harvard Univ.
135:1–30.
Sues, H.-D., and Baird, D. (1998). Procolophonidae (Amniota: Parareptilia) from the
Wolfville Formation (Upper Triassic: Carnian) of Nova Scotia, Canada. J. Vert.
Paleont. 18:525–532.
Sues, H.-D., and Reisz, R. R. (1998). Origin and early evolution of herbivory in terrestrial
Sues, H.-D., Olsen, P. E., Scott, D. M., and Spencer, P. S. (2000). Cranial osteology of
Hypsognathus fenneri, a latest Triassic procolophonid reptile from the Newark
Supergroup of eastern North America. J. Vert. Paleont. 20.
iguana (Iguanidae) and Uromastix aegyptius (Agamidae). J. Morph. 148:363–390.
Troyer, K. (1982). Transfer of fermentative microbes between generations in a
herbivorous lizard. Science 216:540–542.
Turnbull, W. D. (1970). Mammalian masticatory apparatus. Fieldiana, Geol. 18:149–356.
Watson, D. M. S. (1954). On Bolosaurus and the origin and classification of reptiles. Bull.
Mus. Comp. Zool., Harvard Univ. 111:297–449.
Weishampel, D. B., and Norman, D. B. (1989). Vertebrate herbivory in the Mesozoic:
jaws, plants, and evolutionary metrics. In Paleobiology of the Dinosaurs, ed. J. O.
Farlow, Jr., pp. 87–100. Geol. Soc. Amer. Special Paper 238.
Williston, S. W. (1911). A new family of reptiles from the Permian of New Mexico. Amer. J.
Sci. 31:378–398.
Wing, S. L., and Sues, H.-D. (rapporteurs) (1992). Mesozoic and early Cenozoic terrestrial
ecosystems. In Terrestrial Ecosystems through Time, ed. A. K. Behrensmeyer et al.,
pp. 327–416. Chicago: University of Chicago Press.
paul m. barrett
Prosauropod dinosaurs and iguanas: speculations

on the diets of extinct reptiles
Introduction
A thorough knowledge of extant taxa is required when attempting

paleobiological interpretations of extinct animals (Lauder 1995; Witmer
1995). Skeletal material, though informative, can provide only a limited
amount of information on soft-tissue and behavioral characteristics. It is
not possible to identify with certainty the actual diet of any extinct
animal because we cannot observe the animal feeding. Fossilized gut con-
tents (enterolites) and feces (coprolites) may provide some dietary infor-
mation, but convincing enterolites are rare and coprolites are almost
never associated with the animal that produced them. In some cases,
direct comparison of the feeding adaptations of extinct taxa with closely
related extant forms (‘Extant Phylogenetic Bracket’ approach sensu
Witmer 1995) is possible, allowing these interpretations to be made with
some confidence (e.g., between extinct and extant mammals). Some
groups, however, have no present-day representatives for relevant com-
parison. For example, study of extant birds and crocodilians provides
only limited insight into the paleobiology of non-avian dinosaurs. In
cases where the phylogenetic approach to assessing function in extinct
organisms is not feasible, there is a tendency to rely on ahistorical bio-
mechanical modeling (Weishampel 1995). Such modeling is often based
on the identification and study of suitable extant analogues. The choice of
extant taxa for comparison, however, is often limited, and many studies
are forced to utilize animal models that may be distantly related to the
extinct taxon under consideration. This chapter deals with one such
model — the relationship between tooth form and diet in herbivorous rep-
tiles.
[42]
Speculations on prosauropod diets 43
Ever since the original description of Iguanodon (Mantell 1825), the

structure of the more posterior teeth of extant iguanine lizards has been
used as a paradigm for deducing the diet of extinct reptiles. Mantell
noticed that the teeth of both Iguanodon and Iguana had crowns that were
mesiodistally expanded and coarsely serrated along their mesial and
distal edges. As extant iguanines were thought to be exclusively herbivor-
ous, Cuvier (in Mantell 1825) reasoned, on the basis of tooth form, that
Iguanodon was a gigantic herbivorous reptile. Since then, all extinct rep-
tiles with this characteristic type of tooth structure have been interpreted
as herbivores (e.g., Galton 1985a, 1986; Crompton and Attridge 1986).
In this chapter, special attention will be paid to the comparisons that
have been made between the dentitions of prosauropod dinosaurs and
extant iguanine lizards. The diet of prosauropods has been the subject of
considerable controversy, not because authors have disagreed over the
dental features present in these animals, but because the interpretations
based on these features have differed considerably. Prosauropods were a
group of medium-sized to large saurischian dinosaurs with a worldwide
distribution during the Late Triassic and Early Jurassic (Galton 1990).
This group is currently thought to be monophyletic (Sereno 1989; Galton
1990; Upchurch 1993; Gauffre 1995), although some authors have argued
that prosauropods are paraphyletic with respect to sauropods (Gauthier
1986; Benton 1990). Prosauropods were taxonomically diverse and
numerically abundant, and include well-known forms like Plateosaurus
and Massospondylus.
Institutional abbreviations used in this chapter are: BMNH, The
Natural History Museum, London; BP/1, Bernard Price Institute for Palae-
ontological Research, University of the Witswatersrand, Johannesburg;
FMNH, Field Museum, Chicago; IVPP, Institute of Vertebrate Paleontol-
ogy and Paleoanthropology, Academia Sinica, Beijing; MCZ, Museum of
Comparative Zoology, Harvard University; UCMZ, University Museum
of Zoology, University of Cambridge.
Previous work on prosauropod diets
Historical review
The diet of prosauropods has prompted considerable interest among
paleontologists. Almost every conceivable mode of life has been postu-
lated for prosauropods, ranging from active carnivores (Swinton 1934) to
exclusively herbivorous forms (Galton 1984a, 1985a, 1986). This diversity
44 paul m. barrett
of opinion seems, in some regard, to follow the prevailing views on pro-

sauropod classification. Earlier authors tended to place the Prosauropoda
within the Theropoda, perhaps predisposing these authors to view pro-
sauropods as carnivores (Huene 1932; Swinton 1934; Lull 1953). Swinton
(1934:58) regarded Plateosaurus as ‘[an] enemy of unusual power and ter-
rifying aspect’ based on the large size of the animal, the structure of the
large manus claw and the serrated tooth form; he further suggested that
these features may have been adaptations for piscivory. Lull (1953:118) con-
cluded that Anchisaurus (‘Yaleosaurus’) was ‘an alert, active dinosaur,
preying upon the smaller vertebrates of his generation, as the powerful
claws and well developed teeth imply.’ Huene (1932) also commented on
the prosauropod dentition, describing that of Ammosaurus as ‘carnivor-
ous.’ More recent phylogenetic hypotheses, which allied the prosauro-
pods with the sauropods in the Sauropodomorpha (Charig, Attridge and
Crompton 1965; Romer 1966), opened the way for reinterpretations of
prosauropod feeding behaviour and brought the prosauropods into
line with the undoubtedly herbivorous sauropods. Colbert (1962:70–71)
regarded Plateosaurus as exclusively herbivorous, feeding on ‘soft vegeta-
tion’, while consideration of the dentition of Anchisaurus (‘Yaleosaurus’) led
him to the conclusion that ‘Yaleosaurus . . . was turning from a diet of
animals, and depending for much of its nourishment on the plants of
those distant Triassic days.’ However, the common association of prosau-
ropod skeletons with recurved, finely serrated teeth, like those of carniv-
orous archosaurs, led some authors (Charig et al. 1965; Romer 1966) to
suggest that some prosauropods (melanorosaurids and Teratosaurus in
particular) were carnivorous. These associations are now regarded as for-
tuitous; the recurved teeth were probably shed by carnivorous archosaurs
feeding on prosauropod carcasses (Galton 1984a, 1985a, 1986). Further-
more, Teratosaurus has been shown to represent a poposaurid archosaur
(Galton 1985b), rather than a carnivorous prosauropod dinosaur. Charig et
al. (1965) argued that there was no reason to try to ‘shoe-horn’ all prosau-
ropods into a particular ecological category, and speculated that prosau-
ropods covered a wide range of ecological preferences. They proposed
that the group may have included herbivorous, carnivorous, and omnivo-
rous representatives, but this idea was largely ignored by subsequent
authors. However, despite this diversity of opinion, a consensus seems to
have been reached sometime in the early 1970s that prosauropods were
primarily herbivorous (e.g., Galton 1973, 1976).
Following a detailed study of the southern African prosauropod Mas-
a)
b) c)
Figure 3.1. Prosauropod teeth in labial view. a) Massospondylus (BP/1/4376). Scale bar
equals 10 mm. b) Thecodontosaurus (inset with detail of the cutting edge
showing denticles). Scale bars each equal 1 mm. c) Sellosaurus. Scale bar equals
5 mm. (b and c from Galton 1985a; courtesy and copyright of Lethaia.)
sospondylus, Cooper (1981) challenged this view and suggested that prosau-
ropods were scavenger-predators. Kermack (1984) also cast doubts on pro-
sauropod herbivory, concluding that Thecodontosaurus was an omnivore.
Countering these views, Galton (1984a, 1985a, 1986) argued forcefully in
support of prosauropod herbivory. The main arguments in this debate are
reviewed below.
Scavenger-predators or herbivores?
The single most important feature employed in deducing prosauropod
diets, used by all of the authors mentioned above, has been the form of the
teeth. The following description of the prosauropod dentition is based on
the detailed observations of Galton (1984a, 1985a, 1986), supplemented by
personal observations on a wide range of prosauropod taxa. In general,
prosauropod maxillary and dentary teeth have crowns that are mesiodis-
tally expanded (clearly separating the crown from the root), taller than
wide in labial or lingual view, and symmetrical in mesial or distal view
(Figure 3.1). The tooth crowns are coarsely serrated along their mesial and
distal edges with the serrations projecting at an angle of about 45° to
46 paul m. barrett
the edge of the tooth crown. They are labiolingually compressed, not
recurved, and lack cingula. Wear is usually absent, but small apical wear
facets have been reported in Massospondylus (Gow, Kitching and Raath
1990), and high-angle mesial and distal wear facets are present on isolated
teeth referred to Plateosaurus (Cuny and Ramboer 1991) and Yunnanosaurus
(Galton 1985a; see below). Prosauropod maxillary and dentary teeth can
be distinguished from those of contemporary ornithischians in that the
teeth of the latter are more triangular in labial or lingual view (Sereno
1986) and asymmetrical in mesial or distal view (Galton 1984b).
Cooper (1981) listed eight characters in support of the scavenger-preda-
tor hypothesis. Tooth form was one of these characters, and Cooper
(1981:814) characterized the dentition as ‘serrated, carnivorous.’ Galton
(1984a, 1985a, 1986) attempted to refute this particular interpretation by
comparing the teeth of prosauropod dinosaurs with those of undoubted
carnivores and herbivores. He demonstrated that the teeth of prosauro-
pods do not resemble those of archetypal carnivores (theropod dinosaurs,
‘thecodontian’ archosaurs, and varanid lizards) because their crowns are
not recurved, are more compressed labiolingually and mesiodistally
expanded, and, perhaps most importantly, differ in the form of their ser-
rations. The teeth of carnivorous forms possess very fine serrations, which
project from the crown at an angle of around 90° to the edge of the tooth
(Figure 3.2a,b). In contrast, the serrations on prosauropod teeth are
coarser (i.e., larger relative to crown length), less numerous, and project at
an angle of around 45° from the edge of the tooth crown (see above and
Figure 3.1). Galton noted that prosauropod teeth are very similar to those
of extant iguanine lizards (notably Iguana; Figure 3.3a). Previous work on
iguanine diets (Hotton 1955; Montanucci 1968) appeared to indicate that
Iguana was exclusively herbivorous, and Galton concluded that prosauro-
pods were also herbivores, using their teeth to cut and tear plant material
in the same way as iguanine lizards (Throckmorton 1976). Galton’s views
have subsequently been accepted by most paleontologists.
Kermack (1984) noted that the ‘leaf-shaped’ teeth of Thecodontosaurus
indicated an herbivorous diet, but noted the absence of other adaptations
to herbivory, such as the presence of a ventrally displaced jaw joint. Unable
to reconcile the simple ‘carnivorous’ jaw action of Thecodontosaurus with
the ‘herbivore-like’ tooth structure, she advocated a compromise, propos-
ing that Thecodontosaurus was omnivorous. Kermack (1984:110) envisioned
Thecodontosaurus ‘eating largely soft vegetable matter but supplementing
a)0 b)0
c)0
Figure 3.2. Tooth crowns of theropod dinosaurs. a,b) Tooth crown of indeterminate
theropod (BMNH R5226) showing the recurvature and fine serrations
characteristic of carnivorous dentitions. a) ?Labial view of tooth crown. Scale
bar equals 10 mm. b) Detail of the serrations on the distal cutting edge of the
tooth. Scale bar equals 2 mm. c) Premaxillary tooth of Troodon formosus in
lingual view (from Currie 1987; courtesy and copyright of Journal of Vertebrate
Paleontology). Scale bar equals 1 mm.
48 paul m. barrett
a) b)
Figure 3.3. Teeth of Iguana ?delicatissima (UCMZ R8917) in labial view. a) Maxillary
tooth. b) Premaxillary tooth. Scale bars each equal 1 mm.
this by killing prey when possible, in a similar way to Iguana or Uromastix

today’ [my italics], a possibility not considered by Galton.
The other arguments put forward by Cooper (1981) in support of the
scavenger-predator hypothesis were all effectively refuted by Galton
(1984a, 1985a, 1986), obviating the need for further discussion here.
Diet and tooth structure in iguanine lizards
Finding modern analogues for extinct reptilian herbivores is difficult due

to the paucity of present-day herbivorous reptiles. Today, reptilian herbi-
vores are represented by a small number of lizards and turtles (King 1996).
Turtles lack teeth and crop vegetation using a self-sharpening keratinous
beak, limiting study of the relationship between tooth form and diet to
herbivorous lizards, many of which are referable to the Iguanidae. The
majority of iguanids are insectivorous whereas members of the Iguaninae
(e.g., Iguana, Ctenosaura) are thought to be exclusively herbivorous
(Hotton 1955; Montanucci 1968). Several authors have argued for a close
relationship between tooth form and diet in iguanid lizards. Cuspidation
is poorly developed in insectivores whereas herbivores possess cuspidate
Table 3.1. Food items taken by free-living and captive

iguanine lizards
Iguana iguana Dipsosaurus dorsalis Uromastix spp.
Small mammals Small mammals (rarely) Small mammals

Baby mice Insects Insects
Birds Eggs (rarely) Fruit
Insects Chopped meat (rarely) Vegetables
Dog food Vegetables
Chopped meat Fruit
Fruit Leaves, shoots
Vegetables
Dandelion flowers
Mealworms
Eggs (rarely)
Leaves, shoots
Sources: Hotton (1955); Montanucci (1968); Wallach and Boever

(1983); Frey (1986).
or serrated teeth (Hotton 1955; Montanucci 1968). Interestingly, a number

of herbivorous fishes possess coarsely serrated teeth very similar to those
of iguanines. The surgeonfish Acanthurus nigrofuscus uses these teeth to
tear or incise algae from rocks, and Ctenochaetus striatus employs them to
rake particles from algal stands (Purcell and Bellwood 1993).
Contrary to many published accounts, ‘herbivorous’ iguanines are not
exclusively herbivorous. Captive iguanines are known to eat a wide
variety of foods, ranging from fruit to dog food (Frye 1974 [cited in
Wallach and Boever 1983]; Frey 1986; Table 3.1), and show severe growth
disturbances if meat is excluded from their diet (J. Pearson, pers. comm.).
These observations indicate that free-living iguanines, of necessity, must
include animal protein in their diets to avoid various vitamin-deficiency
disorders. Previous authors often deduced the diet of free-living igua-
nines from dissection and inspection of stomach contents. These dissec-
tions have shown a preponderance of plant material in the gut with rare
animal remains such as insect cuticle. However, these dissections are
likely to be misleading for two reasons (E. N. Arnold, pers. comm.). First,
plant material is very resistant to digestion and has a longer passage time
through the gut than meat. Second, plant material is nutritionally
poorer, on a weight for weight basis, than animal protein. Therefore, a
greater volume of vegetation must be consumed in order to provide the
same amount of nutrients as a much smaller quantity of meat. Other
50 paul m. barrett
factors also need to be taken into consideration. The diet of iguanines has
been shown to change during ontogeny (Montanucci 1968), and there has
been a general disregard of the selective pressures exerted on juveniles by
the relationship between diet and dentition, as most authors have con-
centrated on adults in their studies. Furthermore, little if any attention
has been paid to seasonal variation in the abundance of various food items
and the impact of these changes. Is it possible that, depending on envi-
ronmental conditions, these animals can spend extended periods as strict
herbivores or carnivores? More field data are needed on the dietary prefer-
ences of free-living iguanines, but anecdotal evidence shows them
preying on small rodents, eggs, and hatchling birds in addition to their
more generally accepted herbivorous diet (Table 3.1 and references cited
therein). Such observations are likely to provide additional surprises: for
example, giant tortoises on Aldabara are known to feed on dead tortoises
and human feces (E. N. Arnold, pers. comm.), whereas mammalian herbi-
vores like horses, red deer, sheep and cattle have been observed eating
nesting seabirds, chicks, and discarded hamburgers (Furness 1988a,
1988b, 1989; S. Finney, pers. comm.). It seems more reasonable to interpret
Iguana and other ‘herbivorous’ iguanine lizards as opportunistic or facul-
tatively omnivorous rather than strictly herbivorous.
The serrated tooth structure of ‘herbivorous’ iguanines may be an
adaptation to herbivory in the context of continued carnivory or omni-
vory, rather than an a priori adaptation to herbivory (a conclusion reached
independently by E. N. Arnold [pers. comm.]). This hypothesis was tested
by using a comparative approach to iguanine diets (Figure 3.4). Data on
iguanine diets were obtained from a number of sources (Hotton 1955;
Montanucci 1968; Wallach and Boever 1983; Frey 1986). Unfortunately,
the only available data are either qualitative or anecdotal; good quantita-
tive information is lacking. These data were plotted onto a cladogram of
iguanine interrelationships (de Queiroz 1987), together with the tooth
form of the taxa under consideration. Iguana and Cyclura are the most
derived iguanines (de Queiroz 1987) and are also the taxa in which vegeta-
tion makes up the greatest part of the diet (Figure 3.4). Less derived igua-
nines, such as Ctenosaura, are known to eat less vegetation (Hotton 1955;
Montanucci 1968). Iguana possesses the most cuspidate teeth whereas
other iguanines, which eat more insects, have teeth with fewer cusps.
Thus there is a correlation between diet, tooth structure, and phylogeny.
However, it is notable that significant quantities of animal prey are still
Figure 3.4. Correlation between iguanine phylogeny, tooth form, and diet. The tooth
of a basal, insectivorous iguanid (Basiliscus) is shown at the left. Phylogeny
based on de Queiroz (1987), iguanine tooth form based on de Queiroz (1987)
and Montanucci (1968), and diet based on various sources including Hotton
(1955), Montanucci (1968), and Frey (1986). Drawings are not to scale.
52 paul m. barrett
taken by iguanines with cuspidate teeth. Indeed, Montanucci (1968) sug-

gested that the cuspidate teeth of Ctenosaura are well-suited for piercing
insect cuticle.
The posterior maxillary and dentary teeth are those that are most com-
monly illustrated in publications dealing with iguanine diets, and are
also the teeth that form the basis for the ‘herbivorous tooth-form’ para-
digm. In contrast, the structure of the premaxillary, anterior maxillary,
and anterior dentary teeth is rarely mentioned in descriptions of iguanine
feeding behavior. Hotton (1955:91) briefly described the structure of these
teeth, but did not consider them in his analysis of the relationship
between diet and dentition because ‘they are similar in all genera under
consideration.’ Montanucci (1968) and Throckmorton (1976) only men-
tioned these teeth in passing, and concentrated on the structure of the
posterior teeth. The anterior teeth are rather different from the posterior
maxillary and dentary teeth. Anterior tooth crowns are not mesiodistally
expanded and are recurved, terminating in a sharp point (Figure 3.3b).
They are not labiolingually compressed, like the distal teeth, but are sub-
conical in cross-section. Finally, the anterior teeth bear very fine serra-
tions on the distal edges of the tooth crown, quite unlike the coarse
serrations found on the posterior teeth. Hotton (1955) noted that Crotaph-
ytus has ‘high-crowned’ anterior teeth and suggested that these teeth may
be useful for the quick seizure of prey items. The pointed and recurved
anterior teeth seem ideally suited for this purpose, and the use of these
teeth in prey capture seems likely in all iguanines that possess this differ-
entiated dentition.
Implications for prosauropod diet
Prosauropods as omnivores
If the evidence in favor of regarding ‘herbivorous’ iguanines as opportu-
nistic or facultative omnivores is accepted, the utility of iguanine tooth
form as a paradigm for inferring the diet of extinct reptiles needs to be
reassessed. The possession of iguanine-like teeth can no longer be used as
a reliable indicator of herbivory sensu stricto. Indeed, if we are to use igua-
nine lizards as a paradigm for establishing diet, we should conclude that
other animals possessing this type of dentition were also omnivorous.
The striking similarities in the dentitions of prosauropod dinosaurs
and iguanine lizards indicate that it is reasonable to regard prosauropods
as opportunistic or facultative omnivores, a possibility mentioned by
several other authors (Kermack 1984; Gow, Kitching, and Raath 1990;
Zhang and Yang 1994). Prosauropods have recurved premaxillary teeth
and anterior maxillary teeth that sometimes bear very fine serrations
along their distal edges (e.g. Massospondylus [Gow et al. 1990; MCZ 8893],
Jingshanosaurus [Zhang and Yang 1994]) (Figure 3.5). These teeth could
have functioned in a similar way to the anterior teeth of iguanines in the
capture of small prey (e.g., reptiles, mammals, invertebrates). An unde-
scribed anchisaurid skeleton from the Lower Jurassic McCoy Brook For-
mation of Nova Scotia provides some support for this hypothesis. This
specimen was found with an in situ gastric mill that contained a badly
worn maxilla of the sphenodontid Clevosaurus (Shubin, Olsen, and Sues
1994; H.-D. Sues, pers. comm.).
Several other features of prosauropods support the contention that
they may have had more varied diets than usually supposed. As Cooper
(1981) noted, there is no reason to suggest that the large manual ungual of
prosauropods could not have been used in dismembering carrion in addi-
tion to its other functions. Furthermore, supposedly ‘herbivorous’ fea-
tures including the ventrally displaced jaw joint, precise occlusion, and a
well-developed coronoid eminence, are poorly developed or absent in
many prosauropod dinosaurs (Galton 1976; Barrett 1998). It is possible
that different prosauropod taxa had different diets, as originally sug-
gested by Charig et al. (1965).
The degree to which prosauropods exploited animal material prob-
ably depended on environmental factors, age, and ecological opportu-
nities. Possible environmental and ontogenetic factors are discussed
below.
Environmental factors
Prosauropod skeletal remains have been recovered from a variety of
paleoenvironmental settings, ranging from lacustrine deposits with
abundant plant material to deposits formed under arid conditions with
ephemeral water sources and sparse vegetation (see Table 3.2). The major-
ity of prosauropod fossils have been found in sedimentary rocks from
depositional settings that are best described as ‘semi-arid’ or ‘seasonally
wet.’ These environments were typically floodplains or alluvial fans with
ephemeral surface water (braided streams and playa lakes) and occasional
permanent water bodies. In some prosauropod-bearing strata, there is
good evidence of rapid climatic fluctuations (e.g., Fleming Fjord Forma-
tion of East Greenland [Jenkins et al. 1994]). The majority of prosauropod
54 paul m. barrett
Figure 3.5. Premaxillary teeth of Massospondylus (MCZ 8893) in labial view. Scale bar
equals 10 mm.
Table 3.2. Paleoenvironments of some of the principal prosauropod-bearing formations
Formation/locality Taxa present Age Paleoenvironment Source(s)
Portland Formation, Anchisaurus Sinemurian– The lower part of the formation consists of several Smoot (1991)
Newark Supergroup, Ammosaurus Pliensbachian lacustrine cycles. Wet periods alternate with much
Connecticut, USA drier periods. The upper part is characterized by
ephemeral streams, suggestive of a much drier
climate
Navajo Sandstone, Ammosaurus Pliensbachian– Vertebrate remains are found in interdune deposits. Winkler et al. (1991)
Arizona, USA Toarcian These interdunes were relatively mesic with respect
to the surrounding dune systems
Kayenta Formation, Massospondylus Hettangian Broad, well-drained floodplain with abundant Harshbarger, Repenning and
Arizona, USA surface water surrounded by sand dunes and Irwin (1957); Colbert (1981);
highlands Clark and Fastovsky (1986)
McCoy Brook Formation, cf. Ammosaurus Hettangian Seasonally wet H.-D. Sues (pers. comm.)
Nova Scotia, Canada
Fleming Fjord Formation, Plateosaurus Norian Cyclic deposition in a shallow or ephemeral lake Jenkins et al. (1994)
East Greenland
Lower Elliot Formation, Euskelosaurus late Carnian or Semi-arid climate with meandering channel and Smith et al. (1993)
South Africa Blikanasaurus early Norian floodplain deposits. Perennial and ephemeral
Melanorosaurus water sources
Upper Elliot Formation, Massospondylus Hettangian – Semi-arid to arid climate with flood basin Smith et al. (1993)
South Africa; ?Sinemurian deposition and later flood-fans and dunes. Braided
Clarens Formation, Sinemurian streams, playa and dunes present
South Africa;
Forest Sandstone Hettangian –
Formation, Zimbabwe Sinemurian
Knollenmergel, Plateosaurus late Norian Climate semi-arid with periodic heavy rains. The Sander (1992)
Germany sediments were deposited on a broad floodplain
Fissure-fillings, Thecodontosaurus ?Norian Climate semi-arid with infrequent heavy rains Robinson (1957); S. E. Evans
Wales (pers. comm.)
Los Colorados Formation, Riojasaurus Norian Warm, humid, lacustrine deposition Bonaparte (1972)
Argentina Coloradisaurus
Lower Lufeng Formation, Lufengosaurus Hettangian– Lacustine deposits indicating a warm, wet Zhang and Yang (1994); Z. Luo
Yunnan, China Yunnanosaurus Pliensbachian environment with abundant vegetation (pers. comm.)
Jingshanosaurus
paleoenvironments would probably have had a sparse cover of vegetation

for much of the time, though there may have been more pluvial periods,
either seasonally or due to longer-term climatic variations, during which
times more vegetation would have been available.
Unlike the semi-arid environments of today (e.g., savannah, pampas),
which can support large (albeit migratory) populations of mammalian
herbivores, the environments inhabited by prosauropod dinosaurs
lacked the grasses that represent the bulk of the primary production in
such regions today. The vegetation available to prosauropods is likely to
have been xerophytic or confined to areas where water was permanently
available either on or below the surface (Gow et al. 1990). It seems unlikely
that prosauropods could have maintained an exclusively herbivorous
diet under such environmental conditions, particularly given the high
biomass of many prosauropod communities (Galton 1985a), and the
ability to exploit and utilize animal prey or carrion would have been
advantageous. Dietary preferences may have shifted with changes in
climate. For example, more animal food may have been taken during dry
periods and more vegetation taken when the climatic conditions were
wetter. Interestingly, the prosauropod with the most ‘herbivore-like’
teeth (Yunnanosaurus) comes from one of the few environments in which
water availability was probably not limiting (Table 3.2).
Ontogenetic factors
Another possibility is an ontogenetic shift in diet. Juvenile iguanines have
been reported to take a higher proportion of animal protein than adults.
In Ctenosaura, the juveniles have a larger number of cusps on their teeth
than adults, but eat less vegetation (Montanucci 1968). This observation
contradicts the view that higher levels of tooth cuspidation are an a priori
adaptation to herbivory because adults, which take more vegetation in
their diets, would be expected to have more serrations on their teeth than
juveniles. However, fewer cusps on the teeth of adults may simply reflect
the longer functional life of the tooth. More work is needed in this area.
Prosauropods may also have displayed an ontogenetic shift in diet, with
juveniles taking more animal prey than adults. However, no significant
changes in tooth structure could be identified in a growth series of Massos-
pondylus (BP/1/4376, BP/1/4779, BP/1/5241). The teeth of the juvenile speci-
men (BP/1/4376) possess one or two fewer denticles per tooth crown than
the other, larger, specimens, but the increase in the number of denticles on
adult teeth probably reflects the larger size of adult teeth rather than a
58 paul m. barrett
functional change. This observation does not preclude an ontogenetic

change in prosauropod diets. On the basis of histological evidence, it has
been suggested that juvenile Massospondylus grew at higher rates than
adults (Chinsamy 1993). If this were the case we might expect juveniles to
have higher nutritional demands than adults. These demands may have
been met by an increased dependence on animal material in juvenile diets.
The teeth of Yunnanosaurus
Yunnanosaurus is a large prosauropod dinosaur from the Dark Red Beds of

the Lower Jurassic Lower Lufeng Formation of Yunnan (Young 1951). The
structure of its teeth is unique among prosauropods (Young 1942; Galton
1985a) in that the tooth crowns are broadly spatulate, labially convex, and
generally lack denticles (one or two denticles may be present at the apices
of the teeth [Young 1942; pers. obs.]). This unusual pattern, similar to that
of sauropods, may indicate that Yunnanosaurus relied upon vegetation
much more heavily than other prosauropods. Although the teeth of Yun-
nanosaurus are sauropod-like, no other cranial characters support the
assignment of this genus to the Sauropoda. Known sauropod skulls
possess a suite of autapomorphies (transversely broad snout, contact
between the posterior end of the ascending process of the maxilla and the
lacrimal, retraction of the external nares, and anteroposterior shortening
of the nasals and frontals; Upchurch 1995, 1998; Upchurch and Barrett,
this volume), which are absent in Yunnanosaurus (Young 1942; pers. obs.).
Simmons (1965) briefly mentioned several isolated teeth from the
Lower Lufeng Formation, which he referred to Yunnanosaurus. These teeth
(referred to hereafter as ‘Yunnanosaurus’) have large wear facets along their
mesial and distal edges, extremely similar to the wear on some sauropod
teeth, which can only have been formed by tooth-to-tooth contact during
occlusion (Galton 1985a; Figure 3.6a). These wear facets indicate the pres-
ence of precise occlusion in ‘Yunnanosaurus’. The teeth of all other prosau-
ropods lack wear (except in rare instances), indicating that tooth-to-tooth
occlusion was absent. The condition seen in ‘Yunnanosaurus’ would repre-
sent a significant change toward herbivory in the masticatory apparatus
of prosauropods (Galton 1985a).
However, referral of these teeth to Yunnanosaurus is doubtful (Barrett
1998). The teeth of ‘Yunnanosaurus’ have crowns with an almost cylindrical
cross-section. The crowns of teeth definitely referable to Yunnanosaurus are
labiolingually compressed (IVPP V94, V505). Also, the provenance of the
a)
b)
Figure 3.6. a) Tooth (FMNH CUP 2051) referred to Yunnanosaurus by Simmons (1965)
in mesial, labial, and distal views showing wear facets. b) Posterior teeth of
Yunnanosaurus (IVPP V94) in labial view. Scale bars each equal 10 mm.
‘Yunnanosaurus’ teeth is unknown in most cases; although it is known that

they were collected from the Lower Lufeng Formation, the exact localities
and horizons (Dull Purplish or Dark Red Beds) have not been recorded,
and no material unquestionably identifiable as Yunnanosaurus can be
directly associated with these teeth (Simmons 1965). Furthermore, tooth
wear has not been described in any of the type specimens of Yunnanosaurus
spp. (Young 1942, 1951) and was not observed by me in a recent examina-
tion of several specimens of Yunnanosaurus (IVPP V94, V505; Figure 3.6b).
The general absence of tooth wear in prosauropods, the absence of
such wear in several specimens unquestionably referable to Yunnanosau-
rus, the prevalence of this type of wear in sauropods (Calvo 1994; Barrett
60 paul m. barrett
and Upchurch 1995; Upchurch and Barrett, this volume), and the striking
similarity between the teeth of ‘Yunnanosaurus’ and those of sauropods,
indicate that the ‘Yunnanosaurus’ teeth described by Simmons (1965) and
Galton (1985a) should be referred to the Sauropoda (Barrett 1998). The age
of the Lower Lufeng Formation has been contentious, but recent work
indicates that it is Early Jurassic (Hettangian–Pliensbachian) in age (Luo
and Wu 1994, 1995). Basal sauropods are known from beds of this age in
China and elsewhere (McIntosh 1990; Upchurch 1995). Kunmingosaurus,
from the Wuding Basin of Yunnan Province (Dong 1992), and Zizhongosau-
rus, from the Da’znzhai Formation of Sichuan Province (Dong, Zhou, and
Zhang 1983), are both broadly contemporaneous with the ‘Yunnanosaurus’
teeth. If the latter teeth are indeed referable to an early sauropod they
would represent the first sauropod remains to be recovered from the
Lower Lufeng Formation and would also represent some of the oldest sau-
ropod specimens. Furthermore, these teeth would indicate that precise
tooth-to-tooth occlusion appeared early in the evolutionary history of the
Sauropoda (see Upchurch and Barrett, this volume).
Prosauropod diets and paleoecology
The preceding discussion indicates that prosauropod dinosaurs had

rather catholic diets. Plant material probably formed the major constitu-
ent of the diet, which was supplemented with animal protein. The rela-
tive proportions of food items in the diet could have varied due to
environmental and ontogenetic factors (see above). The evolution of large
body size in prosauropods may be correlated with a shift towards an her-
bivorous diet, as has been suggested for the sauropods (Farlow 1987).
Increased body size allows slower passage times through the gut and a
longer period of gut fermentation — two processes that may have allowed
these animals to eat tough, hard-to-digest vegetation (Farlow 1987). This
suggestion is in accordance with our current knowledge of prosauropod
paleoenvironments (Table 3.2) where plant material may have been tough
and xerophytic. Alternatively, the larger prosauropods (e.g., melanoro-
saurids, Plateosaurus) may have relied more heavily on vegetation than
their smaller relatives.
Two distinct prosauropod assemblages were present in the Karoo
Basin of southern Africa; one in the Lower Elliot Formation (which is late
Carnian or early Norian in age; Olsen and Galton 1984) consisting of Eus-
kelosaurus, Blikanasaurus and Melanorosaurus, and one in the Upper Elliot,
Clarens, and Forest Sandstone formations (which is Hettangian–Sinemu-

rian in age; Olsen and Galton 1984) containing the single genus Massos-
pondylus (Haughton 1924; Kitching and Raath 1984). The deposition of
these strata took place during a period of increasing aridity. The strata of
the Lower Elliot Formation display evidence of some permanent water
sources and ephemeral surface water. The Upper Elliot strata consist of
playa and braided stream deposits, and the Clarens and Forest Sandstone
formations indicate drier conditions with playa and sand dunes (Smith,
Eriksson, and Botha 1993). Haughton (1924) and Charig et al. (1965) sug-
gested that the change in the prosauropod assemblages was initiated by
the changes in climate, though these authors believed that differences in
the locomotory capabilities and preferred habitats of the various prosau-
ropod taxa accounted for the differences between the Lower and Upper
Elliot assemblages. Here, it is suggested that the relatively mesic environ-
ment prevailing during Lower Elliot times may have supported denser
vegetational cover than the drier environments present during Upper
Elliot/Clarens times and may have offered more opportunities for prosau-
ropod diversification. Unfortunately, few plant fossils are recorded from
these formations (Gow et al. 1990), and thus vegetational cover must be
inferred from climatic evidence.
Many Mid-Jurassic to Early Cretaceous dinosaurian assemblages are
characterized by a remarkable diversity of sauropods, and some localities
have yielded up to six sauropod genera (Weishampel 1990). This diversity
has been attributed to variation in the feeding adaptations of sauropods,
including cranial and postcranial characters, which may have allowed
niche partitioning between the various sauropod taxa (Barrett and
Upchurch 1995; Upchurch and Barrett, this volume). Several prosauropod
assemblages also display some degree of taxonomic diversity (Table 3.2).
In these cases, it does not seem possible to separate the taxa ecologically as
prosauropod feeding mechanisms appear to be remarkably uniform
(Galton 1984a, 1985a, 1986). Furthermore, niche separation on the basis of
differences in body size does not seem to occur as many sympatric taxa
reached similar adult body sizes. However, if prosauropods were faculta-
tively omnivorous, rather than strictly herbivorous, this would have
allowed greater dietary variation within and between genera, reducing
niche competition and allowing the diversification of similar genera
within the same environment.
The supposed generic diversity in at least one prosauropod fauna,
that of the Lower Lufeng Formation, may be artificially high. Young
62 paul m. barrett
(summarized in Young 1951) named six species comprising three genera

of prosauropod dinosaurs, Lufengosaurus, Yunnanosaurus, and Gyposaurus.
Rozhdestvensky (1965) and Galton (1990) argued that Lufengosaurus and
Yunnanosaurus are valid monotypic genera and that Gyposaurus is a sub-
jective junior synonym of Lufengosaurus. Most authors have stated that
these two genera are distributed throughout the Lower Lufeng Forma-
tion in both the Dull Purplish and Dark Red Beds (Young 1951; Wei-
shampel 1990). However, it has been suggested that there are two
distinct prosauropod assemblages, the older Lufengosaurus fauna (Dull
Purplish Beds) and a younger (Dark Red Beds) Yunnanosaurus fauna
(Dong 1992). The latter possibility seems the most likely; the specimens
referable to each genus come from the Dull Purplish Beds (Lufengosaurus)
and Dark Red Beds (Yunnanosaurus), respectively. Furthermore, the strat-
igraphical ranges of each of these genera were extended into the other
unit of the formation by the referral of isolated prosauropod remains to
each of these taxa by Young (1951). Much of this referred material is actu-
ally not identifiable at lower taxonomic levels; those specimens that do
show diagnostic features of either Lufengosaurus or Yunnanosaurus always
come from the horizon from which the holotypes of the type-species of
the respective genus was obtained (P. M. Barrett, unpubl. data). The
recent description of Jingshanosaurus (Zhang and Yang 1994) from the
Dull Purplish Beds implies that two prosauropod taxa may have been
present during the time of deposition of this unit, but only one (Yunnan-
osaurus) was present in the Dark Red Beds. A taxonomic revision of the
prosauropod dinosaurs from the Lower Lufeng Formation is long
overdue and is needed to resolve this rather unsatisfactory situation.
Diet in other extinct reptilian groups
Some of the dental features present in prosauropod dinosaurs and igua-

nine lizards are present in a variety of other dinosaurian taxa, including
basal ornithischians and some theropods, and in several non-dinosaurian
archosaurs such as crocodyliforms.
Ornithischia
The basal ornithischian Lesothosaurus has a differentiated upper dentition
comprising 6 premaxillary teeth and 15 or 16 maxillary teeth (Thulborn
1970, 1971; Sereno 1991). The maxillary teeth have low, (in labial or lingual
view) triangular crowns that bear coarse denticles along their mesial and
Figure 3.7. Teeth of Lesothosaurus in labial view. a) Maxillary teeth (BMNH R8501). b)
Premaxillary teeth (BMNH RUB17). Scale bars each equal 1 mm.
distal edges (Figure 3.7a). The tooth crowns become progressively

broader in the more posterior maxillary teeth. In contrast, the premaxil-
lary teeth are recurved and lack the large denticles present on the maxil-
lary teeth (Figure 3.7b). The fourth and fifth premaxillary tooth crowns
bear very fine serrations along their mesial cutting edges, and the sixth
premaxillary tooth bears serrations on both mesial and distal margins
64 paul m. barrett
(Thulborn 1970, 1971; Sereno 1991; Figure 3.7b). The similarity of these
teeth to those of prosauropods and iguanines (see above), and the absence
of many adaptations to herbivory (such as a ventrally displaced jaw joint,
precise occlusion, relative movements of the jaws), indicate that basal
ornithischians may have been facultatively omnivorous.
Hypsilophodontid and heterodontosaurid ornithopods retain
recurved premaxillary teeth (Galton 1974; Crompton and Charig 1962).
However, the skulls of hypsilophodontids and other euornithopods
possess a pleurokinetic hinge that allowed the production of a transverse
power stroke during jaw closure (Norman and Weishampel 1985). This
jaw mechanism, and the resulting tooth wear, are most probably asso-
ciated with an almost exclusively herbivorous diet. Various other fea-
tures, such as a ventrally displaced jaw joint, also indicate that
euornithopods were primarily herbivorous. The premaxillary teeth may
have been used for the prehension of plant food or, in rare instances,
capture of small prey. Iguanodontians lack premaxillary teeth, have well-
developed pleurokinetic hinge systems, and complex dental batteries.
These features, and many others, indicate that these dinosaurs were
exclusively herbivorous. The diet of heterodontosaurids is more proble-
matic. The jaw action and structure of the cheek teeth (Weishampel 1984;
Crompton and Attridge 1986) appear to indicate an herbivorous diet, but
the shapes of the premaxillary teeth and of the dentary caniniform tooth
suggest that facultative omnivory may have been a possibility. Molnar
(1977) proposed that the caniniform tooth was used in intraspecific
combat and display, as it is in primitive cervids like Muntiacus. This func-
tion seems entirely feasible. However, the caniniform teeth of Heterodonto-
saurus are finely serrated, in a manner reminiscent of theropod teeth,
whereas the caniniform teeth of cervids lack serrations and possess very
sharp carinae (pers. obs.). The presence of serrations on the caniniforms of
Heterodontosaurus may indicate that they were used during feeding. Inter-
estingly, the cervids with enlarged caniniform teeth are known to take
occasional animal prey (see references cited in Jarman 1974). The enig-
matic Early Cretaceous ornithischian Echinodon has a caniniform anterior
maxillary tooth (Galton 1978), but the crown of this tooth is not serrated
(BMNH 48209; pers. obs.). This tooth may have performed similar func-
tions to those postulated for the dentary caniniform of Heterodontosaurus
(see above).
Pachycephalosaurs retained recurved premaxillary teeth and lacked
many obvious adaptations to herbivory. The more posterior teeth bear a
strong resemblance to those of iguanines, prosauropods, and basal orni-

thischians. These observations indicate that pachycephalosaurs may also
have been facultative omnivores. Stegosaurs and ankylosaurs have simple
maxillary and dentary teeth and lack a specialized masticatory apparatus,
and some retained slightly recurved premaxillary teeth. On the basis of
dental features alone, some of these animals might be regarded as faculta-
tive omnivores. However, the large body size, barrel-shaped abdomen,
heavily armored bodies, quadrupedal habits, and the lack of a manipula-
tive manus in derived thyreophorans indicate that they had an exclu-
sively herbivorous diet.
Theropoda
Although most theropods possessed a typically ‘carnivorous’ dentition
(recurved, labiolingually compressed tooth crowns with serrated cutting
edges), there are several exceptions to this rule. Ornithomimosaurs are
either toothless or retain a poorly developed dentition in the anterior
portion of the mandible (Barsbold and Osmólska 1990). Oviraptorosaurs
are completely toothless (Barsbold, Maryanska, and Osmólska 1990).
These groups are usually considered to be carnivorous, either swallowing
small prey whole (ornithomimosaurs) or using strong adductor muscles
to crush molluscs against the roof of the mouth prior to swallowing (ovi-
raptorosaurs; Barsbold 1983). Cracraft (1971) noted several similarities
between the jaw apparatus of caenagnathid oviraptorosaurs and dicyno-
dont therapsids, although he did not explicitly suggest that caenagnath-
ids were herbivorous. Although they accepted the functional similarities
between caenagnathids and dicynodonts, Barsbold et al. (1990) still con-
sidered caenagnathids carnivorous. However, several authors have
argued that some of these groups were either herbivorous or omnivorous
(Paul 1988; Smith 1992; Sues 1997). As all of these taxa typically lack teeth
they will not be further discussed here.
Pelecanimimus, a basal ornithomimosaur from the Early Cretaceous of
Spain, possesses over 200 small teeth (Pérez-Moreno et al. 1994). The teeth
are confined to the anterior portions of the upper and lower jaws. The
anterior teeth are recurved, and the posterior teeth have slender, blade-
like crowns. Pérez-Moreno et al. (1994) argued that this arrangement was
intermediate between the dentitions of troodontids, which possess dentic-
ulate teeth with moderate interdental spaces (Currie 1987), and the more
derived ornithomimids, which lack any teeth. The large number of closely
packed tooth crowns in Pelecanimimus was interpreted as the functional
66 paul m. barrett
equivalent of a single cutting edge, similar to that provided by a kerati-

nous beak (Pérez-Moreno et al. 1994). If this functional interpretation is
accepted, herbivory, omnivory, or carnivory are equally likely for this
animal. Extant turtles, which also lack teeth, use the beak for a variety of
dietary strategies.
Troodontid theropods are usually thought to be small predators
(Osmólska and Barsbold 1990). However, Holtz and Chandler (1994) have
argued, on the basis of denticle size and serration density, that Troodon
may have been omnivorous or even herbivorous. In most theropods, den-
ticle size and serration density fall on the same allometric growth curve.
In contrast, troodontid denticles are relatively larger and more widely
spaced (Figure 3.2c). This is also the case in basal ornithomimosaurs and
therizinosaurs. The structure of the tooth crowns in these groups was
found to be more similar to that of ‘herbivorous’ dinosaurs and iguanines
than to that of undoubted carnivores (Holtz and Chandler 1994). The sim-
ilarity of troodontid and therizinosauroid teeth was also noted by Russell
and Dong (1994).
The affinities of therizinosaurs have been the subject of some debate.
Some authors have united them with the prosauropods (Sereno 1989),
while others have suggested that they were relicts of an alleged prosauro-
pod–ornithischian transition (Paul 1984). However, recent studies
suggest that they are derived members of the Theropoda (Russell and
Dong 1994; Clark, Perle, and Norell 1994; Sues 1997) as originally sug-
gested by Perle (1979). Paul (1984) argued that therizinosaurs may have
been herbivorous based on the inferred presence of ‘cheeks,’ the retrover-
sion of the pubis (allowing for a larger gut to process plant food), the
inferred presence of a keratinous beak, and the similarity of therizinos-
auroid teeth to those of prosauropods and basal ornithischians. The
coarsely serrated, mesiodistally expanded tooth crowns indicate that
omnivory is also a possibility.
Non-dinosaurian archosaurs
Most crocodilians have simple conical teeth, with tooth crowns that are
not expanded relative to the width of the root and taper to a point. There
is typically little variation in tooth form along the tooth row (Romer
1956). By analogy with extant crocodilians, extinct forms have been inter-
preted as having an exclusively carnivorous diet. However, Chimaerasu-
chus paradoxus, a recently described crocodyliform from the Early
Cretaceous of China, has molariform maxillary teeth that are strikingly
similar to those of tritylodontid synapsids (Wu, Sues, and Sun 1995). The
molariform tooth crowns possess several rows of low, curved cusps and
are mesiodistally and labiolingually expanded. Chimaerasuchus also has
procumbent premaxillary teeth. This combination of dental characters
led the authors to conclude that ‘the structure of the maxillary molari-
forms is consistent with a specialised diet including fibrous material and
may indicate at least facultative herbivory in Chimaerasuchus. The pro-
cumbent premaxillary . . . teeth were presumably involved in seizing food
items’ (Wu et al. 1995). This combination of dental features seems ideally
suited to omnivory. The mandible of Chimaerasuchus was capable of prop-
alinal movements, which may have been involved in processing food (Wu
et al. 1995). However, the lack of wear on the teeth indicates that food pro-
cessing would not have been intensive. Propalinal jaw actions are gener-
ally found in herbivores and are associated with the grinding or shearing
of plant material (King 1996). However, this does not preclude the devel-
opment of propaliny in an omnivore, which would take at least some veg-
etation in its diet.
Another unusual crocodilyform, Malawisuchus from the Early Creta-
ceous of Malawi, also has an unusual, mammal-like dentition and
appears to have been capable of propalinal jaw movements (Clark, Jacobs,
and Downs 1989; Gomani 1997). The anterior teeth of this form are conical
and recurved whereas the posterior maxillary and dentary teeth each have
a single large central cusp and a cuspidate cingulum. It has been proposed
that the molariform posterior maxillary and dentary teeth were
employed in the processing of prey items (using proal motion of the man-
dible to produce shearing between the upper and lower dentitions) and
that the anterior teeth (‘caniniforms’) were used for prey capture (Clark et
al. 1989; Gomani 1997). King (1996) argued that Malawisuchus may have
been herbivorous due to the possession of molariform teeth and the
ability to move the jaw fore-and-aft. However, although the teeth of Mal-
awisuchus are superficially mammal-like, they bear a stronger resem-
blance to the carnassial teeth of carnivorous mammals than to the
grinding teeth of herbivores. Furthermore, although proal jaw actions
are more common in herbivores, there are some instances of propaliny in
animals such as Sphenodon (Robinson 1976), which have a diet of insects
and other small prey. Therefore, fore-and-aft jaw motion does not pre-
clude a carnivorous diet. Indeed, in some cases it may be advantageous to
allow some translational movements of the mandible at the jaw joint in
carnivorous animals in order to get a firmer grip on prey (by using a
68 paul m. barrett
palinal movement to dig the teeth further into the prey), or to allow
shearing between the upper and lower dentitions (to cut through flesh
and bone). It seems likely that Malawisuchus was a specialized predator, as
originally inferred by Clark et al. (1989), rather than a herbivore.
The fossilized gut contents of the Late Permian archosauromorph
reptile Protorosaurus provide a further indication that tooth form does not
always reflect diet (Munk and Sues 1993). The teeth of Protorosaurus are
recurved and conical (i.e., ‘carnivorous’), but the abdominal cavity in two
specimens contains an in situ gastric mill and many ovules of the conifer
Ullmannia and pteridosperms. Although the ovules may have been
ingested accidentally, the combined presence of the ovules and a gastric
mill support the suggestion that Protorosaurus was an omnivore (Munk
and Sues 1993).
Evolution of dinosaurian herbivory
All postulated dinosaurian outgroups (Pseudolagosuchus, Marasuchus,

Lagerpeton, Pterosauria, Scleromochlus, Crurotarsi; Bennett 1996; Gower and
Wilkinson 1996; Novas 1996) are carnivorous. This leads logically to the
assumption that the ‘ancestral’ dinosaur was a carnivore (implying reten-
tion of the primitive dinosauriform condition). Carnivory was supposedly
retained in the Theropoda, whereas herbivory appeared twice in the evolu-
tionary history of the Dinosauria — once at the base of Ornithischia and
once at the base of Sauropodomorpha (e.g., Sereno 1997). This view is sup-
ported by the observation that both Herrerasaurus and Staurikosaurus possess
a classical carnivore dentition and other cranial features (e.g., intramandib-
ular joint, Sereno and Novas 1994) indicative of carnivory. These taxa were
regarded as ‘basal dinosaurs’ that represented successive outgroups to a
monophyletic clade of Saurischia plus Ornithischia (Brinkman and Sues
1987; Benton 1990; Novas 1992). Precladistic views of dinosaurian evolution
present a similar picture of dietary evolution, with ornithischians, sauro-
podomorphs and theropods independently derived from different carniv-
orous ‘thecodontian’ taxa (e.g., Charig et al. 1965; Charig 1979).
In an attempt to test the assumption that carnivory represents the
ancestral dinosaurian condition, inferred dinosaurian diets (carnivory,
omnivory, herbivory) were mapped on to a cladogram of dinosaurian
interrelationships using MacClade 3.01 (Maddison and Maddison 1992;
Figure 3.8). The cladogram used for this purpose is based on the analysis
Figure 3.8. Outline phylogeny of the Dinosauria (sources: Gauthier 1986; Sereno
1986, 1989; Holtz 1994) with distribution of inferred dietary habits mapped on
the cladogram (see text for discussion).
70 paul m. barrett
by Gauthier (1986), with ornithischian relationships derived from Sereno

(1986). The ingroup relationships of theropods are based on the recent
study by Holtz (1994), and prosauropods are regarded as the sister-group
of sauropods (see references mentioned in the Introduction of this
chapter). Dietary information on the various dinosaurian clades has been
taken from the relevant sections in Weishampel, Dodson, and Osmólska
(1990), and has been further modified in accordance with the suggestions
presented in the sections above.
If sauropodomorphs and ornithischians are considered primitively
herbivorous, the diet of stem-saurischians and stem-dinosaurs is ambigu-
ous. Given the topology of recent cladograms for dinosaurian interrela-
tionships (Gauthier 1986; Benton 1990; Novas 1996; Figure 3.8), the
assumption that carnivory was primitive for Dinosauria seems surpris-
ing, as does the assumption that herbivory appeared twice within Dino-
sauria. This scenario requires two evolutionary steps (independent
acquisition of herbivory in Ornithischia and Sauropodomorpha).
However, an alternative scenario, with herbivory as the primitive dino-
saurian condition, is equally parsimonious (requiring the acquisition of
herbivory at the base of Dinosauria, and a reversal to carnivory at the base
of Theropoda). Moreover, reassessment of the phylogenetic position of
Herrerasaurus and Staurikosaurus now appears to place these taxa within
Theropoda (Sereno et al. 1993; Novas 1994), thereby removing them from
their position basal to all Dinosauria, and simultaneously decreasing
their utility in the identification of the ancestral dinosaurian diet. Recod-
ing prosauropods and various ornithischian taxa as omnivorous (see dis-
cussion above and Figure 3.8) renders ambiguous the determination of
diet in the sauropodomorph stem-lineage and various lineages within
Ornithischia, and the primitive dinosaurian diet remains unresolved.
This exercise demonstrates that the ancestral dinosaurian diet cannot
be resolved by consideration of the diets of the terminal taxa within Dino-
sauria or by outgroup comparison alone. More information on the diet of
basal saurischians and basal dinosaurs is needed before the primitive
dinosaurian diet can be confidently inferred. In this case, parsimony
cannot distinguish between the competing hypotheses of primitive dino-
saurian omnivory, herbivory, or carnivory, suggesting that it may not be
wise to rely solely on parsimony-based arguments when attempting to
reconstruct the evolution of biological systems, especially if all of the taxa
under consideration are now extinct. Other phylogenetic approaches,
such as the Extant Phylogenetic Bracket method (Witmer 1995; Weisham-
pel 1995), are useful in deducing features of extinct organisms, but the
latter approach relies on the use of extant organisms as outgroups, a
luxury that is denied to groups where most of the ingroup members, as
well as the immediate outgroups, are extinct.
Although parsimony indicates that, primitively, herbivory, carnivory
or omnivory were equally probable, other factors indicate that the
hypothesis of primitive herbivory can be rejected. Obligate herbivory
requires a substantial number of cranial, dental, and physiological mod-
ifications (see Sues and Reisz [1998] for a recent review), and it is consid-
ered biologically improbable that a committed herbivore would revert to
carnivory. By contrast, the suggestion that carnivory or omnivory was the
ancestral dinosaurian condition is supported by the presence of ‘carnivo-
rous’ features (e.g., recurved, serrated premaxillary teeth) and the lack of
derived ‘herbivorous’ features (e.g., a ventrally displaced jaw joint) in the
basal members of all major dinosaurian clades. This observation adds
support to the suggestion that many of the adaptations commonly
regarded as characteristic of herbivorous feeding complexes may have
originated in the presence of continued carnivorous or omnivorous
habits.
The distribution of diets on the cladogram (Figure 3.8) indicates that
obligate herbivory may have had several independent origins within
Dinosauria, rather than two as usually supposed. For example, herbivory
may have evolved at least three times within Ornithischia. Similarly,
omnivory may either have had multiple origins within Dinosauria or may
have been primitive for the group as a whole. In either case, omnivory
appears to be more widespread than previously acknowledged.
Conclusions
Perhaps tooth form, at least in the case of reptilian herbivores, is not the
reliable predictor of diet that it was once thought to be. The examples
listed above reflect the difficulty of relating tooth form to diet in many
taxa. Omnivory is not recognized, or even suggested, in many extinct taxa
due to the tendency of most authors to place animals at either end of the
carnivory/herbivory spectrum. This is due, in part, to the application of an
unsatisfactory paradigm for herbivory (iguanine tooth structure). Another
reason for the reluctance of authors to infer omnivory is that dental corre-
lates specific to omnivory have never been identified. Perhaps the combi-
nation of recurved anterior teeth and more ‘herbivore-like’ posterior teeth
72 paul m. barrett
could represent such a correlate, at least in the major taxa (Dinosauria,

Lepidosauria, Crocodylomorpha) mentioned above. The anterior teeth
would serve in the procurement and dismemberment of prey items or
carrion, or the procurement of plant material, whereas the posterior teeth
would have been suited to the shearing of both plant and animal fodder.
Members of these taxa that are closer to the herbivorous end of the dietary
spectrum may be expected to lose their premaxillary teeth, or to modify
them to aid in the prehension or shearing of plant material. Other features
of the dentition in these taxa may display more extensive modifications
for herbivory, such as the possession of dental batteries, the presence of
extensive tooth wear (associated with modifications of jaw motion), and
the development of molariform teeth.
Some authors have argued that it may be unrealistic to compare pro-
sauropods, or for that matter other taxa, with iguanine lizards, due to dif-
ferences in body size and remote phylogenetic relationship (Gow et al.
1990). However, the absence of other extant analogues almost forces us to
make such comparisons. This approach does at least allow some empiri-
cally based paleobiological inferences to be made. Also, the detailed simi-
larities between the dentitions of iguanine lizards and the other taxa
discussed above suggest that these comparisons can be meaningful.
If many of the ‘reptilian’ taxa currently thought to be carnivorous or
herbivorous could be identified as omnivores, this would have important
implications for those working on the trophic dynamics of late Paleozoic
and Mesozoic terrestrial ecosystems.
Acknowledgments
This manuscript benefited from many discussions with Drs P. Upchurch

and D. B. Norman. I would also like to thank the following for discussion
on the various issues raised during the preparation of this chapter: Drs A.
B. Arcucci, E. N. Arnold, S. E. Evans, D. J. Gower, Z. Luo, H.-D. Sues, M. A.
Taylor and D. M. Unwin, and S. Finney and J. Pearson. Drs D. B. Weisham-
pel, R. R. Reisz, P. M. Galton, D. B. Norman, P. Upchurch, A. E. Friday, H.-
D. Sues and M. J. Benton provided helpful reviews of the manuscript. Dr
A. C. Milner and S. Chapman (Natural History Museum, London), R.
Symonds (University Museum of Zoology, Cambridge), Professor Dong
Zhiming (Institute of Vertebrate Paleontology and Paleoanthropology,
Beijing), W. Simpson (Field Museum of Natural History, Chicago), C. R.
Schaff (Museum of Comparative Zoology, Harvard) and Drs B. Rubidge
and C. Gow (Bernard Price Institute for Palaeontological Research, Johan-

nesburg) granted access to specimens in their care. Many thanks to the
Lethaia Foundation and Dr P. M. Galton for permission to reproduce the
illustrations used in Figure 3.1b,c, and to Dr P. J. Currie and the Society of
Vertebrate Paleontology for permission to use the drawing in Figure 3.2c.
Figure 3.8 was created by Dr P. Upchurch. Sincere thanks to Dudley
Simons for photographic assistance. The Dinosaur Society and Cam-
bridge Philosophical Society provided funds for visits to the People’s
Republic of China, and Trinity College (Cambridge) supported a research
trip to the United States. This work was initiated while in receipt of a
NERC studentship (GT4/93/128/G) and continued as a Research Fellow of
Trinity College, Cambridge.
References
Barrett, P. M. (1998). Herbivory in the Non-Avian Dinosauria. Unpublished Ph.D.

dissertation, University of Cambridge.
Barrett, P. M., and Upchurch, P. (1995). Sauropod feeding mechanisms: their bearing on
palaeoecology. In 6th Symposium on Mesozoic Terrestrial Ecosystems and Biotas, Short
Papers, ed. A. Sun, and Y. Wang, pp. 107–110. Beijing: China Ocean Press.
Barsbold, R. (1983). [Carnivorous dinosaurs from the Cretaceous of Mongolia]. Sovm.
Sov.- Mong. Paleontol. Eksped. Trudy 19:1–117 (in Russian with English abstract).
Barsbold, R., Maryanska, T., and Osmólska, H. (1990). Oviraptorosauria. In The
Dinosauria, ed. D. B. Weishampel, P. Dodson, and H. Osmólska, pp. 249–258.
Berkeley: University of California Press.
Barsbold, R., and Osmólska, H. (1990). Ornithomimosauria. In The Dinosauria, ed. D. B.
Weishampel, P. Dodson, and H. Osmólska, pp. 225–244. Berkeley: University of
California Press.
Bennett, S. C. (1996). The phylogenetic position of the Pterosauria within the
Archosauromorpha. Zool. J. Linn. Soc. Lond. 118:261–309.
Benton, M. J. (1990). Origin and interrelationships of dinosaurs. In The Dinosauria, ed. D.
B. Weishampel, P. Dodson, and H. Osmólska, pp. 11–30. Berkeley: University of
California Press.
Bonaparte, J. F. (1972). Los tetrapodos del sector superior de la Formacion Los Colorados,
La Rioja, Argentina (Triásico Superior). Opera Lilloana 22:1–183.
Brinkman, D. B., and Sues, H.-D. (1987). A staurikosaurid dinosaur from the Upper
Triassic Ischigualasto Formation of Argentina and the relationships of the
Staurikosauridae. Palaeontology 30:493–503.
Calvo, J. O. (1994). Jaw mechanics in sauropod dinosaurs. GAIA 10:183–194.
Charig, A. (1979). A New Look at the Dinosaurs. London: British Museum (Natural History).
Charig, A. J., Attridge, J., and Crompton, A. W. (1965). On the origin of the sauropods
and the classification of the Saurischia. Proc. Linn. Soc. Lond. 176:197–221.
Chinsamy, A. (1993). Bone histology and growth trajectory of the prosauropod dinosaur
Massospondylus carinatus Owen. Modern Geol. 18:319–329.
Clark, J. M., and Fastovsky, D. E. (1986). Vertebrate biostratigraphy of the Glen Canyon
74 paul m. barrett
Group in northern Arizona. In The Beginning of the Age of Dinosaurs, ed. K. Padian,
pp. 285–301. Cambridge and New York: Cambridge University Press.
Clark, J. M., Jacobs, L. L., and Downs, W. R. (1989). Mammal-like dentition in a Mesozoic
crocodylian. Science 244:1064–1066.
Clark, J. M., Perle, A., and Norell, M. A. (1994). The skull of Erlikosaurus andrewsi, a Late
Cretaceous ‘segnosaur’ (Theropoda: Therizinosauridae) from Mongolia. Am.
Mus. Novitates 3115:1–39.
Colbert, E. H. (1962). Dinosaurs. Their Discovery and Their World. New York: E. P. Dutton &
Co., Inc.
Colbert, E. H. (1981). A primitive ornithischian dinosaur from the Kayenta Formation of
Arizona. Bull. Mus. N. Arizona 53:1–61.
Cooper, M. R. (1981). The prosauropod dinosaur Massospondylus carinatus Owen from
Zimbabwe: its biology, mode of life and phylogenetic significance. Occas. Pap.
Natl. Mus. Rhodesia, B, Nat. Sci. 6:689–840.
Cracraft, J. (1971). Caenagnathiformes: Cretaceous birds convergent in jaw mechanism
to dicynodont reptiles. J. Paleont. 45:805–809.
Crompton, A. W., and Attridge, J. (1986). Masticatory apparatus of the larger herbivores
during Late Triassic and Early Jurassic times. In The Beginning of the Age of
Dinosaurs, ed. K. Padian, pp. 223–236. Cambridge and New York: Cambridge
University Press.
Crompton, A. W., and Charig, A. J. (1962). A new ornithischian from the Upper Triassic
of South Africa. Nature 196:1074–1077.
Cuny, G., and Ramboer, G. (1991). Nouvelles donnees sur la faune et l’age de Saint-
Nicolas-de-Port. Rev. Paléobiol. 10:69–78.
Currie, P. J. (1987). Bird-like characteristics of the jaws and teeth of troodontid
theropods (Dinosauria, Saurischia). J. Vert. Paleont. 7:72–81.
Dong, Z. (1992). The Dinosaurian Faunas of China. Berlin: Springer-Verlag.
Dong, Z., Zhou, S., and Zhang, Y. (1983). [Dinosaur remains of Sichuan Basin.]
Palaeontol. Sin., n. s., C 23:1–145 (in Chinese with English abstract).
Frey, F. L. (1986). Feeding and nutritional diseases of reptiles. In Zoo and Wild Animal
Medicine, 2nd Edition, ed. M. E. Fowler, pp. 139–151. Philadelphia: W. B.
Saunders and Company.
Furness, R. W. (1988a). Predation on ground nesting seabirds by populations of red deer
Cervus elephas and sheep Ovis. J. Zool. 216:565–573.
Furness, R. W. (1988b). The predation of tern chicks by sheep. Bird Study 35:199–202.
Furness, R. W. (1989). Not by grass alone. Natural History 1989(12):8–12.
Galton, P. M. (1973). The cheeks of ornithischian dinosaurs. Lethaia 6:67–89.
Galton, P. M. (1974). The ornithischian dinosaur Hypsilophodon from the Wealden of the
Isle of Wight. Bull. Brit. Mus. (Nat. Hist.), Geol. 25:1–152.
Galton, P. M. (1976). Prosauropod dinosaurs of North America. Postilla 169:1–98.
Galton, P. M. (1978). Fabrosauridae, the basal family of ornithischian dinosaurs
(Reptilia: Ornithopoda). Paläont. Z. 52:138–159.
Galton, P. M. (1984a). Cranial anatomy of the prosauropod dinosaur Plateosaurus from
the Knollenmergel (Middle Keuper, Upper Triassic) of Germany. I. Two
complete skulls from Trossingen/Württ. with notes on the diet. Geol. Palaeontol.
18:139–171.
Galton, P. M. (1984b). An early prosauropod dinosaur from the Upper Triassic of

Nordwürttemberg, West Germany. Stuttgarter Beitr. Naturk., Ser. B 106:1–25.
Galton, P. M. (1985a). Diet of prosauropods from the Late Triassic and Early Jurassic.
Lethaia 18:105–123.
Galton, P. M. (1985b). The poposaurid thecodontian Teratosaurus suevicus V. MEYER, plus
referred specimens mostly based on prosauropod dinosaurs, from the Middle
Stubensandstein (Upper Triassic) of Nordwürttemberg. Stuttgarter Beitr. Naturk.,
Ser. B 116:1–29.
Galton, P. M. (1986). Herbivorous adaptations of Late Triassic and Early Jurassic
dinosaurs. In The Beginning of the Age of Dinosaurs, ed.K. Padian, pp. 203–221.
Galton, P. M. (1990). Basal Sauropodomorpha–Prosauropoda. In The Dinosauria, ed. D. B.
California Press.
Gauffre, F.-X. (1995). Phylogeny of prosauropod dinosaurs. J. Vert. Paleont. 15(suppl. to
3):31A.
Gauthier, J. A. (1986). Saurischian monophyly and the origin of birds. Mem. Calif. Acad.
Sci. 8:1–55.
Gomani, E. M. (1997). A crocodyliform from the Early Cretaceous Dinosaur Beds,
northern Malawi. J. Vert. Paleont. 17:280–294.
Gow, C. E., Kitching, J. W., and Raath, M. A. (1990). Skulls of the prosauropod dinosaur
Massospondylus carinatus Owen in the collections of the Bernard Price Institute for
Palaeontological Research. Palaeontol. Afr. 27:45–58.
Gower, D. J., and Wilkinson, M. (1996). Is there any consensus on basal archosaur
phylogeny? Proc. R. Soc. Lond. B 263:1399–1406.
Harshbarger, J. W., Repenning, C. A., and Irwin, J. H. (1957). Stratigraphy of the
uppermost Triassic and the Jurassic rocks of the Navajo County. Prof. Pap. U.S.
Geol. Surv. 291:1–74.
Haughton, S. H. (1924). The fauna and stratigraphy of the Stormberg Series. Ann. S. Afr.
Mus. 12:323–497.
Holtz, T. H., Jr. (1994). The phylogenetic position of the Tyrannosauridae: implications
for theropod systematics. J. Paleont. 68:1100–1117.
Holtz, T. H., Jr., and Chandler, C. L. (1994). Denticle morphometrics and a possibly
different life habit for the theropod dinosaur Troodon. J. Vert. Paleont. 14 (suppl. to
3):30A.
Hotton, N. III. (1955). A survey of the adaptive relationships of dentition to diet in the
North American Iguanidae. Am. Mid. Nat. 53:88–114.
Huene, F. von. (1932). Die fossile Reptil-Ordnung Saurischia, ihre Entwicklung und
Geschichte. Monogr. Geol. Paläontol. 1, 4:1–361.
Jarman, P. J. (1974). The social organisation of antelope in relation to their ecology.
Behaviour 58:215–267.
Jenkins, F. A., Jr., Shubin, N. H., Amaral, W. W., Gatesy, S. M., Schaff, C. R.,
Clemmensen, L. B., Downs, W. R., Davidson, A. R., Bonde, N., and Osbæck, F.
(1994). Late Triassic continental vertebrates and depositional environments of
the Fleming Fjord Formation, Jameson Land, East Greenland. Medd. Grønland,
Geosci. 32:1–25.
Kermack, D. (1984). New prosauropod material from South Wales. Zool. J. Linn. Soc. Lond.
82:101–117.
76 paul m. barrett
King, G. M. (1996). Reptiles and Herbivory. London: Chapman and Hall.

Kitching, J. W., and Raath, M. A. (1984). Fossils from the Elliot and Clarens formations
(Karoo Sequence) of the north-eastern Cape, Orange Free State and Lesotho, and
a suggested biozonation based on tetrapods. Palaeontol. Afr. 25:111–125.
Lauder, G. V. (1995). On the inference of function from structure. In Functional
Morphology in Vertebrate Paleontology, ed. J. J. Thomason, pp. 1–18. Cambridge and
New York: Cambridge University Press.
Lull, R. S. (1953). Triassic life of the Connecticut Valley. Revised. Bull. Conn. State Geol. Nat.
Hist. Surv. 81:1–331.
Luo, Z., and Wu, X-c. (1994). The small tetrapods of the Lower Lufeng Formation,
Yunnan, China. In In the Shadow of the Dinosaurs, ed. N. C. Fraser and H.-D. Sues,
Luo, Z., and Wu, X-c. (1995). Correlation of vertebrate assemblages of the Lower Lufeng
Formation, Yunnan, China. In Sixth Symposium on Mesozoic Terrestrial Ecosystems
and Biotas, Short Papers, ed. A. Sun and Y. Wang, pp. 83–88. Beijing: China Ocean
Press.
Maddison, W. P., and Maddison, D. R. (1992). MacClade, version 3.01. Sunderland, MA:
Sinauer.
Mantell, G. A. (1825). Notice on the Iguanodon, a newly discovered fossil reptile, from the
sandstone of Tilgate Forest, in Sussex. Phil. Trans. R. Soc. Lond. 115:176–186.
McIntosh, J. S. (1990). Sauropoda. In The Dinosauria, ed. D. B. Weishampel, P. Dodson,
and H. Osmólska, pp. 345–401. Berkeley: University of California Press.
Molnar, R. E. (1977). Analogies in the evolution of combat and display structures in
ornithopods and ungulates. Evol. Theory 3:165–190.
Montanucci, R. R. (1968). Comparative dentition in four iguanid lizards. Herpetologica
24:305–315.
Munk, W., and Sues, H.-D. (1993). Gut contents of Parasaurus (Pareiasauria) and
Protorosaurus (Archosauromorpha) from the Kupferschiefer (Upper Permian) of
Hessen, Germany. Paläontol. Z. 67:169–176.
Norman, D. B., and Weishampel, D. B. (1985). Ornithopod feeding mechanisms: their
bearing on the evolution of herbivory. Am. Nat. 126:151–164.
Novas, F. E. (1992). Phylogenetic relationships of the basal dinosaurs, the
Herrerasauridae. Palaeontology 35:51–62.
Novas, F. E. (1994). New information on the systematics and postcranial skeleton of
Herrerasaurus ischigualastensis (Theropoda: Herrerasauridae) from the
Ischigualasto Formation (Upper Triassic) of Argentina. J. Vert. Paleont.
13:400–423.
Novas, F. E. (1996). Dinosaur monophyly. J. Vert. Paleont. 16:723–741.
Olsen, P. E., and Galton, P. M. (1984). A review of the reptile and amphibian assemblages
from the Stormberg of southern Africa, with special emphasis on the footprints
and the age of the Stormberg. Palaeontol. Afr. 25:87–110.
Osmólska, H., and Barsbold, R. (1990). Troodontidae. In The Dinosauria, ed. D. B.
California Press.
Paul, G. S. (1984). The segnosaurian dinosaurs: relics of the prosauropod–ornithischian
transition? J. Vert. Paleont. 4:507–515.
Paul, G. S. (1988). Predatory Dinosaurs of the World: A Complete Illustrated Guide. New York:
Simon and Schuster.
Pérez-Moreno, B. P., Sanz, J. L., Buscalioni, A. D., Moratalla, J. J., Ortega, F., and Rasskin-
Gutman, D. (1994). A unique multi-toothed ornithomimosaur dinosaur from
the Lower Cretaceous of Spain. Nature 370:363–367.
Perle, A. (1979). [Segnosauridae – a new family of theropods from the Late Cretaceous of
Mongolia.] Sovm. Sov.- Mong. Paleontol. Eksped. Trudy 8:45–55 (in Russian).
Purcell, S. W., and Bellwood, D. R. (1993). A functional analysis of food procurement in
two surgeonfish species, Acanthurus nigrofuscus and Ctenochaetus striatus
(Acanthuridae). Environ. Biol. Fishes 37:139–159.
Queiroz, K. de (1987). Phylogenetic systematics of iguanine lizards. Univ. Calif. Publ. Zool.
118:1–203.
Robinson, P. L. (1957). The Mesozoic fissures of the Bristol Channel area and their
vertebrate faunas. Zool. J. Linn. Soc. Lond. 43:260–283.
Robinson, P. L. (1976). How Sphenodon and Uromastyx grow their teeth and use them. In
Morphology and Biology of Reptiles, ed. A. d’A. Bellairs and C. B. Cox, pp. 43–64.
London: Academic Press.
Romer, A. S. (1956). Osteology of the Reptiles. Chicago: University of Chicago Press.
Romer, A. S. (1966). Vertebrate Paleontology. 3rd Edition. Chicago: University of Chicago
Press.
Rozhdestvensky, A. K. (1965). [Growth changes and some problems of systematics of
Asian dinosaurs.] Paleontol. Zh. 1965:95–109 (in Russian).
Russell, D. A., and Dong, Z. (1994). The affinities of a new theropod from the Alxa
Desert, Inner Mongolia, People’s Republic of China. Can. J. Earth Sci.
30:2107–2127.
Sander, P. M. (1992). The Norian Plateosaurus bonebeds of central Europe and their
taphonomy. Palaeogeogr. Palaeoclimatol. Palaeoecol. 93:255–299.
Sereno, P. C. (1986). Phylogeny of the bird-hipped dinosaurs (Order Ornithischia). Nat.
Geog. Res. 2:234–256.
Sereno, P. C. (1989). Prosauropod monophyly and basal sauropodomorph phylogeny. J.
Vert. Paleont. 9(suppl. to 3):38A.
Sereno, P. C. (1991). Lesothosaurus, ‘fabrosaurids’, and the early evolution of Ornithischia.
J. Vert. Paleont. 11:168–197.
Sereno, P. C. (1997). The origin and evolution of dinosaurs. Ann. Rev. Earth Planet. Sci.
25:435–489.
Sereno, P. C., Forster, C. A., Rogers, R. R., and Monetta, A. M. (1993.) Primitive dinosaur
skeleton from Argentina and the early evolution of Dinosauria. Nature 361:64–66.
Sereno, P. C. and Novas, F. E. (1994). The skull and neck of the basal theropod
Herrerasaurus ischigualastensis. J. Vert. Paleont. 13:451–476.
Shubin, N. H., Olsen, P. E., and Sues, H.-D. (1994). Early Jurassic small tetrapods from
the McCoy Brook Formation of Nova Scotia, Canada. In In the Shadow of the
Dinosaurs, ed. N. C. Fraser and H.-D. Sues, pp.242–250. Cambridge and New
York: Cambridge University Press.
Simmons, D. J. (1965). The non-therapsid reptiles of the Lufeng Basin, Yunnan, China.
Fieldiana, Geol. 15:1–93.
Smith, D. (1992). The type specimen of Oviraptor philoceratops, a theropod dinosaur from
the Upper Cretaceous of Mongolia. N. Jb. Geol. Paläontol., Abh. 186:365–388.
Smith, R. M. H., Eriksson, P. G., and Botha, W. J. (1993). A review of the stratigraphy and
sedimentary environments of the Karoo-aged basins of southern Africa. J. Afr.
Earth Sci. 16:143–169.
78 paul m. barrett
Smoot, J. P. (1991). Sedimentary facies and depositional environments of early Mesozoic

Newark Supergroup basins, eastern North America. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 84:369–423.
Sues, H.-D. (1997). On Chirostenotes, a Late Cretaceous oviraptorosaur (Dinosauria:
Theropoda) from western North America. J. Vert. Paleont. 17:698–716.
Sues, H.-D., and Reisz, R. R. (1998). Origins and early evolution of herbivory in
Swinton, W. E. (1934). The Dinosaurs. A Short History of a Great Group of Extinct Reptiles.
London: Thomas Murby and Company.
iguana (Iguanidae) and Uromastix aegyptius (Agamidae). J. Morphol. 148:363–390.
Thulborn, R. A. (1970). The skull of Fabrosaurus australis, a Triassic ornithischian
dinosaur. Palaeontology 13:414–432.
Thulborn, R. A. (1971). Tooth wear and jaw action in the Triassic ornithischian dinosaur
Fabrosaurus. J. Zool., Lond. 164:165–179.
Upchurch, P. (1993). The Anatomy, Phylogeny, and Systematics of the Sauropod Dinosaurs.
Unpublished Ph.D. Dissertation, University of Cambridge.
Upchurch, P. (1995). The evolutionary history of the sauropod dinosaurs. Phil. Trans. R.
Soc. Lond., B 349:365–390.
Upchurch, P. (1998). The phylogenetic relationships of sauropod dinosaurs. Zool. J. Linn.
Soc. Lond. 124:43–103.
Wallach, J. D., and Boever, W. J. (1983). Diseases of Exotic Animals. Philadelphia: W. B.
Saunders and Company.
Weishampel, D. B. (1984). Evolution of jaw mechanisms in ornithopod dinosaurs. Adv.
Anat. Embryol. Cell Biol. 87:1–110.
Weishampel, D. B. (1990). Dinosaurian distribution. In The Dinosauria, ed. D. B.
California Press.
Weishampel, D. B. (1995). Fossils, function, and phylogeny. In Functional Morphology in
Vertebrate Paleontology, ed. J. J. Thomason, pp. 34–54. Cambridge and New York:
Cambridge University Press.
Weishampel, D. B., Dodson, P., and Osmólska, H. (eds.) (1990). The Dinosauria. Berkeley:
University of California Press.
Winkler, D. A., Jacobs, L. L., Congleton, J. D., and Downs, W. R. (1991). Life in a sand sea:
biota from Jurassic interdunes. Geology 19:889–892.
Witmer, L. M. (1995). The extant phylogenetic bracket and the importance of
reconstructing soft tissues in fossils. In Functional Morphology in Vertebrate
Paleontology, ed. J. J. Thomason, pp. 19–33. Cambridge and New York: Cambridge
University Press.
Wu, X.-c., Sues, H.-D., and Sun, A. (1995). A plant-eating crocodyliform reptile from the
Lower Cretaceous of China. Nature 376:678–680.
Young, C. C.(1942). Yunnanosaurus huangi Young (gen. et sp. nov.), a new Prosauropoda
from the Red Beds at Lufeng, Yunnan. Bull. Geol. Soc. China 22:63–104.
Young, C. C. (1951). The Lufeng saurischian fauna. Palaeontol. Sin., n.s.,C 13:1–96.
Zhang, Y., and Yang, Z. (1994). [A New Complete Osteology of Prosauropoda in the Lufeng Basin,
Yunnan, China: Jingshanosaurus.] Kunming: Yunnan Publishing House of
Science and Technology (in Chinese with English abstract).
paul upchurch and paul m. barrett
The evolution of sauropod feeding mechanisms
Introduction
Sauropods were gigantic, long-necked, herbivorous dinosaurs,

which dominated many Jurassic and Cretaceous terrestrial faunas. First
appearing in the fossil record during the Early Jurassic, they rapidly
achieved a nearly global distribution, with apparent peaks in diversity
and abundance in the Kimmeridgian, Hauterivian–Barremian, and Cam-
panian–Maastrichtian (McIntosh 1990; Weishampel 1990; Hunt et al.
1994; Barrett 1998).
Early discoveries of poorly preserved sauropod material were inter-
preted as the remains of marine crocodiles (Owen 1841, 1842). Better pre-
served specimens (Cetiosaurus) from the Middle Jurassic of Oxfordshire,
however, allowed Phillips (1871:294) to recognize that these animals were
terrestrial dinosaurs which probably ate plants. This discovery was soon
overshadowed by spectacular material from the western United States
(Cope 1877a,b; Marsh 1877a,b, 1878, 1879), prompting Marsh (1878) to
create the new dinosaurian subgroup Sauropoda. The retracted nostrils
in the skull of Diplodocus (Marsh 1884), and the gigantic size of these
animals, led many workers to conclude that sauropods must have been
aquatic (Cope 1878; Marsh 1883; Hatcher 1901; Hay 1908). This ‘aquatic’
hypothesis had an important effect on interpretations of sauropod
feeding habits and mechanisms. If such gigantic animals were restricted
to food available in or near bodies of water, they probably ate soft aquatic
vegetation or invertebrates (Hay 1908; Holland 1924; Haas 1963). This
view was seemingly supported by the ‘weak and inefficient’ masticatory
apparatus of most sauropods, especially the ‘tooth-comb’ of Diplodocus,
which could have been used for straining items from the water.
[79]
80 paul upchurch and paul m. barrett
Recent studies have provided compelling evidence (from both skeletal

structure and trackways) that sauropods were largely terrestrial (Bakker
1971; Coombs 1975; Gillette and Lockley 1989; Dodson 1990). This reinter-
pretation has led to the suggestion that sauropods were high-browsers
that ate vegetation from tree canopies.
Previous studies recognized a twofold division of sauropod feeding
mechanisms: (1) nipping/cropping of coarse vegetation by sauropods
with heavy spatulate dentitions; (2) ‘raking’ of softer vegetation by sauro-
pods with more delicate dentitions (Bakker 1971; Norman 1985; Dodson
1990). Whereas these studies provided important insights into sauropod
paleoecology, more detailed work has indicated that sauropod feeding
was more varied and complex (Barrett and Upchurch 1994, 1995; Calvo
1994a).
Institutional abbreviations in this chapter are: AMNH, American
Museum of Natural History, New York; BMNH, The Natural History
Museum, London; CMNH, Carnegie Museum of Natural History,
Pittsburgh; FMNH, Field Museum, Chicago; CIT, Chengdu Institute of
Technology, Chengdu; HMN, Museum für Naturkunde der Humboldt-
Universität, Berlin; IVPP, Institute for Vertebrate Paleontology and Paleo-
anthropology, Academia Sinica, Beijing; LCM, Leicester City Museum
and Art Gallery, Leicester; MACN, Museo Argentino de Ciencias Natu-
rales, Buenos Aires; OUMZ, Oxford University Museum of Zoology,
Oxford; PMU, Palaeontological Museum, University of Uppsala; USNM,
National Museum of Natural History, Washington D.C.; ZDM, Zigong
Dinosaur Museum, Zigong.
Materials and methods
The sauropod specimens and sources of data that form the basis for this
study are summarized in Tables 4.1 and 4.2. In this analysis, we survey a
number of morphological features that reflect aspects of sauropod
feeding mechanisms. Tooth shape and position, tooth wear, structure of
the jaw joint, orientation of jaw muscles, neck length, and forelimb
length are reviewed for each of the major sauropod groups. In addition,
data on dental microwear were collected using a scanning electron micro-
scope (SEM). Neck flexibility was assessed on the basis of an examination
of intervertebral articulation and the suggestion that shortening of the
cervical ribs facilitates greater lateral flexibility. ‘Maximum browse
heights’ are based on the sum of neck and forelimb length; this provides
Table 4.1. Summary of feeding mechanism characters in sauropod crania
Taxon Specimen/source Dentition Macrowear Microwear Jaw joint Jaw muscle angle
‘Vulcanodontidae’
Barapasaurus Jain et al. 1975, 1979 Spatulate, denticulate None – ? ?
Kotasaurus Yadagiri 1988 Spatulate, denticulate None – ? ?
Euhelopodidae
Euhelopus PMU M2983; Mateer Spatulate Mesial and distal – ? Steep
and McIntosh 1985
Mamenchisaurus Russell and Zheng 1994 Spatulate, denticulate None – Short Steep
Omeisaurus IVPP unnumbered, Spatulate, denticulate Mesial and distal – Short? Steep
ZDM T5703; He et al.
1988
Shunosaurus IVPP 7261–5, ZDM Spatulate, denticulate Mesial and distal – Short Steep
T5401–3; Zhang 1988,
Zheng 1991
‘Cetiosauridae’
Patagosaurus MACN 124, MACN Spatulate Mesial and distal – ? Steep?
CH933, MACN CH934;
Bonaparte 1986
Camarasauridae
Camarasaurus CMNH 11338, 21751, Spatulate Mesial and distal Pits (in adults), Expanded Steep
AMNH 467; Fiorillo coarse transverse rostrocaudally
1991, 1996, Calvo 1994a, scratches
Madsen et al. 1995
Brachiosauridae
Brachiosaurus HMN SII, S66, S116, Cone-chisel/spatulate, Apical, – Short Steep
WJ470; Janensch some denticles subtriangular
1935–36
Titanosauroidea
Malawisaurus Jacobs et al. 1993 ‘Peg’-like None – ? Steep?
Unnamed, Brazil Kellner 1996 ‘Peg’-like Apical, chisel- - ? ?
shaped
Undescribed, FMNH Field No. ‘Peg’-like Apical, chisel- Pits, subparallel ? ?
Madagascar 93–73/95029, 93-06/ shaped scratches
95035, 93074
Diplodocidae
Apatosaurus CMNH 11162; Berman ‘Peg’-like ? – ? Shallow
and McIntosh 1978
Diplodocus AMNH 969, CMNH ‘Peg’-like Apical, present on Fine subparallel Long Shallow
11161, USNM 2672; labial surfaces only scratches,
McIntosh and Berman rostrocaudally
1975, Berman and oriented
McIntosh 1978
Dicraeosauridae
Dicraeosaurus HMN dd416, 429; ‘Peg’-like Apical, labial and – ? Shallow
Janensch 1935–36 lingual
Nemegtosauridae
Nemegtosaurus Nowinski 1971 ‘Peg’-like Apical, labial and – Slightly
lingual elongate Less steep
Note: Catalogue numbers are provided for those specimens examined by the authors.
Table 4.2. Summary of feeding mechanism characters in sauropod postcrania
Neck/trunk Cervical
Taxon Specimen/source ratio ribs Fl/Hl ratio Body length Browse height
Barapasaurus Jain et al. 1975, 1979 ? ? ? ⬃15 m ?
Kotasaurus Yadagiri 1988 ⬃1.4 ? 0.80 10.5 m ⬃3.5–4 m
Vulcanodon Cooper 1984 ? ? 0.78 ⬃10 m ?
Euhelopodidae
Euhelopus Wiman 1929 2.1 Long ? ? ?
Mamenchisaurus CIT holotype; Young and Chao 1972 3.5 Long ? 22 m 10–11 m
Omeisaurus ZDM T5701, T5703–5; He et al. 1988 3.8 Long 0.85–0.88 20 m 10–11 m
Shunosaurus ZDM T5401–2; Zhang 1988 1.35–1.45 Short 0.64 11 m 4.5 m
‘Cetiosauridae’
Cetiosaurus LCM 468.1968, OUMZ J13605–13690 ⬃1.20 Long? 0.84 ⬃15 m 5m
Camarasauridae
Camarasaurus CMNH 11338, 11939; Osborn and Mook 1921, 1.25 (juvenile), Long 0.79 (adult), 18 m ⬃8.5 m
Gilmore 1925, McIntosh et al. 1996 1.45 (adult) 0.87 (juvenile)
Brachiosauridae
Brachiosaurus HMN SII, XV2; Janensch 1950, 1961 2.38 Long? 1.1 25 m 14 m
Titanosauroidea
Opisthocoelicaudia Borsuk-Bialynicka 1977 ? ? 0.83 ? ?
Diplodocidae
Apatosaurus CMNH 563, 3018; Gilmore 1936 2.04–2.17 Short 0.70–0.71 23 m 8m
Barosaurus Lull 1919, McIntosh 1990 2.5–3? Short ? 24 m 12 m
Diplodocus CMNH 84/94, AMNH 223, 655, 5855, USNM 2.13 Short ⬃0.70 ⬃25 m 9–10 m
10865; Hatcher 1901, Gilmore 1932
Dicraeosauridae
Amargasaurus Salgado and Bonaparte 1991 1.36 ? 0.70 10 m 4m
Dicraeosaurus HMN O, M, m; Janensch 1929, 1936, 1961 1.24 Short 0.64 10 m 4m
Notes:
Catalogue numbers are provided for those specimens examined by the authors.
Abbreviations: Fl, forelimb length; Hl, hindlimb length.
Evolution of sauropod feeding 85
an estimate for the height of the head if the sauropod stood in a quadru-
pedal stance with its neck vertical.
The interrelationships of sauropods are based on the cladogram pro-
posed by Upchurch (1995), with the modification that the Titanosaur-
oidea are now considered the sister-group to the Brachiosauridae (Wilson
and Sereno 1994; Calvo and Salgado 1995; Salgado, Coria, and Calvo 1997;
Upchurch 1998). These relationships, and corresponding classification,
are summarized in Figure 4.1.
General comments on sauropod morphology
Reviews of sauropod anatomy can be found in McIntosh (1990), Upchurch

(1993, 1994), and Barrett (1998). Some particular features, common to all
sauropods, deserve special mention because of their relevance to feeding
mechanics.
Sauropod skulls were akinetic. Contacts between the various cranial
elements are usually interdigitating sutures or simpler butt/overlap
joints held together by ligaments. No transverse or longitudinal hinge
joints have been detected. Streptostyly was not possible because the quad-
rate had fixed contacts with the squamosal, quadratojugal, and pterygoid
(Figure 4.2; see also Haas 1963).
In basal Theropoda and Ornithischia, as well as all prosauropods, the
rostral end of the skull is narrow and acute in dorsal view. In sauropods,
the snout has become wide and rounded in outline. This modification
may be related to the cropping of vegetation (see below).
All sauropod skulls, including that of Shunosaurus, display some
degree of narial retraction (Upchurch 1994, 1995). Most sauropods have
laterally facing external nares placed anterodorsal to the antorbital fenes-
tra; in diplodocoids, however, the external nares face upwards and lie
anterodorsal to the orbit (Figure 4.2). Bakker (1986) suggested that narial
retraction reflects the presence of a fleshy proboscis, as occurs in elephants
and tapirs. Such a proboscis could have played a role in food-gathering in
sauropods. It seems likely, however, that sauropods lacked the facial mus-
culature that gives rise to the proboscis in certain mammals (Coombs
1975). Alternative explanations for narial retraction include the require-
ment to elongate the preorbital part of the skull (see below) and/or the
need to protect sensitive nasal structures during feeding. At present,
however, the significance of narial retraction in sauropods remains
obscure.
Figure 4.1. Evolutionary relationships and systematic classification of Sauropoda

based on Upchurch (1995, 1998) and Salgado et al. (1997). Nodes are labelled as
follows: A, Sauropoda; B, unnamed node; C, Eusauropoda; D, Euhelopodidae;
E, unnamed node; F, Neosauropoda; G, Diplodocoidea; H, Nemegtosauridae;
I, unnamed node; J, Dicraeosauridae; K, Diplodocidae; L,
Camarasauromorpha; M, Titanosauriformes; N, Titanosauroidea; O,
Titanosauridae. Paraphyletic taxa are labelled: p, ‘Vulcanodontidae;’ q,
‘Cetiosauridae.’
Figure 4.2. Sauropod skulls in lateral view. A, Euhelopus (after Mateer and McIntosh 1985); B, Camarasaurus (based on CMNH 11338); C, Brachiosaurus
(after Janensch 1935–36); D, Diplodocus (based on CMNH 11161); E, Dicraeosaurus (after Janensch 1935–36); F, Nemegtosaurus (after Nowinski 1971).
Abbreviations: a, antorbital fenestra; d, dentary; en, external naris; it , infratemporal fenestra; m, maxilla; o, orbit; pm, premaxilla; q,
quadrate; st, supratemporal fenestra. Scale bars each equal 50 mm.
The ‘lateral plate’ is a sheet of bone developed from the labial margins
of the premaxillae, maxillae, and dentaries (Upchurch 1995: fig. 12). This
structure covers approximately the basal third of each tooth crown labi-
ally. It is most prominent at the rostral end of the jaws, diminishes in
height posteriorly, and finally disappears close to the posteriormost
tooth. The lateral plate may have braced the teeth against anteriorly and
laterally directed forces generated during cropping or raking.
The absence of a buccal emargination, or ridges on the mandible and
upper jaw, indicates that sauropods lacked a muscular cheek. This places
a limit on the amount of oral food processing: material would have been
lost through the sides of the mouth every time foliage was chopped. The
absence of a fleshy cheek would allow a wider gape, enhancing the
amount of food collected in each mouthful. The presence of a robust
hyoid apparatus in many sauropods, including Brachiosaurus (Janensch
1935–36), Camarasaurus (Gilmore 1925), and Omeisaurus (Dong, Zhou, and
Zhang 1983), indicates that a powerful tongue was present. The tongue
could have been used to guide and manipulate food, reducing some of the
losses from the sides of the mouth.
The mandible of sauropods, unlike those of other dinosaurs, increases
in depth and robustness towards the symphysis (Figure 4.2). This feature
conveys additional strength to the tooth-bearing portion of the jaws
(Barrett and Upchurch 1994), and is probably related to the demands of
foliage cropping and raking.
The extremely elongate sauropod neck is one of this group’s most fun-
damental feeding adaptations. The neck is usually interpreted as having
played a role in high-browsing (Bakker 1971), but it may also have
increased the horizontal feeding range (Martin 1987). Examination of
intervertebral articulations suggests that, in the majority of sauropods,
greatest lateral and vertical flexibility was found in the region nearest the
skull. The middle portion of the neck was perhaps the least flexible, with
long overlapping cervical ribs greatly restricting the amount of lateral
motion. The cervical vertebrae near the base of the neck, which possess
dorsoventrally compressed centra and transversely broad zygapophyses
set far from the midline, appear to have allowed the neck to bend in a ver-
tical plane. There is, however, some debate concerning, first, the ability of
sauropods to raise their necks into a subvertical position and, second, if
neck raising was possible, whether this was habitually used as part of a
food-gathering strategy (Dodson 1990 and references therein). Alexander
(1985) discussed the mechanical problems involved in supporting and

raising the neck of Diplodocus, and concluded that the neck could have
been raised if the notch created by neural spine bifurcation was filled by
muscle or an elastin ligament, and if the area lateral to the neural spines
held muscles (as must have been the case in those sauropods that lack
neural spine bifurcation). Neck elevation would have been even more
likely if the mean density of its tissues had been lowered by air-filled sacs
occupying the cervical pleurocoels during life. Hemodynamic considera-
tions have also been used as an argument against sauropod neck raising
(Dodson 1990 and references therein). However, mechanisms for coping
with such problems, albeit on a smaller scale, are present in extant taxa
(Choy and Altman 1992). Therefore, given the nature of the cervical inter-
vertebral joints, and the apparent absence of convincing engineering and
physiological counter-arguments, we consider that neck elevation did
form part of the food-gathering strategy employed by many sauropods.
The extent to which a ‘gastric mill’ played a role in the sauropod
digestive system is controversial. Calvo (1994b) claimed that the associa-
tion between gastroliths and sauropod skeletal material was only reli-
able in the case of two specimens from Argentina (aff. Rebbachisaurus).
Christiansen (1996), however, cites the presence of gastroliths in close
association with Apatosaurus, Barosaurus, Camarasaurus alenquerensis,
Dicraeosaurus, Diplodocus and Seismosaurus. A gastric mill is also present in
the abdominal region of an undescribed brachiosaurid from the Isle of
Wight (S. Hutt, pers. comm.). Calvo (1994b) predicted that sauropods
that utilized little or no oral processing (such as brachiosaurids and
diplodocoids) would be more likely to need a gastric mill than those with
substantial oral processing (such as camarasaurids). The ‘known’ distri-
bution of gastroliths is consistent with Calvo’s prediction, since all of the
sauropods listed above are either brachiosaurids or diplodocoids, with
the possible exception of Camarasaurus alenquerensis. The taxonomic
status of the latter, from the Upper Jurassic of Portugal, is problematic,
and this taxon possesses at least one diplodocoid synapomorphy
(Upchurch 1995). Therefore, whereas microbial fermentation (facilitated
by a long digestive tract, slow passage rates, and presumably low mass-
specific basal metabolic rates) may have satisfied a sauropod’s energy
requirements (Farlow 1987), gastric mills were probably used to enable
mechanical breakdown of food in at least some of the forms that were not
capable of oral processing.
Sauropod feeding-related characters
The ‘Vulcanodontidae’ are a paraphyletic assemblage of basal sauropods
from the Early Jurassic (McIntosh 1990; Upchurch 1995). Vulcanodon,
Barapasaurus, and especially Kotasaurus are known from reasonably com-
plete postcrania, but cranial material is extremely rare.
Skull A pair of partial lower jaws from the Lower Lufeng Formation
(Hettangian–Pliensbachian: Luo and Wu 1995) of the Wuding Basin,
Yunnan, have been referred to Kunmingosaurus (Dong 1992). These speci-
mens possess a lateral plate and a ventrally expanded symphysis. Even if
the assignment of these mandibular remains to Kunmingosaurus is incor-
rect, they mark the occurrence of some ‘typical’ sauropod feeding adapta-
tions early in the Jurassic.
Dentition A tooth crown (IVPP, unnumbered) from the Lower

Lufeng Formation has been referred to Kunmingosaurus (Z. Dong, pers.
comm.). It is not clear whether this tooth was derived from the upper or
lower jaw, and it is thus impossible to distinguish the mesial and distal
margins of the crown. This specimen retains several plesiomorphic char-
acter-states, such as a labiolingually narrow base and only moderate
mesiodistal expansion of the crown. It also possesses derived states
present in other sauropods, including a shallowly concave lingual
surface, a median ridge within the lingual concavity, and a groove on the
labial surface which runs parallel to either the mesial or distal margin.
The apex of the tooth lacks any distal curvature, although it is lingually
inclined as in most sauropods. Both margins of the crown are coarsely ser-
rated, with up to nine denticles on one, and at least three on the other
(damaged) margin.
Some finely serrated recurved teeth were found in association with the
skeleton of Vulcanodon, but these are likely to have been shed by scaven-
gers (Cooper 1984). The tooth crowns of Barapasaurus and Kotasaurus more
closely resemble the typical spatulate type found in euhelopodids, cetio-
saurids and camarasaurids: the crowns have become more robust, with
labiolingually thickened bases; the lingual surface is strongly concave;
and, on the labial face, grooves extend parallel to the mesial and distal
margins (Jain et al. 1979; Yadagiri 1988). As in Kunmingosaurus, the tooth
crowns of Barapasaurus and Kotasaurus retain denticles.
Tooth wear No wear has been observed on any vulcanodontid tooth.
Jaw joint and musculature No information available at present.
Postcrania The neck of Kotasaurus consists of 12 or 13 cervical verte-

brae and was approximately 2.3 m long (Yadagiri 1988). The length of the
cervical ribs is not known in this form.
The forelimb/hindlimb ratios in Vulcanodon and Kotasaurus (Table 4.2)
should be treated with caution: several limb-bones of Vulcanodon are only
partially preserved, and Yadagiri (1988) used Vulcanodon as a model to esti-
mate limb proportions in Kotasaurus. Nevertheless, current data indicate
that vulcanodontids possessed long forelimbs compared with non-
sauropods (where the forelimb/hindlimb ratio is usually 0.6 or less).
Quadrupedality may be partially linked to this increase in forelimb
length, but other dinosaurian quadrupeds (stegosaurs, ankylosaurs, and
ceratopians) have forelimb/hindlimb ratios lower than those of sauro-
pods (Barrett 1998). Therefore, we interpret forelimb elongation, and the
concomitant elevation of the anterior part of the thorax, as a reflection of
the adaptation towards higher browsing (see below).
Feeding mechanism Compared with prosauropods and other ‘basal’

dinosaurs, vulcanodontids have increased the lengths of the neck and fore-
limbs, suggesting an early adaptation to browsing at a greater height above
ground. Relative to many more derived sauropods, however, vulcanodon-
tids were only small or medium-sized and possessed short necks; most early
sauropods probably browsed no higher than about 3–4 m (Table 4.2). Barap-
asaurus, with its greater body size (with a length of about 15 m), may have
been able to browse at higher levels. The little that is known about vulcano-
dontid skulls suggests that they fed by cropping vegetation. The absence of
tooth wear implies that mechanical breakdown of food did not occur in the
mouth, although more complete material may contradict this hypothesis.
Euhelopodidae
The Euhelopodidae represent a monophyletic group of sauropods
endemic to the Jurassic and Early Cretaceous of China (Figure 4.1).
Skull Well-preserved skull material is known for Euhelopus, Omei-

saurus and Shunosaurus, and information on the palate and mandible of
Mamenchisaurus is now available (Russell and Zheng 1994). These skulls
possess the typical sauropod features described above, and, in many
Figure 4.3. Sauropod teeth. A, Euhelopus (after Wiman 1929); B, Patagosaurus (after
Bonaparte 1986); C, Camarasaurus (after Osborn and Mook 1921); D, Brachiosaurus
(after Janensch 1935–36); E, undescribed titanosaurid; F, Diplodocus (based on
CMNH 11161); G, Dicraeosaurus (after Janensch 1935–36). All teeth are in labial
view except A and D, which are in lingual view. ‘W’ marks the position of wear
facets. Scale bars each equal 10 mm.
respects, are intermediate in shape between the short, high skulls of

camarasaurids and brachiosaurids and the long, low skulls of diplodo-
coids (Figure 4.2).
Dentition Most euhelopodids have spatulate teeth (Figure 4.3). In

Shunosaurus, Euhelopus, and Mamenchisaurus, the tooth crowns are
slender, with moderate mesiodistal expansion. In contrast, the teeth
referred to Omeisaurus are much more robust, with shorter crowns and
greater mesiodistal expansion. The denticles in Shunosaurus are small
and confined to the area immediately adjacent to the apex of the tooth
crown, whereas they are large and continue to a point about one-third
of the way down the crown in Omeisaurus. Denticles are absent in Euhelo-
pus.
Tooth wear The teeth of Euhelopus possess prominent, high-angle

wear facets on the mesial and distal margins, extending from the apex to
the point where the crown attains its greatest mesiodistal width (e.g.,
PMU M2983; Figure 4.3). These facets have a characteristic ‘step-flush’
appearance: a step is formed between the enamel and dentine at the base
of the wear facet, but this step decreases in height and finally disappears
completely close to the apex of the tooth crown. This type of wear is pro-
duced by tooth–tooth contact. As the opposing tooth rows interlock
during jaw closure, the margins of the opposing crowns shear past each
other. The areas immediately adjacent to the apices suffer the highest
amount of wear, as these are the first to be worn away following eruption
and continue to be worn throughout the functional life of the tooth. As a
result, the enamel–dentine interface is worn down to a flush contact close
to the apex. The step pattern, caused by differential wear of enamel and
dentine, persists on the ‘lower’ part of the wear facet because this area is
worn later in the functional life of the tooth. Similar wear is also present
on several teeth of Shunosaurus and Omeisaurus. Some teeth of Euhelopus
and Omeisaurus are very heavily worn, resulting in ‘shoulders’ on the
crown margins (Figure 4.3). Small terminal wear facets may be present on
the teeth of Shunosaurus (Zhang 1988).
It is difficult to assess the distribution of wear facets along the tooth
rows of euhelopodids. In the most complete skull of Shunosaurus (ZDM
T5403), the lingual surface of the upper tooth row is obscured by the man-
dible, and the entire lower tooth row is hidden by the upper teeth. The
most complete specimens of Euhelopus and Omeisaurus do not preserve
entire tooth rows. Wear is absent on all of the teeth of Mamenchisaurus
sinocanadorum (Russell and Zheng 1994).
No information on the microwear of euhelopodid teeth is available at
present.
Jaw joint and musculature The jaw joint of euhelopodids is short and
limited anteriorly and posteriorly by low transverse ridges on the articu-
lar. Some caution is necessary in the interpretation of the glenoid fossa in
Omeisaurus (ZDM T5703) because its shape may have been affected by
preparation. The jaw joint of Shunosaurus is almost level with the tooth
row, whereas those of Omeisaurus, Euhelopus and Mamenchisaurus are offset
ventrally.
The structure of the palate and temporal region suggests that the
adductor jaw musculature of euhelopodids was steeply inclined relative
to the long axis of the skull. This, combined with the structure of the jaw
joint and tooth wear, implies that jaw motion was strictly orthal.
Postcrania Euhelopodids can be divided into short-necked (Shuno-

saurus and Datousaurus) and long-necked forms (Euhelopus, Mamenchisau-
rus, and Omeisaurus). The former condition, in which there are 12 or 13
short cervicals, is the plesiomorphic character-state (Upchurch 1995). In
Shunosaurus, the cervical ribs are short, only extending posteriorly for
the distance of a single cervical vertebra, and do not come into close
contact with each other (PU, pers. obs.). This, combined with the shape
and horizontal orientation of the cervical zygapophyseal articular facets,
suggests that the neck was capable of more lateral flexibility than in most
other sauropods. In long-necked euhelopodids, neck elongation has been
achieved by lengthening individual cervicals, and by increasing the
number of cervical vertebrae from 13 to 17 (Upchurch 1994). The cervical
ribs are greatly elongated (often over 3 m long in Mamenchisaurus), with
the rib shafts forming overlapping ‘bundles’ (Young and Chao 1972; He,
Li, and Cai 1988; Russell and Zheng 1994).
Incomplete preservation of the known specimens means that fore-
limb/hindlimb ratios are not available for Euhelopus and Mamenchisaurus.
Shunosaurus has relatively short forelimbs, whereas those of Omeisaurus are
approximately 85% of hindlimb length (Table 4.2).
Feeding mechanism Euhelopodids utilized a simple orthal jaw

action. The presence of well-developed mesial and distal wear facets indi-
cates a powerful shearing action, with more oral processing than in
vulcanodontids and prosauropods. Small body size, a short neck, abbrevi-
ated cervical ribs, the nature of the cervical zygapophyseal articulations,
and the shortened forelimbs, all imply that Shunosaurus was a low-level
browser, obtaining food over a wide area by moving the neck laterally as
well as vertically. The long-necked euhelopodids were probably special-
ized for high-browsing, obtaining foliage at heights perhaps as great as 10
m to 11 m.
‘Cetiosauridae’
The taxa included within this family vary according to different authors
(compare McIntosh 1990 and Upchurch 1995). Here, the ‘Cetiosauridae’
form an assemblage of Middle and possibly Late Jurassic sauropods,
which is paraphyletic with respect to the Neosauropoda (Figure 4.1).
Skull No complete cetiosaurid skull has been described, although

well preserved cranial material is available for ‘Cetiosaurus mogrebiensis’
(McIntosh 1990). A few isolated teeth from the Middle Jurassic of England
have been referred to Cetiosaurus (e.g., BMNH R3377), but this referral is
dubious. Therefore, this study is based on the partial snout, dentary, and
isolated teeth of Patagosaurus (Bonaparte 1986). These elements suggest
that the cetiosaurid skull resembled those of euhelopodids and Camara-
saurus.
Dentition The teeth of Patagosaurus have mesiodistally expanded

spatulate crowns. Denticles are absent.
Tooth wear In Patagosaurus, there are well-developed mesial and

distal wear facets (MACN 124, MACN CH933, MACN CH934; Bonaparte
1986), resembling those found in euhelopodids (Figure 4.3). Heavily worn
teeth develop ‘shoulders.’
No work on the microwear of cetiosaurid teeth has been reported.
Jaw joint and musculature No information on the structure of the jaw

joint is currently available. Our limited knowledge of the cetiosaurid
skull suggests that jaw adductor muscles were steeply inclined, but this
requires confirmation by better preserved specimens.
Postcrania Cetiosaurus (LCM 468.1968) has 12 or 13 short cervicals,

resulting in a relatively short neck (Table 4.2). The cervical ribs of this
specimen are fragmentary, but probably had long overlapping shafts.
Material not incorporated into the mounted skeleton of LCM 468.1968
includes many long, slender portions of distal cervical rib shafts. Further-
more, the preserved proximal ends of the ribs possess a long anterior
process, which is typically associated with a long distal shaft in sauro-
pods. The structure of the neck in Patagosaurus is less well known, but
probably closely resembled that in Cetiosaurus.
The forelimb/hindlimb ratio in Cetiosaurus is approximately 0.84.
Feeding mechanism Our limited knowledge of cetiosaurid skulls

indicates a jaw action similar to that in euhelopodids. The moderate body
size and neck length, restricted lateral movement of the neck, and rela-
tively long forelimbs, suggest that cetiosaurids were high-browsers,
perhaps obtaining food 5–6 m above ground level.
Camarasauridae
Only Camarasaurus is known in sufficient detail to provide information on
the feeding adaptations of this group.
Skull The cranial structure of Camarasaurus is extremely well

known (Madsen, McIntosh, and Berman 1995). The skull is robust, some-
what taller (relative to its anteroposterior length) than those of euhelo-
podids (Figure 4.2).
Dentition The teeth of Camarasaurus possess mesiodistally

expanded spatulate crowns. Denticles are normally absent, even in
unworn or unerupted teeth. The apices of the crowns are slightly curved
distally and lingually.
Tooth wear Wear facets are present, but their shape and size depend
in part on the position of the tooth along the tooth row (Carey and
Madsen 1972). For example, wear is heaviest on more anteriorly placed
teeth, especially those in the upper jaw. Camarasaurus teeth frequently
display mesial and distal wear facets similar to, but usually larger than,
those found in euhelopodids and cetiosaurids (Figure 4.3). The formation
of ‘shoulders’ is also very common. Small terminal wear facets occur occa-
sionally.
Three studies have examined dental microwear in Camarasaurus (Fio-
rillo 1991, 1996; Calvo 1994a). Fiorillo (1991) described coarse scratches and
pitting on the wear facets of teeth in adult Camarasaurus, which he
ascribed either to low-level browsing (more grit is present in the lower
levels of tree canopies, and the amount of grit in the diet has a strong
influence on tooth microwear) or feeding on coarse vegetation. Fiorillo
(1996) examined the microwear on teeth belonging to a juvenile specimen
of Camarasaurus, and found fine scratches and no pits. This result indi-
cates that the size of microwear scratches and pitting is not height depen-
dent (at least in sauropods), since the juvenile Camarasaurus would have
been browsing at heights lower than those exploited by adults. These dif-
ferences in microwear might indicate ontogenetic variation in diet. Calvo
(1994a) observed pits and scratches, and noted that the majority of the
latter were oriented parallel to the labiolingual axis of the tooth.
Jaw joint and musculature The jaw articulation of Camarasaurus is

offset ventrally. The shallow concave articular surface is anteroposteriorly
expanded and slopes downwards anteromedially. This articular surface is
significantly longer anteroposteriorly than the distal condyle of the quad-
rate, indicating that propalinal motions were possible (Calvo 1994a;
Barrett and Upchurch 1995). White (1958) and Calvo (1994a) also noted the
lateromedial expansion of the articular surface, and argued that Camara-
saurus would have been capable of a transverse power stroke. Lateral

movements of the mandibular rami, however, require the presence of a
complex jaw musculature such as the masseter ‘sling’ of mammals. A
system equivalent to a masseter sling has never been found in any saurop-
sid, and the cranial structure of Camarasaurus does not support the pres-
ence of such a structure. Furthermore, the size and position of the teeth in
Camarasaurus would make transverse motions of the lower jaw impossible
when the jaws were close to occlusion. Therefore, the production of a
transverse power stroke in Camarasaurus is considered highly unlikely.
The orientation of the adductor jaw musculature suggests that jaw
motion was essentially orthal, with some propaliny possible if the inter-
nal and external adductor muscles were contracted ‘out of phase.’ It is not
clear yet whether the longitudinal component in jaw motion was proal,
palinal, or some combination of the two.
Postcrania The neck of Camarasaurus is relatively short compared

with overall body size (Table 4.2). The ribs of the middle cervicals possess
long, overlapping distal shafts (Gilmore 1925).
Camarasaurus possessed relatively long forelimbs (Table 4.2). Some of
the observed range in forelimb/hindlimb ratio (i.e., 0.79 in adults to 0.87
in juveniles) may reflect ontogenetic variation, but it is also possible that
the adult values are inaccurate because of the incorrect association of dis-
articulated material (Osborn and Mook 1921).
Feeding mechanism The jaw action of Camarasaurus was apparently

more complex than those employed by euhelopodids and cetiosaurids.
The structure of the jaw joint, orientation of the adductor jaw muscles,
and tooth wear are consistent with an orthal shearing action and a moder-
ate amount of propaliny. The presence of labiolingually directed micro-
wear scratches is unexpected, but does not necessarily indicate a
transverse motion of the jaws. A food-gathering technique that involved
tugging/raking of foliage from branches might cause material to be
pulled transversely across the wear facets. Alternatively, some food might
have been dragged transversely across the wear facets as the upper and
lower teeth sheared past each other. Perhaps raking was employed during
the initial phase of each food-gathering episode, with orthal shearing and
some propalinal motion being used to finally sever material from the rest
of the branch. Camarasaurus may have fed on coarser vegetation than
other sauropods, a view that is consistent with the greater amount of oral
processing implied by dental macrowear. The neck of Camarasaurus is not
particularly elongate, but the large body size of this animal allowed
browsing at heights of 8–9 m above ground level.
Brachiosauridae
Although several genera have been referred to the Brachiosauridae (McIn-
tosh 1990; Upchurch 1994, 1995), only Brachiosaurus itself is known ade-
quately from both skull and postcranial material.
Skull The skull of Brachiosaurus is relatively tall, with a steeply

arching internarial bar. The region rostral to the external nares forms a
long, flattened ‘muzzle,’ which is much more prominent than in any
other sauropod (Figure 4.2).
Dentition Brachiosaurid teeth are spatulate, but the crowns are

only slightly wider mesiodistally than the roots. This more slender form
of spatulate tooth has been termed a ‘cone-chisel’ tooth by Calvo (1994a).
A few denticles are sometimes present near the apex of unworn teeth
(HMN WJ470; Janensch 1935–36).
Tooth wear In well-preserved jaws of Brachiosaurus, a few of the teeth

possess typical mesial and distal wear facets or even ‘shoulders.’ The
majority of worn teeth, however, bears a large subtriangular facet situ-
ated at the crown apex (Figure 4.3). These facets face ventrolingually and
dorsolabially on the upper and lower teeth, respectively. The prevalence
of these apical facets, and the reduced mesial and distal wear, is the result
of a modification in the relative positions of the tooth rows: instead of
‘interdigitating,’ the upper and lower teeth are aligned, allowing a more
precise shearing bite (Calvo 1994a).
No microwear studies of teeth of Brachiosaurus have been reported.
Jaw joint and musculature A low ridge extends transversely across the
posterior part of the articular glenoid, greatly restricting fore-and-aft
motion of the lower jaw. The adductor jaw musculature, especially in the
temporal region, would have been steeply inclined with respect to the
long axis of the lower jaw, implying a simple orthal jaw action.
Postcrania Although Brachiosaurus possesses only the plesiomorphic

number of cervical vertebrae (12 or 13), elongation of individual cervicals
produces a long neck (Table 4.2). Janensch (1950) and Paul (1988) recon-
structed the cervical ribs of Brachiosaurus as relatively short: each rib
extending backwards to overlap the anterior portion of the succeeding
vertebra. The cervical ribs of the best preserved skeletons of Brachiosaurus,
however, are badly broken. Given that each rib possesses a long anterior
projection, it is likely that the distal shafts were longer than existing
reconstructions would imply (see ‘Cetiosauridae’).
The forelimbs of Brachiosaurus are extremely long (Table 4.2), resulting
in a substantial elevation of the cranial portion of the thorax.
Feeding mechanism Brachiosaurus seems to have developed a more

precise shearing bite than other ‘broad-toothed’ sauropods. Therefore,
food-gathering would have involved more ‘cutting’ through foliage than
tearing and ripping. Reduction in the prominence of mesial and distal
wear facets implies less oral processing. The elongation of the neck and
forelimbs, and the extremely large body size, indicate that Brachiosaurus
was specialized for browsing at heights somewhat greater than those
achieved by other sauropods.
Titanosauroidea
The Titanosauroidea are represented by a relatively large number of valid
genera. However, paradoxically, fragmentary preservation of most
known specimens means that we know less about the anatomy and rela-
tionships of this group than any other sauropod clade (McIntosh 1990).
Skull No complete skull is available for any titanosauroid, though a

few cranial fragments (Table 4.1) offer tantalizing clues to its form and
function. The best preserved tooth-bearing elements are the premaxilla
and dentary of Malawisaurus (Jacobs et al. 1993) and possibly a maxilla
from the Upper Cretaceous of India (Huene and Matley 1933:23–24). Until
recently, reconstructions of the titanosauroid skull resembled the long,
low, crania of diplodocoids (Huene 1929). This view was based on material
assigned to Antarctosaurus wichmannianus, which is more likely to have
belonged to a diplodocoid (Jacobs et al. 1993; Upchurch 1999). The pre-
maxilla of Malawisaurus indicates that the titanosauroid skull was actu-
ally short and high, as in Camarasaurus and Brachiosaurus. Such an
interpretation is also consistent with the revised phylogenetic position of
this clade (Figure 4.1).
Dentition Titanosauroid teeth resemble those of Brachiosaurus in

being relatively slender and possessing nearly parallel mesial and distal
margins, but they have developed additional apomorphies such as the
absence of the lingual concavity and a decrease in relative size (Figure 4.3).
The lingual surface of the crown may be mildly or strongly convex, pro-
ducing an elliptical or cylindrical horizontal cross-section, respectively.
The crown terminates in a bluntly rounded subtriangular apex, which is

often curved lingually.
Tooth wear Wear facets are present on the teeth of Titanosaurus

(Huene 1929) and undescribed titanosaurs from Madagascar (C. A.
Forster, pers. comm.), Brazil (Kellner 1996) and Morocco (Sereno et al.
1996). These teeth, although found in association with titanosauroid
material, were not recovered in contact with jaw elements. The mesial
and distal high-angle wear facets present on spatulate teeth have not been
observed on any titanosauroid tooth. Instead, there is a high-angle wear
facet at the apex of each crown, resulting in a ‘chisel’-shaped tip (Figure
4.3). Assuming these isolated teeth curve lingually (as in most sauropods,
but see ‘Nemegtosauridae’), it appears that wear facets are found on the
labial surface of some teeth and the lingual surface of others. This could
be interpreted as resulting from a precise occlusion between upper and
lower tooth crowns, producing wear on the labial surfaces of lower teeth
and the lingual surfaces of upper teeth.
The microwear on the teeth of the undescribed Madagascan titano-
saurid includes coarse scratches (30–40 ␮m wide) extending parallel to
the long axis of the crown (Figure 4.4). These scratches often extend along
the entire length of the wear facet. There are also randomly distributed
pits. This wear pattern could have been caused by durable food particles,
or grit, crushed between the teeth during occlusion. If the lingual and
labial wear facets occurred on the upper and lower teeth, respectively, the
scratches could have been caused by tooth–tooth contact.
Jaw joint and musculature No information on the titanosauroid jaw

joint is available at present. If the skull was short and tall, the adductor
jaw musculature would have been steeply inclined.
Postcrania Information on the neck structure of titanosauroids is

mainly derived from disarticulated materials, although nearly complete
cervical series are known from Brazil (Powell 1987). Saltasaurus has rela-
tively short cervicals in comparison with those of the unnamed Brazilian
form (Powell 1992), suggesting that relative neck length varied between
different titanosauroid groups. The cervical ribs are partially preserved in
the Brazilian titanosaur; the anterior processes are very prominent and at
least some of the distal shafts were elongate (Powell 1987: pl. 1).
The best preserved articulated postcranial skeleton of a titanosauroid
belongs to Opisthocoelicaudia (Borsuk-Bialynicka 1977; Upchurch 1995).
This specimen, which possesses virtually complete articulated limbs,
Figure 4.4. Scanning electron micrograph of the tooth crown of an unnamed

titanosaurid from Madagascar (FMNH field no. 93–73/95029). This picture was
obtained from a high-resolution peel of the tooth. The micrograph is oriented
with the apex of the crown towards the top of the page. Although most of the
microwear is located on the surface of the wear facet (rhomboidal area in the
top half of the picture), some wear striae are also present on the ‘unworn’
crown surface. Scale bar equals 1 mm.
demonstrates that at least some titanosauroids had relatively long fore-

limbs (Table 4.2).
Feeding mechanism Calvo (1994a) suggested that titanosaurids uti-

lized orthal jaw motions with precise tooth–tooth contact in order to
produce a shearing bite. Although this view was based on a study of
Nemegtosaurus (which is now considered a diplodocoid; Upchurch 1995,
1999), the data obtained from undoubted titanosauroids support Calvo’s
hypothesis. Maximum browse heights cannot be estimated for most tita-
nosauroids, although the structure of the neck and forelimbs suggests
that some forms fed at ‘intermediate’ heights, approximately 6–8 m
above ground level. Given the observed variation in cervical elongation
and body size, and the absence of information concerning cranial struc-
ture, it is likely that future discoveries will reveal considerable diversity in
titanosauroid feeding mechanics.
Diplodocoidea
The Diplodocoidea comprise three families, the Nemegtosauridae,
Dicraeosauridae, and Diplodocidae (Upchurch 1995). The skulls of
diplodocoids share a number of characteristics relevant to inferring
feeding (Barrett and Upchurch 1994), which will be discussed together
before examining each family individually.
Skull The skull is long and low, with an elongate preorbital region.
In dorsal view, the mandible is subrectangular in outline, whereas in
other sauropods the dentaries curve gradually round towards the sym-
physis (Upchurch 1998: fig. 5).
Dentition The tooth crowns are ‘peg’-like in all diplodocoids. The

teeth of nemegtosaurids and dicraeosaurids resemble those of titanosaur-
oids, whereas diplodocids possess a unique slender tooth form. The teeth
are restricted to the extreme rostral ends of the jaws, producing a ‘tooth-
comb.’ As in Titanosauriformes, the relative positions of the teeth have
been modified in diplodocoids so that one tooth on the upper jaw contacts
one or two teeth on the lower jaw (Salgado and Calvo 1997; Upchurch 1999).
Postcrania No postcranial material belonging to nemegtosaurids

has been recovered to date. The cervical ribs of dicraeosaurids and diplo-
docids are short and do not form overlapping ‘bundles.’ The forelimbs in
both taxa are shorter than those of other sauropods (Table 4.2).
Nemegtosauridae
Nemegtosaurids are a poorly known group represented by two isolated
skulls from the Upper Cretaceous of Mongolia.
Skull The nemegtosaurid skull shares several derived character-

states with dicraeosaurids and diplodocids (see above), but also retains
some plesiomorphic features. For example, the snout is elongated antero-
posteriorly and is more rectangular in dorsal view, but the posterior part
of the skull has undergone less modification than in other diplodocoids
(Upchurch 1999).
Dentition In Nemegtosaurus (Nowinski 1971), and probably Quaesito-

saurus (Kurzanov and Bannikov 1983), the teeth are not strongly procum-
bent. The upper teeth curve lingually, but the lower ones are highly
unusual because they curve slightly labially.
Tooth wear No tooth wear was reported in the description of Quae-

sitosaurus (Kurzanov and Bannikov 1983), but extensive wear is present
in Nemegtosaurus (Nowinski 1971). Apical wear facets occur on the labial
and lingual surfaces of lower and upper teeth, respectively. There is also
a small amount of wear on the mesial and distal margins of some
crowns.
No studies of the microwear on nemegtosaurid teeth have been
reported.
Jaw joint and musculature The glenoid area, on the dorsal surface of
the articular, is slightly elongated anteroposteriorly (Nowinski 1971).
Elongation of the snout and the possibility that the quadrate sloped a
little anteroventrally, suggest that the adductor jaw muscles were ori-
ented at a shallower angle to the long axis of the skull. These features
indicate that at least a moderate amount of propalinal movement could
have been achieved.
Feeding mechanism The tooth wear observed in nemegtosaurids sug-

gests that a shearing bite was employed. The presence of mesial and distal
wear could reflect some oral processing, or a less precise tooth-to-tooth
contact than that found in titanosauroids and dicraeosaurids. The role
played by propalinal movements is not entirely clear, although it may
have widened the gape and facilitated correct relative positioning of the
upper and lower tooth rows during the bite phase.
Dicraeosauridae
The Dicraeosauridae are represented by two genera, Dicraeosaurus and
Amargasaurus, both of which are known from partial skull and postcranial
material.
Skull The skull material assigned to Dicraeosaurus (Janensch

1935–36) and Amargasaurus (Salgado and Bonaparte 1991; Salgado and
Calvo 1992) is fragmentary. Nevertheless, the available specimens indicate
that the dicraeosaurid skull was similar in proportions and structure to
those of diplodocids. The snout region is elongate, and the quadrates
probably sloped anteroventrally (see below).
Dentition The slender ‘peg’-like teeth of Dicraeosaurus (Figure 4.3) are

slightly compressed labiolingually, with a gently convex lingual surface.
Both labial and lingual surfaces bear grooves paralleling each of the mesial
and distal margins. The crown tapers towards the apex and curves lin-
gually. The mesial and distal margins of the crown form acute edges.
Tooth wear None of the worn teeth assigned to Dicraeosaurus has

been found in situ, making it impossible to determine the true orientation
of wear facets. Assuming that all of the isolated teeth curve lingually (but
see Nemegtosauridae), it appears that small subtriangular wear facets are
present on the labial surfaces of some crown apices and the lingual sur-
faces of others. This wear pattern is very similar to that observed in tita-
nosauroids and was probably formed in the same way.
No microwear studies of dicraeosaurid teeth have been reported.
Jaw joint and musculature The jaw joint is not preserved in either
Dicraeosaurus or Amargasaurus. The phylogenetic relationships of dicraeo-
saurids, however, suggest that the glenoid was probably at least some-
what anteroposteriorly elongated. It is likely that the adductor jaw
musculature was oriented at a low angle to the long axis of the skull: the
preorbital region is elongated and the anteroventral slope of the quadrate
is implied by the very long, anteriorly directed basipterygoid processes
(Salgado and Calvo 1997; Upchurch 1999). Propaliny may, therefore, have
played an important role in the jaw action.
Postcrania Dicraeosaurids possess 11 or 12 short cervical vertebrae,

producing a rather short neck (Table 4.2). The cervical vertebrae bear
extremely tall neural spines (Janensch 1936; Salgado and Bonaparte 1991),
which would have impeded raising the head far above the level of the
shoulders.
Feeding mechanism The tooth wear indicates that dicraeosaurids

probably used a precision shear bite. The role played by propalinal move-
ments is not fully understood (see Nemegtosauridae above). The short
neck and forelimbs restricted vertical and increased lateral motion of the
neck, and the comparatively small body size (Table 4.2) suggest that
dicraeosaurids were specialized low-level browsers. Rearing up into a
bipedal stance could have increased browse height, but the absence of
pleurocoels in the cervical and dorsal centra seems to indicate that
strength of the vertebral column was more important than lightness.
Diplodocidae
The Diplodocidae contain a number of well-known sauropods, most of
which come from the Upper Jurassic of North America.
Skull The skull of Diplodocus longus (CMNH 11161) is particularly

well preserved and forms the basis for much of the following discussion.
Although fragmentary, the cranial material assigned to Apatosaurus
closely resembles that of Diplodocus (Berman and McIntosh 1978).
In Diplodocus, the ventral margin of the maxilla and dorsal margin of
the dentary are formed into thin sheets of bone that overlap each other
when the jaws are fully or nearly closed. This combination of structures,
which has not been observed in other sauropods, could be regarded as a
bony equivalent of a fleshy cheek: it would have prevented loss of food
from the lateral margins of the mouth (see ‘Feeding mechanism’ below).
The quadrate slopes anteroventrally, so that the temporal region is situ-
ated above and behind the jaw joint (Figure 4.2). As in dicraeosaurids, the
occipital condyle is directed ventrally rather than posteroventrally rela-
tive to the long axis of the skull (Berman and McIntosh 1978). This
orientation could be regarded as an inevitable by-product of other mod-
ifications to the posterior part of the skull (Salgado and Calvo 1997;
Upchurch 1999), but might also be related to an alteration in the habitual
posture of the head (relative to the neck) during feeding. There is a long
retroarticular process in Diplodocus.
Dentition The teeth of Diplodocus are highly unusual compared with

those of other sauropods. The long slender crowns are elliptical in cross-
section and lack any trace of labial and lingual grooves. The mesial and
distal margins are parallel to each other for virtually the entire length of
the tooth, rapidly converging at the apex to form a blunt triangular
point. There is little lingual curvature of the crowns. The teeth are dis-
tinctly ‘procumbent’ (i.e., their long axes are directed rostroventrally and
rostrodorsally in upper and lower teeth respectively). The upper teeth are
a little longer, and somewhat more robust, than the lower teeth.
Tooth wear The teeth of Diplodocus frequently display heavy wear

(contra Haas 1963). No wear has been observed on the mesial and distal
margins, but a prominent facet is often present at the apex (Figure 4.3).
Wear is somewhat heavier on the upper teeth than on the lower ones.
These apical wear facets are planar elliptical surfaces oriented at approxi-
mately 45° to the long axis of the tooth. Unlike all other wear patterns
observed in the Sauropoda, the wear facets on the teeth of Diplodocus are
found on the labial sides of both upper and lower teeth (Barrett and
Upchurch 1994; contra Calvo 1994a). The base of each facet (furthest from
the apex) has a ‘step’ between the enamel and dentine (see above), whereas
the top of the facet (near the tip of the crown) is ‘flush.’
Fiorillo (1991) examined the microwear on the teeth of Diplodocus. Fine
scratches were present, the majority being oriented ‘perpendicular to the
contact between the enamel and dentine’ (Fiorillo 1991:24). No pits were
found. Calvo (1994a), and our own SEM study, confirmed the presence of
fine scratches extending labiolingually across the wear facets (Figure 4.5).
In addition, there are pit-like areas (Figure 4.5), substantially larger than
those typically formed by grit or phytoliths, which represent the results
of extremely heavy wear.
Jaw joint and musculature The dorsal surface of the articular glenoid
is strongly elongated anteroposteriorly, and there are no transverse ridges
to hinder longitudinal movements of the quadrate over this surface
(Barrett and Upchurch 1994). The elongate preorbital region, and the
anteroventrally sloping quadrate, indicate that the jaw adductor muscles
were oriented at a relatively low angle to the long axis of the skull. Diplod-
ocus also has a long retroarticular process, which may reflect an increased
role for M. depressor mandibulae in the jaw action; this muscle may have
contributed to retraction of the lower jaw during a palinal movement
and/or could have been involved in the evolution of a wider gape (see
below).
Postcrania Diplodocids have increased the number of cervical verte-

brae from 13 to 15 (Diplodocus and Apatosaurus) or 16 (Barosaurus). Individual
vertebrae are elongate, especially in Diplodocus and Barosaurus (Upchurch
1995). As a result of these modifications, the necks of diplodocids are rela-
tively long (Table 4.2), although they do not reach the extremes present in
Figure 4.5. Scanning electron micrograph of the wear facet on a tooth of Diplodocus. This picture was obtained from a high-resolution peel of a
premaxillary tooth belonging to specimen CMNH 11161. The micrograph is oriented so that the labial margin is situated towards the bottom of
the page. The sketch diagram (not to scale) indicates the area of the wear facet detailed in the micrograph, and also shows the principal axes of
the tooth. The majority of striae extend parallel to the labiolingual axis of the wear facet. Scale bar equals 0.1 mm.
some euhelopodids. As in dicraeosaurids, the cervical ribs are very short,

especially in Apatosaurus.
Feeding mechanism Diplodocids clearly possessed a radically differ-

ent feeding mechanism compared with those in other sauropods. The
shape and size of the teeth makes them unsuitable for an orthal shearing
bite. The orientation of the wear facets could not have been formed by the
‘horizontal slicing’ action proposed by Calvo (1994a), or indeed any kind
of tooth–tooth contact. Several attempts have been made to explain this
wear pattern in terms of tooth–food contact. Holland (1924) proposed
that Diplodocus was molluscivorous, plucking mussels from shell-beds.
Contact between the labial surfaces of the tooth apices, and either a rock
surface or other mussels, could have caused the labial wear. As the teeth
slid past the shell of the target mussel, however, some wear should also
have been formed on the lingual surfaces. Furthermore, labial wear facets
would have weakened the animal’s grip on the mussel shell if the latter
was held between the tips of the teeth as suggested by Holland (1924: fig.
4). Bakker (1986) proposed that labial wear was produced during bark
stripping. While this hypothesis may explain the orientation of the wear
facets, it seems unlikely for several reasons. First, the slender, procum-
bent teeth of Diplodocus do not seem well suited to withstand the forces
generated by levering up portions of bark. Second, the margins of the
wear facets are blunt and rounded, not sharp and chisel-like as they
would need to be if used to penetrate and strip bark. Third, a diet of bark
is very poor in nutrients (extant mammals only strip bark when other
food sources become scarce).
Norman (1985) and Dodson (1990) have suggested that diplodocids fed
by raking vegetation from branches. Barrett and Upchurch (1994) pro-
posed a detailed model of unilateral branch stripping to explain labial
wear and other features of the skull of Diplodocus. In unilateral branch
stripping, one tooth row is used to strip foliage from a stem, while the
other acts as a guide and stabilizer. Whether the upper or lower tooth row
is used for stripping depends on the position of the animal’s head relative
to the stem, which, in turn, may reflect browse height (Barrett and
Upchurch 1994: fig. 5). Thus the labial wear on the upper and lower tooth
rows would be formed independently.
Unilateral branch stripping explains several features of the skull of
Diplodocus. For example, given that the microwear was formed by tooth-
to-food contact, the labiolingual orientation of the microscopic scratches
indicates that the abrasive material moved parallel to the surfaces of the
wear facets, in a direction lying within a parasagittal plane. The maxillary

and dentary ‘pseudo-cheek’ would have prevented stripped foliage from
escaping through the sides of the mouth during branch stripping. The
elongate preorbital region would have allowed a longer portion of stem to
be stripped in a single action. Palinal motion of the lower jaws (not proal,
contra Barrett and Upchurch 1994) may have played two important roles:
an increase in gape (facilitating placement of a foliage-bearing stem in the
mouth) and fine adjustments of the relative positions of the tooth rows in
order to produce a smooth stripping action.
The unilateral branch stripping model has implications for our under-
standing of feeding in Diplodocus. The orientation of the wear on the
upper teeth suggests that stems were often raked at a steep upward angle
relative to the long axis of the skull (Barrett and Upchurch 1994: fig. 5A).
This angle, coupled with the greater robustness of the upper teeth and
their heavier wear, suggests that high-browsing played an important role
in food-gathering. The wear on the lower tooth row indicates that some
food was acquired at lower browse heights (Barrett and Upchurch 1994).
The elaborate nature of the proposed feeding mechanism for Diplodocus
also implies more precise, and perhaps selective, food-gathering.
Diplodocids are often considered to have been high-browsers (Bakker
1971, 1986). The elongate neck would have allowed them to browse at
maximum heights of 10 m (Apatosaurus and Diplodocus) or 12 m (Barosaurus)
(Table 4.2). The shortened forelimbs and long sacral neural spines may
have enabled these forms to rear up into a bipedal stance (Bakker 1986),
increasing maximum browse height by as much as 2 m. The greater
lateral flexibility of the neck, coupled with shortened forelimbs, could be
interpreted as adaptations to food-gathering at lower levels. The high-
and low-browsing models of diplodocid feeding are not mutually exclu-
sive; indeed, such variation in vertical feeding range is consistent with
unilateral branch stripping and increased food selectivity.
Implications for sauropod paleoecology
Many faunas, particularly in the Late Jurassic and Early Cretaceous, con-
tained several sympatric sauropod taxa (Dodson 1990; Weishampel 1990).
It is difficult to explain such diversity if sauropods are regarded as
uniform high-browsers with simple jaw actions (Dodson 1990) because
we might expect some ecological partitioning between similar organisms
sharing the same environment. The data presented above indicate that
the feeding mechanisms of sauropods were not as uniform as previously

suggested. Possible niche separation between sauropods is illustrated in
the following ‘case study.’
To date, fourteen sauropod genera have been described from the
Upper Jurassic Morrison Formation of western North America (McIntosh
1990), although no more than six have been recovered from any single
locality (Dodson et al. 1980). Several localities include representatives of
the ‘Cetiosauridae’ (Haplocanthosaurus), Brachiosauridae (Brachiosaurus),
Camarasauridae (Camarasaurus) and Diplodocidae (Apatosaurus, Barosaurus,
and Diplodocus). Each of the major sauropod groups in the Morrison For-
mation would have possessed a distinctive jaw mechanism. The camara-
saurids, brachiosaurids, and probably the cetiosaurids (assuming the
unknown skull of Haplocanthosaurus resembled that of Patagosaurus), had
robust skulls and ‘heavy,’ spatulate dentitions suitable for feeding
on coarse vegetation. Subtle variation in jaw action, such as the
raking/tearing and precise cutting methods for foliage severing in camar-
asaurids and brachiosaurids, respectively, could be related to differences
in preferred ‘target’ plants. Diplodocids possessed a highly unusual food-
gathering system based on unilateral branch stripping. The relatively
small head, narrow mouth, delicate tooth-comb and absence of oral pro-
cessing imply that diplodocids fed on smaller and less coarse food items;
they may have deliberately selected the more nutritious portions of
plants, such as new shoots. This greater selectivity is consistent with the
hypothesis that diplodocids fed over a wider vertical range than other
sauropod groups. (The more nutritious parts of plants are scarcer and
more widely distributed.)
Estimates of ‘maximum’ browse height provide some additional evi-
dence for ecological partitioning of sauropods, although there may have
been considerable overlap between genera. It seems likely that high-
browsers would have often fed at levels below their maximum browse
heights, and juveniles may have had to compete for resources at the same
level as adult non-sauropods (such as ornithopods and stegosaurs). Even
the adults of medium-sized sauropods may have browsed at similar
heights as subadults of ultra-high browsers. Nevertheless, there does
seem to have been some separation of sauropods on the basis of browse
height. Brachiosaurids, and possibly Barosaurus, could have obtained food
up to 14 m above ground level; camarasaurids and the remaining diplo-
docids were restricted to maximum browse heights of 10 m or less; and
Haplocanthosaurus may not have fed above 5–6 m. The extent to which sau-
ropods overlapped or were separated by browse height depends on a

number of factors which are difficult to gauge: the relative costs and
benefits of neck raising and lowering, and of bipedalism; the extent of
overlap in preferred food items; the spatial distribution of preferred food
items.
Several other dinosaur faunas, such as those in the Lower Shaximiao
Formation (Middle Jurassic) of Sichuan (China) or Tendaguru beds (Late
Jurassic) of Tanzania, contain three or more sympatric sauropod genera
(Weishampel 1990). In each case, some ecological partitioning on the basis
of feeding mechanism seems possible.
Phylogenetic patterns in sauropod feeding mechanisms
Ideally, we would like to understand the selective pressures and con-

straints that influenced sauropod evolution. Since food-gathering is such
an important component in survival, it seems probable that patterns in
feeding-mechanism evolution will reflect significant events in the history
of the Sauropoda. In order to identify such patterns, relevant character-
states have been mapped onto a simplified version of the cladogram pre-
sented earlier (Figure 4.6). Missing data inevitably produce some
ambiguity in character-state distribution; the pattern discussed below is
based on the delayed transformation optimization (DELTRAN) in PAUP
3.1.1 (Swofford 1993), which tends to favor convergence over reversal when
ambiguity occurs. Thus the proposed sequence of apomorphy acquisi-
tion, and the identification of convergence versus reversal, should be
treated as provisional.
Node A (Sauropoda) is characterized by an apomorphic lengthening of
the forelimbs. This derived state is retained by all later sauropod groups,
with partial reversals in Shunosaurus and the dicraeosaurid–diplodocid
clade. Brachiosaurus has elongated forelimbs beyond the derived condition
that unites all sauropods. Relative forelimb length may be affected by
many factors, but the current distribution of this character within the
sauropods suggests some correlation with browse height.
Several derived states combine to form the derived character-state
known as ‘spatulate teeth,’ which occurs at node B. These apomorphic
features include the development of a relatively larger and more mesio-
distally expanded crown, the presence of one or two grooves on the labial
surface, and the formation of a lingual concavity. These features persist
throughout much of subsequent sauropod evolution but are lost at least
Figure 4.6. Feeding mechanism characters mapped onto a simplified sauropod

cladogram (see text for details). The patterns within the branches of the
cladogram show the distribution of different types of jaw action (see key).
Abbreviations: cv, cervical; d, dentary; ect, ectopterygoid; emf, external
mandibular fenestra; m, maxilla; oc, occipital condyle; pr, process; ra,
retroarticular. For convenience, ‘cetiosaurids’ are represented as a single
terminal taxon.
twice independently (see below). Most of these derived states can be inter-
preted as part of an adaptation to cropping coarse vegetation, although
the functional significance of the lingual concavity is not understood.
Node B may also be characterized by an increase in body size from 10 m to
15 m or more, representing a shift in sizes beyond that achieved by the
largest prosauropods. Body size remains variable within the Sauropoda,

with independent reductions in Shunosaurus, dicraeosaurids, and some
titanosauroids, and independent occurrences of gigantism in Brachiosau-
rus, Seismosaurus and the titanosauroid Argentinosaurus. Changes in sauro-
pod body size probably reflect a number of selective pressures, including
those affecting food acquisition (such as browse height) and food process-
ing (Farlow 1987).
Several important apomorphies are currently clustered at node C
(Eusauropoda). This distribution may be misleading, however, because
adequate vulcanodontid material is not available. It is likely, for example,
that the acquisition of the robust mandibular symphysis and the increase
in cervical number from 10 to 12 occurred below node C. Other apomor-
phic states, including the presence of mesial and distal tooth wear, seem
to be genuinely absent in vulcanodontids. The derived character-states
diagnosing the Eusauropoda are either retained or further elaborated in
later sauropods; reversals to the original plesiomorphic conditions have
not been identified.
Within the Euhelopodidae, derived states reflect specializations for dif-
ferent browsing heights. Shunosaurus developed a more flexible neck,
shortened forelimbs, and smaller body size, as part of a low-browsing
strategy. Other euhelopodids seem to have invaded a high-browsing
niche: extremely long necks are achieved through the acquisition of an
additional four cervical vertebrae and elongation of each cervical vertebra.
Particularly elongate necks occurred independently two or perhaps three
times elsewhere in the Sauropoda. Evidence in support of such conver-
gence is provided not only by phylogenetic distribution, but also by the
different methods used to achieve neck elongation: all groups exploit
elongation of individual cervical vertebrae, but Brachiosaurus retains the
plesiomorphic cervical number (12 or 13), euhelopodids ‘create’ new cervi-
cals, and, in diplodocids, two dorsal vertebrae were integrated into the cer-
vical series (Upchurch 1994). It is likely that titanosauroid necks also varied
considerably in relative proportions, but the fragmentary preservation of
known specimens currently obscures most of the useful details. There
seems to have been repeated selective pressures favoring the development
of long necks during sauropod evolution, perhaps reflecting competition
between high-browsers or an ‘arms race’ with target food plants.
The only potential apomorphy at node E (cetiosaurids⫹neosauropods)
is the loss of denticles on tooth crowns. The distribution of this character-
state is complex: an independent loss of serrations occurs in Euhelopus,
whereas small denticles are variably present in Brachiosaurus. The coarse

serrations present in early sauropod teeth, as in prosauropods and orni-
thischians, may have facilitated the slicing/cutting of plant matter as tooth
crowns slid past each other. Why cetiosaurids and neosauropods should
have lost tooth serrations is not understood at present.
The external mandibular fenestra is reduced in eusauropods and
finally obliterated in neosauropods (node F). This probably reflects a re-
arrangement of the adductor jaw musculature relative to vulcanodontids
and prosauropods, but its functional significance is not clear.
The current lack of postcranial remains for nemegtosaurids means
that only cranial apomorphies can cluster at node G (Diplodocoidea).
Several apomorphies (Figure 4.6) mark a major shift in feeding mecha-
nism relative to the cropping action present in other eusauropods. In par-
ticular, the acquisition of small ‘peg’-like teeth arranged in a tooth-comb
indicates the exploitation of a more delicate food source. Alterations to
the shape of the mandible and the relative positions of teeth on opposing
jaws would facilitate a more precise cutting action during feeding. This
‘precise shear bite’ is retained in dicraeosaurids, and an analogous system
is acquired independently in Titanosauriformes. Although delayed trans-
formation favors this ‘convergence hypothesis,’ it is equally parsimonious
to suggest that the precision shear bite is a synapomorphy uniting the
Neosauropoda, which underwent subsequent reversal in Camarasaurus. A
similar ambiguity arises with respect to the elongation of the jaw joint
and the utilization of propalinal jaw motion. Either slight propaliny
was acquired independently in Camarasaurus and diplodocoids, or this
derived character-state unites all neosauropods and is reversed in Titano-
sauriformes. Only the discovery of skull material belonging to basal
camarasauromorphs and diplodocoids will resolve these issues.
Further modifications in the diplodocoid skull occur at node I. In par-
ticular, the role of propalinal jaw movements would have been enhanced
by further elongation of the snout and the pronounced anteroventral
orientation of the quadrate. Various adaptations for low-browsing were
acquired, including shortened forelimbs and greater lateral flexibility of
the neck. This implies that low-browsing was retained (with further mod-
ifications) by dicraeosaurids, whereas diplodocids superimposed a rein-
vented high-browsing strategy in order to achieve a wide vertical feeding
range.
From the phylogenetic perspective, many components of the diplo-
docid feeding mechanism were acquired prior to unilateral branch strip-
ping. For example, the elongate snout, anteroventrally oriented quad-

rate, and elongate articular glenoid, are also present in dicraeosaurids
and are incipiently developed in nemegtosaurids. Although these fea-
tures probably played an important role in unilateral branch stripping,
they represent co-opted structures which were originally acquired as part
of a precision shear jaw action. The slender ‘pencil-like’ teeth and maxil-
lary-dentary ‘pseudo-cheek,’ however, may represent apomorphies spe-
cifically related to branch stripping.
The Camarasauromorpha (node L) are characterized by the develop-
ment of a short, high skull. In basal forms, such as Camarasaurus, this
modification may relate to a more vertical orientation of the adductor
musculature, reflecting a powerful orthal shearing action. In the titano-
sauriform clade, however, there is a ‘trend’ towards the independent
development of a more precise shear bite. This system included many of
the apomorphies acquired by diplodocoids, but differs by the absence of a
tooth-comb and propaliny. Some of the similarities between the diplodo-
coid and titanosauriform feeding mechanisms are the result of inevitable
correlations between characters. For example, a more precise contact
between opposing tooth rows automatically results in a reduction in
mesial and distal wear and an increase in apical wear (Upchurch 1999).
Why these two neosauropod clades should converge in some aspects of
their feeding mechanics and diverge in others is not entirely clear. One
possibility is that the similarities reflect a more selective diet, while the
dissimilarities relate to preferences for different plants or parts of plants.
Current evidence implies that feeding mechanisms within the Titano-
sauroidea were themselves rather variable. For example, there is some
indication that the skull of ‘advanced’ titanosaurs (e.g., Saltasaurus) had
independently acquired some ‘diplodocoid’ features (Salgado and Calvo
1997; Upchurch 1999). The extent of such convergence, and patterns in
apomorphy distribution, however, are obscured by the absence of both
well-preserved titanosauroid cranial material and a well-supported phy-
logeny for this group.
The evolution of feeding mechanisms within the Sauropoda displays a
complex pattern of convergence and reversal. Basic components of the
jaw action (e.g., tooth shape and position, orientation of adductor jaw
muscles, structure of jaw joint) seem relatively stable and long-lived.
Most of the homoplasy in sauropod jaw mechanisms occurs as a result of
the independent acquisition of precision shear bites in titanosauriforms
and diplodocoids. In contrast, postcranial characters, such as relative
neck and forelimb lengths, appear to be more variable during the course
of sauropod evolution. There is at least a partial correlation between
browse height and certain postcranial characters: elongate but inflexible
necks, long forelimbs, and large body size are linked to high-browsing
whereas more flexible necks, shortened forelimbs, and smaller body size
are related to low-browsing. The reasons underlying these different levels
of homoplasy are not well understood. This pattern could be an artefact
created by the relatively poor preservation of many sauropod skulls; it
may also represent a genuine phenomenon reflecting different selection
pressures on skulls and postcrania.
The distribution of apomorphies on the sauropod cladogram, if taken at
face value, implies three major phases in evolution of feeding mechanisms
in sauropod dinosaurs. The first phase, occurring at nodes A–C, involved
many modifications that would facilitate the cropping of large amounts of
low-quality, coarse vegetation. The other two phases of ‘rapid’ evolution
occur independently in the Diplodocoidea and Titanosauriformes; these
apomorphies seem to reverse previous ‘trends’ insofar as they reflect more
selective feeding on less coarse material. This interpretation has a number
of important implications: (1) the initial radiation of sauropods in the Late
Triassic and Early Jurassic may be related not only to the exploitation of a
high-browsing niche, but also to the ability of these animals to process
coarse, nutrient-poor vegetation; (2) most of the important innovations in
sauropod feeding-mechanism evolution were acquired at a relatively early
stage (i.e., by the time the major lineages had diverged in the Middle or
early Late Jurassic); and (3) the ‘decline’ of ‘broad-toothed’ sauropods,
during the Cretaceous, may have been triggered by some selection pres-
sure(s) against forms that relied on large quantities of low-quality food.
Thus fluctuations in sauropod diversity may have been strongly influenced
by the evolution of feeding mechanisms. Such patterns should be treated
with caution because of our current ignorance of much of sauropod evolu-
tion. For example, pattern (2) would be contradicted by the discovery that
titanosauroids developed a wide variety of different feeding mechanisms
during the Cretaceous. Nevertheless, current evidence suggests that
feeding mechanisms played a central role in the origin and subsequent evo-
lutionary history of sauropods.
Conclusions and prospectus
Without well-preserved, definite gastrointestinal contents, it will remain

difficult to determine which plants, or parts of plants, formed the diets of
different sauropod taxa. Nevertheless, the data presented above place

limits on the probable food-gathering strategies employed by sauropods.
Current conclusions include: (1) euhelopodids, cetiosaurids, and camara-
saurids, used more oral processing than vulcanodontids and prosauro-
pods; (2) a precise shearing bite was developed by brachiosaurids,
titanosauroids, nemegtosaurids and dicraeosaurids; (3) diplodocids were
specialized for branch stripping and may have been more selective
feeders; (4) sauropods were probably specialized to exploit resources at
different heights above ground level. Whatever interpretation we place
on the data presented above, it seems clear that sauropod feeding mecha-
nisms were more varied and complex than many previous studies have
assumed. This diversity offers some insights into why many Jurassic and
Cretaceous terrestrial faunas were able to support several sympatric sau-
ropod taxa.
A variety of approaches could be used to expand this study. In particu-
lar, the tooth microwear in euhelopodids, brachiosaurids, and nem-
egtosaurids require attention. Furthermore, variation in macro- or
microscopic wear patterns has not been examined systematically in teeth
from different positions in the jaw or through ontogeny. A more rigorous
engineering approach to the study of forces in the sauropod skull and
postcranial skeleton should yield insights into the relationship between
structure and function. There still exists no detailed understanding of the
biomechanical significance of many of the cranial and vertebral characters
that separate sauropods at the familial or generic level. By presenting our
interpretations of sauropod feeding mechanisms here, we hope to encour-
age further investigation of these spectacular terrestrial herbivores.
Acknowledgments
This research was supported by The Leverhulme Trust, Bristol University

(PU), a NERC studentship (GT4/93/128/G) and Trinity College, Cam-
bridge (PMB). Grants from The Royal Society and The Dinosaur Society
enabled us to collect data on Chinese sauropods. We are very grateful to
the following for allowing access to the material in their care: Angela
Milner and Sandra Chapman (BMNH), John Martin (LCM), Wolf-Dieter
Heinrich and the late Hermann Jaeger (HMN), Mary Dawson, David
Berman, Luo Zhexi and Norman Wuerthele (CMNH), Mark Norell
(AMNH), Michael Brett-Surman (USNM), Cathy Forster (State University
of New York, Stony Brook), Philip Powell (OUMZ), He Xin-Lu (CIT), Dong
Zhiming (IVPP), Ouyang Hui (ZDM), Jose Bonaparte (MACN), and John
Peel (PMU). Lorraine Cornish (BMNH) gave expert advice on tooth

casting. Ian Marshall (Department of Earth Sciences, Cambridge) and Bill
Lee (Department of Zoology, Cambridge) greatly assisted our SEM
studies. This paper was greatly improved as a result of helpful reviews by
Peter Dodson and Jeff Wilson. We are extremely grateful to Xu Xing
(IVPP) for his assistance during our research in China. Finally, we thank
Hans-Dieter Sues for inviting us to contribute to this volume.
References
Alexander, R. M. (1985). Mechanics of posture and gait in some large dinosaurs. Zool. J.
Linn. Soc. London 83:1–25.
Bakker, R. T. (1971). The ecology of the brontosaurs. Nature 229:172–174.
Bakker, R. T. (1986). The Dinosaur Heresies. New York: William Morrow.
Barrett, P. M. (1998). Herbivory in the Non-Avian Dinosauria. Unpublished Ph.D.
Dissertation, University of Cambridge.
Barrett, P. M., and Upchurch, P. (1994). Feeding mechanisms of Diplodocus. GAIA
10:195–204.
Barrett, P. M., and Upchurch, P. (1995). Sauropod feeding mechanisms: their bearing on
palaeoecology. In Sixth Symposium on Mesozoic Terrestrial Ecosystems and Biotas, Short
Papers, eds. A. Sun and Y. Wang, pp. 107–110. Beijing: China Ocean Press.
Berman, D. S, and McIntosh, J. S. (1978). Skull and relationships of the Upper Jurassic
sauropod Apatosaurus (Reptilia, Saurischia). Bull. Carnegie Mus. Nat. Hist. 8:1–35.
Bonaparte, J. F. (1986). Les Dinosaures (Carnosaures, Allosauridés, Sauropodes,
Cétiosauridés) du Jurassique moyen de Cerro Condor (Chubut, Argentine). Ann.
Paléont. 72:325–386.
Borsuk-Bialynicka, M. (1977). A new camarasaurid Opisthocoelicaudia gen. n. sp. n. from
the Upper Cretaceous of Mongolia. Palaeont. Pol. 37:5–64.
Calvo, J. O. (1994a). Jaw mechanics in sauropod dinosaurs. GAIA 10:183–194.
Calvo, J. O. (1994b). Gastroliths in sauropod dinosaurs. GAIA 10:205–208.
Calvo, J. O., and Salgado, L. (1995). Rebbachisaurus tesonei sp. nov. A new Sauropoda from
the Albian-Cenomanian of Argentina; new evidence on the origin of the
Diplodocidae. GAIA 11:13–33.
Carey, M. A., and Madsen, J. H. Jr. (1972). Some observations on the growth, function
and differentiation of sauropod teeth from the Cleveland-Lloyd Quarry. Proc.
Utah Acad. Sci., Arts, Lett. 49: 39–43.
Choy, D. S. J., and Altman, P. (1992). The cardiovascular system of Barosaurus: an
educated guess. The Lancet 340:534–536.
Christiansen, P. (1996). The evidence for and implications of gastroliths in sauropods
(Dinosauria, Sauropoda). GAIA 12:1–7.
Coombs, W. P. (1975). Sauropod habits and habitats. Palaeogeogr. Palaeoclimat., Palaeoecol.
17:1–33.
Cooper, M. R. (1984). A reassessment of Vulcanodon karibaensis Raath (Dinosauria:
Saurischia) and the origin of the Sauropoda. Palaeontol. Afr. 25:203–231.
Cope, E. D. (1877a). On Amphicoelias, a genus of saurian from the Dakota epoch of
Colorado. Proc. Am. Phil. Soc. 17:242–246.
Cope, E. D. (1877b). On a gigantic saurian from the Dakota epoch of Colorado. Paleontol.
Bull. 25:5–10.
Cope, E. D. (1878). A new species of Amphicoelias. Am. Nat. 12:563–565.
Dodson, P. (1990). Sauropod paleoecology. In The Dinosauria, ed. D. B. Weishampel, P.
Dodson, and H. Osmólska, pp. 402–407. Berkeley: University of California Press.
Dodson, P., Behrensmeyer, A. K., Bakker, R. T., and McIntosh, J. S. (1980). Taphonomy
and paleoecology of the Upper Jurassic Morrison Formation. Paleobiology
6:208–232.
Dong, Z. (1992). Dinosaurian Faunas of China. Berlin: Springer-Verlag.
Dong, Z., Zhou, S., and Zhang, Y. (1983). [The dinosaurian remains from Sichuan Basin,
China.] Palaeont. Sin., n.s. C 162 (23):1–145 (in Chinese with English summary).
Fiorillo, A. R. (1991). Dental microwear on the teeth of Camarasaurus and Diplodocus:
implications for sauropod paleoecology. In Fifth Symposium on Mesozoic Terrestrial
Ecosystems and Biotas, Extended Abstracts, ed. Z. Kielan-Jaworowska, N. Heintz, and
H. A. Nakren, pp. 23–24. Contr. Paleont. Mus., Univ. Oslo 364.
Fiorillo, A. R. (1996). Further comments on the patterns of microwear and resource
partitioning in the Morrison Formation sauropods Diplodocus and Camarasaurus.
J. Vert. Paleont. 16(suppl. to 3):33A.
Gillette, D. D., and Lockley, M. G. (1989). Dinosaur Tracks and Traces. Cambridge and New
Gilmore, C. W. (1925). A nearly complete articulated skeleton of Camarasaurus, a
saurischian dinosaur from The Dinosaur National Monument. Mem. Carnegie
Mus. Nat. Hist. 10:347–384.
Gilmore, C. W. (1932). On a newly mounted skeleton of Diplodocus in the United States
National Museum. Proc. U.S. Natl. Mus. 81:1–21.
Gilmore, C. W. (1936). Osteology of Apatosaurus with special reference to specimens in
the Carnegie Museum. Mem. Carnegie Mus. Nat. Hist. 11:175–300.
Haas, G. (1963). A proposed reconstruction of the jaw musculature of Diplodocus. Ann.
Carnegie Mus. Nat. Hist. 36:139–157.
Hatcher, J. B. (1901). Diplodocus Marsh, its osteology, taxonomy and probable habits, with
a restoration of the skeleton. Mem. Carnegie Mus. Nat. Hist. 1:1–64.
Hay, O. P. (1908). On the habits and the pose of the sauropodous dinosaurs, especially
Diplodocus. Am. Nat. 42:672–681.
He, X-L., Li, C., and Cai, K. J. (1988). [The Middle Jurassic dinosaur fauna from Dashanpu,
Zigong, Sichuan: sauropod dinosaurs (2) Omeisaurus tianfuensis.] Chengdu: Sichuan
Publishing House of Science and Technology (in Chinese).
Holland, W. J. (1924). The skull of Diplodocus. Mem. Carnegie Mus. Nat. Hist. 9:379–403.
Huene, F. von (1929). Los Saurisquios y Ornithisquios del Crétaceo Argentino. Ann. Mus.
La Plata 3:1–196.
Huene, F. von, and Matley, C. A. (1933). The Cretaceous Saurischia and Ornithischia of
the central provinces of India. Mem. Geol. Surv. India 26:1–74.
Hunt, A. P., Lockley, M. G., Lucas, S. G., and Meyer, C. A. (1994). The global sauropod
fossil record. GAIA 10:261–279.
Jacobs, L. L., Winkler, D. A., Downs, W. R., and Gomani, E. M. (1993). New material of an
Early Cretaceous sauropod dinosaur from Africa. Palaeontology 36:523–534.
Jain, S. L., Kutty, T. S., Roychowdhury, T., and Chatterjee, S. (1975). The sauropod
dinosaur from the Lower Jurassic Kota Formation of India. Proc. R. Soc. Lond. B
188:221–228.
Jain, S. L., Kutty, T. S., Roychowdhury, T., and Chatterjee, S. (1979). Some characteristics
of Barapasaurus tagorei, a sauropod dinosaur from the Lower Jurassic of Deccan,
India. Proc. Fourth Intl. Gondwana Symp., Calcutta 1977, 1:204–216.
Janensch, W. (1929). Die Wirbelsäule der Gattung Dicraeosaurus. Palaeontographica, Suppl.
7(1), 2(1):37–133.
Janensch, W. (1935–36). Die Schädel der Sauropoden Brachiosaurus, Barosaurus und
Dicraeosaurus aus den Tendaguru-Schichten Deutsch-Ostafrikas.
Palaeontographica, Suppl. 7(1), 2(2):147–298.
Janensch, W. (1936). Ein aufgestelltes Skelett von Dicraeosaurus hansemanni.
Palaeontographica, Suppl. 7(1), 2(2):299–308.
Janensch, W. (1950). Die Wirbelsäule von Brachiosaurus brancai. Palaeontographica, Suppl.
7(1), 3(2):27–93.
Janensch, W. (1961). Die Gliedmaszen und Gliedmaszengürtel der Sauropoden der
Tendaguru-Schichten. Palaeontographica, Suppl. 7(1), 3(4):177–235.
Kellner, A. W. A. (1996). Remarks on Brazilian dinosaurs. Mem. Queensland Mus.
39:611–626.
Kurzanov, S. M., and Bannikov, A. F. (1983). [A new sauropod from the Upper Cretaceous
of Mongolia.] Paleontol. Zh. 1983:91–97 (in Russian).
Lull, R. S. (1919). The sauropod dinosaur Barosaurus Marsh. Mem. Connecticut Acad. Art. Sci.
6:1–42.
Luo, Z., and Wu, X.-C. (1995). Correlation of vertebrate assemblage of the Lower Lufeng
Formation, Yunnan, China. In Sixth Symposium on Mesozoic Terrestrial Ecosystems
and Biotas, Short Papers, ed. A. Sun and Y. Wang, pp. 83–88. Beijing: China Ocean
Press.
Madsen, J. H., McIntosh, J. S., and Berman, D. S. (1995). Skull and atlas-axis complex of
the Upper Jurassic sauropod Camarasaurus Cope (Reptilia: Saurischia). Bull.
Carnegie Mus. Nat. Hist. 31:1–115.
Marsh, O. C. (1877a). Notice of a new gigantic dinosaur. Am. J. Sci. (3) 14:87–88.
Marsh, O. C. (1877b). Notice of some new dinosaurian reptiles from the Jurassic
Formation. Am. J. Sci. (3) 14:514–516.
Marsh, O. C. (1878). Principal characters of American Jurassic dinosaurs. Part I. Am. J. Sci.
(3) 16:411–416.
Marsh, O. C. (1879). Principal characters of American Jurassic dinosaurs. Part. II. Am. J.
Sci. (3) 17:86–92.
Marsh, O. C. (1883). Principal characters of American Jurassic dinosaurs. Part VI.
Restoration of Brontosaurus. Am. J. Sci. (3) 26:81–85.
Marsh, O. C. (1884). Principal characters of American Jurassic dinosaurs, Part VII. On the
Diplodocidae, a new family of the Sauropoda. Am. J. Sci. (3) 27:161–167.
Martin, J. (1987). Mobility and feeding of Cetiosaurus (Saurischia: Sauropoda) — why the
long neck? In Fourth Symposium on Mesozoic Terrestrial Ecosystems, Short Papers, ed. P.
J. Currie and E. H. Koster, pp. 154–159. Drumheller: Tyrrell Museum of
Palaeontology.
Mateer, N. J., and McIntosh, J. S. (1985). A new reconstruction of the skull of Euhelopus
zdanskyi (Saurischia, Sauropoda). Bull. Geol. Inst. Univ. Uppsala 11:125–132.
McIntosh, J. S., and Berman, D. S. (1975). Description of the palate and lower jaw of the
sauropod dinosaur Diplodocus (Reptilia, Saurischia) with remarks on the nature
of the skull of Apatosaurus. J. Paleont. 49:187–199.
McIntosh, J. S., Miller, W. E., Stadtman, K. L., and Gillette, D. D. (1996). The osteology of
Camarasaurus lewisi (Jensen, 1988). Brigham Young Univ. Geol. Stud. 41:73–115.
Norman, D. B. (1985). The Illustrated Encyclopaedia of Dinosaurs. London: Salamander
Books Ltd.
Nowinski, A. (1971). Nemegtosaurus mongoliensis n. gen., n. sp. (Sauropoda) from the
uppermost Cretaceous of Mongolia. Palaeont. Pol. 25:57–81.
Osborn, H. F., and Mook, C. C. (1921). Camarasaurus, Amphicoelias and other sauropods of
Cope. Mem. Am. Mus. Nat. Hist. 3: 247–287.
Owen, R. (1841). A description of a portion of the skeleton of the Cetiosaurus, a gigantic
extinct saurian reptile occurring in the oolitic formations of different portions
of England. Proc. Geol. Soc. London 3, Pt. II, No. 80:457–462.
Owen, R. (1842). Report on British fossil reptiles. Pt. II. Rep. Brit. Ass. Adv. Sci. 11:60–204.
Paul, G. S. (1988). The brachiosaur giants of the Morrison and Tendaguru with a
description of a new subgenus, Giraffatitan, and a comparison of the world’s
largest dinosaurs. Hunteria 2(3):1–14.
Phillips, J. (1871). Geology of Oxford and the Valley of the Thames. Oxford.
Powell, J. E. (1987). Morfológia del esqueleto axial del los dinosaurios titanosauridos
(Saurischia, Sauropoda) del estado de Minas Gerais, Brasil. An. 10 Congr. Brasileiro
Paleont. 155–171.
Powell, J. E. (1992). Osteologia de Saltasaurus loricatus (Sauropoda-Titanosauridae) del
Cretácico Superior des Noroeste Argentino. In Los Dinosaurios y su entorno biotico,
ed. J. L. Sanz, and A. D. Buscalioni, pp. 165–230. Act. Seg. Curso Paleontol. en Cuenca,
Instituto ‘Juan de Valdes.’
Russell, D. A., and Zheng, Z. (1994). A large mamenchisaurid from the Junggar Basin,
Xinjiang, People’s Republic of China. Can. J. Earth Sci. 30:2082–2095.
Salgado, L., and Bonaparte, J. F. (1991). Un nuevo sauropodo Dicraeosauridae,
Amargasaurus cazaui gen. et. sp. nov., de la Formacion la Amarga, Neocomiano de
la Provincia del Neuquen, Argentina. Ameghiniana 28:333–346.
Salgado, L., and Calvo, J. O. (1992). Cranial osteology of Amargasaurus cazaui Salgado and
Bonaparte (Sauropoda, Dicraeosauridae) from the Neocomian of Patagonia.
Ameghiniana 29:337–346.
Salgado, L., and Calvo, J. O. (1997). Evolution of titanosaurid sauropods. II: The cranial
evidence. Ameghiniana 34:33–48.
Salgado, L., Coria, R. A., and Calvo, J. O. (1997). Evolution of titanosaurid sauropods. I:
Phylogenetic analysis based on the postcranial evidence. Ameghiniana 34:3–32.
Sereno, P. C., Dutheil, D. B., Larochene, M., Larsson, H. C. E., Lyon, G. H., Magwene, P.
M., Sidor, C. A., Varricchio, D. J., and Wilson, J. A. (1996). Predatory dinosaurs
from the Sahara and Late Cretaceous faunal differentiation. Science 272:986–991.
Swofford, D. L. (1993). PAUP: Phylogenetic Analysis Using Parsimony, Version 3.1.1.
Champaign, IL: Natural History Survey.
Upchurch, P. (1993). The Anatomy, Phylogeny, and Systematics of the Sauropod Dinosaurs.
Unpublished Ph.D. Dissertation, University of Cambridge.
Upchurch, P. (1994). Sauropod phylogeny and palaeoecology. GAIA 10:249–260.
Upchurch, P. (1995). The evolutionary history of sauropod dinosaurs. Phil. Trans. R. Soc.
Lond. B 349:365–390.

Soc. 124:43–103.
Upchurch, P. (1999). The phylogenetic relationships of the Nemegtosauridae. J. Vert.
Paleont. 19(1):106–125.
Weishampel, D. B. (1990). Dinosaurian distribution. In The Dinosauria, ed. D. B.
California Press.
White, T. E. (1958). The braincase of Camarasaurus lentus Marsh. J. Paleont. 32:477–494.
Wilson, J. A., and Sereno, P. C. (1994). Higher-level phylogeny of sauropod dinosaurs. J.
Vert. Paleont. 14(suppl. to 3):52A.
Wiman, C. (1929). Die Kreide-Dinosaurier aus Shantung. Paleontol. Sin. C 6:1–67.
Yadagiri, P. (1988). A new sauropod Kotasaurus yamanpalliensis from Lower Jurassic Kota
Formation of India. Rec. Geol. Surv. India 11:102–127.
Young, C. C., and Chao, H. C. (1972). [Mamenchisaurus hochuanensis sp. nov.] Inst. Vert.
Paleontol. Paleoanthropol. Monogr. (Series A) 8:1–30 (in Chinese).
Zhang, Y. (1988). [The Middle Jurassic dinosaur fauna from Dashanpu, Zigong, Sichuan:
Sauropod Dinosaurs (1) Shunosaurus.] Chengdu: Sichuan Publishing House of
Science and Technology (in Chinese).
Zheng, Z. (1991). [Morphology of the braincase of Shunosaurus.] Vert. PalAsiat. 29:108–118
(in Chinese with English summary).
david b. weishampel and coralia-maria jianu
Plant-eaters and ghost lineages: dinosaurian

herbivory revisited
Introduction
Biotic interactions, whether they are in the form of hosts and par-
asites, flowers and pollinators, or seeds and dispersers, are often consid-
ered the sine qua non of coevolution, a process that some have argued has
an overarching influence on the structure of life on Earth (Futuyma and
Slatkin 1983; Margulis and Fester 1991; Thompson 1994). Whether true
or not, the identification of coevolutionary relationships can be rather
difficult from a paleontological perspective, particularly when consider-
ing groups of extinct organisms well removed from those of the present
day.
Several attempts have been made to understand plant–herbivore
interactions during the Mesozoic, a time when plant-eaters interacted
with a mixture of ‘pteridophyte’, ‘gymnosperm’, and angiosperm plants.
Virtually all of these studies have begun by grouping co-occurring faunas
and floras (e.g., Benton 1984; Hotton 1986; Tiffney 1986; Farlow 1987; Coe
et al. 1987; Fleming and Lips 1991). Co-occurrence is either in terms of taxa
that may have associated with one another or the features that reflect this
plant–herbivore interaction (tooth shape, foliage type, etc.). In more
expanded form, these groupings are then examined in light of their
paleobiogeographic and temporal distributions to understand their coev-
olutionary patterns (e.g., Bakker 1978; Wing and Tiffney 1987).
The discovery of patterns of ecological interactions and/or coevolution
between plants and herbivores during the Mesozoic is particularly
important in view of the great radiation of angiosperms beginning in the
latter half of the Early Cretaceous and extending through the end of that
period (Doyle and Donoghue 1986; Crane 1989). In view of these changes
[123]
124 david b. weishampel and coralia-maria jianu
in the plant realm, not only in diversity, but also in physiognomy and life
histories (Crane 1987; Upchurch and Wolfe 1987), contemporary herbi-
vores surely confronted new feeding opportunities and perhaps prob-
lems associated with their digestion. Consequently, it is important to
evaluate changes in taxonomic diversity and feeding systems among con-
temporary primary consumers of these angiosperms and other plants.
An early attempt to identify such changes was made by Weishampel
and Norman (1989); this study examined the evolutionary dynamics of
herbivorous tetrapods across the Mesozoic. This work used species
of a variety of monophyletic higher taxa, including Procolophonoidea,
Synapsida (e.g., Dicynodontia, Tritylodontoidea, and Multituberculata),
Rhynchosauria, Aetosauria (Stagonolepididae), and Dinosauria and,
while each group was treated as monophyletic, there was no attempt to
incorporate more detailed phylogenetic information beyond combining
species into higher taxa. Inferences were made about the feeding behav-
ior of members of these clades on the basis of body size, tooth structure,
and jaw mechanics, among other anatomical features (see below).
From these kinds of data (taxon-based and function-based), Weisham-
pel and Norman (1989) estimated speciation, extinction, and species-
based net-profit-and-loss rates. These metrics were compared with
similar dynamics among terrestrial plants as documented by Niklas,
Tiffney and Knoll (1980, 1985), in order to identify the evolutionary
dynamics among these herbivores and plants through the Mesozoic. For
example, many ornithischians — the most taxonomically and functionally
diverse group among all Mesozoic herbivorous tetrapods — appear to
show a rise in species origination toward the end of the Early Cretaceous
and another increase toward the end of the Late Cretaceous, a pattern
which was interpreted as roughly mirroring the radiation of early angio-
sperms.
If this ornithischian pattern proves to be real, then it suggests that
ornithischians may have had a close adaptive if not coevolutionary con-
nection with emerging angiosperms. Expressed in another way, might
angiosperms have been a kind of forcing factor in the trophic diversifica-
tion of ornithischians? In order to answer this question, we need as
complete as possible a census of ornithischians and other Mesozoic herbi-
vores. The fossil record provides the ultimate basis for this census (i.e.,
raw species counts), but because taphonomic biases are unpredictable, it
is unclear what these numbers may mean. One way to approach a more
accurate census is to include diversity data that come not only from the
Dinosaurian herbivory revisited 125
Figure 5.1. Phylogenetic relationships of the major clades within Ornithischia

(redrawn from Sereno 1986).
actual census of Mesozoic herbivores and their feeding styles, but also
from the identification of ghost lineages.
Taxa and feeding groups
In order to examine the relationship between taxa known directly from

the fossil record and the ghost lineages inferred from them on estimates
of diversity among Mesozoic herbivores, we have concentrated on the two
major clades of dinosaurian herbivores, Ornithischia and Sauropodomor-
pha. These clades are well represented throughout the Mesozoic, species
diversity is often very high, and phylogenetic analyses of most of the taxa
within these clades are now emerging.
Five major clades are known within Ornithischia (Figure 5.1), with
additional clusters of species intercalated among them. The first major
ornithischian clade under consideration is Thyreophora, consisting of
Ankylosauria and Stegosauria, with several individual stem species. The
Ankylosauria (Figure 5.2a) consist of quadrupedal, heavily armored dino-
saurs, which are known primarily from the Cretaceous (although from as
early as the Middle Jurassic) of North America, Europe, Asia and Australia.
Ranging in length from 5 m to 9 m, these herbivores browsed within 1–2 m
off the ground and used the cutting edges of their scoop-shaped beak to
crop foliage and perhaps fructifications to be delivered to the small,
simple, triangular teeth that lined the jaws (Coombs and Maryanska 1990).
a)
b)
Figure 5.2. Thyreophora. a) Skull of the nodosaurid ankylosaur Panoplosaurus mirus

(without lower jaw). b) Skull of the stegosaur Stegosaurus stenops. Scale bar
equals 10 cm.
These teeth bear prominent denticles and are somewhat separated from
each other; chewing appears to have taken place by simple jaw closure (i.e.,
an orthal power stroke). Weishampel and Norman (1989) identified this
kind of dentition as indicative of puncture-crushing mastication. Tooth
wear is often present, but not systematically developed. The oral cavity
was bound by muscular cheeks (as in all ornithischians except Lesothosau-
rus diagnosticus), and there may have been a long flexible tongue. The enor-
mous abdominal region suggests a large fermentation compartment.
The other major thyreophoran clade, Stegosauria (Figure 5.2b), is
known primarily from the Middle Jurassic through the Early Cretaceous
of Europe, Asia, North America and Africa. These heavily built, 3–9 m
long quadrupeds bore a variety of spines and/or plates along the ridge of
their back. Stegosaurs had relatively small but elongate heads with
simple, spatulate, denticulate teeth with relatively large gaps between
teeth. These teeth occasionally display obliquely inclined wear surfaces,
suggesting at least some sort of orthal puncture crushing like that seen in
Figure 5.3. Ornithopoda. Skull of the hadrosaurid Edmontosaurus regalis. Scale bar
equals 20 cm.
ankylosaurs (Galton 1985, 1990). Browsing was likely within a meter

above the ground although it is possible that some stegosaurs may have
been able to browse at higher (2–3 m) levels by assuming a tripodal
posture.
Ornithopoda (Figure 5.3), the longest-lived ornithischian clade (Early
Jurassic through the end of the Cretaceous), is known worldwide.
Ranging in length from 1 m to 12 m, these typically bipedal herbivores
had closely packed teeth with well-developed medium-angle shear sur-
faces that extended continuously along the length of the dentition. Con-
sequently, ornithopods can be characterized as having a transverse
grinding style of mastication (Weishampel 1984; Norman 1984; Weisham-
pel and Witmer 1990; Sues and Norman 1990; Norman and Weishampel
1990; Weishampel and Horner 1990). Browsing appears to have been con-
centrated within 1–2 meters off the ground, although in the large iguano-
dontians (e.g., Iguanodon, most hadrosaurids) it may have extended up to
4 m.
Known nearly exclusively from the Cretaceous from the Northern
Hemisphere, ceratopsians (Figure 5.4a) were mostly quadrupedal herbi-
vores. These so-called horned dinosaurs ranged in length from 2 m to 9 m
and browsed no more than 2 m above the ground. The front of the jaws
was equipped with a hooked parrot-like beak and the cheek teeth were
closely packed, worn into a continuous vertical chewing surface (Ostrom
1964; Dodson and Currie 1990). Only in species of Psittacosaurus is the
occlusal surface worn at a lower angle (approximately 60°) and this basal
ceratopsian is known to possess gastroliths (Sereno 1990).
Pachycephalosaurs (Figure 5.4b), like their marginocephalian sister-
group Ceratopsia, are known only from the Cretaceous of the Northern
Hemisphere. Ranging from 2 m to 5 m in length, these domed-headed
a)
b)
Figure 5.4. Marginocephalia. a) Skull of the ceratopsid ceratopsian Centrosaurus

apertus. Scale bar equals 20 cm. b) Skull of the pachycephalosaur Stegoceras
validum. Scale bar equals 5 cm.
ornithischians were foragers on low-growing plants (1 m); the small, tri-

angular, denticulate teeth had relatively large gaps between them and an
orthal power stroke (Maryanska 1990). Finally, the large abdominal
region suggests the presence of a large fermentation chamber.
The other great clade of dinosaurian herbivores, Sauropodomorpha
(Figure 5.5), is known worldwide from the Late Triassic through the end
of the Cretaceous. The major clades of sauropodomorphs include prosau-
ropods, euhelopodids, camarasaurids, brachiosaurids, titanosauroids,
and diplodocoids, with other taxa interpolated among them (Upchurch
1995, 1998). Although new research suggests that many were clearly not
restricted to plant-eating (Barrett, this volume), sauropodomorphs
ranged in length from 2.5 m to perhaps as much as 30 m (McIntosh 1990;
Dodson 1990a). The dentition consisted of relatively widely spaced,
Figure 5.5. Phylogenetic relationships of the major clades within Sauropodomorpha

(redrawn from Upchurch 1995, 1998).
Figure 5.6. Sauropodomorpha. Skull of the diplodocid Diplodocus longus. Scale bar
equals 20 cm.
simple peg-like or spatulate teeth often restricted to the front of the snout
(Figure 5.6; Calvo 1994; Barrett and Upchurch 1994). Tooth wear is known
in several sauropodomorphs (Fiorillo 1991) and indicates that foliage was
stripped from branches but not extensively chewed. Mechanical break-
down of foliage instead appears to have been accomplished by gastroliths
in a muscular gizzard. Browsing was at high levels (perhaps up to 15 m).
Thus we can characterize the various clades of ornithischians and sau-
ropodomorphs in the following way (Weishampel and Norman 1989):
stegosaurs, ankylosaurs, and pachycephalosaurs are envisioned as orthal
pulpers (denticulate crowns, large gaps between teeth, and an orthal
power stroke), ceratopsians excluding Psittacosaurus spp. as orthal slicers
(closely packed teeth and a vertical power stroke that produced high-angle
wear facets), ornithopods and probably Psittacosaurus spp. as transverse

grinders (closely packed teeth that moved transversely against one
another to produce medium-angle tooth wear), and sauropodomorphs as
gut processors (simple teeth, little tooth wear, and gastroliths).
Ghost lineages
Norell and Novacek (1993) described ghost lineages as those portions of

the history of a taxon for which there is no direct fossil record. These
ghost lineages logically come from the phylogenetic continuity between
ancestors and paired descendants at cladogenesis and can be identified
through the stratigraphic calibration of phylogeny. The older of the two
resulting species provides evidence that some of the history leading to the
sister species is missing from the stratigraphic record.
Ghost lineages are most accurately identified when the phylogeny of
the group of interest is well understood and the stratigraphic occur-
rence of the descendant sibling species is relatively precise. For the
groups under consideration here (ornithischians and sauropodo-
morphs), both phylogeny and stratigraphic distribution are reasonably
well known or at least the focus of current research. Consequently, it is
possible to include information not only on raw species diversity but
also on ghost lineage occurrences to increase the quality of our species
census. In this way, we will be better able to test the relationship
between the evolution of dinosaurian herbivory and the radiation of
angiosperms.
Raw and ghost diversity patterns
Because they were the most diverse and abundant among Mesozoic herbi-
vores and because their phylogeny is reasonably well understood, orni-
thischians and sauropodomorphs are the best subjects for analyzing
possible adaptive and/or coevolutionary relationships with contem-
porary plants. The species-level cladogram for Ornithischia used in this
study is based in part on Sereno (1986) and includes information from
Sues and Galton (1987), Coombs and Maryanska (1990), Sereno and Dong
(1992), Weishampel, Norman and Grigorescu (1993), Forster et al. (1993),
Sampson (1995), Weishampel (1996), and Forster (1996). Resolution of
species on this cladogram are generally unproblematic as most genera are
monospecific. Where multispecific genera are encountered, they are
positioned as unresolved sister-taxa, sometimes with a star-burst effect.

The total number of species on this ornithischian cladogram is nearly 140,
ranging in age from the Early Jurassic (Hettangian–?Sinemurian)
through the end of the Cretaceous (late Maastrichtian; Weishampel
1990a). The species-level cladogram is available from the authors on
request.
The species-level cladogram for Sauropodomorpha is more proble-
matic, and we rely on Galton (1990) and Upchurch (1995, 1998) with inter-
polation of additional species from McIntosh (1990). As in ornithischians,
single species are most commonly allotted to each sauropodomorph
genus. The total number of sauropodomorph species on this cladogram is
nearly 100, ranging in age from the Late Triassic (Norian) through the end
of the Cretaceous (late Maastrichtian; Weishampel 1990a). The species-
level sauropodomorph cladogram is available from the authors on
request.
These phylogenetic relationships are then calibrated against the earli-
est occurrence of each species. Ghost lineages of both ornithischians and
sauropodomorphs are rather scattered throughout each clade, but are an
integral part of all higher taxa. Figure 5.7 provides a detail taken from
Iguanodontia, one of the most diverse clades of Ornithopoda.
In order to assess the relationship of ghost lineages to estimates of
ornithischian diversity, we sampled actual species and their ghost lin-
eages at 2.5-million-year intervals from the end of the Cretaceous back
through to the earliest occurrence of both major clades of dinosaurian
herbivores. This 2.5-million-year interval was chosen in order to insure
that all species as well as their ghost lineages will be sampled in view of
Dodson’s (1990b) calculation that dinosaurian species have a temporal
range of approximately 5 million years.
The raw count of ornithischian species (Figure 5.8) reveals a modest
diversity peak in the Early Jurassic, a second much larger series of peaks in
the Late Jurassic, another peak at the end of the Early Cretaceous, and
finally the largest peak of the series at the end of the Cretaceous, much as
was identified by Weishampel and Norman (1989). Among sauropodo-
morphs, raw diversity levels are much more modest than for ornithis-
chians. There is a peak in the Late Jurassic, another less substantial peak
in the mid-Cretaceous, and another just prior to the end of the Cretaceous
(Figure 5.9).
In contrast, when the phylogeny of these same species is calibrated
against stratigraphy, the pattern of ornithischian ghost lineages indicates
Figure 5.7. Ghost-lineage diagram for Iguanodontia. The circles represent the stratigraphic occurrences of iguanodontian species (Tenontosaurus
tilletti, Ouranosaurus nigeriensis, and Telmatosaurus transsylvanicus are indicated). The vertical lines represent relationships among species, and the
striped horizontal lines constitute ghost lineages.
20
15
10
0
⫺70
⫺80
⫺90
⫺100
⫺110
⫺120
⫺130
⫺140
⫺150
⫺160
⫺170
⫺180
⫺190
⫺200
⫺210
⫺220
30
20
10
0
⫺70
⫺80
⫺90
⫺100
⫺110
⫺120
⫺130
⫺140
⫺150
⫺160
⫺170
⫺180
⫺190
⫺200
⫺210
⫺220
Figure 5.8. Ornithischian diversity through the Mesozoic, based on raw species data
from the fossil record (above) and based on ghost-lineage data (below).
a more-or-less gradual rise in diversity from the Early to Late Jurassic

(represented principally by stegosaurs), followed by an abrupt increase in
the Late Jurassic (principally driven by stegosaurs and basal iguanodon-
tians), followed by a reduction in diversity that levels across the mid-Cre-
taceous (broadly distributed among all ornithischian clades except
stegosaurs), followed by the greatest diversity increase at the end of the
Cretaceous, in which the summed species diversity is more than twice
8
6
4
2
0
⫺70
⫺80
⫺90
⫺100
⫺110
⫺120
⫺130
⫺140
⫺150
⫺160
⫺170
⫺180
⫺190
⫺200
⫺210
40
30
20
10
0
⫺70
⫺80
⫺90
⫺100
⫺110
⫺120
⫺130
⫺140
⫺150
⫺160
⫺170
⫺180
⫺190
⫺200
⫺210
Figure 5.9. Sauropodomorph diversity through the Mesozoic, based on raw species
data from the fossil record (above) and based on ghost-lineage data (below).
that since the Late Jurassic and an order of magnitude greater than before
the Late Jurassic (Figure 5.8). This latter is a consequence of high levels of
cladogenesis among hadrosaurid ornithopods and ceratopsid ceratop-
sians. For sauropodomorphs (Figure 5.9), there is a small peak at the end
of the Triassic, another slightly higher peak at the beginning of the Juras-
sic (together based on cladogenesis of prosauropods), followed by a
decline through most of the Middle Jurassic, then the greatest peak at the
end of the Jurassic, during which time diversity jumped approximately
sevenfold (driven by increased cladogenesis in brachiosaurids, camara-
saurids, titanosauroids and diplodocoids). Thereafter, diversity levels
gradually decline through the end of the Mesozoic, with a final abrupt
decline from the Campanian through the Maastrichtian.
Feeding group diversity
However interesting these taxonomic data may be, it is really the diver-
sity of feeding groups that is ultimately useful in assessing herbi-
vore–plant interactions and/or coevolution. Consequently, we have taken
the feeding styles that have been identified among known species and
deduced ‘ghost feeding styles’ from their ghost lineages. As described pre-
viously, these feeding styles include orthal pulping (Lesothosaurus, thyre-
ophorans, pachycephalosaurs), orthal slicing (nearly all ceratopsians),
transverse grinding (ornithopods, Psittacosaurus), and gut processing (sau-
ropodomorphs, Psittacosaurus). Using these categories, ornithischian and
sauropodomorph feeding styles were mapped and optimized onto the
cladogram for Ornithischia using the DELTRAN optimization option of
PAUP (Swofford 1985) in order to produce the minimal resolution of
feeding styles on the tree given the available data. Because of the stric-
tures of DELTRAN and lack of appropriate information from the fossils
themselves, we also have ghost lineages with unresolved feeding styles.
Among others, this includes the feeding style of the ghost lineage leading
from the base of Ornithischia to Neoceratopsia.
Sampling of optimized feeding styles was again at the same 2.5 million
years as in the ghost lineage analyses described previously.
Ghost-lineage diversity of these feeding styles is indicated in Figure
5.10. From the end of the Late Triassic until the Late Jurassic, gut proces-
sors, orthal pulpers, and transverse grinders existed at relatively low
diversity, although gut processors were at least twice as diverse as the
other two feeding styles. All three groups show an abrupt diversity
increase in the Late Jurassic, gut processors preceding orthal pulpers and
transverse grinders by approximately 10 million years. Late Jurassic diver-
sity of gut processors is twice that of both other groups combined. There-
after, diversity levels decreased among all three groups, throughout the
remainder of the Mesozoic in gut processors, and through the middle of
the Early Cretaceous in orthal pulpers and transverse grinders. Orthal
pulping then appears to have undergone a second wave of diversification
in the Late Cretaceous. In contrast, transverse grinders have a modest
peak at the end of the Early Cretaceous, followed by a decline through the
first half of the Late Cretaceous and a final rise at the end of the Mesozoic.
The last feeding style, orthal slicing, is restricted to the Late Cretaceous.
This short-lived mode of oral processing goes from a diversity of zero to
eight species to zero again in the course of 25 million years.
Figure 5.10. Feeding style diversity among herbivorous dinosaurs through the Mesozoic, based on ghost-lineage data.
In order to reduce the effects of different sample sizes and the patchi-
ness of certain sampling intervals, we have also transformed these data
into percentage contributions of feeding styles across these same inter-
vals (Figure 5.11). The Late Triassic began solely with gut processors,
which gradually gave way to orthal pulpers and transverse grinders
through the Early Jurassic. From the Middle through the Late Jurassic,
gut processors reverse this pattern, with a concomitant percentage
decrease in orthal pulpers and transverse grinders. Transverse grinders
maintain a fairly stable diversity percentage (approximately 20–30%)
through the Late Cretaceous. At the same time, orthal pulpers show a
modest increase in diversity in the mid-Cretaceous. In contrast, gut pro-
cessors steadily decline from the Early Cretaceous onward. Finally, the
rise in orthal slicers in the Late Cretaceous contributes a large percent
diversity increase, apparently involving reciprocal percent declines in
orthal pulpers and gut processors but not transverse grinders.
Discussion
The use of ghost lineage diversity, rather than raw species counts, pro-
vides a more accurate picture of minimal diversity levels at any time in the
geological past. Due to the phylogenetic continuity of species within
clades, it is possible to deduce minimal levels of species diversity — as well
as feeding styles — beyond those species physically available to us from the
fossil record. Without this additional ghost-lineage-based information,
raw species counts and evolutionary metrics derived from them can
provide only a partial picture of changes in herbivore diversity across the
Mesozoic. For example, raw species counts from Ornithischia underrep-
resent minimal diversity levels overall by 75% and by as much as 95% for
individual intervals (Figure 5.8). Even at their best, these raw counts do
only as well as 50% representation when sample size is low and the clade is
young.
Incorporation of information from ghost lineages affects not only rela-
tive diversity levels at any given moment, but, as a consequence, also
affects overall patterns of change within a given clade. Norman and Wei-
shampel (1989) indicated three separate intervals of diversification among
transverse grinders (as represented by ‘advanced ornithopods,’ now
referred to as euornithopods; Weishampel 1990b), with the middle peak
arguably associated with the origin of angiosperms. Based on ghost-
lineage diversity, however, this middle peak virtually disappears. In its
Figure 5.11. Feeding style diversity among herbivorous dinosaurs through the Mesozoic, expressed as percentages.
place, there are only modest fluctuations in diversity that intercede

between the extreme Late Jurassic and Late Cretaceous peaks.
Given these revised metrics, what is the relationship of the Early Creta-
ceous origin of angiosperms to changes in feeding styles? Species diver-
sity of both orthal pulpers and perhaps transverse grinders based on
ghost lineages increases only modestly if at all through this time when
angiosperms appear to have begun their initial radiation. This pattern
exists in both untransformed and percentage analyses (cf. Figures
5.10–5.11). If we regard diversity patterns as a signal for important evolu-
tionary events linking plants and herbivores, then fluctuations in ghost-
lineage diversity are less than spectacular, given the major radiation
among angiosperms. Instead, it is during the Late Cretaceous that diver-
sity is radically altered, in greatest part due to the origin and diversifica-
tion of orthal slicers, indicated in both ghost diversity counts and
percentages, but also increases in diversity among transverse grinders
and slightly less so among orthal pulpers, indicated in ghost diversity
counts but not percentages.
Whatever coevolutionary interactions may have existed between flow-
ering plants (angiosperms) and ornithischian dinosaurs, they are not
apparent at the onset of the evolutionary radiation of angiosperms, as
least as far as diversity patterns are concerned. Whether coevolutionary
links existed, but lagged behind the Early Cretaceous diversification of
flowering plants is equally unclear. One might imagine that the Late Cre-
taceous diversification of orthal pulpers, transverse grinders, and orthal
slicers (based on untransformed ghost diversity) were reflected in changes
in Late Cretaceous angiosperms, such as alterations in their physiog-
nomy, as per shifts in stature, or in threshold effects such as floral domi-
nance or shifting biogeographic realms (but see Wing, Hickey and
Swisher 1993). Conversely, changes in food resources provided by other
tracheophytes might explain these diversity patterns among contempo-
rary herbivores. However, these are merely suggestions and at present
nothing explicitly links Mesozoic plants and herbivores in a coevolution-
ary way beyond these untested scenarios.
What we have presented here is, in part, a revision of earlier work on
the trophic diversity of herbivorous dinosaurs (Weishampel and Norman
1989), as well as an advocacy of the use of phylogeny to provide additional
information on such diversity. Ornithischians and sauropodomorphs
appear to have been little affected by the initial radiation of angiosperms,
but whether they took advantage — in an evolutionary sense — of changes
in the distribution of these plants at the end of the Cretaceous is unclear.

However, it is these Late Cretaceous patterns that need further investiga-
tion — from the point of view of both diversity fluctuations and also the
ebb and flow of features that relate to feeding in herbivores and resistance
to predation by plants.
Finally, species occurrences in the fossil record imply, via their phylog-
eny, unseen aspects of diversity that are reflected in their ghost lineages.
Thus ghost lineages provide information on clade diversity beyond that
available from raw species counts and help provide a clearer picture of
minimal diversity levels, in this case among dinosaurian herbivores.
Acknowledgments
We thank H.-D. Sues for inviting us to contribute to this volume. We

also thank L. M. Witmer, M. B. Meers, R. E. Heinrich, Z. Csiki, and M. A.
Norell for discussions and comments on the ideas contained in this
study.
At one stage or another, this research was supported by grants from the
National Science Foundation (INT-8619987, EAR-8719878, EAR-9004458,
SBR-9514784), the National Geographic Society, the National Research
Council, and The Dinosaur Society.
References
Bakker, R. T. (1978). Dinosaur feeding behaviour and the origin of flowering plants.
Nature 274:661–663.
Barrett, P. M. and Upchurch, P. (1994). Feeding mechanics of Diplodocus. GAIA
10:195–204.
Benton, M. J. (1984). Tooth form, growth, and function in Triassic rhynchosaurs
(Reptilia: Diapsida). Palaeontology 27:737–776.
Calvo, J. O. (1994). Jaw mechanics in sauropod dinosaurs. GAIA 10:183–194.
Coe, M. J., Dilcher, D. L., Farlow, J. O., Jarzen, D. M., and Russell, D. A. (1987). Dinosaurs
and land plants. In The Origins of Angiosperms and Their Biological Consequences, ed.
E. M. Friis, W. G. Chaloner, and P. R. Crane, pp. 225–258. Cambridge and New
Coombs, W. P., and Maryanska, T. (1990). Ankylosauria. In The Dinosauria, ed. D. B.
California Press.
Crane, P. R. (1987). Vegetational consequences of angiosperm diversification. In The
Origins of Angiosperms and Their Biological Consequences, ed. E. M. Friis, W. G.
Chaloner, and P. R. Crane, pp. 107–144. Cambridge and New York: Cambridge
University Press.
Crane, P. R. (1989). Patterns of evolution and extinction in vascular plants. In Evolution

and the Fossil Record, ed. K. C. Allen and D. E. G. Briggs, pp. 153–187. Washington,
DC: Smithsonian Institution Press.
Dodson, P. (1990a). Sauropod paleoecology. In The Dinosauria, ed. D. B. Weishampel, P.
Dodson, P. (1990b). Counting dinosaurs: how many kinds were there? Proc. Natl. Acad. Sci.
U.S.A. 87:7608–7612.
Dodson, P., and Currie, P. J. (1990). Neoceratopsia. In The Dinosauria, ed. D. B.
California Press.
Doyle, J. A., and Donoghue, M. J. (1986). Seed plant phylogeny and the origin of
angiosperms: an experimental cladistic approach. Bot. Rev. 52:321–431.
Fiorillo, A. R. (1991). Dental microwear on the teeth of Camarasaurus and Diplodocus:
implications for sauropod paleoecology. In Fifth Symposium on Mesozoic Terrestrial
Ecosystems and Biota, ed. Z. Kielan-Jaworowska, N. Heintz, and H. A. Nakrem, pp.
23–24. Oslo: University of Oslo.
Fleming, T. H., and Lips, K. R. (1991). Angiosperm endozoochory: were pterosaurs
Cretaceous seed dispersers? Am. Nat. 138:1058–1065.
Forster, C. A. (1996). Species resolution in Triceratops: cladistic and morphometric
approaches. J. Vert. Paleont. 16:259–270.
Forster, C. A., Sereno, P. C., Evans, T. W., and Rowe, T. (1993). A complete skull of
Chasmosaurus mariscalensis (Dinosauria: Ceratopsidae) from the Aguja Formation
(late Campanian) of west Texas. J. Vert. Paleont. 13:161–170.
Futuyma, D. J., and Slatkin, M. (eds.) (1983). Coevolution. Sunderland, MA: Sinauer.
Galton, P. M. (1985). Diet of prosauropod dinosaurs from the Late Triassic and Early
Jurassic. Lethaia 6:67–89.
Galton, P. M. (1990). Basal Sauropodomorpha – prosauropods. In The Dinosauria, ed. D.
B. Weishampel, P. Dodson, and H. Osmólska, pp. 320–344. Berkeley: University
of California Press.
Hotton, N., III. (1986). Dicynodonts and their role as primary consumers. In The Ecology
and Biology of Mammal-Like Reptiles, ed. N. Hotton, III, P. D. MacLean, J. J. Roth,
and E. C. Roth, 71–82. Washington, DC: Smithsonian Institution Press.
Margulis, L., and Fester, R. (eds.) (1991). Symbiosis as Source of Evolutionary Innovation.
Cambridge, MA: MIT Press.
Maryanska, T. (1990). Pachycephalosauria. In The Dinosauria, ed. D. B. Weishampel, P.
Niklas, K. J., Tiffney, B. H., and Knoll, A. H. (1980). Apparent changes in the diversity of
fossil plants. Evol. Biol. 12:1–89.
Niklas, K. J., Tiffney, B. H., and Knoll, A. H. (1985). Patterns in vascular land plant
diversification: an analysis at the species level. In Phanerozoic Diversity Patterns,
ed. J. W. Valentine, pp. 97–128. Princeton: Princeton University Press.
Norell, M. A., and Novacek, M. J. (1993). Congruence between suprapositional and
phylogenetic patterns: comparing cladistic patterns with the fossil record.
Cladistics 8:319–337.
Norman, D. B. (1984). On the cranial morphology and evolution of ornithopod

dinosaurs. In The Structure, Development and Evcolution of Reptiles, ed. M. J. W.
Ferguson, pp. 521–547. Symp. Zool. Soc. Lond. 52.
Norman, D. B., and Weishampel, D. B. (1990). Iguanodontidae and related ornithopods.
In The Dinosauria, ed. D. B. Weishampel, P. Dodson, and H. Osmólska, pp.
510–533. Berkeley: University of California Press.
Ostrom, J. H. (1964). A functional analysis of jaw mechanics in the dinosaur Triceratops.
Postilla 88:1–35.
Sampson, S. D. (1995). Two new horned dinosaurs from the Upper Cretaceous Two
Medicine Formation of Montana, with a phylogenetic analysis of the
Centrosaurinae (Ornithischia: Ceratopsidae). J. Vert. Paleont. 15:743–760.
Sereno, P. C. (1986). Phylogeny of the bird-hipped dinosaurs (Order Ornithischia). Natl.
Geogr. Res. 2:234–256.
Sereno, P. C. (1990). Psittacosauridae. In The Dinosauria, ed. D. B. Weishampel, P. Dodson,
Sereno, P. C., and Dong, Z. (1992). The skull of the basal stegosaur Huayangosaurus taibaii
and a cladistic analysis of Stegosauria. J. Vert. Paleont. 12:318–343.
Sues, H.-D., and Galton, P. M. (1987). Anatomy and classification of the North American
Pachycephalosauria (Dinosauria: Ornithischia). Palaeontographica A 198:1–40.
Sues, H.-D., and Norman, D. B. (1990). Hypsilophodontidae, Tenontosaurus,
Dryosauridae. In The Dinosauria, ed. D. B. Weishampel, P. Dodson, and H.
Osmólska, pp. 498–509. Berkeley: University of California Press.
Swofford, D. L. (1985). PAUP: Phylogenetic Analysis Using Parsimony. Champaign, IL:
Illinois Natural History Survey.
Thompson, J. N. (1994). The Coevolutionary Process. Chicago: University of Chicago Press.
Tiffney, B. H. (1986). Evolution of seed dispersal syndromes according to the fossil
record. In Seed Dispersal, ed. D. R. Murray, pp. 273–305. North Ryde, NSW:
Academic Press of Australia.
Upchurch, G. R., and Wolfe, J. A. (1987). Mid-Cretaceous – early Tertiary vegetation and
climate: evidence from fossil leaves and wood. In The Origins of Angiosperms and
Their Biological Consequences, ed. E. M. Friis, W. G. Chaloner, and P. R. Crane, pp.
75–105. Cambridge and New York: Cambridge University Press.
Upchurch, P. (1995). The evolutionary history of sauropod dinosaurs. Phil. Trans. R. Soc.
Lond. B 349: 365–390.
Soc. 124:43–103.
Weishampel, D. B. (1984). Evolution of jaw mechanisms in ornithopod dinosaurs. Adv.
Anat. Embryol. Cell Biol. 87: 1–110.
Weishampel, D. B. (1990a). Dinosaurian distributions. In The Dinosauria, ed. D. B.
California Press.
Weishampel, D. B. (1990b). Ornithopoda. In The Dinosauria, ed. D. B. Weishampel, P.
Weishampel, D. B. (1996). Fossils, phylogeny, and discovery: a cladistic study of the
history of tree topologies and ghost lineage durations. J. Vert. Paleont. 16:191–197.
Weishampel, D. B., and Horner, J. R. (1990). Hadrosauridae. In The Dinosauria, ed. D. B.
California Press.
Weishampel, D. B., and Norman, D. B. (1989). The evolution of occlusion and jaw
mechanics in late Paleozoic and Mesozoic herbivores. In Paleobiology of the
Dinosaurs, ed. J. O. Farlow, Jr., pp. 87–100. Geol. Soc. Am. Special Paper 238.
Weishampel, D. B., and Witmer, L. M. (1990). Heterodontosauridae. In The Dinosauria,
ed. D. B. Weishampel, P. Dodson, and H. Osmólska, pp. 486–497. Berkeley:
University of California Press.
Weishampel, D. B., Norman, D. B., and Grigorescu, D. (1993). Telmatosaurus
transsylvanicus from the Late Cretaceous of Romania: the most basal
hadrosaurid. Palaeontology 36:361–385.
Wing, S. L., and Tiffney, B. H. (1987). The reciprocal interaction of angiosperm evolution
and tetrapod herbivory. Rev. Palaeobot. Palynol. 50: 179–210.
Wing, S. L., Hickey, L. J., and Swisher, C. C. (1993). Implications of an exceptional fossil
flora for Late Cretaceous vegetation. Nature 363: 342–344.
john m. rensberger
Dental constraints in the early evolution of

mammalian herbivory
Introduction
The processing of plant food in mammals requires specialized

mechanisms in order to sustain high rates of nutrient assimilation, and
the evolution of herbivory in mammals can potentially be traced by exam-
ining changes in these structures. Conditions suitable for the preserva-
tion of soft tissues are exceedingly rare (Schaal and Ziegler 1993), and thus
we cannot hope to trace the evolution of herbivory using structures of the
stomach and intestinal tract. However, oral comminution of plant
material is a nearly universal behavior among mammalian herbivores,
owing to the problem of breaking down the plant cell walls composed of
cellulose (Janis and Fortelius 1988). Fortunately, dental structures special-
ized for this purpose are well represented in the fossil record. One
major group of Mesozoic mammals, the multituberculates, had already
acquired masticatory mechanisms allowing herbivory (Hahn 1971; Krause
1982). However, the main mammalian radiation that led to modern herbi-
vores probably occurred immediately following the extinction of the
non-avian dinosaurs at the end of the Cretaceous. That event must have
opened many niches that previously had been inaccessible to mammals,
judging from the high taxonomic and morphological rates of evolution
across the Cretaceous–Paleocene boundary (Sloan 1987).
This chapter examines the functional changes in the shapes of teeth in
the two dominant groups of early Paleocene ungulates in North America,
compares these with common Eocene ungulates, and considers the influ-
ence that tissue strength had on the directions of these changes. The
results indicate that the changes were non-linear, apparently because of
mechanical constraints that had to be overcome before the more efficient
mastication characteristic of modern ungulates emerged.
[144]
Teeth of early mammalian herbivores 145
Institutional abbreviations in this chapter are: MORV, Museum of

the Rockies (fossil vertebrate collection), Bozeman, MT; UWBM, Burke
Museum of Natural History and Culture, Seattle, WA. Other abbrevia-
tions: EDJ, enamel–dentine junction; SEM, scanning electron micro-
graph.
Tribosphenic mechanisms
The cheek teeth of the basal placental and marsupial insectivores that
gave rise to the Cenozoic carnivores and herbivores have a tribosphenic
cusp arrangement. Tribosphenic teeth have functional components that
are found in later Cenozoic herbivores but the implementation and effi-
ciencies of these functions are quite different.
Shearing mechanism The tribosphenic dentition consists of a

complex arrangement of relatively sharp cusps that function to fracture
food in several ways — through shearing, compressing or puncturing the
food particles. These teeth have cusps with steeply inclined sides that
slide past one another in a shearing movement during chewing. The teeth
of the arctocyonid Protungulatum, a very primitive ungulate that occurs
near the Cretaceous–Tertiary boundary (Sloan and Van Valen 1965), still
retain much of the tribosphenic geometry, with cusps of moderately high
aspect ratio and rather steeply inclined shearing surfaces (Figure 6.1). A
fundamental part of this mechanism is a series of crests that extend along
the sides of adjacent cusps (Figure 6.2). Like the cusps, the opposing
occlusal edges of the crests glide past one another in close proximity
during chewing movements (Figure 6.1).
These parts of the tribosphenic tooth, the cusp surfaces and crests, and
the arrangement of the jaw muscles that move them can be called a shear-
ing mechanism, but its effect on the food material is complex, producing
tensile and compressive as well as shearing stresses. During the initial
closing movement of the opposing tooth surfaces, the food is compressed.
However, as the crest edges pass one another, the opposing edges push the
food trapped between them in opposite directions, creating tensile and
shearing stresses within the material, although these stresses are difficult
to measure in the food.
If the food is brittle and not very fibrous, it may be fractured by the
external compressive load before the edges pass one another in shear.
When an object is compressed, tensile stresses are generated in directions
at right angles to the direction of the compressive stresses (Rensberger
146 john m. rensberger
Figure 6.1. Buccal view of occluding P2 –M3 and P£–M™ in the maxilla and dentary of
arctocyonid Protungulatum donnae (MORV 5000 and 5001) from stream channel
at top of the Upper Cretaceous Hell Creek Formation. Arrows indicate
directions of shearing movement between opposing premolars and between
crests of anterior side of upper molar and posterior side of lower molar
trigonid. Scale bar (lower right of P2) equals 2 mm.
1995: figs. 9.8, 9.16), and these tensile stresses may be high enough to
produce fracture. For example, if the object is a seed with a hard outer
shell and is subjected to vertical compression, vertical lines of fracture
caused by horizontal tensile stresses will tend to form as the seed’s diame-
ter expands.
Puncture-crushing mechanism Observations of chewing activity in

various living mammals (opossum, tree shrew, bat, primates, cat, and pig)
indicate that a distinct type of mechanism, puncture-crushing, is used
during early stages of processing a food object. The jaws are moved
together in a rapid series of bites in which the opposing cusps do not
reach full occlusion (Atkinson and Shepherd 1961; Ahlgren 1966; Cromp-
ton and Hiiemae 1970; Kallen and Gans 1972; Herring and Scapino 1973;
Hiiemae and Kay 1973; Luschei and Goodwin 1974; Hiiemae and Thexton
1975). Puncture-crushing movements utilize the height and sharpness of
the cusps, which open holes in the food material or fracture it if it is
brittle.
Figure 6.2. Occlusal views of (A) right P2–M3 (MORV 5000) and (B) right P£–M™
(MORV 5001) in Protungulatum donnae. Arrows point to crests on (A) anterior
side of paracone that shear against crests on (B) posterior side of trigonid. Scale
bar (occlusal view, lower right) equals 2 mm.
Compressive mechanism In addition to the compression generated by

the previously described shearing mechanism, tribosphenic teeth have a
structure specialized solely for compression, the talonid basin of the
lower molar that opposes the tip of the protocone of the upper molar. The
opposition of protocone tip against concave surface of the talonid basin
efficiently flattens food caught between these structures, bursting mem-
branes and cell walls.
Efficiency
A measure of efficiency of mechanisms for oral food processing is the

amount of increased food surface area produced per chewing stroke. The
increase in surface area for materials that are fractured by shearing activity
is equivalent to twice the sum of the lengths of the cut edges multiplied by
the depth of the occlusal stroke or thickness of the material (Rensberger
1973:520). Edge lengths are measured in the plane normal to the chewing
direction — in other words, in the occlusal plane in mammals with orthal
(vertical) chewing movement.
For small food objects, a compressive mechanism can be more efficient
than a shearing one. In compression, a small three-dimensional food par-
ticle is flattened into a geometry approaching a two-dimensional surface.
For example, a spherical object compressed to one-fifth its original diam-
eter attains a surface of about twice the original surface area. By compari-
son, the same sphere cut in half by a shearing edge produces only about
1.5 times the original surface area. Frugivores, which utilize food consist-
ing largely of non-fibrous materials with cells that are easily burst and
spread across surfaces, thus tend to have relatively broad, shallow occlusal
surfaces to maximize the amount of food that is compressed during each
stroke. Primates (including humans) have relatively large occlusal basins
and hence have emphasized the compressive mechanisms in tribosphenic
teeth. Not unexpectedly, the rate of comminution of easily fractured food
materials has been found to correlate in humans more strongly with
occlusal area than with dental length measures (Duke and Lucas 1985).
Constraints Because of these different geometric optima, the mech-

anisms for shearing and compression compete for space in a given denti-
tion. Furthermore, because efficiency in shearing mechanisms depends
on total edge length, efficiency in mammals with vertical chewing move-
ments is limited by the width (or length) of the tooth and the number of
crests that can be arranged across that space. Each crest must lie adjacent
to a space for receiving the occluding crest on the opposing tooth, further
limiting the number of crests that can occupy a given area. Crest density is
also limited by the strength of the dental materials: thinner crests frac-
ture under lower stresses than thicker crests composed of the same
material. Consequently, larger mammals with orthal chewing and
greater bite force require absolutely thicker crests and intervening spaces
and cannot simply pack an increased number of thin crests into the space
available in a given jaw size.
Changes in occlusal geometry
The evolution of herbivory in Cenozoic mammals must have faced these

constraints, and new occlusal geometries that appeared may be closely
Figure 6.3. Relationship of taper angle of paracone on M2 with molar width (as
indicator of relative body size) in Late Cretaceous and early Paleocene
insectivores and ungulates. In some specimens of Procerberus and Protungulatum
the tooth may be an M1.
tied to them. One conspicuous change in occlusal geometry found in both

major groups of early Paleocene ungulates, arctocyonids and peripty-
chids, was a reduction in cusp sharpness.
Relationship of cusp bluntness to body size How cusp sharpness varies

among taxa can be examined by measuring the angle formed by the
slopes of the surfaces on either side of a cusp tip. When cusp angle and
tooth size are compared across a range of Late Cretaceous and early Paleo-
cene mammals (Figure 6.3), it is apparent that cusp shapes become
blunter as size increases.
In the early arctocyonid Protungulatum (Figures 6.1–6.2), the molar
cusps are already of somewhat lower aspect ratio than in Cretaceous insec-
tivorous taxa. In a specimen of the larger and stratigraphically slightly
younger arctocyonid Eoconodon (Figure 6.4), the protoconids are lower in
height relative to their basal widths than those in Protungulatum.
Figure 6.4. Occlusal (A) and buccal (B) views of P™–M£ of Eoconodon coryphaeus, UWBM
59709, from the base of the lower Paleocene Nacimiento Formation (Puercan).
Scale bar (occlusal view, lower left) equals 5 mm.
Another change was the reduction of occlusal crests. Crests still

connect the molar cusps in Protungulatum and functioned as edges that
sheared against crests of the opposing teeth (Figure 6.2), but these are
greatly reduced in Eoconodon (Figure 6.4A) to the extent that the occlusal
surface is more strongly dominated by cusps of circular cross-section. The
replacement of crests by circular cusps is even more pronounced in the
periptychid condylarths of early Paleocene age. Typical adult dentitions
of Ectoconus and Periptychus have the cusp tips worn away to an extreme
extent, forming broad, flat surfaces (Figure 6.5) dominated by circular
enamel borders. The cusp shapes in Eoconodon and the periptychids con-
trast with those of most later Cenozoic ungulates, which have teeth
dominated by linear edges associated with more horizontal chewing
directions (Rensberger 1973, 1986, 1988). An examination of the direction
of the striae on the wear surfaces of periptychids indicates a strictly orthal
Figure 6.5. Occlusal (A) and buccal (B) views of right P£–M£ of Periptychus coarctatus,
UWBM 59686, from base of Nacimiento Fm., Puercan, early Paleocene. Scale
bar (occlusal view, lower right) equals 5 mm.
(vertical) chewing direction, lacking the horizontal movements that char-

acterize most later ungulates (Rensberger 1986).
This association between cusp bluntness, body size, and the loss of
crests may be explainable as responses to the relative efficiencies of com-
pressive and shearing mechanisms, the physiological demands of increas-
ing body size, and the strength of the dental materials. As early ungulates
became larger, the primitive, primarily orthal chewing motion remained
unmodified. Because of the different geometric optima of the mechanisms
for compression and shearing, the compressive mechanism was more effi-
cient (as noted above) under vertical chewing movements and, if tender
plant parts were available in sufficient quantity, could meet the increasing
nutritional demands of larger bodies if the diameters of the compressive
areas of the teeth simply increased in proportion to the increase in linear
dimensions of body size.
On the other hand, the shearing mechanism was more severely con-
strained as body size increased because the additional crest length
needed to keep pace would have required either radical changes in the
crest geometry and chewing direction or disproportionate increases in
tooth size. For example, simply increasing the height of a cusp, which
would increase the length of a crest running from the tip to the base of
the cusp, does not increase the effective length of the crest, which is
measured in a direction perpendicular to the shearing direction. Crest
lengths in tribosphenic teeth have as their upper limits the more or less
horizontal distances between major cusps, which are already situated on
the margins of the teeth. For the crests to function in shearing, there
must be space for the opposing crests to occlude precisely between adja-
cent crests, and generating crests in the interior of a tooth requires major
geometric changes. Although dramatic increases in the number of crests
do occur in later Cenozoic herbivores, they are always associated with
horizontal chewing movements in which ‘grinding’ occurs parallel to
the occlusal plane and which thus avoid the need for intercrest occlusion
with opposing crests.
Load In addition to the constraints on shearing mechanisms as

body size increased, another factor limiting the avenues of change was the
increasing load borne by the teeth. Chewing function in mammals
depends on convergence in small occlusal areas of the forces generated by
the jaw musculature, and larger body size results in larger muscles and
larger forces. Without reduction in the stress-concentrating mechanisms
of sharp cusps and crests, this would result in much larger stresses within
the dental materials.
An additional advantage of enlarged compressive regions in the teeth
is that the brittle materials forming the teeth are much stronger in
response to compressive than to tensile stresses. The broader and flatter
compressive surfaces generate lower tensile stresses than narrower,
sharper structures (see below). This may be the most important factor
influencing the early evolutionary modifications of cusp shape.
One can see the effect of stresses on cusps of lower height-to-breadth
Figure 6.6. Relationship of cusp shape to stress magnitude in enamel. Finite

element models of tooth cusps of different heights but same base size,
subjected to the same load. Outer layer of elements of model has higher
Young’s Modulus than inner elements to simulate enamel layer surrounding
dentine.
ratios by modeling the stress regimes in cusps of different aspect ratios

while keeping the loading forces and diameters of the cusps constant.
Figure 6.6 illustrates three finite-element models of cusps of different
heights but with the same basal cross-section and subjected to the same
load. As the cusp acquires a lower aspect ratio, the maximum tensile
stresses decrease. Reduction of aspect ratio (increasing cusp bluntness)
and reduction in crests (structures that increase aspect ratios in small
areas) is a simple way to minimize the tensile stresses within the teeth as
body and muscle size increase.
Cusp bluntness and dietary behavior Although blunt cusps increase

the amount of surface normal to the vertical chewing direction and thus
have the effect of enhancing the efficiency of the compression mecha-
nism, selection in this direction at the same time increasingly restricts the
range of food materials that can be efficiently processed because fibrous
plant tissues require shearing mechanisms for comminution. Owing to
the limited availability of easily processed, soft plant parts that are low in
cellulose, as body size continued to increase and demand larger volumes
of nutrients, the selective advantage of any innovation that would allow
the acquisition of shearing mechanisms and higher occlusal stresses must
have been considerable.
Table 6.1. Comparison of selected mechanical properties for various hard tissues
Material Work of fracture (J m⫺2) Young’s Modulus (stiffness) (GN m⫺2)
Bone 1710c 13.5a

Dentin 550–270d e 11—16b
Enamel 200–13d f 20—90b
Notes:
a Currey (1984), measured in longitudinal direction.
b Waters (1980).
c Currey (1979); bone measured is cow femur. Currey (1984) indicated maximum is more
typically 2800.
d Rasmussen et al. (1976).
e Lower value when tensile stress is parallel to tubule direction.
f Lower value when tensile stress is perpendicular to prism direction.
Enamel reinforcement
Acquisition of stronger enamel appears to have been one of those innova-

tions that allowed elevation of occlusal stresses and evolution of new
shearing mechanisms. Enamel is probably the densest and most brittle
material of the skeleton and consequently most easily fractured (see Table
6.1). The only other material in the mammalian skeleton that approaches
the weakness of enamel to tensile stresses is the dense bone of the ear
region, which is seldom subjected to high loads.
Fracture propagation To understand how enamel could be restruc-

tured to become more resistant to stresses, one must consider how frac-
tures propagate. Fracture in stiff, brittle materials is promoted by the
increasing magnitude of stress concentrated at the tip of a propagating
crack. As a crack lengthens, an increasing proportion of the tensile stress
trajectories in the material bends around the crack tip, increasing the
stress concentration at the tip (e.g., Rensberger 1995: fig. 9.2).
It is clear that a microstructural condition that reduces the concentra-
tion of stress at the crack tip would help prevent the fracture from increas-
ing in length. Prism decussation (zones of prisms with a common
orientation alternating with zones of prisms with a different orientation)
is a common feature of mammalian enamel and accomplishes this goal in
several ways.
Several attributes of the stress field surrounding a propagating crack
can be exploited by the microstructure to inhibit crack expansion. The
stress concentration at the crack tip varies directly with the sharpness of
the tip. In addition to the stresses parallel to the nominal forces acting on
a tooth or other object, other stresses aligned in directions normal to the
nominal stress direction exist ahead of the tip of a propagating crack.
Because prism boundaries are microstructurally the weakest regions in
prismatic enamel, if prisms running oblique to the main crack plane exist
in front of a propagating crack, these ‘ahead’ stresses will tend to open
small cracks between the obliquely running prisms in front of the main
crack tip. As the main crack reaches these oblique cracks, the latter have
the effect of enlarging the tip of the main crack and thereby reducing the
stress concentration (Rensberger 1997: fig. 4) or simply causing the crack
to turn and follow the new interface direction. That direction is less favor-
able to crack propagation because it is less normal to the stresses acting to
pull the crack faces apart (Pfretzschner 1988).
There are other factors associated with prism decussation that inhibit
crack propagation (Rensberger 1995), including the complexity of the
new surface generated by the crack. Decussating prisms cause fractures
passing through the enamel to deviate from the main plane of the frac-
ture and form larger, more rugose surfaces. Regardless of the material,
new surface generated by cracks subtracts energy and tends to inhibit
further propagation (Griffith 1921).
Enamel structure in early placental herbivores Although prism decussa-

tion is common among Cenozoic mammals, it was a rare attribute in the
earlier Paleocene forms (Koenigswald, Rensberger and Pfretzschner
1987). Periptychid condylarths, which were abundant during the early
Paleocene, lacked it altogether. A sample of enamel from the early Paleo-
cene Hemithlaeus shows prisms extending in parallel from the
enamel–dentine junction (EDJ) to the outer surface of the tooth without
bending at all (Figure 6.7).
Prisms in a much larger early Paleocene periptychid, Periptychus, are
also uniformly parallel throughout their course, and extend from the EDJ
in the same outward–upward direction (Figure 6.8). There is, however, a
narrow interval in the inner part of the enamel, at about one-fifth the dis-
tance to the outer surface, where the prisms bend cervically as well as out
of the radial plane (i.e., cervically and circumferentially). In Figure 6.8,
the greater electron emission (lighter gray color) on one side of the bend is
due to the out-of-plane slope of the fracture surface. Although this is not
prism decussation, this sigmoidal bend is probably a crack-resisting
mechanism because a crack plane cannot remain perfectly flat while
passing through the enamel along the easiest path, the straight route fol-
Figure 6.7. Enamel structure of a lower premolar of Hemithlaeus kowalevskianus,

UWBM 59469, Puercan (early Paleocene). SEM (scanning electron micrograph)
of lightly etched fractured surface. EDJ (enamel–dentine junction) at left,
outer surface at right, occlusal direction toward top. Scale bar (near lower
right) equals 20 ␮m.
lowing prism boundaries. In addition to forcing the crack to follow a less

favorable direction, the out-of-plane bend slightly increases the surface
area of the crack plane and thus helps resist the growth of the crack. The
vertical continuity of this bend and the extreme uniformity of the prism
direction elsewhere in the enamel are not known in any later mammal of
this or larger size. This appears to be a very simple acquisition that
slightly improved the strength of the enamel.
Figure 6.8. SEM of the structure of the buccal enamel of a P¢ of Periptychus coarctatus,
UWBM 59671, Puercan (early Paleocene). Occlusal direction toward top, EDJ at
left. Scale bar (lower right) equals 50 ␮m.
Small divergences in prism direction also occur in the early Paleocene

arctocyonid Eoconodon (Figure 6.9). Most prisms tend to arise at the EDJ
and follow almost the same occlusolateral direction as in the peripty-
chids. However, some bundles of prisms occupying regions ranging up to
100 ␮m or more in diameter bend in a slightly cervical or slightly circum-
ferential direction. The diverging prisms sometimes maintain the new
direction until they reach the outer surface. This decussation is very rudi-
mentary. The diverging prisms usually form small angles with neighbor-
ing prisms of about 30° to 40° and the decussation is scattered and
irregular. The small angle and the irregularity contrasts with the larger
angles of decussation and regular spacing of decussating zones found in
extant medium-sized or larger mammals (Kawai 1955; Koenigswald
1997).
Specimens of Loxolophus, another early Paleocene arctocyonid, show a
slightly greater degree of regularity in prism decussation. In UWBM
59725, the nominal prism direction forms an angle of about 50° with the
EDJ, but there are bundles that run more directly outward (angled about
70° or so with the EDJ). The interspaces between these divergent bundles
Figure 6.9. SEM of the structure of the enamel on the lingual side of the metaconid
of M¡ of Eoconodon coryphaeus, UWBM 59657, Puercan (early Paleocene).
Occlusal toward upper left, outer surface at right, and dentine and EDJ at
lower left. Scale bar (bottom, left of center) equals 50 ␮m.
are sometimes subequally spaced along the occlusal–cervical axis of the

tooth (Figure 6.10). This pattern, though not consistently developed, is
suggestive of those in more advanced mammals with clearly decussating
prisms. In some regions (Figure 6.10), the widths of the decussating
bundles of prisms are narrow, but the widths are 130 ␮m or more in others
(Figure 6.11). In some regions no repeating pattern is evident. The angle of
divergence in prism direction is typically 30° to 40°.
Enabling effect of prism decussation
At certain body sizes, the available volume of easily processed plant

parts, such as fruits and seeds, would no longer be sufficient to sustain
growth and maintenance functions, even though they are rich in
digestible nutrients. At that point, occlusal mechanisms allowing
access to the tougher but more abundant and evenly distributed plant
parts would be required. Blunt teeth with only simple compressive
mechanisms and lacking crests would be unable to process efficiently
the required volumes of the more fibrous plants. In order to attain
Figure 6.10. Enamel structure of upper premolar of Loxolophus hyattianus, UWBM

59725, Puercan (early Paleocene). SEM of section near tip of paracone, EDJ at
left, tip of cusp toward top, outer surface at right. Scale bar (lower left) equals
50 ␮m.
sharper occlusal structures, a stronger enamel reinforcement than

those present in the early Paleocene mammals discussed above would
become necessary.
When data from Eocene ungulates are added to the graph of cusp
angle (Figure 6.12), the smaller of these later taxa, such as Homogalax,
Hyracotherium, and Ectocion, fall close to the regression line for Cretaceous
and Paleocene taxa (Figure 6.3). However, the larger of the Eocene and
Figure 6.11. Enamel structure of upper premolar of Loxolophus hyattianus, UWBM

59725, Puercan (early Paleocene). SEM of section 2 mm below tip of paracone
on opposite side of cusp from Figure 6.10. EDJ at right, occlusal toward top,
outer surface at right. Scale bar (lower left) equals 50 ␮m.
Oligocene taxa returned to small cusp angles and appear to be free of the
body-size constraint. The narrow angles in the occlusal structures of the
large forms like Coryphodon (Figure 6.13), together with the extensive jaw
musculature required to move such massive jaws, indicates that high
peak stresses were occasionally concentrated in the occlusal structures.
Nevertheless the steep sides of the lophs and cusps were critical in main-
Figure 6.12. Relationship of cusp angle to width of upper molar in Eocene ungulates
contrasted with Cretaceous/early Paleocene taxa. See also definitions of
symbols in Figure 6.3. The cusp angle was measured on the paracone of M2
except that where the paracone is diminished or merged into a loph, the
following was substituted: Hyrachyus and Hyracodon (taper angle of ectoloph
near metacone); Coryphodon, Pyrotherium (taper angle of metacone). Two angles
were measured on Uintatherium: the larger angle was measured on the
paracone, the smaller angle on the metacone.
taining sharp edges on these shearing structures (Figure 6.13) as the occlu-
sal surfaces wore down.
A key factor that seems to have facilitated the reacquisition of shearing
mechanisms was toughening of the brittle enamel by prism decussation.
Prism decussation is much better developed in the Eocene ungulates than
in the early Paleocene taxa. The only two ungulates illustrated in Figure
6.12 with molar widths greater than 7 mm that lack moderate to strong
prism decussation are Periptychus and Eoconodon, but these taxa lived
during the early Paleocene, at a time when no mammals had acquired
anything but incipient enamel reinforcement.
It is noteworthy that none of the ungulates shown in Figure 6.12 had
acquired the strong translatory chewing movements that characterize
Figure 6.13. Occlusal (a) and buccal (b) views of linear shearing edges and narrow
cusp angles on an upper molar of Coryphodon sp., UWBM 59329, Wasatchian
(early Eocene). Scale bar (below occlusal view) equals 10 mm.
most successful ungulate lineages of the later Cenozoic. Coryphodon, Uin-

tatherium, and Pyrotherium retained the strongly orthal chewing move-
ment of primitive mammals. However, the rather transverse lophs of the
lower occlusal surfaces struck the lophs of the upper surfaces off-center,
causing shearing along a plane making an angle of roughly 45° with the
vertical direction. (Note the inclined angles of the shearing surfaces on
the lophs of Coryphodon; Figure 6.13.) The shearing surfaces in Pyrotherium
and Uintatherium are mechanically similar.
The protoloph and metaloph of the upper molars in the early rhino-
cerotoids Hyrachyus and Hyracodon are also rather transverse, yet these
lophs do not interdigitate between the lophs of the upper dentition but
move across them in nearly horizontal planes. The ectoloph forms a steep
anteroposteriorly aligned shearing mechanism that functioned during
the early phase of jaw closure (Rensberger 1992: fig. 8).
During the later Cenozoic, horses and various selenodont artiodactyls
exploited the advantages of predominantly transverse chewing move-
ments, which allowed them to employ multiple, closely spaced enamel
edges and heavy emphasis on shearing for more efficient processing of
fibrous plant fodder. Interestingly, enamel reinforcement in the advanced
horses and artiodactyls is accomplished to a great extent by a different
type of microstructure (Figure 6.14; Pfretzschner 1992, 1993; Kozawa and
Suzuki 1995). This enamel functionally consists of thin, flat sheets of
hydroxyapatite crystallites, which in alternating laminae extend in oppos-
ing directions. In the modern horse, the prisms, which form each alternate
layer, have become so flattened and aligned so completely in the same
plane, and the interprismatic matrix has become so similarly flattened as
an alternating layer, that almost the only distinction between prismatic
and interprismatic enamel is the roughly 75° difference in alignment of
the crystallites in the alternating layers (Rensberger 1996). This type of
enamel is the most plywood-like of all known hard tissues in vertebrates.
Summary and conclusions
Hadrosaurid dinosaurs and late Cenozoic ungulate mammals, including

horses and bovids, had elaborate dental mills for processing vegetation at
high rates. However, the evolution of dental structures in early ungulates
was impeded by constraining factors related to their derivation from
those of insectivorous mammals.
The masticatory system in insectivorous mammals utilizes high,
Figure 6.14. Enamel structure of M£ of Phenacodus primaevus, UWBM 39030,

Wasatchian (early Eocene). SEM of polished and etched radial section, lingual
side of metaconid. EDJ at left margin of micrograph, outer surface at right,
occlusal toward top. Prism decussation zones (lighter and darker bands) run
outward and slightly cervically from near the EDJ, but disappear in outer third
of enamel thickness. Scale bar (lower left) equals 100 ␮m.
sharp cusps that interdigitate with opposing cusps during largely vertical
chewing movements. These complex teeth contained three mechanisms
for fracturing food. The insertion of cusps between and against opposing
cusps provides a shearing mechanism, but, at the same time, the move-
ment of the protocone of the upper molars into the talonid basin in the
lowers adds a compressive mechanism, and the height and sharpness of
the cusps adds a puncture-crushing mechanism.
In the early Paleocene ungulates, the shearing component was greatly
reduced as the ratio of cusp height to cusp width was reduced and crests
were lost. These changes seem to provide maximal functional efficiency,
given the limitation of space, and made compression the dominant mech-
anism. The compressive mechanism, which depends on areas of surface in
the plane normal to the chewing direction, would have been able to keep
up with the increased nutrient requirements of larger body size by
increasing its linear dimension (radius) in proportion to linear increases
in body size, whereas the shearing mechanisms, which depend on lengths
of crests, would have needed to increase at a higher rate.
A second factor influencing the shift toward compressive mechanisms

appears to have been enamel strength. Modeling shows that the stresses
in tooth enamel induced by chewing activity are lower in low, wide cusps
than in high, sharp cusps. It is clear that prism decussation makes the
enamel more resistant to fracture, and, by early Eocene times, moderate
to strong prism decussation had appeared in most ungulates with upper
molars of at least 7 mm in width. From this size upward in early Eocene
ungulates, the earlier trend toward cusp bluntness is reversed, occlusal
structures again have higher aspect ratios, and crests reappear.
The early increase in cusp angle and loss of crests with increased body
size, followed by later reduction of cusp angle after prism decussation
became established in larger mammals, implies that enamel brittleness
was an important constraint on the evolution of herbivory during the
early Cenozoic. This constraint, together with the different mechanics of
shearing and compressive mechanisms, must have combined to limit the
dietary diversity and/or the rates of nutrient intake in many early Ceno-
zoic herbivores.
Acknowledgments
I am grateful to Dr Hans-Dieter Sues for inviting me to participate in the

symposium on the evolution of herbivory and to Drs Donald L. Rasmus-
sen and John R. Horner who, respectively, collected and lent the fine jaws
of Protungulatum donnae for this study. Drs Mikael Fortelius, John Hunter,
and Tracy Popowics kindly reviewed the manuscript and contributed
comments and suggestions that resulted in its significant improvement.
Beverly Witte prepared the Paleocene and Eocene fossil material from the
UWBM collection, and she, Ben Witte, and Mesfin Asnake collected many
of the specimens used in this project.
References
Ahlgren, J. (1966). Mechanisms of mastication. Acta Odont. Scand. 24:5–109.

Atkinson, H. F., and Shepherd, R. W. (1961). Temporomandibular joint disturbances and
the associated masticatory patterns. Aust. Dent. J. 6:219–222.
Crompton, A. W., and Hiiemae, K. M. (1970). Molar occlusion and mandibular
movements during occlusion in the American opossum, Didelphis marsupialis L.
Zool. J. Linn. Soc. 49:21–47.
Currey, J. D. (1979). Mechanical properties of bone with greatly differing functions. J.
Biomechanics 12:313–319.
Currey, J. D. (1984). The Mechanical Adaptations of Bones. Princeton: Princeton University

Press.
Duke, D. A., and Lucas, P. W. (1985). Chewing efficiency in relation to occlusal and other
variations in the human dentition. British Dent. J. 159:401–403.
Griffith, A. A. (1921). The phenomena of rupture and flow in solids. Phil. Trans. R. Soc.
Lond. A 221:163.
Hahn, G. (1971). The dentition of the Paulchoffatiidae (Multituberculata, Upper
Jurassic). Mem. Serv. Geol. Portugal (N.S.) 17:1–39.
Herring, S. W., and Scapino, R. P. (1973). Physiology of feeding in miniature pigs. J.
Morph. 141:427–460.
Hiiemae, K. M., and Kay, R. F. (1973). Evolutionary trends in the dynamics of primate
mastication. In Craniofacial Biology of the Primates, ed. M. R. Zingeser, pp. 28–64.
Symposium 4th Int. Congr. Primat. Basel: Karger.
Hiiemae, K. M., and Thexton, A. J. (1975). Consistency and bite size as regulations of
mastication in cats. J. Dent. Res. 54(A):194.
Janis, C. M., and Fortelius, M. (1988). On the means whereby mammals achieve
increased functional durability of their dentitions, with special reference to
limiting factors. Biol. Rev. 63:197–230.
Kallen, F. C., and Gans, C. (1972). Mastication in the little brown bat (Myotis lucifugus). J.
Morph. 136:385–420.
Kawai, N. (1955). Comparative anatomy of the bands of Schreger. Okimimas Folia Anat.
Jap. 27:115–131.
Koenigswald, W. von. (1997). Brief survey of enamel diversity at the schmelzmuster level
in Cenozoic placental mammals. In Tooth Enamel Microstructure, ed. W. von
Koenigswald and P. M. Sander, pp. 137–161. Rotterdam: Balkema.
Koenigswald, W. von, Rensberger, J. M., and Pfretzschner, H.-U. (1987.) Changes in the
tooth enamel of early Paleocene mammals allowing increased diet diversity.
Nature 328:150–152.
Kozawa, Y., and Suzuki, K. (1995). Appearance of new characteristic features of tooth
structure in the evolution of molar teeth of Equoidea and Proboscidea. In Aspects
of Dental Biology: Palaeontology, Anthropology and Evolution, ed. J. Moggi-Cecchi, pp.
27–31. International Institute for the Study of Man.
Krause, D. W. (1982). Jaw movement, dental function, and diet in the Paleocene
mulituberculate Ptilodus. Paleobiology 8:265–281.
Luschei, E. S., and Goodwin, G. M. (1974). Patterns of mandibular movement and jaw
activity during mastication in the monkey. J. Neurophysiol. 37:954–966.
Pfretzschner, H.-U. (1988). Structural reinforcement and crack propagation in enamel.
In Teeth Revisited: Proceedings of the VIIth International Symposium on Dental
Morphology, ed. D. E. Russell, J.-P. Santoro, and D. Sigogneau-Russell, pp.
133–143. Mém. Mus. natn. Hist. nat., Paris, série C, 53.
Pfretzschner, H.-U. (1992). Enamel microstructure and hypsodonty in large mammals.
In Structure, Function and Evolution of Teeth, ed. P. Smith and E. Tchernov, pp.
147–162. London and Tel Aviv: Freund Publishing House Ltd.
Pfretzschner, H.-U. (1993). Enamel microstructure in the phylogeny of the Equidae. J.
Vert. Paleont. 13:342–349.
Rasmussen, S. T., Patchin, R. E., Scott, D. B., and Heuer, A. H. (1976). Fracture properties
of human enamel and dentin. J. Dent. Res. 55:154–164.
Rensberger, J. M. (1973). An occlusion model for mastication and dental wear in
herbivorous mammals. J. Paleont. 47:515–528.

Rensberger, J. M. (1988). The transition from insectivory to herbivory in mammalian
teeth. In Teeth Revisited: Proceedings of the VIIth International Symposium on Dental
Morphology, ed. D. E. Russell, J.-P. Santoro, and D. Sigogneau-Russell, pp.
351–365. Mém. Mus. natn. Hist. nat., Paris, série C, 53.
Rensberger, J. M. (1992). Relationship of chewing stress and enamel microstructure in
rhinocerotoid cheek teeth. In Structure, Function and Evolution of Teeth, ed. P. Smith
and E. Tchernov, pp. 163–183. London and Tel Aviv: Freund Publishing House
Ltd.
Rensberger, J. M. (1995). Determination of stresses in mammalian dental enamel and
their relevance to the interpretation of feeding behaviors in extinct taxa. In
Functional Morphology in Vertebrate Paleontology, ed. J. J. Thomason, pp. 151–172.
Rensberger, J. M. (1996). Differences in enamel fracture behavior and the evolution of
dental strength in horses. J. Vert. Paleont. 16 (suppl. to 3):60A.
Rensberger, J. M. (1997). Mechanical adaptation in enamel. In Tooth Enamel
Microstructure, ed. W. von Koenigswald and P. M. Sander, pp. 237–257.
Rotterdam: Balkema.
Schaal, S., and Ziegler, W. (1993). Messel — An Insight into the History of Life and of the Earth.
Oxford: Oxford University Press.
Sloan, R. E. (1987). Paleocene and latest Cretaceous mammal ages, biozones,
magnetozones, rates of sedimentation, and evolution. Geol. Soc. Amer. Special Paper
209:165–200.
Sloan, R. E., and Van Valen, L. (1965). Cretaceous mammals from Montana. Science
148:220–227.
Waters, N.E. (1980). Some mechanical and physical properties of teeth. In The Mechanical
Properties of Biologic Materials, ed. J. F. V. Vincent and J. D. Currey, pp. 95–135. Symp.
Soc. Exp. Biol. no. 34. Cambridge and New York: Cambridge University Press.
christine m. janis
Patterns in the evolution of herbivory in large

terrestrial mammals: the Paleogene of North
America
Introduction
This chapter examines the evolution of herbivory among the larger

terrestrial mammals (ungulates and ungulate-like mammals) of the Early
Tertiary (Paleogene, encompassing the Paleocene, Eocene, and Oligocene
epochs). There are sound reasons for examining paleoecological trends in
large mammals separately from small ones, both in terms of the allometry
of the ecological attributes of living mammals (see Brown 1995) and in
terms of paleontological and taphonomic biases (see Fortelius et al. 1996).
The term ‘herbivory’ is used to refer to diets that include primarily
plant material, as opposed to animal material. In this chapter, I especially
emphasize the evolution of the more restricted form of herbivory, foli-
vory, meaning the consumption of only the fibrous, structural parts of the
plants, such as stems and leaves. Folivorous mammals today include both
browsers and grazers. (There were no grasslands in the Early Tertiary to
support the latter dietary habit.) These mammals have some sort of sym-
biotic association with micro-organisms in the gut for cellulose fermenta-
tion. Folivores are usually contrasted with frugivores, that is animals that
eat primarily fruit and other non-structural parts of the plant. There is
also a category of ‘frugivorous/folivorous’ mammals, which take a
mixture of fruit, seeds and leaves but do not select fibrous food requiring
fermentation.
Exclusively herbivorous mammals must be above a certain body size
(approximately one kilogram) because plant material does not contain
enough nutrition per unit volume to support the metabolic rate of
smaller endotherms (Kay 1975). Some small (i.e., less than 1 kg in body
mass) extant mammals, such as voles and other specialized rodents, can
[168]
North American Paleogene large herbivores 169
be exclusively herbivorous by virtue of being ‘granivorous,’ specialized

for taking grass seeds. Granivory may involve feeding on species of
grasses, but in terms of the type of nutrition afforded the animal (i.e.,
feeding on the reproductive parts of plants rather than structural parts) it
is more akin to frugivory than to grazing folivory (see Eisenberg 1981 for a
discussion of mammalian feeding types).
The data in this chapter all come from the North American fossil
record, although evolutionary trends in the higher latitudes of the Old
World appear to be similar and are briefly discussed. Dietary preferences
of extinct mammals are deduced from their dental features. In the case of
the ungulates, at least, there is a diversity of modern taxa of known
dietary habits so that the correlation between dental morphology and
diet can be known with some assurance. The extinct ungulate-like
mammals have no precise modern analogs, and thus their dental
morphologies are more difficult to interpret precisely. However, as will be
discussed, there are good biomechanical reasons for assigning particular
diets to particular dental configurations. Patterns in the evolution of
body size are also examined and discussed, since Paleogene mammals
radiated into a wider variety of body sizes than were prevalent in the Mes-
ozoic, when no mammal was bigger than a house cat. Ungulates have
always been mostly of large body size compared with most other
mammals, and the story of the evolution of herbivory in ungulates is in
part the story of the evolution of larger body size.
The global climate and distribution of vegetation of the world was
quite different in the Paleogene. Specifically, early Paleogene continents
at temperate latitudes, such as North America, had much warmer mean
annual temperatures (with concomitant lesser degrees of seasonality)
than in the present day (Figure 7.1), with tropical-like forest habitats
extending to within the confines of the Arctic Circle (Wolfe 1978, 1985).
Following a peak in mean annual temperatures in northern latitudes at
the early-middle Eocene boundary, temperatures started to decline,
reaching a low in the early Oligocene. Concomitant with this climatic
deterioration, tropical forests became more restricted towards the lower
latitudes, and the major vegetational type in North America changed to
that of a more seasonal deciduous type (summarized for North America in
Wing 1998). The evolution of mammalian herbivory is examined against
this background of climatic and vegetational changes.
This survey examines the generic richness through Paleogene times of
ungulates and ungulate-like mammals distinguished both by dental
170 christine m. janis
Figure 7.1. (A) Temporal ranges of lineages of North American ungulates and
ungulate-like mammals. Key to shading: Cross-hatched lines, non-ungulates
and archaic ungulates; diamond-patterned lines, modern ungulate orders. (B)
Paleotemperatures in the North Sea for the Paleogene (modified from
Burchardt 1978).
structure (⫽ dietary type) and body size. Many of these general dietary
trends have been discussed by other workers (e.g., Wing and Tiffney 1987;
Collinson and Hooker 1991; Maas and Krause 1994; Gunnell et al. 1995).
However, this chapter represents a new approach by attempting to quan-
tify patterns of change in dental morphologies and also examining pat-
terns in the evolution of body size, specifically focusing on ungulates and
ungulate-like mammals. Jernvall, Hunter, and Fortelius (1996, 2000),
convergently with the inception of this work, adopted a somewhat
similar approach to the quantification of dental patterns, but their work
has a broader geographic range (including all of the Northern Hemi-
sphere) and uses faunal data at a coarser level of resolution. This chapter
not only provides a quantification of some generally known trends but
arrives at some novel conclusions, such as the speculation (based on the
dietary habits of the ungulates) that the initial climatic event in the
middle Eocene of the continental interior of the conterminous United
States was drying rather than cooling.
Overview of the evolution of Early Tertiary herbivorous

mammals
Although mammals have been a component of the terrestrial vertebrate

fauna since the end of the Triassic, during the Mesozoic (the famed first
two-thirds of their evolutionary history [Lillegraven, Kielan-Jaworowska,
and Clemens 1979]) they were almost exclusively small (under 1 kg) and
insectivorous or omnivorous. At this small body size they would have
been unable to be exclusively herbivorous. Multituberculate mammals
have been considered as Mesozoic herbivores, but despite their complex
teeth, the dental structure and body size of most taxa would dictate a
more omnivorous diet (Krause 1986). The recently discovered Cretaceous
zhelestids, purported ungulate ancestors from the early Late Cretaceous
of Uzbekistan (Archibald 1996), were likewise too small to have been
exclusively herbivorous.
It was not until the Cenozoic, following the demise of the non-avian
dinosaurs at the end of the Cretaceous, that mammals radiated into a
larger variety of ecomorphological types, including larger body sizes.
Thus the larger mammals in the Paleocene had the opportunity to
include vegetation as a significant portion of their diet. However,
although there was a great Paleocene diversification of taxa considered
herbivorous, such as many lineages of ‘condylarths,’ or archaic ungulates,
their dentitions and patterns of dental wear suggest that they were
mainly omnivorous or generalist feeders (Janis 1979; Rensberger 1986;
Hunter 1997). Some larger (up to bear-size during the Paleocene) ungu-
late-like taxa such as pantodonts and uintatheres had teeth apparently
more adapted for shearing foliage; but the teeth of these animals were
rather small in comparison with their bodies, quite unlike the ‘batteries’
of shearing teeth seen in modern folivorous ungulates, and had very
simple shearing lophs (Janis 1979). Their teeth may have been adapted to
masticate only soft, perhaps aquatic, types of vegetation. No analogs to
these herbivores are known among modern mammals.
Terrestrial herbivores with ridged or lophed teeth adapted for a more
extensive folivorous diet are not generally apparent until the Eocene,
with some late Paleocene ‘heralds’ such as the more lophodont ‘condy-
larth’ Meniscotherium and the taxonomically enigmatic ungulate Arctosty-
lops. This difference between Paleocene and Eocene forms in lineages
generally regarded as ‘herbivorous’ is also apparent in the primates, in
which there is also no evidence of folivorous taxa until the Eocene (Janis
1979; Collinson and Hooker 1991).
The start of the Eocene marked the first appearance of many modern
mammalian orders, such as the perissodactyl and artiodactyl ungulates,
and the true primates (Euprimates). (However, note that the earliest per-
issodactyls may be late Paleocene in age [Dashzeveg and Hooker 1997].)
The appearance of these new groups did not immediately result in extinc-
tions among the more archaic forms, at least in North America (Maas et al.
1995) (although early Eocene extinctions of archaic taxa were more
marked in Europe [Russell 1975]). There was a rapid diversification of per-
issodactyls in the early Eocene; artiodactyl diversification occurred a little
later, during the middle to late Eocene. Although many of the early
members of these orders, such as the hyracotheriine horses and the dicho-
bunid artiodactyls, had teeth suggestive of frugivory and selective brows-
ing rather than true folivory (Janis 1979, 1990a; Collinson and Hooker
1991), by late Eocene times almost all perissodactyls and artiodactyls were
larger animals, more clearly specialized for folivory, with the exception of
the suiform (pig-like) artiodactyls that had divergently specialized for
omnivory.
The Eocene thus appears to have been a critical time for the evolution
of true herbivory in mammals. The early Eocene represented the earliest
appearance of true folivores, and by the close of the epoch the diversity of
trophic types of herbivorous mammals more closely resembled that seen
in the modern fauna. The Eocene is also perhaps the most interesting
Cenozoic epoch in terms of climatic changes, the late early to early middle
Eocene representing the warmest time in the Cenozoic (see Figure 7.1),
with the later Eocene temperatures of the higher latitudes plummeting
in the transition between the ‘Hot House’ types of global environments of
the Mesozoic and early Cenozoic and the ‘Ice House’ global environments
of the later Cenozoic (Berggren and Prothero 1992 and references therein).
This chapter attempts to explain some of the patterns in the diversity of
ungulates, and of their dental features, during the Paleogene, and to
relate them to changing global climates and the evolution of terrestrial
herbivory.
Materials and methods
Taxonomic data
This chapter only examines North American mammals. The North Amer-
ican Paleogene record, especially in the Paleocene, is probably the best
available in the world (Savage and Russell 1983). Certainly, the data are
easier to obtain and to compile, due in part to the limited number of polit-
ical boundaries in comparison with other continents. North America was
isolated for much of the Paleogene, and due to its latitudinal position was
strongly affected by the late Eocene climatic changes (see Janis 1993 and
references therein). Similar, but not identical, patterns of faunal change
among herbivorous mammals, especially ungulates, were seen on other
northern continents (Jernvall et al. 1996, 2000), as will be discussed later.
The data on the number of taxa present in any one interval of time
come from the contributions in Janis, Scott, and Jacobs (1998), specifically
from chapters by Archibald, Cifelli and Schaff, Colbert and Schoch,
Coombs, Effinger, Honey et al., Kron and Manning, Lander, Lucas, Lucas
and Schoch (a and b), Lucas, Schoch, and Williamson, MacFadden, Mader,
Prothero (a–d), Stucky, Wall, Webb, and Wright (see references). Among
the taxa of archaic ungulates (‘condylarths’) considered (data from
Archibald 1998), clades with dentitions indicative of a more carnivor-
ous/omnivorous diet, such as arctocyonids and mesonychids, were
excluded from consideration. The taxonomic level of investigation was
that of the genus; although species-level studies may be more appropriate
for addressing some questions, the generic level may be considered as
more appropriate for this broad-based paleoecological study (see, for
example, Sepkoski and Kendrick 1993).
The Appendix to this chapter lists the taxa present in each subdivision
of time, together with their dental morphology and body size category
(see below). In a few instances, taxa were considered present in an interval
if they were recorded from the time intervals preceding and succeeding
that particular interval but not actually recorded from the interval itself.
Time intervals
The units of time selected for study follow the divisions of the North
American land mammal ‘ages’ that were adopted in Woodburne (1987),
and followed (with some updating revisions) in Janis et al. (1998). The land
mammal ‘ages’ and their subdivisions are not of equal duration, nor are
they equally well-sampled for fossil localities. As can be seen in Figures
7.3–7.6, units in the Paleocene may be less than a million years in length,
whereas those in the later Eocene can be several million years long (for
example, the entire Duchesnean Land Mammal age; it should be noted
that the Duchesnean is a particularly poorly known interval of time
[Lucas, 1992]). An attempt to investigate these sources of bias in these data
is described elsewhere (Janis 1997–98). In summary, when the data are
adjusted for the pattern of generic richness, there is only a slight dampen-
ing of the overall diversity pattern in select places, and, when time inter-
val number is removed, the correlation between the number of localities
and the number of genera is trivial. Possible effects of these biases are dis-
cussed later in the text.
Dental structure and dietary categories

The form of the cheek teeth of ungulates can be finely subdivided into
many different detailed patterns (Jernvall 1995; Jernvall et al. 1996, 2000),
but, for the purposes of this chapter, only three main types were consid-
ered (Figure 7.2). ‘Bunodont’ teeth represent the primitive quadrituber-
cular form of tooth seen in ungulates, as well as in many other
omnivorous mammals (Janis and Fortelius 1988). These are low-crowned,
rounded-cusped teeth, present today in mammals with an omnivo-
rous/frugivorous type of diet (e.g., bears, most pigs, and many primates
including humans), that act to compress and pulp non-brittle, non-
fibrous food in a pestle-and-mortar fashion (Lucas and Luke 1984).
The form of the cheek teeth in more herbivorous mammals shows an
increase in the height of the low, isolated cusps, and a tendency to join
them together to form high, cutting ridges, termed ‘lophs’ that act to
shear and shred flat, tough, fibrous food such as leaves (Lucas and Luke
Figure 7.2. Cheek tooth patterns in herbivorous mammals, seen in occlusal and
anterior views. Each example is represented by an upper left molar; shading
indicates areas of exposed dentine (on a moderately worn tooth). (A) Bunodont
tooth, indicative of an omnivorous/frugivorous type of diet. Examples in
modern mammals include bush pig, bear, humans, and many other primates.
(B) Semi-lophed tooth ( ‘bunolophodont’ or ‘bunoselenodont’), indicative of a
frugivorous/selective folivorous type of diet (i.e., non-fibrous parts of plant).
Examples in modern mammals include mouse deer and tree hyraxes (plus leaf-
eating primates like colobines). (C) Lophed tooth, indicative of a folivorous
diet (i.e., fibrous parts of plant). Examples in modern animals include
perissodactyls such as horses and rhinos (C i, ‘lophodont’) and ruminant
artiodactyls such as deer and antelope (C ii, ‘selenodont’). Lophed teeth can be
low-crowned (brachydont, as in browsing mammals like deer and giraffe) or
hypsodont (as in predominantly grazing mammals like horses and cows).
Bunodont and semi-lophed teeth are almost invariably brachydont (with
certain members of the extinct order Taeniodonta as the only known
exception).
1984; Janis and Fortelius 1988). Such lophed teeth have developed from
bunodont teeth on many occasions (Janis and Fortelius 1988; Janis 1995;
Jernvall et al. 1996, 2000). Among the variety of possible patterns two basic
ones are found among extant ungulates: the ‘selenodont’ teeth of rumi-
nant artiodactyls (in which the lophs extend in a predominantly mesio-
distal direction) and the ‘lophodont’ teeth of perissodactyls and rock
hyraxes (in which the lophs extend in a predominantly labiolingual
direction). Here no distinction is made between the different types of

lophed teeth (although up to thirteen different types can be distin-
guished among the diversity of living and fossil ungulates [Jernvall
1995]); all are considered functionally equivalent and indicative of a foliv-
orous diet. Lophed teeth may additionally be subdivided into brachydont
(low-crowned) and hypsodont (high-crowned). Many modern ungulates
are hypsodont, although this condition was less common in the Paleo-
gene. Today hypsodonty is associated with grazing, as high-crowned
teeth will take longer to be worn down by the abrasive elements in grass,
and render the dentition more durable. Hypsodonty is a good indication
of a more abrasive diet, although the correlation may not be so much with
grass per se as with open habitats, as dust and grit on both grass and low-
level browse will tend to abrade the teeth (Janis 1995).
A third type of tooth form considered here is termed ‘semi-lophed’,
which is intermediate between the bunodont and fully lophed forms.
Some ridges are present, but the cusps are not fully incorporated into
lophs (see Figure 7.2). Again this term encompasses a great deal of varia-
tion in form that has occurred throughout ungulate evolution (Jernvall et
al. 1996, 2000). Although a tooth of this form would be a necessary inter-
mediate evolutionary stage in the acquisition of fully lophed teeth from
an ancestral bunodont condition, it is also clearly adapted for a particular
type of diet. Extant mammals that are selective frugivore/folivores, select-
ing predominantly the non-fibrous parts of the plant, have this form of
cheek tooth (for example, the ‘bunoselenodont’ teeth of tragulids or the
‘bunolophodont’ teeth of tree hyraxes; Janis 1990a, 1995), and this type of
tooth is also present in leaf-eating monkeys. Extinct mammals with this
type of cheek tooth pattern are considered to have had a similar type of
diet.
Very few extant terrestrial mammals have teeth that can be classified as
semi-lophed. In the present-day African fauna, only three taxa fall into
this category: the water chevrotain (Hyemoschus aquaticus), the tree hyrax
(Dendrohyrax dorsalis), and the pygmy hippo (Choeropsis liberensis). Among
the diversity of North American Paleogene ungulates, however, many of
the archaic (‘condylarth’) families have members that could be considered
semi-lophed, as do many primitive perissodactyls and artiodactyls. I have
also included in this grouping mammals that do not fall within the clade
Ungulata (Prothero, Manning, and Fischer 1988) but that nevertheless are
‘ungulate-like’ in their general structure and dentition; these include tae-
niodonts, tillodonts, pantodonts, and dinoceratans (uintatheres, which
may or may not be true ungulates; see Lucas and Schoch 1998b). The Tae-
niodonta contain the only taxa that have teeth that may be considered
‘semi-lophed hypsodont.’ Their teeth have been classified as ‘semi-
lophed’ for the purposes of this study, but actually the structure of their
cheek teeth is quite different from those of ungulates and clearly indepen-
dently derived.
It is a fundamental assumption of this chapter that dental patterns are
evolutionarily plastic. The various types of lophed teeth have indepen-
dently evolved from bunodont forms in numerous lineages, in marsupi-
als as well as in placentals. I thus assume that the types of crown structure
represent dental adaptations to the prevailing vegetational conditions,
and do not merely reflect phylogenetic affinities. That is to say that a
mammal possesses lophed cheek teeth because of its dietary habits, not
because of its ancestry (although the precise form of the pattern of the
lophs will most likely reflect its evolutionary history).
Body masses and size distributions

Body masses were estimated from dental measures (length of m2) pro-
vided in the chapters in Janis et al. (1988), which were then used in con-
junction with the algorithms in Janis (1990b), or from estimates made
directly by the authors of the contributions.
The body size categories were chosen to reflect natural groupings of
mammals of different ecomorphological types. Although they funda-
mentally correspond to a logarithmic scale, the categories do not corre-
spond to precise logarithmic means (such as 1–10 kg, 10–100 kg, etc.).
There is no reason to assume that such numerical means represent biolog-
ically significant junctures. Additionally, as all ungulates are ‘large’ by
mammalian standards (as the median mass of modern mammals ranges
around one kilogram [Eisenberg 1981]) a traditional logarithmic scale
would group the majority of mammals only into two size categories (i.e.,
10–100 kg, and 100–1000 kg). The categories that have been chosen here
appear to reflect natural ecomorphological ‘breaks,’ similar to those rec-
ognized by various other authors (Jarman 1974; Bertram and Biewener
1990; Fortelius, Van der Made, and Bernor 1996; Fortelius et al. 1996; Lewis
1997).
Size category 1 (‘rabbit-sized’) includes small terrestrial herbivores
under 5 kg (among modern ungulates, encompassing small ruminants
such as Asian tragulids and dwarf antelope such as dik-dik, and hyraxes).
Size category 2 (‘dog-sized’) ranges from 5 kg to 25 kg (encompassing
small antelope and deer such as duikers and roe deer, and peccaries). Size
category 3 (‘antelope-sized’) ranges from 25 kg to 150 kg (including the
majority of antelope and deer, plus llamas and pigs). Size category 4
(‘horse-sized’) ranges from 150 kg to 500 kg (including not only equids,
but larger bovids and deer, such as cows and wapiti, the okapi, as well as
tapirs and the pygmy hippo). Size category 5 (‘rhino-sized’) includes taxa
over 500 kg (including not only rhinos, and other large herbivores such as
hippos and elephants, but also very large artiodactyls such as giraffes and
camels). This final size category might be better split into two if consider-
ing modern and Neogene mammals, separating the megaherbivores (over
1000 kg, or perhaps over 2000 kg) from the merely large ones; however,
few Paleogene mammals (those under consideration here) attain weights
in excess of 500 kg, and only a handful (some later Eocene uintatheres and
brontotheres) would have weighed more than one metric ton.
Results
Tooth types
Figure 7.3 illustrates the generic diversity of ungulate and ungulate-like
mammals with different tooth types over the duration of the Paleogene,
and Figure 7.4 expresses these same data as percentage of the total fauna.
The modern fauna, represented by the generic richness of dental types in
present-day African ungulates (including subungulates), is shown to the
right of the plot. Today there are few terrestrial herbivores with buno-
dont or semi-lophed teeth (although many arboreal primates have teeth
of this type); the vast majority of African ungulates have lophed teeth,
with about two-thirds of the genera having hypsodont teeth.
The modern situation contrasts rather sharply with the conditions in
the Paleogene, especially during the early part of that time interval. At the
start of the Paleocene, all of the ungulate taxa were bunodont; some early
Paleocene condylarths actually had dental structures indicative of insecti-
vory rather than of generalist omnivory (Hunter 1997). Even by the end of
the Paleocene, bunodont taxa still comprised around 50% of the fauna.
The slight fall in the numbers of bunodont taxa in the middle Paleocene,
with a subsequent rise in the latest Paleocene and early Eocene, may
reflect changing higher-latitude temperatures (see Figure 7.1). Semi-
lophed taxa made their first appearance in the late early Paleocene, and
increased in numbers through the epoch, but fully lophed taxa (e.g., arc-
tostylopids) were not apparent until the very end of the Paleocene.
Figure 7.3. Generic diversity within dental structural types among Paleogene ungulates and ungulate-like mammals. Key to
shading: tile or clam-shell pattern, bunodont; dashed pattern, semi-lophed (brachydont); vertical striping, semi-
lophed (hypsodont); light cross-hatching, lophed (brachydont); heavy cross-hatching, lophed (hypsodont). Column
on right-hand side shows pattern among present-day African ungulate genera for comparison.
Figure 7.4. Percentages of generic diversity within dental structural types among Paleogene ungulates and ungulate-like mammals.
Key to shading as for Figure 7.3.
Maas and Krause (1994) showed a category of ‘herbivorous/folivorous’

mammals (of all taxonomic categories, not just ungulates) existing at
quite high generic diversities throughout the Paleocene, but their defini-
tions of diets as defined by dental morphology must be different from the
ones that I use here. I have defined folivorous taxa strictly by the posses-
sion of true dental lophs, whereas Maas and Krause (1994:108) defined
these dietary types as having teeth with ‘greater emphasis on shearing
crests.’ They did not state which taxa were actually included in this cate-
gory.
Herbivores from the early and early middle Eocene (Wasatchian and
Bridgerian Land Mammal ages) retained a similar diversity of tooth types
to those of the latest Paleocene, with the numbers of lophed taxa hovering
between 10% to 20% of the fauna, and the bunodont taxa maintaining
their position at 40% to 50% of the fauna. The decrease in the percentage
of bunodont taxa during the Wasatchian and into the early Bridgerian
reflects an increase in the numbers of semi-lophed taxa rather than any
real change in the numbers of bunodont taxa. Note that the Paleocene
and early Eocene lophed taxa were predominantly of small body size, in
contrast to the size of lophed ungulates in the modern fauna (see Figure
7.7). The early Eocene increase in generic diversity reflects the appearance
of the modern ungulate orders Artiodactyla and Perissodactyla, in partic-
ular the radiation of equid and tapiroid perissodactyls.
A marked change occurred later in the middle Eocene, commencing in
the late Bridgerian and peaking in the late Uintan and Duchesnean. The
diversity of ungulate genera showed a significant increase, reflecting the
diversification of ceratomorph (rhino-like) perissodactyls and of seleno-
dont artiodactyls. Part of the magnitude of this increase may be due to the
time-averaging effects in the longer intervals of the late Uintan and the
Duchesnean, but there is clearly a change in the taxonomic diversity and
composition of the ungulate fauna at this time. Most significantly, the
numbers of bunodont taxa decreased dramatically, to around 10% of the
herbivore fauna. This decrease represented the loss not only of archaic
bunodont ungulates, such as ‘condylarths,’ but also the loss of the more
primitive artiodactyls (Figure 7.7). However, the difference in the faunal
percentages was not made up at this time by the expansion of taxa with
lophed teeth, as might be expected from the fact that lophed ungulates
form the predominant dental type in the modern fauna. Rather, it was by
the expansion of semi-lophed taxa, which comprised around 50% of the
large herbivore fauna for the remainder of the epoch. The percentages of
lophed taxa increased slightly towards the end of the Eocene, with the
first hypsodont taxa appearing at the end of the middle Eocene, but
lophed taxa did not come to comprise more than 50% of the fauna until
the Oligocene. Note also that generic diversity of ungulates decreased
from a peak in the late middle Eocene to late Oligocene levels approach-
ing those of the early Eocene (Figure 7.3).
By the end of the Oligocene, the diversity of tooth types in the ungu-
late fauna approached the modern condition, with the exception of the
low numbers of hypsodont taxa. The rise in the number of bunodont taxa
in the Oligocene reflects the radiation of the larger, more derived suiform
artiodactyls, such as anthracotheres, entelodonts, and peccaries (see
Figure 7.7).
Body sizes
The distribution of body sizes through the Paleogene reveals a rather dif-
ferent type of information about diversity patterns (Figures 7.5–7.6).
Among present-day African ungulates, all sizes are fairly equally repre-
sented; the smallest number of genera are the ‘dog-sized’ ones in category
2, and the largest numbers are in the ‘antelope-sized’ category 3. This,
again, is very different from the situation during the early Paleogene. At
the start of the Paleocene, all the terrestrial herbivores were small, buno-
dont condylarths in size categories 1 and 2 (see also Figure 7.7). Later in the
early Paleocene larger forms appeared, to comprise around 50% of the
herbivorous fauna by the late Paleocene. These larger taxa were the ungu-
late-like mammals: taeniodonts, tillodonts, pantodonts and dinoceratans
(Figure 7.7). Interestingly enough, these animals almost all fall into size
categories 4 and 5, so that the Paleocene herbivore communities were
made up of small, bunodont animals and large, semi-lophed ones, with
almost no taxa of medium size. Note, however, that these larger taxa were
absolutely rare as faunal components (R. Stucky, pers. comm.). The single
genus representing size category 3 is also the one hypsodont semi-lophed
taeniodont.
By the early Eocene, the number and proportion of smaller herbivores
increased, a reflection of the appearance of the perissodactyls and artio-
dactyls. No very large herbivores were present, reflecting the extinction of
the barylambdid pantodonts (the hippo-like Eocene pantodont, Corypho-
don, was not as large as the more ground-sloth-like Paleocene barylamb-
dids). The number of taxa in size category 3 increased slightly, reflecting
the evolution of larger lophed tapiroids (see Figure 7.7). By the middle
Figure 7.5. Generic diversity within body size classes among Paleogene ungulates and ungulate-like mammals. Key to shading: lightest tone, Size 1
(‘rabbit-sized’, ⬍5 kg); medium-light tone, Size 2 (‘dog-sized’, 5–25 kg); cross-hatching, Size 3 (‘antelope-sized’, 25–150 kg); medium-dark tone,
Size 4 (‘horse-sized’, 150–500 kg); checkered tone, Size 5 (‘rhino-sized’, ⬎500 kg). Column on right-hand side shows present-day African
ungulate genera for comparison.
Figure 7.6. Percentages of generic diversity within body size classes among Paleogene ungulates and ungulate-like mammals. Key to shading as
for Figure 7.5.
Figure 7.7. Generic diversity within different body size classes, broken down by
dental structural type. (A) Bunodont taxa. (B) Semi-lophed taxa. (C) Lophed
taxa. Key to shading as for Figure 7.5.
Eocene, very large terrestrial herbivores reappeared, represented by

rhinos (amynodonts) and the superficially rhino-like brontotheres and
dinoceratans. Note, however, that through the early and middle Eocene
the most prevalent component of the fauna was made up by taxa in size
category 2, representing semi-lophed artiodactyls and lophed perissodac-
tyls (Figure 7.7).
During the late Eocene and into the Oligocene, the numbers of the
smallest herbivores, in size category 1, declined severely, representing pri-
marily the decline and extinction of small, primitive artiodactyls, encom-
passing all three tooth types (Figure 7.7). The number of taxa in the
largest size categories remained fairly constant, as did those in size cate-
gory 2, but a significant increase was seen in the number of taxa in the
medium-sized category 3. This increase came almost entirely from the
increase in numbers of lophed ungulates, such as tylopod (oreodont,
camelid, and protoceratid) and ruminant (traguloid) artiodactyls, and
equid (anchitheriine) and rhinocerotoid (hyracodont) perissodactyls.
Note that in the present-day fauna almost all the taxa in category 3 are
fully lophed (the bunodont bush pig is the only exception).
Discussion
Patterns emerging from the ungulate diversity data
The paucity of Paleocene lophed taxa

The considerable herbivore diversity during the Paleocene comprised a
mixture of small, bunodont ungulate taxa, and large, semi-lophed non-
ungulate herbivorous taxa. It was not until the late Paleocene that lophed
or semi-lophed ungulates (also small) appeared in the fossil record, and
lophed ungulates did not represent a significant component of the fauna
until the Eocene. As previously discussed, the skulls and teeth of the
Paleocene large ungulate-like herbivores show some radical differences
from those of extant large herbivores, with an apparently small tooth
volume for their body size (Janis 1979). It thus seems unlikely that these
animals had the type of folivorous diet (fibrous, with a large volume of
intake) found in present-day large folivorous ungulates. It appears more
likely that the type of terrestrial folivory encountered today was not prev-
alent until the Eocene.
Almost all the Paleogene bunodont ungulates were small, or small-to-
intermediate in size (i.e., under 25 kg). Extant bunodont ungulates (pigs)
are mainly medium-sized or of medium-large size, as was characteristic

for the North American fauna by the Oligocene. Similarly, the lophed
ungulate taxa were initially of fairly small size: it was not until the late
middle Eocene that larger lophed taxa (rhinocerotoids) first emerged.
This suggests that the warmer world of the early Paleogene supported a
larger diversity of smallish terrestrial herbivores at temperate latitudes
than found in modern faunas at any latitude, perhaps indicative of a very
different climatic and vegetational regime from anywhere in the world
today (see discussion below).
The diversity increase of the late middle Eocene

The general pattern of generic diversity in the northern latitudes rather
closely follows the paleotemperature curve (Figure 7.1) for mammals
(Stucky 1990, 1992) and for amphibians and reptiles (Hutchison 1982).
There was an increase in both diversity and temperature over the Paleo-
cene/Eocene boundary, with a peak in the late early to early middle
Eocene, followed by a decline to the end of the Paleogene. Gingerich
(1984) especially noted how the diversity of primates in the northern lati-
tudes closely followed the paleotemperature curve for the entire Ceno-
zoic. The faunal changes of the middle and late Eocene of North America
were marked by the extinction of many archaic Eocene taxa, such as the
archaic ungulates (‘condylarths’) and ungulate-like taxa considered here,
and also among other taxa such as plesiadapiforms (early ?primates), true
primates (adapiforms and omomyids), archaic carnivorous mammals
(‘miacoid’ carnivorans and creodonts), and archaic insectivorous
mammals (‘proteutherians’), as well as non-eutherian taxa such as multi-
tuberculates and didelphoid marsupials (Savage and Russell 1983). While
many archaic mammals became extinct, many of the more modern
groups of mammals (for example, the modern families of carnivorans and
artiodactyls) made their first appearance in the late Eocene. But the
overall pattern was a decline in diversity.
However, the diversity of large terrestrial herbivores did not closely
follow this overall trend. Diversity increased through the Paleo-
cene/Eocene boundary, and up to the middle Eocene, followed by an
initial decline in the Bridgerian (early middle Eocene), as with the major-
ity of other mammals. However, there was a marked, further increase in
diversity in the later middle Eocene, peaking in the late Uintan and
Duchesnean, and then declining back to early Eocene values in the late
Eocene (Chadronian). The absolute magnitude of this increase (a doubling
of generic diversity between early middle and late middle Eocene) may be
inflated; for example, many late middle Eocene primitive artiodactyl taxa
are known from a single or very few occurrences (see Stucky 1998).
However, the primary pattern of increase is real, even when taxa known
from restricted occurrences are eliminated (Janis 1997–98). A peak in
generic diversity of ungulates in the late Eocene (rather than the late
middle Eocene), is reported by Jernvall et al. (1996, 2000), for Europe and in
Asia, as well as in North America. (However, note that the data in this
chapter are more up-to-date than those used by Jernvall et al., which were
derived from Savage and Russell 1983 [J. Hunter, pers. comm].)
In North America, the late middle to late Eocene (Uintan–Chadronian)
marked the appearance of many new taxa, such as tapirid and rhinocero-
toid perissodactyls, and families of suiform and selenodont (tylopod plus
ruminant) artiodactyls. The majority of these new taxa probably repre-
sent an immigration event from Asia at this time (see Janis 1993). Thus,
regardless of the nature of the events during the later Eocene that were
detrimental to archaic mammals, and to arboreal modern mammals such
as primates in the higher latitudes, they were favorable to the diversifica-
tion of terrestrial herbivores.
The late middle Eocene increase in semi-lophed taxa

One might imagine that the supposed cooling and drying of the later
Eocene (see Berggren and Prothero 1992 and references therein) would
represent a situation advantageous to modern types of ungulates —
mainly large (size categories 3–5), folivorous animals with hypsodont
lophed teeth, living in a cool and dry world today. However, the ungulate
taxa of this type did not form the initial diversification at the start of the
‘Ice House’ world in the late middle Eocene. Rather the bulk of the diver-
sity of Uintan and Duchesnean taxa was made up of small (under 25 kg),
semi-lophed, primitive (‘dichobunids’) and early selenodont artiodactyls
(camelids and traguloids), and small to medium-sized (10–150 kg) equid
and tapiroid perissodactyls, almost none of them hypsodont. Jernvall et
al. (2000) show a similar late Eocene rise (and subsequent Oligocene
decline) in their ‘one-lophed’ taxa (broadly corresponding to my category
of ‘semi-lophed’).
There is no equivalent to this type of diversity of terrestrial herbivor-
ous taxa in present-day tropical forests (see also Andrews 1992). Perhaps
the closest equivalent today would be in the forests of tropical Central and
South America, where there is a diversity of ‘ungulate-like’ caviomorph
rodents such as agoutis and pacas. In contrast to Old World tropical

forests, where the only small terrestrial ungulates are tragulids (sparse in
both taxonomic diversity and absolute numbers), the New World tropical
forests have a greater abundance of these large rodents, which have been
perceived as playing the role of an ungulate-like ecomorph (Dubost 1968).
However, these present-day tropical forests also contain a large diversity
of arboreal herbivores, and thus cannot be precisely compared with the
late Eocene of North America when the numbers of arboreal herbivores,
such as primates, were in rapid decline.
The types, and body size diversity, of lophed taxa held fairly steady
from the late Eocene through to the end of the Oligocene. But there was a
large change in the number and size diversity of semi-lophed taxa over
this time period. The large taxa, the brontotheres and dinoceratans (size
category 5), were all extinct by the end of the Eocene. Among the smaller
ones, the ‘dichobunid’ artiodactyls (mostly size category 1) almost all went
extinct (with the exception of the leptochoerids), and the more derived
selenodont ones, such as ruminants and camelids (mostly size category 2),
were replaced by related, fully lophodont forms. By the end of the Oligo-
cene, the taxonomic diversity of semi-lophed taxa in North America was
as reduced as it is in present-day Africa, and the remaining taxa occupied
only the two smallest size categories. Jernvall et al. (2000) note a reappear-
ance of a small number of taxa with a single loph in the late Oligocene and
Miocene (apparently represented by the re-immigration of chalicotheres,
now representing animals in size category 4 rather than category 2,
although I would probably term these taxa as ‘fully lophed’).
Another interesting point noted by Jernvall et al. (1996, 2000) is the dif-
ference in the patterns of diversity and disparity in ungulate dental types
between the Paleogene and the Neogene. Whereas ungulate generic
diversity rose again in the Miocene, following the fall in the late Paleo-
gene, and the disparity of dental types remained high (both bunodont
forms and forms with four or more dental lophs are apparent), the diver-
sity of tooth forms never reached the level seen in the Paleogene. Only
seven crown types (out of a total range of 28) are present among extant
ungulates, whereas, during the later Eocene, up to a dozen or so different
crown types were present at any one time interval. Jernvall et al. interpret
these data as indicating a greater number of environmental opportunities
for ungulates in later Eocene habitats than is found today, although it is
not clear how many of these ‘niches’ have been taken over by rodents and
other mammals in today’s ecosystems. The pattern of dental diversity
among present-day small herbivores, such as rodents and lagomorphs,

really needs to be included in order to complete this picture. Rodent
diversity and abundance greatly increased in the late middle Eocene in
North America, with the appearance of forms with complexly lophed
cheek teeth (eomyids and zapodids), and, furthermore, lagomorphs made
their first appearance at this time (Webb 1989). In Europe, there was also a
diversification of rodents with teeth indicative of a folivorous diet (theri-
domyids) during the late Eocene (Collinson and Hooker 1991). It is
unlikely that small rodents or lagomorphs would have been in direct
competition with larger ungulates for the same resources; rather, this
increase in rodent diversity most likely reflects a climatic shift, as dis-
cussed later.
The interaction between size categories and dental types

Among present-day ungulates, the majority of African taxa are in the
middle size range (category 3), and the majority also have lophed teeth.
However, it is perhaps surprising to see that, among Paleogene terrestrial
herbivores, size category 3 was almost completely occupied by lophed
herbivores (Figure 7.7). Paleocene faunas, almost completely lacking in
lophed taxa (and the true lophed taxa at the end of the epoch being only
small animals), contained small (bunodont) and large (semi-lophed) ter-
restrial herbivores, but no medium-sized ones. In fact, the rise of the
medium-sized terrestrial herbivores largely reflects the rise to predomi-
nance of forms with lophed teeth. The biological significance of this cor-
relation of body size with dental structure (and hence with diet) is not
clear, as the animals in the modern fauna clearly demonstrate that it is pos-
sible to be a lophed folivore at a range of body sizes. It must then be the
case that more omnivorous and selective browsing taxa are somehow
avoiding this size category (as hinted at by the diversity of modern African
animals, but the sample size is small). Again, the biological reasons for
this are unclear.
Evolutionary and ecological implications

As discussed above, these data suggest that full folivory, as indicated by
the presence of taxa with fully lophed teeth, was not seen in North Amer-
ican mammalian communities until the latest Paleocene, and that a radi-
ation of more diverse, larger-bodied folivores did not occur until the late
middle Eocene. Other studies of Paleocene mammals reveal little evi-
dence of folivorous forms (Wing and Tiffney 1987; Stucky 1990; Krause
and Maas, 1990; Collinson and Hooker 1991). The reasons for this must lie
in paleoenvironmental factors. The ten million years of the Paleocene cer-
tainly represented a sufficient amount of time for the evolution of a wide
variety of mammalian sizes and body types. There is no reason to assume
that this time period would have been insufficient for the evolution of
adaptations for folivory. One speculation concerning early Cenozoic eco-
systems is that the demise of the dinosaurs would have resulted in a sig-
nificant change in environmental conditions (Wing and Tiffney 1987;
Stucky 1990). The large herbivorous dinosaurs of the Late Cretaceous may
have browsed the herbage to such levels that an open type of habitat was
created and sustained, much as elephants maintain certain types of open
habitats today. Sudden removal of the large herbivores may have resulted
in a floral ‘rebound,’ creating dense forest with herbage that would be
inaccessible to potential small mammalian herbivores. Alternatively (or
perhaps additionally), a more equable climate in the Paleocene in compar-
ison with that of the Late Cretaceous may have favored a closed canopy
woodland over the evidently more open habitats of the late Mesozoic
(Wolfe and Upchurch 1986; Wolfe 1990). The bolide impact at the Creta-
ceous/Tertiary boundary may also have had a long-term impact on the
North American vegetation, resulting in low floristic diversity (Wing
1998), which may have had an effect on the diversity of terrestrial herbi-
vores. Large seeds in the Paleocene evidently provided food for terrestrial
omnivores (Wing and Tiffney 1987), but the leaves may have been too
fibrous (see Janis 1989).
Given the above speculation, what changes could have caused a
restructuring of the habitat in the Eocene, such that the vegetation was
now accessible to small (under 25 kg) terrestrial ungulates (as demon-
strated by their lophed or semi-lophed teeth)? Warming during the late
Paleocene may have made the vegetational habitat more favorable for
small ungulates; the increase in mean annual temperature and decrease
in temperature seasonality at this time should have resulted in a less sea-
sonal supply of food resources, as there would have been less annual
fluctuation in leaf and fruit production (Wing 1998). The effects of this
warming event may explain the first appearance of fully lophed taxa at
this time. The faunas of the higher latitudes in the early Eocene provide
evidence of a tropical forest-like habitat, with a high diversity of ar-
boreal primates and gliders, but the terrestrial forms were much more
diverse than in the Paleocene (Collinson and Hooker 1987; Rose 1990;
Stucky 1990; Andrews 1992). However, the combination of mammalian
ecomorphs, with a high diversity of both terrestrial frugivores and ar-

boreal insectivores, has no precise counterpart in any present-day
habitat (Collinson and Hooker 1991; Andrews 1992).
The diversity of lophed and semi-lophed ungulates in the early Eocene
of North America suggests a different type of forest habitat. There must
have been a more open forest floor, with a less dense tree canopy, and more
understory growth to provide food for these animals. This inferred
greater spacing of the trees suggests that the higher latitudes during the
early Eocene were not only warmer but also drier than in the Paleocene, or
at least with more seasonally pulsed episodes of rainfall. The apparent
increased palatability of herbage to the mammals also suggests a more
seasonal type of vegetation, where the leaves would be less fibrous and
hence more easily digested (Deinum and Dirven 1975). Gunnell et al.
(1995), in an analysis of the entire North American faunal community, also
concluded that the early Eocene may have been drier than the Paleocene,
and Wolfe (1987) reported a predominance of deciduous plants occurring
with the warming that occurred in North America from the late Paleocene
into the early Eocene.
Although the ‘heyday’ of the evolution of large herbivorous mammals
is often seen as being correlated with the spread of the savanna grasslands
in the Miocene (Webb 1977, 1989; MacFadden, this volume), reflected by a
peak in generic diversity (slightly exceeding late Eocene levels) on all
three northern continents, it should be appreciated that much of the
morphological diversity of ungulate tooth crown types present in the late
Eocene was lost by the end of the Paleogene (Jernvall et al. 1996, 2000). The
late Eocene represented a peak time for the diversification of small-to-
medium-sized ungulates, types of animals that are rare components of
the modern, mainly large-bodied, ungulate fauna.
Why the later part of the Eocene, which was the time of extinction of so
many of the archaic mammals, should favor the diversification of these
smaller, semi-lophed ungulates remains a bit of a mystery. Both paleo-
botanical and faunal diversity data indicate an increase in seasonality
during the later Eocene. The vegetation in higher latitudes was less tropi-
cal in nature in the late middle Eocene than earlier in the epoch (Wolfe
1985; Collinson and Hooker 1987). Increasing seasonal dryness affected
the vegetation in the continental interior (Wing 1998), and, in the late
Eocene, tropical vegetation was replaced by more temperate types, such as
mixed coniferous and deciduous woodlands, and winter frosts may have
been present (Wolfe 1978, 1985; Collinson, Fowler and Boulter 1981). An
increased abundance of small seeds in the late Eocene is suggestive of sea-

sonal plant growth cycles (Wing and Tiffney 1987). A more open, less
forested, type of habitat for the late Eocene at higher latitudes is also sug-
gested by changes in the postcranial structure of rodents and ungulates,
with the appearance of taxa with cursorial adaptations, and the extinc-
tion of arboreal mammals such as many primates (Webb 1978, 1989;
Stucky 1990, 1992). Changes in diversity patterns are seen not only among
the high-latitude mammals, but also among the amphibians and reptiles,
which declined in diversity in the later Eocene (Hutchison 1982).
Increase in seasonality may have made more normally fibrous parts of
plants more readily available, as such environmental conditions result in
greater differentiation in fiber content between the leaves and the stem
(Deinum and Dirvan 1975). This vegetational change would have espe-
cially affected small, selective-feeding artiodactyls, which formed a large
component of the late Eocene diversification (see Janis 1989). However,
the later Eocene has been traditionally interpreted as a time of climatic
cooling (Berggren and Prothero 1992 and references therein). While the
disappearance of most frugivores, and of obligate arboreal mammals such
as primates, coincides with this scenario, it is difficult to envisage a cool,
seasonal climate containing a diversity of tragulid-like, selective-brows-
ing artiodactyls, ungulates that today are confined to tropical environ-
ments. Yet the semi-lophed ‘dichobunids’, tylopods, and ruminants that
diversified the late middle Eocene appear to have been almost identical in
craniodental structure to the extant tragulids. The abrupt drop in diver-
sity of semi-lophed taxa in the Oligocene corresponds with a dramatic
temperature fall in the earliest Oligocene, when mean annual tempera-
tures in North America declined by about 13 °C in as little as one million
years (Prothero and Heaton 1996; Wing 1998).
A possible window on the conditions of the late middle Eocene is pro-
vided by the work of Badgley and Fox (in review), who documented
climatic variables and ecological structure for modern mammalian com-
munities in all of North America. Canonical ordination of climatic and
ecological variables resulted in localities being ordinated in multidimen-
sional space, with a primary/first axis representing seasonality of tem-
perature (high seasonality corresponding to cold climates, low
seasonality corresponding to low climates of today), and a secondary axis
representing moisture stress (high precipitation versus high evapotran-
spiration). The ecological structure (size structure and trophic struc-
tures) of faunas was plotted in the same multidimensional space. Figure
Figure 7.8. Position of present-day North American taxa on multidimensional

environmental axes (from Badgley and Fox, in review), with possible positions
of North American Paleogene taxa superimposed. Circles indicate where
particular modern localities fall in the multidimensional climate space, and
the size of the circle indicates the numbers of modern species in that locality.
Key to the shading of the circles: filled circles, frugivorous taxa; unfilled
circles, herbivorous and granivorous taxa (dotted circles, preponderance of
small, granivorous taxa; wavy line circles, preponderance of large, folivorous
taxa). (A) probable placement in multidimensional climate space of
Paleocene–early Eocene bunodont ungulates and primates. (B) probable
placement in multidimensional climate space of middle-late Eocene ungulates
(semi-lophed). (C) probable placement in multidimensional climate space of
Oligocene and Neogene folivorous ungulates (lophed).
7.8 represents a compilation of several of their figures, and includes a

‘best guess’ estimate of where Paleogene mammals would fall in this
multidimensional space. Note that present-day frugivorous taxa (small
to medium-sized and represented by primates, rodents, bats and procyo-
nid carnivores) fall mainly in the warm/wet quadrant, herbivorous taxa
(including large ungulates and small rodents) fall mainly in the cold/dry
quadrant, whereas granivorous taxa (mainly of small size) fall mainly in
the warm/dry quadrant.
Although it is impossible to make a precise extrapolation from the

modern-day taxa and climates to those of the North American Paleogene,
the following interpretation seems likely. Paleogene primates and other
herbivorous mammals probably had similar environmental demands to
present-day primates, and would have occupied the warm-wet quadrant
of the multidimensional space (Figure 7.8, ‘A’). Likewise the lophed, foliv-
orous herbivorous ungulates of the Oligocene would probably have occu-
pied the cold-dry quadrant of the space, as do the ungulates today (Figure
7.8, ‘C’). The major speculation I advance is that the diversity of small, or
small to medium-sized semi-lophed ungulate taxa of the late middle
Eocene, for which there is no direct modern analog, would be broadly
analogous to the ‘granivorous’ taxa of Badgley and Fox (today largely rep-
resented by rodents), falling in the warm-dry quadrant in the multi-
dimensional space (Figure 7.8, ‘B’).
While the dental structure of the later Eocene small semi-lophed artio-
dactyls resembles that of tragulids, which today inhabit moist, tropical
habitats, it is not too dissimilar (in terms of functional equivalence) to
that of bunolophodont granivorous rodents. That is, these teeth could
equally well be used to masticate seeds in a more arid environment as well
as buds in a more moist one. A change from a tropical, non-seasonal
climate to a colder, seasonal one (as has always been hypothesized for the
later Eocene) would affect selective feeders equally in both terrestrial and
arboreal environments. In contrast, a climatic change predominantly
in the direction of aridity might have significant effects on arboreal
mammals by affecting the structure of the forest canopy and the seasonal-
ity of large fruit production, but have a lesser effect on the availability of
seed and pod resources derived from lower-level plants for terrestrial
selective herbivores.
Therefore, the composition of the late middle Eocene ungulate fauna,
with the increased diversity of the semi-lophed taxa that are most likely
indicative of warm-dry conditions (or that at least contradict the notion of
a cool climate), does not suggest that cooling was a significant issue until
the late Eocene (Chadronian) and, especially, during the Oligocene.
Although paleobotanical data also suggest a drying trend in the later
Eocene of North America (Frederiksen 1995; Wing 1998), there has been,
as far as I am aware, no other proposal that middle Eocene drying
occurred in the absence of significant cooling (although Frederiksen
[1995] did suggest that drying may be the main event). This suggestion of
an initial middle Eocene interval of drying without cooling is at variance
with the paleotemperature data obtained from the benthic foraminifera,

but may better reflect Paleogene conditions in the continental interior.
Alternatively, it may be the case that cooling had an insignificant effect on
the flora and fauna until such a point that a significant frost season was
invoked (C. Badgley, pers. comm.).
This chapter has been concerned with the faunal diversities and prob-
able paleoecological conditions in North America. Despite broad similar-
ities in the Eocene across the northern continents in ungulate taxonomic
and morphological diversity (Jernvall et al. 1996, 2000), details of floral
and faunal evolution may well have been different on different conti-
nents. For example, angiosperm pollen shows that floral diversity is
greatest during the middle middle Eocene in North America, but in the
early early Eocene in western Europe; differing paleogeographic configu-
rations and rates of species origination and extinction may account for
these differences (Frederiksen 1995). Note, however, that Wing, Alroy, and
Hickey (1995) did not support these conclusions for North America; their
data, based on whole plant fossils, show a maximum diversity in the earli-
est Eocene (although these differences between the researchers may relate
to the different areas of North America under study). J. J. Hooker (pers.
comm.) notes that, among ungulate tooth types, taxa with bilophodont
teeth (like those of modern tapirs) were more common in western Europe
during the Eocene than in North America. This difference may merely
represent the phylogenetic composition of the two areas (the diversity of
ceratomorph perissodactyls was considerably different, with tapiroids
being more common in Europe and rhinocerotoids more common in
North America), or it may represent a difference in dental types that
reflected different environmental conditions. The data in this chapter
have generated some interesting speculations about Paleogene changes
in paleoecology and paleoclimates in North America; it remains to be seen
if these patterns hold up globally, or in North America with more detailed
study.
Acknowledgments
I thank Hans-Dieter Sues and Conrad Labandeira for inviting me to par-

ticipate in their symposium ‘Origin and Evolution of Terrestrial Her-
bivory’ at the Sixth North American Paleontological Convention (1996),
and Catherine Badgley, Robert Emry, Jeremy Hooker, John Hunter, and
Richard Stucky for comments on the original manuscript.
References
Andrews, P. (1992). Community evolution in forest habitats. J. Hum. Evol. 18:326–343.

Archibald, J. D. (1996). Fossil evidence for a Late Cretaceous origin of ‘hoofed’ mammals.
Science 272:1150–1153.
Archibald, J. D. (1998). Archaic ungulates (‘Condylarthra’). In The Evolution of Tertiary
Mammals of North America, ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp.
Badgley, C., and Fox, J. L. In review. Ecological structure of North American
mammalian faunas in relation to climate. Journal of Biogeography.
Berggren, W. A., and Prothero, D. R. (1992). Eocene–Oligocene climatic and biotic
evolution: an overview. In Eocene–Oligocene Climatic and Biotic Evolution, ed. D. R.
Prothero and W. A. Berggren, pp. 1–28. Princeton: Princeton University Press.
Bertram, J. E. A., and Biewener, A. A. (1990). Differential scaling of the long bones in the
terrestrial Carnivora and other animals. J. Morphol. 20:157–169.
Brown, J. H. (1995). Macroecology. Chicago: University of Chicago Press.
Burchardt, B. (1978). Oxygen isotope paleotemperatures from the Tertiary period in the
North Sea area. Nature 275:121–123.
Cifelli, R. L., and Schaff, C. R. (1998). Arctostylopida. In The Evolution of Tertiary Mammals
of North America, ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 332–336.
Colbert, M. W., and Schoch, R. M. (1998). Tapiroidea and other moropomorphs. In The
Evolution of Tertiary Mammals of North America, ed. C. M. Janis, K. M. Scott, and L.
L. Jacobs, pp. 569–582. Cambridge and New York: Cambridge University Press.
Collinson, M. E., Fowler, M. K., and Boulter, M. C. (1981). Floristic changes indicate a
cooling climate in the Eocene of southern England. Nature 291:315–317.
Collinson, M. E. and Hooker, J. J. (1987). Vegetational and mammalian fauna changes in
the early Tertiary of southern England. In The Origin of Angiosperms and their
Biological Consequences, ed. E. M. Friis and W. G. Chaloner, pp. 259–303.
Collinson, M. E., and Hooker, J. J. (1991). Fossil evidence of interactions between plants
and plant-eating mammals. Phil. Trans. R. Soc. Lond. B 333:197–208.
Coombs, M. C. (1998). Chalicotherioidea. In The Evolution of Tertiary Mammals of North
America, ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 560 – 568. Cambridge
Dashzeveg, D., and Hooker, J. J. (1997). New ceratomorph perissodactyls (Mammalia)
from the Middle and Late Eocene of Mongolia: their implications for phylogeny
and dating. Zool. J. Linn. Soc. 120:105–138.
Deinum, B., and Dirven, J. G. P. (1975). Climate, nitrogen and grass. 6. Comparison of
yield and chemical composition of some temperate and tropical grass species
grown at different temperatures. Neth. J. Agric. Sci. 23:69–82.
Dubost, G. (1968). Les niches écologiques de forêts tropicales sud-américaines et
africaines, sources de convergences remarquables entre Rongeurs et
Artiodactyles. La Terre et la Vie 22:3–28.
Effinger, J. A. (1998). Entelodontidae. In The Evolution of Tertiary Mammals of North
America, ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 375–380. Cambridge
Eisenberg, J. F. (1981). The Mammalian Radiations. Chicago: University of Chicago Press.
Fortelius, M., Van der Made, J., and Bernor, R. L. (1996). Middle and late Miocene
Suoidea of central Europe and the eastern Mediterranean: evolution,
biogeography, and paleoecology. In The Evolution of Western Eurasian Neogene
Mammal Faunas, ed. R. Bernor, V. Fahlbusch, and H.-W. Mittmann, pp. 348–377.
New York: Columbia University Press.
Fortelius, M., Werdelin, L., Andrews, P., Bernor, R. L., Gentry, A., Humphrey, L.,
Mittman, H.-W., and Viranta, S. (1996). Provinciality, diversity, turnover, and
paleoecology in land mammal faunas of the later Miocene of Western Eurasia. In
The Evolution of Western Eurasian Neogene Mammal Faunas, ed. R. Bernor, V.
Fahlbusch, and H.-W. Mittmann, pp. 414–448. New York: Columbia University
Press.
Frederiksen, N. O. (1995). Differing Eocene floral histories in southeastern North
America and western Europe: influence of paleogeography. Hist. Biol. 10:13–23.
Gingerich, P. D. (1984). Primate evolution. In Mammals: Notes for a Short Course, ed. P. D.
Gingerich and C. E. Badgley, pp. 167–181. Knoxville: University of Tennessee,
Department of Geological Sciences, Studies in Geology, 8.
Gunnell, G. F., Morgan, M. E., Maas, M. C., and Gingerich, P. D. (1995). Comparative
paleoecology of Paleogene and Neogene mammalian faunas: trophic structure
and composition. Palaeogeogr., Palaeoclimatol., Palaeoecol. 115:265–286.
Honey, J. G., Harrison, J. A., Prothero, D. R., and Stevens, M. S. (1998). Camelidae. In The
Evolution of Tertiary Mammals of North America, ed. C. M. Janis, K. M. Scott, and L.
L. Jacobs, pp. 439–462. Cambridge and New York: Cambridge University Press.
Hunter, J. P. (1997). Adaptive Radiation of Early Paleocene Ungulates. Unpublished Ph.D.
dissertation, State University of New York at Stony Brook.
Hutchison, J. H. (1982). Turtles, crocodilians, and champsosaur diversity changes in the
Cenozoic of the north-central region of Western United States. Palaeogeogr.,
Palaeoclimatol., Palaeoecol. 37:149–151.
Janis, C. M. (1979). Aspects of the Evolution of Ungulate Mammals. Unpublished Ph.D.
dissertation, Harvard University. Cambridge, MA.
Janis, C. M. (1989). A climatic explanation for patterns of evolutionary diversity in
ungulate mammals. Palaeontology 32:463–481.
Janis, C. M. (1990a). The correlation between diet and dental wear in herbivorous
mammals, and its relationship to the determination of diets of extinct species.
In Evolutionary Paleobiology of Behavior and Coevolution, ed. A. J. Boucot, pp.
241–259. Amsterdam and New York: Elsevier.
Janis, C. M. (1990b). Correlation of cranial and dental variables with body size in
ungulates and macropodoids. In Body Size in Mammalian Paleobiology: Estimation
and Biological Implications, ed. J. Damuth and B. J. MacFadden, pp. 255–300.
Janis, C. M. (1993). Tertiary mammal evolution in the context of changing climates,
vegetation, and tectonic events. Annu. Rev. Ecol. Syst. 24: 467–500.
Janis, C. M. (1995). Correlation between craniodental morphology and feeding behavior
in ungulates: reciprocal illumination between living and fossil taxa. In
Functional Morphology in Vertebrate Paleontology, ed. J. J. Thomason, pp. 76–98.
Janis, C. M. (1997–98). Ungulate teeth, diets, and climatic changes at the
Eocene/Oligocene boundary. Zoology 100:203–220.
Janis, C. M., Scott, K. M., and Jacobs, L. L. (eds.) (1998). Evolution of Tertiary Mammals of
North America. Cambridge and New York: Cambridge University Press.
Jarman, P. J. (1974). The social organisation of antelopes in relation to their ecology.
Behaviour 48: 213–267.
Jernvall, J. (1995). Mammalian molar cusp patterns: Developmental mechanisms of
diversity. Acta Zool. Fenn. 198:1–61.
Jernvall, J., Hunter, J. P., and Fortelius, M. (1996). Molar tooth diversity, disparity, and
ecology in Cenozoic radiations. Science 274:1489–1492.
Jernvall, J., Hunter, J. P., and Fortelius, M. (2000). Trends in the evolution of molar
crown types in ungulate mammals: evidence from the Northern Hemisphere. In
Function and Evolution of Teeth, ed. M. Teaford, M. Ferguson, and M. M. Smith,
Kay, R. F. (1975). The functional adaptations of primate molar teeth. Am. J. Phys.
Anthropol. 43:195–216.
Krause, D. W. (1986). Competitive exclusion and taxonomic displacement in the fossil
record: the case of rodents and multituberculates in North America. Contrib. Geol.
Univ. Wyoming, Spec. Pap. 3:93–117.
Krause, D. W., and Maas, M. C. (1990). The biogeographic origins of late Paleocene–early
Eocene mammalian immigrations to the western interior of North America. In
Dawn of the Age of Mammals in the Northern Part of the Rocky Mountains, ed. T. M.
Bown and K. D. Rose. Geol. Soc. Am. Spec. Pap. 243:71–105.
Kron, D. G., and Manning, E. (1998). Anthracotheriidae. In The Evolution of Tertiary
Mammals of North America, ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp.
Lander, B. (1998). Oreodontoidea. In The Evolution of Tertiary Mammals of North America,
ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 402–425. Cambridge and New
Lewis, M. E. (1997). Carnivoran paleoguilds of Africa: implications for hominid food
procurement strategies. J. Hum. Evol. 32: 257–288.
Lillegraven, J. A., Kielan-Jaworowska, Z., and Clemens, W. A. (eds.) (1979). Mesozoic
Mammals: The First Two-Thirds of Mammalian History. Berkeley: University of
California Press.
Lucas, P. W., and Luke, D. A. (1984). Chewing it over: basic principles of food breakdown.
In Food Acquisition and Processing in Primates, ed. D. J. Chivers, B. A. Wood, and A.
Bilsborough, pp. 283–302. New York: Plenum Press.
Lucas, S. G. (1992). Redefinition of the Duchesnean land mammal ‘age’, late Eocene of
western North America. In Eocene–Oligocene Climatic and Biotic Evolution, ed. D. R.
Prothero and W. A. Berggren, pp. 88–105. Princeton: Princeton University Press.
Lucas, S. G. (1998). Tillodontia. In The Evolution of Tertiary Mammals of North America, ed.
C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 268–278. Cambridge and New York:
Lucas, S. G., and Schoch, R. M. (1998a). Pantodonta. In The Evolution of Tertiary Mammals
of North America, ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 274–283.
Lucas, S. G., and Schoch, R. M. (1998b). Dinocerata. In The Evolution of Tertiary Mammals of
North America, ed. C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 284–291.
Lucas, S. G., Schoch, R. M., and Williamson, T. E. (1998). Taeniodonta. In The
Evolution of Tertiary Mammals of North America, ed. C. M. Janis, K. M. Scott, and
L. L. Jacobs, pp. 260–267. Cambridge and New York: Cambridge University
Press.
Maas, M. C., and Krause, D. W. (1994). Mammalian turnover and community structure
in the Paleocene of North America. Hist. Biol. 8:91–128.
Maas, M. C., Anthony, M. R. L., Gingerich, P. D., Gunnell, G. F., and Krause, D. W. (1995).
Mammalian generic diversity and turnover in the Late Paleocene and Early
Eocene of the Bighorn and Crazy Mountain Basins, Wyoming and Montana
(USA). Palaeogeogr., Palaeoclimatol., Palaeoecol. 115:181–207.
MacFadden, B. J. (1998). Equidae. In The Evolution of Tertiary Mammals of North America, ed.
C. M. Janis, K. M. Scott, and L. L. Jacobs, pp. 537–559. Cambridge and New York:
Mader, B. J. (1998). Brontotheriidae. In The Evolution of Tertiary Mammals of North America,
Prothero, D. R. (1998a). Oromerycidae. In The Evolution of Tertiary Mammals of North
Prothero, D. R. (1998b). Protoceratidae. In The Evolution of Tertiary Mammals of North
Prothero, D. R. (1998c). Hyracodontidae. In The Evolution of Tertiary Mammals of North
Prothero, D. R. (1998d). Rhinocerotidae. In The Evolution of Tertiary Mammals of North
Prothero, D. R., and Heaton, T. H. (1996). Faunal stability during the Early Oligocene
climatic crash. Palaeogeogr., Palaeoclimatol., Palaeoecol. 127:257–283.
Prothero, D. R., Manning, E., and Fischer, M. (1988). The phylogeny of the ungulates. In
The Phylogeny and Classification of Tetrapods, Vol. 2, ed. M. J. Benton, pp. 201–234.
Systematics Association Special Volume No. 35B. Oxford: Clarendon Press.
Rose, K. D. (1990). Postcranial skeletal remains and adaptations in early Eocene
mammals from the Willwood Formation, Bighorn Basin, Wyoming. In Dawn of
the Age of Mammals in the Northern Part of the Rocky Mountains, ed. T. M. Bown and K.
D. Rose. Geol. Soc. Am. Spec. Pap. 243:107–113.
Russell, D. E. (1975). Paleoecology of the Paleocene–Eocene transition in Europe. In
Approaches to Primate Paleobiology, ed. F. S. Szalay, pp. 28–61. Basel: Karger.
Savage, D. E., and Russell, D. E. (1983). Mammalian Paleofaunas of the World. Reading, MA:
Addison-Wesley Publishing Company.
Sepkoski, J. J., Jr., and Kendrick, D. C. (1993). Numerical experiments with model
monophyletic and paraphyletic taxa. Paleobiology 19:168–184.
Stucky, R. K. (1990). Evolution of land mammal diversity in North America during the
Cenozoic. In Current Mammalogy, ed. H. H. Genoways, pp. 375–432. New York:
Plenum Press.
Stucky, R. K. (1992). Mammalian faunas in North America of Bridgerian to early
Arikareean ‘ages’ (Eocene and Oligocene). In Eocene–Oligocene Climatic and Biotic
Evolution, ed. D. R. Prothero and W. A. Berggren, pp. 464–493. Princeton:
Princeton University Press.
Stucky, R. K. (1998). Eocene bunodont and bunoselenodont Artiodactyla
(‘dichobunids’). In The Evolution of Tertiary Mammals of North America, ed. C. M.
Janis, K. M. Scott, and L. L. Jacobs, pp. 358–374. Cambridge and New York:
Wall, W. P. (1998). Amynodontidae. In The Evolution of Tertiary Mammals of North America,
Webb, S. D. (1977). A history of savanna vertebrates in the New World. Part I: North
America. Annu. Rev. Ecol. Syst. 8:355–380.
Webb, S. D. (1989). The fourth dimension in North American terrestrial mammalian
communities. In Patterns in the Structure of Mammalian Communities, ed. D. W.
Morris, Z. Abramsky and M. R. Willig, pp. 181–203. Special Publication of the
Museum, No. 28. Lubbock: Texas Tech University.
Webb, S. D. (1998). Hornless ruminants. In The Evolution of Tertiary Mammals of North
Wing, S. L. (1998). Tertiary vegetation of North America as a context for mammalian
evolution. In The Evolution of Tertiary Mammals of North America, ed. C. M. Janis, K.
M. Scott, and L. L. Jacobs, pp. 37–65. Cambridge and New York: Cambridge
University Press.
Wing, S. L., and Tiffney, B. H. (1987). The reciprocal interaction of angiosperm evolution
and tetrapod herbivory. Rev. Palaeobot. Palynol. 50:179–210.
Wing, S. L., Alroy, J., and Hickey, L. J. (1995). Plant and mammal diversity in the
Paleocene to Early Eocene of the Bighorn Basin. Palaeogeogr., Palaeoclimatol.,
Palaeoecol. 115:117–155.
Wolfe, J. A. (1978). A paleobotanical interpretation of Tertiary climates in the Northern
Hemisphere. Am. Sci. 66:694–703.
Wolfe, J. A. (1985). Distribution of major vegetational types during the Tertiary. Geophys.
Monogr. 32:357–375.
Wolfe, J. A. (1987). Late Cretaceous–Cenozoic history of deciduousness and the terminal
Cretaceous event. Paleobiology 13:215–226.
Wolfe, J. A. (1990). Palaeobotanical evidence for a marked temperature increase
following the Cretaceous/Tertiary boundary. Nature 343:153–156.
Wolfe, J. A., and Upchurch, G. R. (1987). North American nonmarine climates and
vegetation during the Late Cretaceous. Palaeogeogr., Palaeoclimatol., Palaeoecol.
61:33–77.
Woodburne, M. O. (ed.) (1987). Cenozoic Mammals of North America. Berkeley: University of
California Press.
Wright, D. B. (1998). Tayassuidae. In The Evolution of Tertiary Mammals of North America,
Appendix
Database for herbivorous mammals from the Paleogene of North America

considered in this chapter, arranged by stratigraphic provenance. Columns
(from left to right) represent order, family, genus, dental type and size class
(see text for explanation). Notes: Underlined dental type, hypsodont; n.g.,
unnamed new genus; inc.sed., incertae sedis.
PALEOCENE
Puercan 1
‘Condylarthra’ Hyopsodontidae Litomylus Bunodont 1
Oxyprimus Bunodont 1
Periptychidae Anisonchus Bunodont 1
Conacodon Bunodont 1
Maiorana Bunodont 1
Mimatuta Bunodont 1
Puercan 2
Taeniodonta Onychodectes Bunodont 2
Schochia Bunodont 2
Wortmania Bunodont 2
‘Condylarthra’ Hyopsodontidae Litomylus Bunodont 1
Mioclaenidae Bubogonia Bunodont 2
Choeroclaenus Bunodont 1
Ellipsodon Bunodont 1
Litaletes Bunodont 1
Promioclaenus Bunodont 1
Protoselene Bunodont 1
Tiznatzinia Bunodont 1
Carsioptychus Bunodont 2
Ectoconus Bunodont 2
Escatepos Bunodont 1
Gillisonchus Bunodont 1
Haploconus Bunodont 1
Hemithlaeus Bunodont 1
Oxyacodon Bunodont 1
Puercan 3
Taeniodonta Onychodectes Bunodont 2
Wortmania Bunodont 2
P A L E O C E N E (cont.)
Puercan 3 (cont.)
‘Condylarthra’ Hyopsodontidae Haplaletes Bunodont 1
Litomylus Bunodont 1
Mioclaenidae Choeroclaenus Bunodont 1
Ellipsodon Bunodont 1
Tiznatzinia Bunodont 1
Carsioptychus Bunodont 2
Ectoconus Bunodont 2
Gillisonchus Bunodont 1
Hemithlaeus Bunodont 1
Oxyacodon Bunodont 1
Tinuviel Bunodont 1
Torrejonian 1
Taeniodonta Conoryctella Bunodont 2
Mioclaenidae Ellipsodon Bunodont 1
Mioclaenus Bunodont 2
Periptychus Bunodont 2
Phenacodontidae Tetraclaenodon Bunodont 2
Torrejonian 2
Taeniodonta Conoryctella Bunodont 2
Pantodonta Pantolambda Semi-lophed 4
Torrejonian 2 (cont.)
Phenacodontidae Tetraclaenodon Bunodont 2
Torrejonian 3
Taeniodonta Conoryctes Bunodont 2
Huerfanodon Bunodont 2
Psittacotherium Semi-lophed 2
Pantodonta Pantolambda Semi-lophed 4
Titanoides Semi-lophed 4
Phenacodontidae Ectocion Semi-lophed 2
Phenacodus Bunodont 2
Tetraclaenodon Bunodont 2
Tiffanian 1
Taeniodonta Conoryctes Bunodont 2
Psittacotherium Semi-lophed 2
Pantodonta Caenolambda Semi-lophed 4
Pantolambda Semi-lophed 4
Mioclaenidae Litaletes Bunodont 1
Tiffanian 1 (cont.)
Tetraclaenodon Bunodont 2
Tiffanian 2
Taeniodonta Psittacotherium Semi-lophed 2
Pantodonta Caenolambda Semi-lophed 4
Mioclaenidae Promioclaenus Bunodont 1
Periptychidae Periptychus Bunodont 2
Tiffanian 3
Pantodonta Barylambda Semi-lophed 5
Caenolambda Semi-lophed 4
Haplolambda Semi-lophed 4
Dinocerata Prodinoceras Semi-lophed 4
‘Condylarthra’ Hyopsodontidae Aletodon Bunodont 1
Dorraletes Bunodont 1
Haplaletes Bunodont 1
Mioclaenidae Promioclaenus Bunodont 1
Tiffanian 4
Ignatiolambda Semi-lophed 4
Tiffanian 4 (cont.)
Utemylus Bunodont 1
Mioclaenidae Phenacodaptes Bunodont 1
Tiffanian 5
Ectoganus Semi-lophed 3
Ignatiolambda Semi-lophed 4
Leptolambda Semi-lophed 5
Mioclaenidae Phenacodaptes Bunodont 1
Tiffanian 6
Ectoganus Semi-lophed 3
Tillodonta Esthonyx Semi-lophed 2
Mioclaenidae Apheliscus Bunodont 1
Phenacodaptes Bunodont 1
Tiffanian 6 (cont.)
Arctostylopoidea Arctostylops Lophed 1
Clarkforkian 1
Taeniodonta Ectoganus Semi-lophed 3
Pantodonta Barylambdidae Barylambda Semi-lophed 5
Coryphodontidae Coryphodon Semi-lophed 4
Haplomylus Bunodont 1
Clarkforkian 2
Pantodonta Barylambdidae Caenolambda Semi-lophed 4
Coryphodontidae Coryphodon Semi-lophed 4
Phenacodontidae Copecion Bunodont 2
Ectocion Semi-lophed 2
Clarkforkian 3
Clarkforkian 3 (cont.)
Pantodonta Coryphodontidae Coryphodon Semi-lophed 4
Hyopsodus Bunodont 1
Meniscotherium Lophed 2
EOCENE
Early Wasatchian
‘Condylarthra’ Hyopsodontidae Haplomylus Bunodont 1
Perissodactyla Equidae Hyracotherium Semi-lophed 2
Xenicohippus Semi-lophed 2
‘Tapiroidea’ Cardiolophus Semi-lophed 2
Heptodon Lophed 2
Homogalax Semi-lophed 2
Artiodactyla ‘Dichobunidae’ Bunophorus Bunodont 2
Diacodexis Bunodont 1
‘Homacodon’ n.g. Bunodont 1
Simpsonodus Bunodont 1
Middle Wasatchian
E O C E N E (cont.)
Middle Wasatchian (cont.)
Xenicohippus Semi-lophed 2
Heptodon Lophed 2
Hyrachyus Lophed 3
Paleomoropus Semi-lophed 2
Artiodactyla ‘Dichobunidae’ Bunophorus Bunodont 2
Hexacodus Semi-lophed 1
Simpsonodus Bunodont 1
Late Wasatchian
Stylinodon Semi-lophed 3
Megalesthonyx Semi-lophed 3
Trogosus Semi-lophed 4
Dinocerata Bathyopsis Semi-lophed 4
Prodinoceras Semi-lophed 4
Orohippus Semi-lophed 2
E O C E N E (cont.)
Late Wasatchian (cont.)
Heptodon Lophed 2
Hyrachyus Lophed 3
Selenaletes Semi-lophed 2
Brontotheriidae Eotitanops Semi-lophed 3
incertae sedis Lambdotherium Semi-lophed 3
Artiodactyla ‘Dichobunidae’ Antiacodon Semi-lophed 1
Bunophorus Bunodont 2
‘D.’ waltonensis Bunodont 1
Hexacodus Semi-lophed 1
Early Bridgerian
Taeniodonta Stylinodon Semi-lophed 3
‘Condylarthra’ Hyopsodontidae Hyopsodus Bunodont 1
‘Tapiroidea’ Fouchia Lophed 2
Helaletes Lophed 2
Heptodon Lophed 2
Hyrachyus Lophed 3
Isectolophus Lophed 2
Selenaletes Semi-lophed 2
Brontotheriidae Eotitanops Semi-lophed 3
Palaeosyops Semi-lophed 4
Artiodactyla ‘Dichobunidae’ Antiacodon Semi-lophed 1
Bunophorus Bunodont 2
Helohyus Bunodont 2
Homacodon Semi-lophed 1
Microsus Bunodont 1
E O C E N E (cont.)
Middle Bridgerian
Tillodonta Tillodon Semi-lophed 4
Uintatherium Semi-lophed 5
Phenacodontidae Meniscotherium Lophed 2
Perissodactyla Equidae Orohippus Semi-lophed 2
‘Tapiroidea’ Dilophodon Lophed 2
Helaletes Lophed 2
Hyrachyus Lophed 3
Brontotheriidae Palaeosyops Semi-lophed 4
Artiodactyla ‘Dichobunidae’ Achaenodon Bunodont 3
Antiacodon Semi-lophed 1
Helohyus Bunodont 2
Microsus Bunodont 1
Late Bridgerian
Dinocerata Tethyopsis Semi-lophed 5
Phenacodontidae Phenacodus Bunodont 2
Perissodactyla Equidae Epihippus Semi-lophed 2
‘Tapiroidea’ Desmatotherium Lophed 4
Dilophodon Lophed 2
Hyrachyus Lophed 3
Rhinoceratoidea Amynodon Lophed 5
Brontotheriidae Palaeosyops Semi-lophed 4
Mesatirhinus Semi-lophed 5
Telmatherium Semi-lophed 5
Antiacodon Semi-lophed 1
E O C E N E (cont.)
Late Bridgerian (cont.)
Helohyus Bunodont 2
Microsus Bunodont 1
Neodiacodexis Bunodont 1
Sarcolemur Bunodont 1
Agriochoeridae Protoreodon Semi-lophed 2
Early Uintan
Dinocerata Eobasilus Semi-lophed 5
Tethyopsis Semi-lophed 5
‘Tapiroidea’ Colodon Lophed 3
Dilophodon Lophed 2
New genus Lophed 2
Helaletes Lophed 2
Hyrachyus Lophed 3
Epitriplopus Lophed 2
Triplopus Lophed 2
Uintaceras Lophed 4
Brontotheriidae Metarhinus Semi-lophed 5
Sphenocoelus Semi-lophed 5
Sthenodectes Semi-lophed 5
Chalicotherioidea Eomoropus Lophed 2
Grangeria Lophed 3
Bunomeryx Semi-lophed 1
Helohyus Bunodont 2
‘Helohyus’ n.g. Bunodont 2
Hylomeryx Semi-lophed 1
‘Homacodon’ n.g. Semi-lophed 1
Laredochoerus Bunodont 2
Parahyus Bunodont 3
Suoidea inc.sed. Brachyhyops Bunodont 2
Agriochoeridae Agriochoerus Lophed 3
E O C E N E (cont.)
Early Uintan (cont.)
Protoreodon Semi-lophed 2
Protoceratidae Leptoreodon Semi-lophed 2
Leptotragulus Semi-lophed 2
Oromerycidae Oromeryx Semi-lophed 2
Malaquiferus Semi-lophed 2
Merycobunodon Semi-lophed 2
Protylopus Semi-lophed 2
Camelidae Poebrodon Lophed 2
Late Uintan
Dinocerata Uintatherium Semi-lophed 5
Helaletes Lophed 2
Heteraletes Lophed 3
Hyrachyus Lophed 3
Protapirus Lophed 4
Schizotheriodes Lophed 2
New genus Lophed 2
Amynodontopsis Lophed 5
Megalamynodon Lophed 5
Metamynodon Lophed 5
Triplopus Lophed 2
Uintaceras Lophed 4
Brontotheriidae ‘Diplacodon’ Semi-lophed 5
Eotitanotherium Semi-lophed 5
Metatelmatherium Semi-lophed 5
Notiotitanops Semi-lophed 5
Protitanotherium Semi-lophed 5
Chalicotherioidea Grangeria Lophed 3
Artiodactyla ‘Dichobunidae’ Apriculus Semi-lophed 2
Auxontodon Bunodont 2
Bunomeryx Semi-lophed 1
Hylomeryx Semi-lophed 1
Ibarus Bunodont 1
E O C E N E (cont.)
Late Uintan (cont.)
Mesomeryx Semi-lophed 1
Mytonomeryx Semi-lophed 2
Pentacemylus Semi-lophed 2
Tapochoerus Bunodont 2
Texodus Semi-lophed 1
‘P.’ n.g. A Semi-lophed 1
‘P.’ n.g. B Semi-lophed 1
‘P.’ n.g. C Semi-lophed 1
Protoceratidae Heteromeryx Semi-lophed 2
Leptoreodon Semi-lophed 2
Poabromylus Semi-lophed 2
Toromeryx Semi-lophed 2
Oromerycidae Oromeryx Semi-lophed 2
Malaquiferus Semi-lophed 2
Hypertragulidae Simimeryx Semi-lophed 1
Leptomerycidae Hendryomeryx Semi-lophed 1
Leptomeryx Semi-lophed 2
Duchesnean
Haplohippus Semi-lophed 2
Mesohippus Lophed 3
Heteraletes Lophed 3
Protapirus Lophed 4
Toxotherium Lophed 3
Rhinoceratoidea Amynodontopsis Lophed 5
Hyracodon Lophed 3
Megalamynodon Lophed 5
Penetrigonias Lophed 4
‘Procardurcodon’ Lophed 5
E O C E N E (cont.)
Duchesnean (cont.)
Teleteceras Lophed 4
Triplopus Lophed 2
Uintaceras Lophed 4
Brontotheriidae Duchesneodus Semi-lophed 5
Protitanops Semi-lophed 5
Artiodactyla ‘Dichobunidae’ Discritocheorus Bunodont 2
Leptochoerid n.g. Bunodont 2
Pentacemylus Semi-lophed 2
Anthracotheriidae Heptacodon Semi-lophed 3
‘A.’ n.g. E Lophed 3
‘A.’ n.g. F Lophed 3
‘P.’ n.g. A Semi-lophed 1
‘P’ n.g. D Semi-lophed 2
Merycoidodontidae Aclistomycter Lophed 2
Bathygenus Lophed 2
Leptauchenia Lophed 3
Oreonetes Lophed 2
Prodesmatochoerus Lophed 3
Leptoreodon Semi-lophed 2
Pseudoprotoceras Lophed 2
Oromerycidae Eotylopus Semi-lophed 2
Oromeryx Semi-lophed 2
Hypertragulidae Hypertragulus Semi-lophed 2
Hypisodus Lophed 1
Parvitragulus Semi-lophed 1
Simimeryx Semi-lophed 1
Ruminantia inc.sed. Hidrosotherium Semi-lophed 2
E O C E N E (cont.)
Early Chadronian
Perissodactyla Equidae Mesohippus Lophed 3
Miohippus Lophed 3
Protapirus Lophed 4
Rhinoceratoidea Amphicaenopus Lophed 5
Hyracodon Lophed 3
Subhyracodon Lophed 4
Triplopides Lophed 4
Trigonias Lophed 5
Brontotheriidae Brontops Semi-lophed 5
Megacerops Semi-lophed 5
Menops Semi-lophed 5
Artiodactyla ‘Dichobunidae’ Stibarus Bunodont 1
Entelodontidae Archaeotherium Bunodont 4
Anthracotheriidae Aepinacodon Semi-lophed 4
Bothriodon Semi-lophed 4
Heptacodon Semi-lophed 3
‘A.’ n.g. E Lophed 3
Merycoidodontidae Bathygenus Lophed 2
Oreonetes Lophed 2
‘Leptotragulus’ Semi-lophed 2
Montanatylopus Semi-lophed 3
Camelidae Poebrotherium Lophed 2
‘P.’ franki Lophed 2
E O C E N E (cont.)
Early Chadronian (cont.)
Hypisodus Lophed 1
Middle Chadronian
Miohippus Lophed 3
Protapirus Lophed 4
Hyracodon Lophed 3
Trigonias Lophed 5
Merycoidodontidae Bathygenus Lophed 2
Oreonetes Lophed 2
‘P.’ franki Lophed 2
Hypisodus Lophed 1
E O C E N E (cont.)
Middle Chadronian (cont.)
Late Chadronian
Miohippus Lophed 3
Protapirus Lophed 4
Hyracodon Lophed 3
Trigonias Lophed 5
Leptochoerus Bunodont 2
Merycoidodontidae Leptauchenia Lophed 3
Oreonetes Lophed 2
Paratylopus Lophed 2
Hypisodus Lophed 1
OLIGOCENE
Orellan
Miohippus Lophed 3
Protapirus Lophed 4
Hyracodon Lophed 3
Trigonias Lophed 5
Anthracotheriidae Bothriodon Semi-lophed 4
Elomeryx Semi-lophed 4
Tayassuidae Perchoerus Bunodont 2
Thinohyus Bunodont 2
Merycoidodontidae Eucrotaphus Lophed 3
Oreonetes Lophed 2
Protoceratidae Pseudoprotoceras Lophed 2
Hypisodus Lophed 1
Whitneyan
Miohippus Lophed 3
Protapirus Lophed 4
Hyracodon Lophed 3
O L I G O C E N E (cont.)
Whitneyan (cont.)
Diceratherium Lophed 5
Choerodon Bunodont 5
Megachoerus Bunodont 5
Anthracotheriidae Elomeryx Semi-lophed 4
Tayassuidae Perchoerus Bunodont 2
Thinohyus Bunodont 2
Merycoidodontidae Blickohyus Lophed 3
Eporeodon Lophed 3
Eucrotaphus Lophed 3
Oreonetes Lophed 2
Protoceratidae Protoceras Lophed 3
Paralabis Lophed 3
Pseudolabis Lophed 3
Hypisodus Lophed 1
Nanotragulus Lophed 2
Early Early Arikareean

Perissodactyla Equidae Kalobatippus Lophed 3
Mesohippus Lophed 3
Miohippus Lophed 3
Tapiridae Miotapirus Lophed 3
Protapirus Lophed 4
Rhinoceratoidea Diceratherium Lophed 5
Hyracodon Lophed 3
Artiodactyla ‘Dichobunidae’ Leptochoerus Bunodont 2
Early Early Arikareean (cont.)
Choerodon Bunodont 5
Dinohyus Bunodont 5
Anthracotheriidae Elomeryx Semi-lophed 4
Kukusepasutanka Semi-lophed 4
Tayassuidae Thinohyus Bunodont 2
Merycoidodontidae Blickohyus Lophed 3
Eporeodon Lophed 3
Merycochoerus Lophed 4
Merycoides Lophed 4
Oreodontoides Lophed 3
Sespia Lophed 2
Camelidae Gentilicamelus Lophed 2
Miotylopus Lophed 3
Poebrotherium Lophed 2
Stenomylus Lophed 3
Hypisodus Lophed 1
Leptomerycidae Leptomeryx Semi-lophed 2
Pronodens Lophed 2
Late Early Arikareean

Perissodactyla Equidae Kalobatippus Lophed 3
Miohippus Lophed 3
Tapiridae Miotapirus Lophed 3
Protapirus Lophed 4
Rhinoceratoidea Diceratherium Lophed 5
Artiodactyla ‘Dichobunidae’ Leptochoerus Bunodont 2
Dinohyus Bunodont 5
Anthracotheriidae Arretotherium Semi-lophed 4
Tayassuidae Thinohyus Bunodont 2
Merycoidodontidae Eporeodon Lophed 3
Late Early Arikareean (cont.)
Merycochoerus Lophed 4
Merycoides Lophed 4
Oreodontoides Lophed 3
Sespia Lophed 2
Camelidae Miotylopus Lophed 3
Stenomylus Lophed 3
Leptomerycidae Leptomeryx Semi-lophed 2
Pronodens Lophed 2
b r u c e j. m a c f a d d e n
Origin and evolution of the grazing guild in

Cenozoic New World terrestrial mammals
Introduction
Today grasslands cover about 25% of the world’s land surface and
constitute an enormous food resource that is exploited by invertebrate
and vertebrate grazers. Grazers are defined as herbivores with diets con-
sisting predominantly (⬎90%) of grass and other associated low ground
cover in grassland biomes (Janis and Ehrhardt 1988). In present-day eco-
systems, grazing is a common-place feeding strategy. Although terrestrial
herbivory can be documented in vertebrates over the past 300 million
years since the Carboniferous (Sues and Reisz 1998; Reisz and Sues, this
volume), the origin of the grazing guild in vertebrates is a relatively
recent arrival on the global ecological landscape.
This chapter presents paleontological evidence that terrestrial grazing
guilds have only existed since the middle Tertiary, about 35 million years
ago. I will confine my discussion here to mammals because: (1) although
in terms of biomass and diversity, invertebrate grazers (mostly insects) are
potentially the largest component of terrestrial grazing guilds worldwide
(Dyer et al. 1982), the fossil record of this group is relatively poor; and (2)
extinct mammalian grazers generally have an exceedingly rich fossil
record that can be used to understand the evolution of the grazing guild.
Furthermore, recent studies of extinct mammalian grazers combine a
diverse array of traditional morphological evidence along with some new
techniques, including stable isotopic analyses. Together these techniques
allow a better understanding of the origin and evolution of the grazing
guild. This chapter will focus attention on the fossil record in the New
World; however, the pattern described here also is generally applicable to
the Old World.
[223]
224 bruce j. macfadden
Figure 8.1. Cross-sections of (a) low-crowned or brachydont (human), and (b) high-
crowned or hypsodont (horse) molar teeth showing development of the
various dental tissues and elongated crown in the horse. (From Janis and
Fortelius 1988, reproduced with permission of Cambridge University Press.)
The fossil record of grazing mammals: morphological and

isotopic evidence
Kovalevsky (1873) published a classic monograph on fossil horses, which

provides the foundation for our understanding of the evolution of
grazing in that group. He proposed that the acquisition of high-crowned
teeth in horses during the Miocene was an evolutionary adaptation to
feeding on grasses. The reason for the correlation of high-crowned teeth
with grazing is because grasses are highly abrasive and tend to accelerate
tooth wear relative to softer browse (Figure 8.1). Grasses contain micro-
scopic, dumbbell-shaped particles called phytoliths (Figure 8.2) com-
posed of silica (SiO™). Although phytoliths may also function in support
(e.g., in bamboo), it is fairly well accepted that the primary role of these
particles is to deter herbivory. As such, the ‘cost’ of becoming a grazer is
the evolutionary necessity to acquire progressively higher-crowned teeth.
Thus, even though the teeth of grazers wear very rapidly, they have more
tooth volume to wear down during their lifetime relative to those of a
short-crowned browser feeding on softer plants (Figure 8.1).
There are two general evolutionary morphotypes related to acquiring
high-crowned teeth for grazing. Hypsodont teeth are high-crowned
New World Cenozoic mammalian grazers 225
Figure 8.2. Dumbbell-shaped, abrasive silica particles (phytoliths) preserved in

Miocene grass from Kenya. (From Dugas and Retallack 1993, reproduced with
permission of The Paleontological Society.)
and have determinate growth. In addition to high crowns, hypsodont

mammals further increase the durability of their teeth by infolding the
enamel, which results in increased surface area for this more resistant
kind of dental tissue (Van Valen 1960; Janis and Fortelius 1988). Many
groups of extant or extinct presumed grazers are hypsodont, including
from the New World the classic example of the horse, and also certain
groups of rhinos, proboscideans, and ruminant artiodactyls. The other
kind of high-crowned tooth is termed hypselodont, or ever-growing.
Hypselodont teeth usually have relatively simple enamel patterns on the
occlusal surface, and sometimes (e.g., in notoungulates), enamel is
reduced in thickness or partially lost. Hypselodont teeth are mostly
found in small North American mammals (rodents and lagomorphs) and
are widespread among the extinct South American notoungulates.
Evolution is rarely simple, and there are other contributing factors or

alternate reasons that have been proposed to explain the evolution of
high-crowned teeth in Tertiary mammals. Workers such as White (1959)
and Janis (1988) have suggested that the presence of contaminant grit
ingested from the ground when cropping close to the substrate, or as fine
dust on the grass fodder itself, could have resulted in, or contributed to,
the evolution of high-crowned teeth independent from the spread of
grasslands. Although it is possible that, as in the past (as also happens
today), contaminant grit did get ingested during feeding on grasses and
other low ground cover, this abrasive material by itself is not sufficient to
explain the widespread acquisition of hypsodont teeth during the
Miocene throughout much of the world. Consequently, most workers
still favor as the principal explanation the acquisition of high-crowned
teeth as a coevolutionary response to feeding primarily on grass.
Evidence from modern herbivorous mammals with known diets
further confirms this classic hypothesis of a correlation between diet and
crown height (Janis 1988). As a general rule, extant herbivores with low-
crowned teeth are predominately browsers and species with high-
crowned teeth are predominantly grazers. Although there are some
notable exceptions to this oversimplified rule (e.g., the hippopotamus
Hippopotamus amphibius and the llama Lama glama are both relatively
short-crowned grazers), the model otherwise is generally accepted by
most paleontologists.
Within the past decade, a new technique, stable isotopic analysis of
tooth enamel, has been applied to the study of diets and ecology of extinct
herbivorous mammals. CO£2⫺ is incorporated into the hydroxyapatite
crystal lattice of skeletal tissue (Newesley 1989; McClellan and Kauwen-
bergh 1990), primarily replacing the PO¢2⫺, although lesser amounts also
can weakly bond at other sites (OH⫺). Carbon isotope analysis was ini-
tially applied to study diets of ancient human cultures using different
kinds of skeletal tissues, including organic collagen and apatite, the latter
of which is the mineral phase of bone and teeth (e.g., van der Merwe 1982;
DeNiro 1987). Subsequent paleontological studies of stable isotopes ana-
lyzed mineralized bone, because collagen is rarely preserved in the fossil
record. It was quickly found that the porous nature of bone makes it
prone to post-mortem chemical alteration by ground water in the sedi-
ments. Thus, in the late 1980s, several workers began to concentrate their
efforts on analyzing the stable isotopes of tooth enamel, which consists of
more than 95% mineral hydroxyapatite (Hillson 1986) and is compact in
its crystal structure and relatively resistant to diagenesis (e.g., Lee-Thorp

and van der Merwe 1987; Quade et al. 1992). The analysis of the carbon iso-
topes in tooth enamel has opened the door to a major new line of paleon-
tological investigation.
The stable isotopes of carbon (12C and 13C) are combined (fractionated)
in different proportions depending upon whether a plant photosynthe-
sizes this element using the Calvin, Hatch–Slack, or CAM pathways.
Most (c. 85%) terrestrial plant biomass, including trees, shrubs, forbs, and
cool-growing season grasses, photosynthesize carbon using the Calvin
pathway. These are also called C£ plants because of the formation of initial
3-carbon compounds during photosynthesis. Most tropical and temper-
ate grasses photosynthesize carbon using the Hatch–Slack pathway; these
are called C¢ plants because of the formation of 4-carbon compounds as
the initial photosynthetic product. CAM, or Crassulacean Acid Metabo-
lism, plants consist mostly of succulents such as the Cactacaea and have
␦13C values with a broad range that overlaps pure C£ (infrequently) on the
one hand, to pure C¢ (more frequently) on the other, depending upon the
species. These will not be considered further in this chapter because they
are not central to the origin of grasslands and the grazing guild.
C£ and C¢ plants can be discriminated by their ratios of stable carbon
isotopes, as expressed by the following standard notation:
␦13CPDB (in parts per thousand, ‰) ⫽ [(Rsample/Rstandard)⫺1] ⫻ 1000
where R ⫽ 13C/12C. All measurements of a fossil sample (in this case, tooth
enamel) are compared with the international standard, the Cretaceous
Pee Dee belemnite (PDB).
The two principal photosynthetic pathways of interest here fraction-
ate carbon isotopes in fundamentally different proportions. C£ plants
have mean ␦13C values of ⫺27‰ with a broad range between about ⫺31 to
⫺23‰ (Figure 8.3). In contrast, C¢ plants are isotopically more positive
with mean ␦13C values of ⫺13‰ with a more narrow range between about
⫺15 and ⫺11‰ (Deines 1980; Farquhar, Ehleringer and Hubick 1989;
Boutton 1991). Relative to the ␦13C values of plant foodstuffs, tooth
enamel carbonate of mammalian herbivores is enriched (i.e., less nega-
tive) by about 13‰ (Koch, Zachos and Gingerich 1992; Quade et al. 1992;
Cerling et al. 1997).
Using the two methods (crown height and carbon isotopes) allows a
better discrimination of the feeding ecologies of extinct herbivores. As
shown in the lower left-hand quadrant of Figure 8.4, short-crowned
Figure 8.3. Plot of ␦13C values versus fraction C¢ contained in varied kinds of plants
and the corresponding enriched (more positive) values of tooth enamel. This
model assumes: (1) a mean ␦13C value for C£ plants of ⫺27‰ with a range of
⫾ 4‰; (2) a mean ␦13C value for C¢ plants of ⫺13‰ with a range of ⫾ 2‰; and
(3) relative to plant foodstuffs, fossil tooth enamel ␦13C (hydroxyapatite phase)
is enriched by 13‰ (e.g., Farquhar et al. 1989; Boutton 1991; Cerling et al. 1993;
Quade et al. 1992). (Slightly modified from MacFadden and Cerling 1996,
reproduced with permission of the Society of Vertebrate Paleontology.)
herbivores (hypsodonty indices [HI] ⬍1) with relatively negative (c.

⬍⫺10‰) ␦13C values for their tooth enamel are principally C£ browsers.
Proceeding clockwise, high-crowned herbivores [HI ⬎1] with relatively
negative ␦13C values are C£ grazers. Although this feeding ecology is rel-
atively rare in present-day ecosystems, and confined to regions with cool
growing seasons (high elevations, high latitude, or Mediterranean-type
climate), evidence will be presented below to show that this adaptation
was very common during the middle Tertiary before 7 million years ago.
In the upper right-hand quadrant of Figure 8.4, high-crowned (HI
ranging from 1 to 3) herbivores with relatively positive ␦13C values
(⬎⫺2‰) are C¢ grazers. This feeding ecology is very common in modern
ecosystems. In the lower right-hand corner, short-crowned herbivores
with relatively positive ␦13C values are very rare in modern ecosystems
because C¢ browse is exceedingly rare. (In this case, it also is possible
Figure 8.4. Plot of carbon isotopic values (␦13C) versus tooth crown height to show
how these two data sets can be used to discriminate C£ grazer, C¢ grazer, mixed
and/or CAM feeder and C£ browser. (Although a C¢ browser is theoretically
possible in this matrix, C¢ browse is very rare in nature. Given the
intermediate and wide-ranging ␦13C values for CAM plants, this matrix cannot
discriminate between a specialized succulent browser versus a generalized,
mixed feeder.) Hypsodonty index is the ratio of unworn/little-worn molar
crown height to the anteroposterior molar length. Low-crowned (brachydont)
herbivores have a tooth in which the HI is ⬍1; high-crowned (hypsodont or
hypselodont) herbivores have a HI that is ⬎1. Maximum observed HIs are
usually ⬍3. (From MacFadden and Shockey 1997, reproduced with permission
of The Paleontological Society.)
that the particular species in question was feeding on CAM plants.) This
category also could include a C¢ grazer which is the exception to the
general rule; thus the hippopotamus (Hippopotamus amphibius) is short-
crowned, but primarily a grazer (Owen-Smith 1988) with relatively posi-
tive ␦13C values (⫺4.2‰, N⫽7, from Amboseli Park, Kenya; Bocherens et
al. 1996). Species with intermediate values for both crown height and
carbon isotopes (Figure 8.4, center) represent multiple possibilities, e.g.,

mixed (browse/grass) feeders or specialized CAM feeders.
Miocene savannas of North America and ‘The Great

Transformation’
Studies of fossil horses and other extinct herbivorous mammals of North

America have shown that during the early Miocene there was a significant
increase in hypsodonty and diversification of presumed grazers starting
about 20 million years ago. Diversification of these hypsodont mammals
is exemplified by the horses (Family Equidae), which reached a maximum
diversity of 13 genera during the middle Miocene about 15 million years
ago (Figure 8.5). This was not a localized phenomenon; it occurred, so far
as can be interpreted, throughout Holarctica. At many fossil sites in
North America during this time, 10 to 12 species of horses, many of which
were hypsodont, co-occurred and are therefore believed to have been
sympatric (Figure 8.6). It is not just the horses that are hypsodont or
diverse during this time. Other groups of middle Tertiary hypsodont
North American mammals include the Leporidae (rabbits), five families
of rodents (Castoridae, Mylagaulidae, Geomyidae, Heteromyidae and
Cricetidae), gomphothere proboscideans, Rhinocerotidae, Camelidae,
and Antilocapridae (Figure 8.7). These are joined during the Plio-Pleis-
tocene by hypsodont Cavioidea (three families) and Erethizontidae from
South America. During the Late Pleistocene, the Elephantidae and
Bovidae dispersed into North America from the Old World. Thus the
dental structure and presumed feeding ecology of Late Tertiary herbivor-
ous mammals from North America underwent a dramatic change that
Simpson (1951), in referring to horses, called ‘The Great Transformation.’
The explanation for this major shift in feeding adaptation has been
fairly well accepted by paleontologists since the late nineteenth century.
Although the oldest fossil grasses in North America are known from the
Eocene of Tennessee (Crepet and Feldman 1991), thereafter, during the
middle Tertiary, grasses or their associated remains (e.g., phytoliths) are
rarely encountered as fossils until the Miocene. Accordingly, it has been
speculated that grasslands began to spread as dominant biomes during
the early Miocene and opened up a significant new food resource to be
exploited by herbivores. As mentioned above, because of the highly abra-
sive nature of grasses, the evolutionary cost to potential grazers was the
acquisition of high-crowned teeth. This opportunity was exploited by
Figure 8.5. (A) Chart showing diversity of presumed browsing (shaded) and grazing
(unshaded) horses. (B) Overall diversity of fossil horses during the past 18
million years. (From MacFadden 1992, reproduced with permission of
Cambridge University Press.)
Figure 8.6. Acquisition of high-crowned (hypsodont) teeth in four coexisting genera of three-toed hipparion horses from the Miocene and Pliocene of
North America. To the left of each crown is a view of the occlusal surface. Maximum unworn or little worn crown heights are indicated to right
of crown. During the middle Miocene (Barstovian) hipparions were hypsodont (hypsodonty indices c. 1) and by the Hemphillian they were very
hypsodont (hypsodonty indices ranging from 2 to 3). (From MacFadden 1985, reproduced with permission of The Paleontological Society.)
Figure 8.7. Time chart showing the distribution and taxonomic composition of the
predominant hypsodont mammals in North and South America.
Interpretation of grazing feeding ecology is inferred from either the presence
of high-crowned or ever-growing teeth in the clade, or evidence from carbon
isotopes, as discussed in the text. The grazing adaptations in edentates and
litopterns are questionable. Only the portion of a particular clade’s range with
high-crowned taxa is included here. For example, fossil horses are known in
North America beginning in the Eocene, but only become high-crowned in the
middle Miocene, the latter of which is shown here. (Compiled with slight
modification from several sources: Patterson and Pascual 1972; Webb 1977;
Webb 1978; from MacFadden [1997], reproduced with permission of Elsevier.)
Figure 8.8. (A) Plot showing ␦13C values taken from fossil horse specimens from
North America. (B) Plot of advent of hypsodonty. (Modified from MacFadden
1992 and reproduced with permission of Cambridge University Press.) Note
there is no direct correspondence between shift in ␦13C values and hypsodonty,
suggesting that during the late Miocene horses originally exploited C£-based
grass biomes. The advent of C¢ grasses occurs during the later Miocene
(between about 8–6 million years ago) and corresponds with global
aridification and cooling and more pronounced seasonality. (Slightly modified
from Wang et al. 1994, reproduced with permission of Elsevier.)
horses and other North American hypsodont grazers and was a time of
great morphological change as these taxa exploited a new ‘adaptive zone’
(sensu Simpson 1953). In modern ecological terms, this was the origin of
the terrestrial grazing ‘guild’ (Root 1967) in North America.
The notion of the Great Transformation and the origin of C¢ grassland
biomes in temperate and tropical regions can be tested using stable iso-
topes. If hypsodont herbivorous mammals changed over to feeding on
grass during the Miocene, then, based on modern analogies, there should
be a corresponding shift from a C£ (browsing) to C¢ (grazing) signal
recorded in the carbon isotopic values of their tooth enamel. This is
exactly the hypothesis that was tested several years ago and the results
were very surprising. Wang, Cerling and MacFadden (1994) analyzed the
␦13C values from 40 fossil equid specimens ranging from the Eocene
(including Hyracotherium, or ‘Eohippus’) to the Late Pleistocene Equus.
Not surprisingly, prior to the Miocene, all fossil horses have tooth enamel
carbonate ␦13C values of about ⫺10‰ or less (Figure 8.8), indicating a C£
browsing signal before the advent of grasslands. Surprisingly, in the early
and middle Miocene during the time of rapid increase in hypsodonty
(Figure 8.8B) and presumed spread of grasslands, ␦13C values of hypso-

dont horses remain relatively negative, contrary to what would have been
predicted. The anticipated isotopic shift to a dominantly C¢ signal in hyp-
sodont taxa does not appear until about 7 million years (Figure 8.8A).
This shift is dramatic, with approximately a 10‰ isotopic enrichment
within a relatively short time (i.e., less than one million years) at about 7
million years ago. The explanation for this lag in the advent of hypso-
donty and spread of grasslands was initially perplexing. However, it fits
into a global model that has developed for world’s ecosystems. Carbon
isotopic values from a variety of sources, including terrestrial mammals
and paleosol carbonates, show that prior to 7 million years ago, the
world’s ecosystems were fundamentally C£-based. (There are some fossil
C¢ grasses prior to 7 million years ago, but they apparently were neither
common nor widespread [Tidwell and Nambudiri 1989].) The advent of
C¢-based ecosystems at about 7 million years ago corresponds to the late
Miocene ‘global carbon shift’ (Figure 8.9), including, in terrestrial ecosys-
tems, the beginning of C¢ grasslands as we know them today (Cerling
1992; Cerling et al. 1993, 1997). The most probable explanation for the
changeover to C¢ grasses after 7 million years ago relates to levels of
atmospheric CO™, more pronounced seasonality, and increased aridity. At
high levels of atmospheric CO™, C£ photosynthesis is favored, whereas, at
low levels of atmospheric CO™ and high seasonality (including periods of
high water stress), C¢ plants are favored. C¢ photosynthesis apparently
arose independently numerous times in various plant groups and, from
an evolutionary point of view, it is relatively plastic (Ehleringer et al. 1991,
1997). Returning to early Miocene hypsodont mammals, the current
explanation for the lack of an isotopic shift seems to indicate that these
herbivores were grazing, but during this time they fed on C£ grasses,
which were widespread in terrestrial ecosystems prior to 7 million years
ago. Evidence in support of this hypothesis comes from other paleonto-
logical data. For example, fossil grass parts (anthoecia) have been found
in the tooth enamel infoldings of the 10 million year old hypsodont
rhinoceros Teleoceras from Nebraska (Voorhies and Thomasson 1979;
Figure 8.10).
After 7 million years ago, hypsodont mammals shifted to feeding on
C¢ grasses that became widespread and formed the origins of the modern
grassland biomes. Other than the fact that it seems to have been pro-
found, the overall effect of this shift on the existing grazing guilds is still
unclear. It is known that this was the time of dramatic drop in diversity in
Figure 8.9. ␦13C values of fossil tooth enamel and paleosols from Pakistan and fossil
tooth enamel from North America, and the timing (but not same magnitude of
␦13C) of the carbon shift preserved in oceanic microfossils during the Miocene.
(Slightly modified from Cerling et al. 1993, reproduced with permission of
Macmillian.)
groups such as horses (Figure 8.5; MacFadden 1992; Cerling et al. 1997) and
other terrestrial mammals in North America (Webb 1977, 1983).
South American grasslands and ‘precocious hypsodonty’
South America drifted during most of the Tertiary as an island continent,

and on it evolved a unique fauna of herbivorous mammals. This rich fossil
record reveals that there was an abundant and diverse group of high-
crowned herbivores, the Order Notoungulata, that evolved elongated, or
ever-growing (hypselodont), teeth (Figure 8.11). Based on analogy to taxa
from the Northern Hemisphere, this dental adaptation was presumably
for grazing. Interestingly, the advent of high-crowned presumed grazers
in the South American fossil record predates the origin of ecological
Figure 8.10. Superbly preserved skeletons of the 10-million-year-old hypsodont

rhinoceros Teleoceras from the Ashfall Fossil Beds of Nebraska. (Reproduced
with permission of the Nebraska State Museum, University of Nebraska.)
Figure 8.11. Eurygenium, an example of a South American notoungulate (Family

Notohippidae) with very high-crowned (ever-growing, hypselodont) teeth,
from Salla, a late Oligocene/early Miocene locality in the Andean highlands of
Bolivia. This curious beast appears to have a disproportionately large cranium
relative to its body size, and this is in no small way related to the enormous
teeth, which have crowns that extend dorsally into the palatal region. (From
Shockey 1997, reproduced with permission of the Society of Vertebrate
Paleontology.)
Figure 8.12. Comparisons of the timing of acquisition of high-crowned (either

hypsodont or hypselodont) teeth in the South American Notohippidae
(Notoungulata) from Argentina and Bolivia and the North American Equidae.
(From Shockey 1997, reproduced with permission of the Society of Vertebrate
Paleontology.)
analogs in the Northern Hemisphere by about 10 to 15 million years.

Although some South American high-crowned taxa appear earlier in the
fossil record, they become very diverse and widespread about 35 to 30
million years ago, represented by the notoungulate families Notohippi-
dae, Toxodontidae, Interatheriidae, Mesotheriidae, Archaeohyracidae
and Hegetotheriidae, and possibly by some members of the litoptern
family Macraucheniidae (Figure 8.7). This asynchrony (relative to North
America) in the origin and diversification of high-crowned mammalian
herbivores has been called ‘precocious hypsodonty’ (Patterson and
Pascual 1972; Webb 1978; Simpson 1980). This is illustrated in Figure 8.12,
which compares the earlier acquisition of high crowned teeth in the
South American notoungulates with that of North American horses
(Shockey 1997). The prevailing explanation for precocious hypsodonty is
that grasslands spread earlier (about 30 to 35 million years ago) in South
America than they did in North America (e.g., Stebbins 1981). The poor
fossil record of grasses from South America currently does not offer any
hard evidence to support this idea.
Like the North American counterparts from the Great Transforma-
tion, the advent of grasslands and high-crowned teeth in South American
herbivorous mammals could also be tested with carbon isotopes. MacFad-

den et al. (1994, 1996) showed that, although many notoungulate families
were hypsodont or hypselodont during the middle Tertiary, carbon iso-
topic results indicate that C£ foodstuffs, presumably grasslands, predom-
inated until 7 million years ago. Thereafter, as in North America and
elsewhere in the world (Cerling 1992; Cerling et al. 1993, 1997), C¢ grasses
predominated in low-elevation equatorial and mid-latitudes that sustain
grassland biomes. Using paleosol carbonates from a Miocene–Pliocene
sedimentary sequence in northern Argentina, LaTorre, Quade and McIn-
tosh (1997) showed the details of the carbon isotopic shift recorded glo-
bally elsewhere, thus providing additional evidence of the origin of
C¢-based terrestrial ecosystems.
Antiquity of C£ and C¢ grasses and grazing guild patterns
Studies of hypsodonty and carbon isotopes together reveal a pattern in

which the earliest grasslands of the middle Tertiary were fundamentally
C£-based. It was under this regime that grazing mammal guilds origi-
nated and diversified. When C¢ grasses and grassland biomes became
common after the late Miocene global carbon shift 7 million years ago,
grazers correspondingly shifted to this new type of grass. As exemplified
by horses (Figure 8.5) and also other herbivores, a significant drop in
grazing diversity during the past 7 million years seems related to the
lower productivity of more arid grasslands (Lauenroth 1979). The mix of
modern C£ and C¢ terrestrial grassland communities as we know them
today have existed over the past 7 million years, which, in a geological
context, is a very short period of time.
The study of ancient grazing mammal guilds reveals several patterns
in common, regardless of when and where these guilds existed:
1. Grazing mammals span a wide range of body sizes, from tiny rodents
to large proboscideans. At the larger end of this continuum, it has
been suggested that, relative to medium-sized mammalian grazers,
larger-bodied grazers are better adapted for digesting forage of poor
nutritive value (Schmidt-Nielsen 1975). This relates to factors such as
mass-specific metabolic rate and the longer residence time in the gut
for enzymes to digest nutrients from the grass cells (Owen-Smith
1988).
2. Within a given guild there usually is a continuum of coexisting species
including: (1) mixed feeders with a significant proportion of grass in
their diets; (2) pure grazers, which feed predominantly on grass; and
(3) ‘hypergrazers,’ which are highly specialized grazers. In Late

Tertiary ecosystems, most taxa within this guild and across this
continuum are hypsodont to very hypsodont and ␦13C values of tooth
enamel range from about ⫺6‰ (mixed feeder/grazer) to 0 to ⫹2‰
(hypergrazer). In Plio-Pleistocene terrestrial ecosystems in Florida
between about 3.5 and 0.5 million years ago, carbon isotopic evidence
indicates that extinct species of Equus occupied the hypergrazer niche.
Preliminary evidence suggests that after 0.5 million years ago the Old
World immigrant Bison either displaced Equus as the hypergrazer, or
co-occupied this specialized feeding adaptation (MacFadden and
Cerling 1996). In South America after the late Miocene global carbon
shift at 7 million years ago, but before the Great American Interchange
2 to 3 million years ago, notoungulates such as toxodonts seemed to
have occupied the hypergrazing niche. They were later displaced by
northern immigrants, including Lama and Equus (MacFadden et al.
1996; MacFadden and Shockey 1997).
3. Species within the same ecosystem originating from the same
phylogenetic stock will tend to separate or differentiate their diets as
one way to minimize competition for limited resources. Although we
do not have definite fossil evidence for this process evolving through
time, there are some examples that may serve to illustrate the net
result of this competition. The exceedingly rich middle Pleistocene (1
to 0.5 million years old) mammal fauna from Bolivia contains two
families of herbivores, the Camelidae (Laminae) and Equidae, in which
three species each coexisted. Based solely on dentitions, the lamas
(Palaeolama, Lama, and cf. Vicugna) have relatively short crowns (all HI ⫽
0.7) and the horses range from hypsodont (Hippidion [HI ⫽ 1.9] and
Onohippidium [HI ⫽ 2.2]) to very hypsodont (Equus [HI ⫽ 3.1]). Despite a
lack of niche differentiation within each of these families based solely
upon the crown heights, tooth carbon isotopic results show that these
coexisting species did indeed have different diets. For the lamas, a
small data set suggests that Palaeolama was a C£ browser (␦13C ⫽
⫺11.4‰), whereas cf. Vicugna (␦13C ⫽ ⫺4.5‰), and Lama (␦13C ⫽
⫺3.5‰) were primarily mixed feeders/C¢ grazers. Similarly, for the
horses Hippidion was principally a C£ browser (␦13C ⫽ ⫺8.8‰),
Onohippidium was a C£ browser/mixed feeder (␦13C ⫽ ⫺7.3‰), and
Equus was closest to a pure C¢ grazer (␦13C ⫽ ⫺3.4‰; MacFadden and
Shockey 1997). An interesting question arises from these data: Why did
not the hypsodonty alone discriminate the niche separation in feeding
ecology within these two groups? The answer to this might relate to
the concept of phylogenetic constraint, where once structures (such as
high-crowned teeth) evolved in the ancestors of the species in
question, there is no selective advantage to losing this structure if the

descendant returns to a browsing diet.
Conclusions and summary
‘Grazers’ are defined as animals with diets consisting predominantly

(⬎90%) of grass. Mammalian grazers are characterized by the presence of
high-crowned (hypsodont or hypselodont) cheek teeth, which appear to
be a coevolutionary response to the pervasive presence of highly abrasive
silica phytoliths in grasses. Based on the timing of acquisition of high-
crowned teeth, the rich fossil record of extinct mammalian herbivores
indicates that grazing first developed as a major feeding guild during the
Miocene. Although some fossil grasses have been reported earlier in the
Cenozoic, it seems that the rapid acquisition of high-crowned teeth in
many clades of extinct herbivores was a coevolutionary response to the
advent of widespread grassland communities during the Miocene. The
grazing guild originated and evolved independently in terrestrial com-
munities in South America (primarily Order Notoungulata) and North
America (orders Rodentia, Proboscidea, Artiodactyla and Perissodactyla)
prior to the formation of the Panamanian land bridge.
In addition to dental evidence pertaining to grazing, within the past
decade stable carbon isotopes have rapidly gained acceptance as another
tool to understand ancient herbivore diets, paleoecology, and global
change. Using both tooth structure and carbon isotopes, we can discrimi-
nate among C£ browsers and C¢ grazers, both of which are very common
in present-day ecosystems, and C£ grazers, which are rare in modern eco-
systems.
This integrative approach using morphology and isotopes suggests
that grasslands and the grazing guild probably originated slightly earlier
in South America (c. 35 to 30 million years ago) than in North America
(c. 20 to 15 million years ago). The carbon isotopic data indicate that the
earliest spread of grasslands occurred with C£ grasses (unlike the
common C¢ grasses of today). The physiological change to predominantly
C¢-dominated grassland communities occurred after the late Miocene
‘global carbon shift’ about 7 million years ago. The decline in grazing
diversity after the late Miocene appears related to increased aridity and
lower productivity of grasslands relative those prior to the late Miocene.
(Primary productivity is proportional to factors such as rainfall [Lauen-
roth 1979].) Late Pleistocene herbivore extinctions resulted in a major
drop in overall diversity within the grazing guild as we know it in

modern terrestrial ecosystems. The origin of the grazing adaptation in
the middle Tertiary is a very recent event in the history of terrestrial eco-
systems where herbivory is known to have existed for 300 million years.
Acknowledgments
This research has been supported by National Science Foundation grants

EAR-9506550 and IBN-9528020 and by the National Geographic Society. I
thank Christine M. Janis and Paul Koch for their reviews of this manu-
script. This is University of Florida Contribution to Paleobiology number
493.
References
Bocherens, H., Koch, P. L. Mariotti, A., Geraads, D., and Jaeger, J.-J. (1996). Isotopic
biogeochemistry (13C, 18O) of mammalian enamel from Pleistocene hominid
sites. Palaios 11:306–318.
Boutton, T. W. (1991). Stable carbon isotope ratios of natural minerals: II. Atmospheric,
terrestrial, marine, and freshwater environments. In Carbon Isotope Techniques, ed.
D. C. Coleman and B. Fry, pp. 173–195. San Diego: Academic Press.
Cerling, T. E. (1992). Development of grasslands and savannas in East Africa during the
Neogene. Palaeogeogr. Palaeoclimatol. Palaeoecol. (Global Planet. Change Sect.)
97:241–247.
Cerling, T. E., Harris, J. M., MacFadden, B. J., Leakey, M. G., Quade, J., Eisenmann, V.,
and Ehleringer, J. R. (1997). Global vegetation change through the
Miocene/Pliocene boundary. Nature 389:153–158.
Cerling, T. E., Wang, Y., and Quade, J. (1993). Global ecological change in the late
Miocene: expansion of C¢ ecosystems. Nature 361:344–345.
Crepet, W. L., and Feldman, G. D. (1991). The earliest remains of grasses in the fossil
record. Am. J. Bot. 78:1010–1014.
DeNiro, M. J. (1987). Stable isotopy and archaeology. Am. Sci. 75:182–191.
Dienes, P. (1980). The isotopic composition of reduced organic carbon. In Handbook of
Environmental Isotope Chemistry. 1. The Terrestrial Environment, ed. P. Fritz and J. C.
Fontes, pp. 329–406. Amsterdam: Elsevier.
Dugas, D. P., and Retallack, G. J. (1993). Middle Miocene fossil grasses from Fort Ternan,
Kenya. J. Paleont. 67:113–128.
Dyer, M.I., Detling, J. K., Coleman, D. C., and Hilbert, D. W. (1982). The role of herbivores
in grasslands. In Grasses and Grasslands: Systematics and Ecology, ed. J. R. Estes, R.J.
Tyrl and J. N. Brunken, pp. 255–295. Norman, OK: University of Oklahoma Press.
Ehleringer, J. R., Cerling, T. E. and Helliker, B. R. (1997). C¢ photosynthesis,
atmospheric CO™, and climate. Oecologia 112:285–299.
Ehleringer, J. R., Sage, R. F., Flanagan, L. B., and Pearcy, R. W. (1991). Climate change and
the evolution of C¢ photosynthesis. Trends Ecol. Evol. 6:95–99.
Farquahar, G. D., Ehleringer, J. R., and Hubick, K. T. (1989). Carbon isotopic
discrimination and photosynthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol.
40:503–537.
Hillson, S. (1986). Teeth. Cambridge and New York: Cambridge University Press.
Janis, C. M. (1988). An estimation of tooth volume and hypsodonty indices in ungulate
mammals, and the correlation of these factors with dietary preferences. In Teeth
Revisited: Proceedings of the VIIth International Symposium on Dental Morphology, ed. D.
E. Russell, J.-P. Santoro, and D. Sigogneau-Russell, pp. 367–387. Mém. Mus. Natn.
Hist. Nat, Paris, sér. C, 53.
Janis, C. M., and Ehrhardt, D. (1988). Correlation of relative muzzle width and relative
incisor width with dietary preference in ungulates. Zool. J. Linn. Soc. 92:267–284.
Koch, P. L., Zachos, J. C. and Gingerich, P. D. (1992). Correlation between isotope records
in marine and continental carbon resevoirs near the Palaeocene/Eocene
boundary. Nature 358:319–322.
Kovalevsky, W. (1873). Sur l’Anchitherium aurelianense Cuv. et sur l’histoire
paléontologique des chevaux. (Prèmiere Partie.). Mém. Acad. Imp. Sci. St.-
Pétersbourg, Sér. 7, 20(5):1–73.
LaTorre, C., Quade, J., and McIntosh, W. (1997). The expansion of C4 grasses and global
change in the late Miocene: stable isotope evidence from the Americas. Earth
Planet. Sci. Lett. 146:83–96.
Lauenroth, W. K. (1979). Grassland primary production: North American grasslands in
perspective. In Perspectives in Grassland Ecology, ed. N. R. French, pp. 3–21. Berlin
and New York: Springer-Verlag.
Lee-Thorp, J., and van der Merwe, N.J. (1987). Carbon isotope analysis of fossil bone
apatite. S. Afr. J. Sci. 83:712–715.
MacFadden, B. J. (1985). Patterns of phylogeny and rates of evolution in fossil horses:
hipparions from the Miocene of North America. Paleobiology 11:245–257.
MacFadden, B. J. (1992). Fossil Horses: Systematics, Paleobiology, and Evolution of the Family
Equidae. Cambridge and New York: Cambridge University Press.
MacFadden, B. J. (1997). Origin and evolution of the grazing guild in New World
terrestrial mammals. Trends Ecol. Evol. 12:182–187.
MacFadden, B. J., and Cerling, T. E. (1996). Mammalian herbivore communities, ancient
feeding ecology, and carbon isotopes: a 10 million-year sequence from the
Neogene of Florida. J. Vert. Paleont. 16:103–115.
MacFadden, B. J., and Shockey, B. J. (1997). Ancient feeding ecology and niche
differentiation of Pleistocene mammalian herbivores from Tarija, Bolivia:
morphological and isotopic evidence. Paleobiology 23:77–100.
MacFadden, B. J., Cerling, T. E. and Prado, J. (1996). Cenozoic terrestrial ecosystem
evolution in Argentina: evidence from carbon isotopes of fossil mammal teeth.
Palaios 11:319–327.
MacFadden, B. J., Wang, Y., Cerling, T. E., and Anaya, F. (1994). South American fossil
mammals and carbon isotopes: a 25 million-year sequence from the Bolivian
Andes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 107:257–268.
McClellan, G. H., and Kauwenbergh, S. J. V. (1990). Mineralogy of sedimentary apatites.
In Phosphorite Research and Development, I. ed. A. J. G. Notholt and I. Jarvis, pp.
23–31. London: Geological Society Special Publications.
Newesley, H. (1989). Fossil bone apatite. Appl. Geochem. 4:233–245.

Owen-Smith, N. (1988). Megaherbivores: The Influence of Very Large Body Size on Ecology.
Patterson, B. and Pascual, R. (1972). The fossil mammal fauna of South America. In
Evolution, Mammals, and Southern Continents, ed. A. Keast, F. C. Erk, and B. Glass,
pp. 247–309. Albany: State University of New York Press.
Quade, J., Cerling, T. E., Barry, J. C., Morgan, M. E., Pilbeam, D. R., Chivas, A. R., Lee-
Thorp, J.A., and van der Merwe, N. J. (1992). A 16 million year record of paleodiet
using carbon and oxygen isotopes in fossil teeth from Pakistan. Chem. Geol.
(Isotope Geosci. Sect.) 94:183–192.
Root, R. B. (1967). The niche separation exploitation pattern of the Blue-gray
Gnatcatcher. Ecol. Monogr. 37:317–350.
Schmidt-Nielsen, K. (1975). Animal Physiology: Adaptation and Environment. Cambridge
Shockey, B. J. (1997). Two new notoungulates (Family Notohippidae) from the Salla Beds
of Bolivia (Deseadean: late Oligocene): systematics and functional morphology. J.
Vert. Paleont. 17:584–599.
Simpson, G. G. (1951). Horses: The Story of the Horse Family in the Modern World and through
Sixty Million Years of History. Oxford: Oxford University Press.
Simpson, G. G. (1953). The Major Features of Evolution. New York: Columbia University
Press.
Simpson, G. G. (1980). Splendid Isolation: The Curious History of South American Mammals.
New Haven: Yale University Press.
Stebbins, G. L. (1981). Coevolution of grasses and herbivores. Ann. Missouri Bot. Gard.
68:75–86.
Sues, H.-D., and Reisz, R. R. (1998). Origins and early evolution of herbivory in
Tidwell, W. D., and Nambudiri, E. V. M. (1989). Tomlinsonia thomassonii, gen. et sp. nov., a
permineralized grass from the Upper Miocene Ricardo Formation, California.
Rev. Palaeobot. Palynol. 60:165–177.
van der Merwe, N. J. (1982). Carbon isotopes, photosynthesis, and archaeology. Am. Sci.
70:596–706.
Van Valen, L. (1960). A functional index of hypsodonty. Evolution 14:531–532.
Voorhies, M. R. and Thomasson, J. R. (1979). Fossil grass antothecia within Miocene
rhinoceros skeletons: diet of an extinct species. Science 206:331–333.
Wang, Y., Cerling, T. E., and MacFadden, B. J. (1994). Fossil horses and carbon isotopes:
new evidence for Cenozoic dietary, habitat, and ecosystem changes in North
America. Palaeogeogr., Palaeoclimatol., Palaeoecol. 107:269–279.
Webb, S. D. (1977). A history of savanna vertebrates in the New World. Part I. North
America. Annu. Rev. Ecol. Syst. 8:355–380.
Webb, S. D. (1978). A history of savanna vertebrates in the New World. Part II. South
America and the Great Interchange. Annu. Rev. Ecol. Syst. 9:393–426.
Webb, S. D. (1983). The rise and fall of the late Miocene ungulate fauna of North
America. In Coevolution, ed. M. H. Nitecki, pp. 267–306. Chicago: University of
Chicago Press.
White, T. E. (1959). The endocrine glands and evolution, no. 3: os cementum,
hypsodonty, and diet. Contrib. Mus. Paleontol., Univ. Michigan 13:211–265.
Taxonomic index
Acanthurus 49 Biarmosuchia 36
Aclelstorhinus 36 Bison 240
Aetosauria 124 Blikanasaurus 55, 60
Agriochoeridae 212, 214–21 Bolosauridae 25, 26
Ailuropoda melanoleuca 3 Bolosaurus 25, 26
Alamosaurus 86 Bovidae 178, 230, 233
Allosauridae 69 Brachiosauridae 81, 83, 85, 92, 98, 99, 117,
Amargasaurus 84, 86, 104 128, 134
Amblyrhynchus 51 Brachiosaurus 81, 83, 86, 87, 88, 98, 99, 110, 111,
Ammosaurus 44, 55 113, 114, 117
Amniota 5, 13, 15, 36 Brachylophus 51
Anchisaurus 44, 55 Brontotheriidae 186, 189, 210–213, 215–18
Ankylosauria 65, 69, 125, 126, 129
Anomodontia 20–23, 36 Cactacaea 227
Antarctosaurus 99 Calippus 231
Anthracotheriidae 216–21 Camarasauridae 81, 83, 89, 92, 95–7, 110, 115,
Antilocapridae 230, 233 117, 128
Apatosaurus 82, 84, 86, 89, 106, 108–10 Camarasauromorpha 115
Araeoscelidia 36 Camarasaurus 81, 83, 86, 87–9, 92, 95–7, 99,
Archaeohippus 231 110, 112, 114, 115
Archaeohyracidae 233, 238 Camelidae 186, 188, 213–22, 230, 233, 240
Archaeopteryx 69 Captorhinidae 26–8, 34, 36, 37
Archosauria 36 Captorhinus 26–8
Archosauromorpha 31, 33, 68 Caseasauria 18, 19
Arctocyonidae 145, 149, 157, 173 Caseidae 11, 18, 19, 36, 37
Arctostylopoidea 170, 207 Castoridae 230, 233
Arctostylops 172 Cavioidea 230, 233
Argentinosaurus 113 Centrosaurus apertus 128
Artiodactyla 7, 108–22, 141, 163, 172, 175, 181, Ceratomorpha 181
186, 187, 208–221, 225, 233 Ceratopsia 6, 69, 125, 127, 128
Astrohippus 231 Ceratosauria 69
Cetiosauridae 81, 83, 94, 95, 96, 110, 113, 117
Barapasaurus 81, 83, 86, 90, 112 Cetiosaurus 79, 83, 86, 95, 117
Barasaurus 31 Chalicotheroidea 189, 212, 213
Barosaurus 84, 86, 89, 106, 109, 110 Chimaerasuchus 66, 67
Barylambdidae 182, 207 Chinchilloidea 233
Basilicus 51 Cistecephalus 22
Bauriidae 28 Clevosaurus 53
Belebey 25, 26 Cnemidophorus 12
[245]
246 Taxonomic index
Coloradisaurus 55, 56 Eothyrididae 19, 36

Condylarthra 6, 155, 171–3, 178, 181, 187, Eothyris 19
202–13, 214, 216 Equidae 208–19, 221, 230, 236, 240
Conolophus 51 Equus 234, 240
Cormohipparion 232 Erethizontidae 230, 233
Coryphodon 160–3, 182 Eucynodontia 29, 36
Coryphodontidae 207–10 Euhelopodidae 81, 83, 91–4, 113, 117
Cotylorhynchus 17, 19 Euhelophus 81, 83, 86, 87, 91–3, 94
Cricetidae 230, 233 Euornithopoda 64, 69, 137
Crocodylomorpha 72 Eusauropoda 113, 114
Ctenochaetus 49 Euskelosaurus 55, 60
Ctenosaura 48, 50–2, 57 Exaeretodon 28
Cyclura 50, 51
Cynodontia 28–30, 34, 36 Geomyidae 230, 233
Cynognathidae 36 Glaucosaurus 17, 18
Cynognathus 29 Gomphodontia 28–30, 34, 36
Gomphotheridae 233
Datousaurus 94 Gorgonopsia 36
Dendrohyrax dorsalis 176, 177 Gyposaurus 62
Deltatheridium 149
Desmatodon 14 Haplocanthosaurus 86, 110
Diadectes 14, 15 Hegetotheriidae 233, 238
Diadectidae 5, 13, 14–16, 36 Hemithlaeus 155, 156, 202
Diadectomorpha 13, 14–16 Herrerasaurus 68, 70
Diademodon 29 Heteromyidae 230, 233
Diapsida 26, 31, 33 Heterodontosauridae 69
Dichobunidae 188, 189, 193, 208–21 Heterodontosaurus 64
Dicraeosauridae 82, 84, 102, 103, 104, 105, Hipparion 231, 232
108, 111, 112, 113, 114, 115, 117 Hippidion 231, 240
Dicraeosaurus 82, 84, 86, 87, 89, 92, 104 Hippopotamus 178, 226, 229
Dicynodontia 11, 20–3, 28, 34, 35, 37, 65, Homogalax 159
124 Howesia 33
Diictodon 22 Hyemoschus aquaticus 176
Dinocephalia 20, 36, 37 Hyopsodontidae 202–15
Dinocerata 182, 186, 189, 205–13 Hyperodapedon 32, 33
Dinohippus 231 Hypertragulidae 214–22
Dinosauria 28, 68, 70–2, 124 Hypohippus 231
Diplodocidae 82, 84, 86, 89, 105–10, 111, 112, Hypsognathus 30, 31
113, 114, 117 Hyrachyus 161, 163
Diplodocoidea 102–9, 114, 115, 116, 128 Hyracodon 161, 163
Diplodocus 79, 82, 84, 86, 87, 89, 92, 105–10, Hyracotherium 159, 234
117, 129
Dipsosaurus 49, 51 Ianthasaurus 16, 17, 18
Dromaeosauridae 69 Iguana 43, 46, 48–51
Iguaninae 43, 46–53, 57, 62–6, 71–2
Echinodon 64 Iguanidae 3, 5, 12, 34, 48, 51
Ectocion 159 Iguanodon 43, 127
Ectoconus 150 Iguanodontia 131, 132
Edaphosauridae 5, 16, 17, 18, 36 Interatheriidae 233, 238
Edaphosaurus 5, 16, 17, 18
Edentata 233 Jingshanosaurus 53, 56, 62
Edmontosaurus regalis 127
Elephantidae 191, 230, 233 Kalobatippus 231
Entelodontidae 216–21 Kennalestes 149
Eoconodon 149, 150, 157, 158, 161 Kotasaurus 81, 83, 90, 91
‘Eohippus’ 234 Kunmingosaurus 60, 90
Taxonomic index 247
Labidosaurikos 26–8 Onohippidium 240

Labidosaurus 26, 27, 28 Ophiacodontidae 36
Lagerpeton 68 Opisthocoelicaudia 83, 100
Lama 226, 240 Ornithischia 34, 62, 68, 70, 71, 85, 125, 130,
Lambdotherium 161 135, 137, 139
Laminae 240 Ornithomimosauria 65, 66, 69
Lanthanosuchidae 36 Ornithopoda 6, 64, 110, 127, 131, 134, 135, 137
Lepidosauria 72 Oromerycidae 213–18
Lepidosauromorpha 36 Ouranosaurus nigeriensis 132
Leporidae 230, 233 Oviraptoridae 69
Leptomerycidae 214, 215, 217–22 Oviraptorosauria 65
Leptopleuron 31 Owenetta 24, 30, 31
Leptopleuroninae 31 Owenettidae 31, 36
Lesothosaurus 62, 63, 69, 126, 135 Oxyclaenus 149
Limnoscelidae 36
Limnoscelis 14, 15, 18 Pachycephalosauria 64, 65, 69, 125, 127–9, 135
Loxolophus 149, 157, 159, 160 Palaeolama 240
Lufengosaurus 56, 62 Panoplosaurus mirus 126
Pantodonta 182, 203–9, 210
Macraucheniidae 233, 238 Parahippus 231
Malawisaurus 81, 82, 86, 99 Parareptilia 23, 24, 30, 35
Malawisuchus 67, 68 Pareiasauria 23, 24, 28, 34, 35, 36, 37
Mamenchisaurus 81, 83, 91–4 Patagosaurus 81, 86, 92, 95
Marasuchus 68 Patranomodon 21, 22, 23
Marginocephalia 128 Pelecanimimus 65–6
Massetognathus 29 Periptychidae 149–51, 155, 202–7
Massospondylus 43–6, 53–5, 57, 58, 60, 61 Periptychus 149–51, 155, 161
Megahippus 231 Perissodactyla 7, 172, 175, 188, 208–21, 233,
Melanorosaurus 55, 60 241
Meniscotherium 172 Phenacodontidae 203–11
Merychippus 231 Phenacodus 161, 164
Merycoidodontidae 215–21 Plateosaurus 43, 44, 46, 55–6, 60
Mesosuchus 32, 33 Pliohippus 231
Mesotheriidae 233, 238 Pristerodon 22
Mioclaenidae 202–9 Proboscidea 7, 225, 230, 233, 241
Moradisaurinae 26 Procerberus 149
Moradisaurus 26, 28 Procolophon 30, 31
Moschops 20 Procolophonidae 30, 31, 36
Multituberculata 124, 144, 171, 187 Procolophonoidea 24, 30, 124
Muntiacus 64 Procynosuchidae 36
Mylagaulidae 230, 233 Prosauropoda 5, 42–62, 69, 128
Protoceratidae 186, 213–22
Nannipus 231, 232 Protorosaurus 68
Nemegtosauridae 82, 102, 103, 112, 117 Protorothyrididae 36
Nemegtosaurus 82, 86, 87, 100–5, 114–15, 117 Protungulatum 145–7, 149, 150
Neoceratopsia 135 Pseudhipparion 231
Neohipparion 231, 232 Pseudolagosuchus 68
Neosauropoda 94, 113, 114, 115 Psittacosaurus 127, 129, 130, 135
Notohippidae 233, 237, 238 Pterosauria 68
Notoungulata 225, 233, 236–8, 240, 241 Pyrotherium 161, 163
Nycteroleteridae 36
Quaesitosaurus 86, 103
Octodontoidea 233
Oedaleops 19 Rana hexadactyla 12
Omeisaurus 81, 83, 86, 88, 91–4 Rebbachisaurus 89
Onohippidion 231 Reptilia 23, 25–6, 30–1, 33
248 Taxonomic index
Rhinocerotidae 225, 230, 233 Therapsida 20, 34, 37, 65

Rhinocerotoidea 211–21, 225 Therizinosauroidea 66, 69
Rhynchosauria 31–3, 36, 124 Therocephalia 36
Rhynchosauridae 32, 34 Theropoda 44, 46, 47, 62–8, 70, 85
Riojasaurus 56 Thrinaxodon 29
Rodentia 7, 189, 190, 195, 225 Thrinaxodontidae 36
Ruminantia 215 Thyreophora 65, 125, 126, 135
Tillodonta 170, 176, 182, 206–11
Saltasaurus 100, 115 Titanosauriformes 114–16
Saurischia 43, 68, 70 Titanosauroidea 82, 83, 85, 99–102, 112–17,
Sauromalus 51 128, 134
Sauropoda 5, 69, 79–117, 129 Titanosaurus 100
Sauropodomorpha 5, 44, 68, 70, 125, 128–31, Toxodontidae 233, 238, 240
134, 135, 136, 139 Tragulidae 176, 177, 193
Scelidosaurus 69 Trilophosauridae 33
Scleromochlus 68 Trilophosaurus 33, 36
Scutosaurus 24 Trirachodon 29
Seismosaurus 89, 113 Tritylodontoidea 124
Sellosaurus 45 Troodon 47, 66
Shunosaurus 81, 83, 85, 86, 91–4, 111, 113 Troodontidae 65, 66, 69
Sphenacodontidae 36 Tseajaia 15
Sphenodon 67 Tylopoda 193
Stagonolepididae 28, 124 Tyrannosauridae 69
Staurikosaurus 68, 70
Stegoceras validum 128 Uintatherium 161, 163
Stegosauria 65, 69, 91, 125–7, 129, 133 Ulemosaurus 20
Suminia 20, 21, 23 Ullmannia 68
Suoidea 212, 214, 215, 216 Ungulata 176
Synapsida 5, 11, 16, 18, 20, 28, 34, 35, 66, Uromastix 48, 49
124
Veranopseidae 36
Taeniodonta 175, 177, 182, 202–13, 233 Vicugna 240
Tapinocephalidae 20 Vulcanodon 83, 90, 91
Tapiridae 196, 220, 221 Vulcanodontidae 81, 83, 90, 91, 117
Tapiroidea 208–14, 216–19
Tayassuidae 219–21 ‘Yaleosaurus’ 44
Teleoceras 235, 237 Youngindae 36
Telmatosaurus transylvanicus 132 Yunnanosaurus 46, 55, 57–60, 62
Tenontosaurus tilletti 132
Teratosaurus 44 Zalambdalestes 149
Tetraceratops 36 Zhelestidae 171
Thecodontosaurus 45, 46, 55, 68 Zizhongosaurus 60
Subject index
adapiforms 187 Ashfall Fossil Beds 237

Africa 60, 126, 176, 180, 182–4 Asia 125, 177, 188
African fauna 176, 178, 179, 183, 190 aspect ratios 149
agouti 189 Australia 125
Aldabara 50 autapomorphies 58
algae 49
algal eaters 49 bamboo 224
algorithms 177 barrel shapes 4–5, 11, 17, 22, 32, 35, 65
alluvial conditions 55 Barremian 79
amniotes 1, 5, 9, 10, 12–13, 15, 36, 37 Barstovian 231, 232
amniotic reproduction 13 barylambdid pantodonts 182
amphibians 12 basal amniotes 15
amynodonts 186 basal dinosaurs 66, 68, 70–1, 90–1, 114
analogies 3, 11, 13, 24, 48, 234, 236 basal placentals 145
analogs 172, 236, 238 bears 174, 175
anatomical correlates 11, 44, 71, 169 Beaufort Group 9, 20, 22
differences 89, 91, 99, 124 benthic organisms 196
anchisaurids 53 bilophodont 196
anchitherline 186 biological development 3, 18, 29, 34, 50, 71,
Andean highlands 237 172, 178, 179, 188, 189, 192, 193, 196
angiosperms 6, 123–4, 130, 137, 139, 196 biological modelling 3, 5, 42, 48, 49, 50, 71,
ankylosaurs 65, 91, 127, 129 153, 153
ankylothecodont implants 32, 33 biochemical modelling 42, 49
‘antelope-sized’ 178, 182, 183 biomechanics 4, 6, 42, 152, 153, 154, 155, 169
antelopes 175, 177, 178 biomes 235, 239
anthoecia 235 biotic interaction 123
anthracotheres 182 birds 22, 42
apomorphies 113, 115, 116 Blancan 231, 232
aquatic animals 79 body armour 65
arboreals 37, 178, 179, 189–93 body size 5, 17, 19, 23, 37, 42–4, 60, 65, 72, 79,
archaic forms 172–3, 176, 181, 187, 188, 192 83, 84, 98–9, 112, 116, 124, 149, 152–3,
archosaurs 31, 35, 42, 44, 46 168–9, 171, 177–8, 182, 183, 184, 185, 190,
arctocyonids 145, 146, 149, 157 191
arctostylopids 170 bolide impact 191
Argentina 28, 56, 89, 238, 239 Bolivia 237, 238, 240
Arikareean Age 180, 183, 184, 220–2 bolosaurids 25
Arizona 55 brachydont 175, 176, 178, 179, 224, 229
artiodactyls 6, 7, 163, 170, 172, 176, 178, 181–2, braided streams 53, 61
186–9, 193, 225 Brazil 82, 100
[249]
250 Subject index
Bridgerian 180, 181, 183–4, 187 co-existing genera 232, 239–41

brontheres 178, 186 cog dentition 31, 33
browsers 168, 172, 175, 190–1, 224, 226, colobines 175
228–31, 234–5, 240 comparative studies 12, 24, 42, 43, 49, 50, 52,
buccal emargination 88 70–1, 168, 188–9, 238, 239, 241
buccal views 146, 150–1, 157, 162 ‘condylarths’ 6, 150, 155, 170, 171–3, 176, 178,
bunodont teeth 174, 175–6, 178–9, 181–2, 181–2, 187, 202–13, 214, 216
185, 186–7, 190, 194, 202 conifers 68
bunolophodonts 175, 176, 195 Connecticut 55
bunoselenodonts 175 conspecifics 26
burrows 22 continental drift 236
bush pig 175, 186 continental ecosystems 1–2, 6–7, 9, 14–15,
18–20, 25–26, 31, 60, 187–9, 193–6, 230,
C3 plants 7, 227, 228, 229, 234, 235, 239, 240, 231–3, 234–6, 238–9, 241–2
241 convergence 111, 114, 115
C4 plants 7, 227, 228, 229, 234, 235, 239, 240, coprolites 42
241 correlations 3, 50, 51, 71–2, 111, 148, 168, 169,
caegnathid oviraptorosaurs 65 172, 175, 188–93, 227
caloric energy 2 cows 175, 178
Calvin cycle 227 cranial structures 14, 16, 18, 19, 21, 22–5,
CAM 227, 229, 230 26–9, 85, 87, 94, 95, 115, 193, 237
camels 178 craniodental 26, 33, 34, 37, 61, 68, 71, 88, 96,
Campanian-Maastrichtian 79, 134 97–9, 106, 193, 236, 237
Canada 13, 55 Crassulacean Acid Metabolism 227, 229,
carbon shift 225, 226, 228, 229, 230, 234, 235, 230
236, 239–41 creodonts 187
Carboniferous 5, 9, 12–14, 16, 36, 37, 223 Cretaceous 61, 65, 66, 79, 99, 109, 116, 117,
carnassial teeth 67 123–5, 133, 137, 139, 140, 144, 145, 146,
Carnian Age 55, 60 149, 159, 161, 171, 227
carnivores 1, 9, 22, 43–4, 46, 49, 53, 65–8, 72, crocodilians 42, 66
145, 173, 187 crocodiles 79
carnivory 68, 70, 71, 72 crocodyliforms 62, 66, 67
caseid synapsids 11, 18–19, 23, 34, 37 cuspidation 6, 18, 25, 31, 33–4, 48, 50, 51, 52,
caviomorph rodents 188, 189 57, 66–7, 145, 149, 150, 152, 153, 160, 161,
cellulysis 2, 11, 12, 34, 37 164–5, 174, 176
Cenozoic 6, 145, 148, 150, 152, 155, 163–5, 171,
173, 191, 241 Dark Red Beds 58, 59, 62
Central America 7, 188 Da’znzhai Formation 60
ceratomorph perissodactyls 196 deer 175, 178
ceratopians 91 DELTRAN 135
cervids 64 dental correlations 71, 172–3, 175, 177, 189,
Chadronian 180, 183, 184, 187–8, 195 235–6
China 25, 56, 58–62, 66, 90 dental material 154–65, 225–7, 236, 240
clades 5, 15, 18, 20, 23, 25, 27–8, 30, 33, 35, dental microwear scans 80, 92, 93, 95–101,
68, 70–1, 99, 115, 125, 127–8, 131, 133, 134, 103, 104, 106–8, 129, 130, 154–65, 172,
135, 137, 140, 173, 233 226, 228–36
cladograms 21, 27, 30, 50, 68, 69, 70, 85, 111, dental structure 174–93, 196, 226–30,
112, 116, 130, 131, 135 234–6
Clarendon 231, 232 dentition 5–7, 13, 14, 15–18, 19, 20, 21, 22, 25,
Clarens Formation 55, 61 26–35, 45, 46, 47, 48, 50, 52, 54, 57–9,
Clarkforkian 180, 183, 184, 207–8 60, 62–3, 64–8, 71–2, 79–81, 82, 88, 90,
climate 6, 53, 55–7, 61, 169, 170, 173, 178, 179, 92, 93–101, 102–7, 108, 110–12, 114–15,
190–4, 195–6, 228, 234–5 126–30, 144–5, 148, 149, 150, 151,
coeval strata 9 152–163, 164, 171–7, 196, 228, 229, 230,
coevolution 123, 124, 130, 135, 139, 169, 191, 232–3, 238–41
226, 230, 231, 232 dermal armour 35, 65, 125, 126
Subject index 251
detritivory 9, 11, 12 East Greenland 55

diagenesis 227 ecology 1, 4, 5, 12, 22, 37, 44, 50, 53, 57, 72,
diagrams 36, 132 109–11, 123, 190–6, 223, 226–30, 233–6,
dichobunid artiodactyls 172 240, 241–2
dicynodont therapsids 22 ecomorphological types 171, 177, 190, 191,
diets 26, 33, 42–4, 47, 48, 49, 50–1, 51, 52–3, 192, 193
57–8, 60, 62, 65, 68, 69, 70, 72, 79, 110, ectocranial crests 29
115, 153, 169, 171, 175, 176–7, 181, 190–93, edaphosaurid synapsids 37
226, 234, 235 edentates 233
digestive aids 2, 3, 9, 10–12, 17, 34, 68, 89, EDJ 155, 157, 158, 159, 163
106, 127, 129, 130, 168 elephants 85, 178, 191
digestion 1–3, 10–12, 15, 18, 20, 23, 28, 32, 34, Elliot Formation 55, 60
35, 60, 89, 123 endosymbiosis 2–4, 10–12, 34
dik-dik 177 England 95
dinoceratans 176, 182 enterolites 42, 49, 53, 68, 144
dinosauriform condition 68 environment 53–7, 60–1, 123–4, 173, 188–96,
dinosaurs 1, 5, 6, 28, 33, 37, 46, 52, 53, 57–8, 234–5
61, 62, 66, 79, 164 enzymes 2
diplodocoids 85, 89, 99, 102–3, 115, 128, Eocene 6, 159, 161–96, 230–4
134 eomyids 190
distribution 23, 26, 30–1, 36, 43, 79, 123, 127, epochs 5, 6, 7, 9, 13, 169–96
169, 172–3, 181, 233 equids 178, 186, 188, 234, 238
divergence 172, 173, 181, 190, 191, 192–3 euhelopodids 90, 92, 94–6, 112–3, 117, 128
diversity 1, 3, 5, 6, 9, 43, 61, 79, 124, 125, 130, eupelycosaurs 18
131, 132, 133, 134, 135, 136, 137, 138, 139, Europe 9, 15, 18, 125, 172, 188, 190, 196
140, 144, 157, 169, 172–9, 180, 181–96, evolutionary dynamics 124, 227
223, 230–1, 234–5, 241–2 exclusive herbivores 43, 44, 46, 48, 61, 168
‘dog-sized’ 177, 182, 183 ‘Extant Phylogenetic Bracket’ 42, 70
Duchesnean Land Mammal Age 174, 180, extinct orders 175, 176, 189, 192, 227, 241–2
181, 183, 184, 187, 188, 214–15
duikers 178 facial musculature 85, 88
Dull Purplish Bed Formation 59, 62 facultative herbivores 67
dune systems 55, 61 facultative omnivores 50, 52, 61, 64, 65
faunal changes 172, 173, 190, 191, 196
Early Early Arikareean 220–1 faunivorous forms 4, 5, 10, 11–13, 16–18, 24,
Early Bridgerian 210 26, 29, 33–5, 49, 68, 69
Early Cenozoic 165, 173, 191 feeding habits 9, 10, 12, 14, 16, 20, 22–3, 26,
Early Chadronian 216–17 30, 32–7, 42–9, 53, 60, 68, 69–72, 79–81,
Early Cretaceous 61, 64, 65–6, 109, 123–6, 88, 102–3, 110, 114–17, 123–5, 135, 145,
131, 135, 139 146, 147, 164, 172, 228, 229, 230, 234, 235,
Early Eocene 162–4, 172–3, 179–81, 183–8, 238–42
191–2, 194–6 feeding mechanisms 34, 81, 85, 88–91, 94–5,
Early Jurassic 43, 58, 60, 79, 90, 116, 127, 131, 97, 99, 102–3, 105, 108–17, 124, 135, 145,
133 146, 147, 148, 150, 151, 162–5, 174–7, 181,
Early Mesozoic 35, 36 190, 191, 228, 229
Early Miocene 230, 234–5, 237 feeding styles 34, 135, 136, 137, 138, 139, 148,
Early Norian Age 55, 60 228, 229, 230, 239–41
Early Oligocene 169, 180–5 fermentative symbiosis 12, 17, 34, 35, 60, 128,
Early Paleocene 149–51, 156–9, 160–1, 164, 130, 168
178–80, 181–4 fish eaters 44
Early Paleogene 182, 183, 187 Fleming Fjord Formation 53, 55
Early Permian 9, 14–15, 18–19, 26 flood plains 55, 56
Early Tertiary 6, 171 floral dominance 139, 191
Early Triassic 29, 33 Florida 240
Early Unitan 212–13 floristic diversity 191, 196
Early Wasatchian 208 fluvial settings 53
252 Subject index
folivores 6, 12, 49, 108–10, 113, 129, 168–9, habitats 191–3, 234, 235
172, 175, 176, 181, 186, 188, 190–2, 194–5 hadrosaurids 127, 134, 163
folivory 6, 12, 88, 108–10, 114, 172, 186, 191 Hauterivian-Barremian 79
food processing 2, 4, 10, 11, 12, 34–5, 164 Hell Creek Formation 146
food values 2, 108 Hemingfordian 231
food webs 5, 37, 191–5, 230, 239 Hemphillian 231, 232
foraminifera 196 ‘heralds’ 172
Forest Sandstone Formation 55, 61 herbivores 2–66, 70–2, 79–80, 117, 123–55,
forests 191–3, 195 164–95, 202–23, 226–34, 235–41
fossil studies 3, 28, 174, 223, 226, 230, 231, herbivory 1–24, 31–6, 42–58, 64–72, 123, 127,
233, 234, 235, 236–239 130, 144, 148, 164–9, 173, 223, 224, 238,
fossil grasses 230, 235, 238–41 240–42
fossorial habits 22 heterodonts 29, 30, 64
frogs 12 Hettangian Age 55, 56, 90, 131
frugivores 6, 148, 168, 174, 175–6, 192, 193, Hettangian-Pliensbachian 56, 60, 61, 90
194, 195 high browsers 80, 88, 91, 95, 98–9, 109, 110,
frugivory 172 113, 116, 128, 129
fruit eaters 49, 158, 168 Holarctica 230
functional interpretation 1, 42, 124, 164, 165 holotypes 62
functional morphology 1–10, 47, 48, 49, homologues 13
64–6, 80, 85, 124–7, 151–7 homoplasy 115, 116
horned dinosaurs 127
gastric mills 10, 53, 68, 89, 127 ‘horse-sized’ 178, 183
gastroliths 89, 127, 129 horses 163–4, 172, 175, 224–30, 231–6
generalist feeders 172 ‘Hot House’ types 173
generic diversity 61, 169, 178, 179, 180, 181–5, humans 148, 174, 175, 224, 226
188–9, 230 hyracotherine horses 172
genetic shifts 16, 30 hydrolysis 2
geochemical modelling 61, 226 hyopsodontids 170, 228, 229
geological periods 2–13, 20, 22, 28, 36, 53, 58, hypotheses 6, 11, 24, 27, 30, 36, 44–53, 68, 71,
61, 105, 110, 111, 123–34, 135, 149–61, 79, 124
162–8 hypselodont 225, 229, 236, 237, 238–41
geological strata 22, 53, 60–1, 131 hypsilophodontid ornithopods 64
Germany 56 hypsodont 175–9, 182, 188, 224–41
‘ghost lineages’ 6, 124, 130–40 hyracodont 186
gigantism 113, 128, 129
giraffe 175, 178 ‘Ice House’ 173, 188
glenoid fossa 93, 103–6, 115 iguanodontians 64, 127, 131, 132–3
gliders 191 India 28, 99
global climates 173, 190–1, 234–5 Indian green frog 12
global environments 173, 190–1, 223, 234, insectivores 11, 17, 31, 48–9, 149, 164, 171, 187,
235, 239–42 192
gomphothere proboscideans 230 insectivory 12, 31, 48, 50, 149, 164, 178,
Gondwana 31 179
granivores 169, 194 insects 223
grasses 225, 226, 230, 234, 235 interrelationships 50, 70
grazers 168, 169, 175, 176, 192, 223, 224–36, invertebrates 7, 53, 79, 80, 88, 223
240 isotopic analysis 1, 7, 223–41
grasslands 7, 57, 169, 192, 223–42 isotopic shifts 234–5, 236–40
grazing herbivores 6, 7, 168, 169, 192,
226–29, 234–42 jaw features 10, 20–3, 31–4, 46, 53, 64, 67, 72,
Great American Interchange 240 80–98, 100–15, 128, 129, 145, 152, 160,
Great Transformation 230, 234–40 163
Greenland 53 Jurassic 43, 53, 61, 79, 90, 94, 105, 109–17,
gut processors 135, 136, 137–8 125–7, 131–9
gymnosperms 123 juveniles 12, 16, 30, 31, 50, 57, 58, 110
Subject index 253
Karoo Basin 60 Lower Paleocene 150

Kayenta Formation 55 Lower Shaximiao Formation 111
Kenya 225
keratinous beaks 20–3, 33, 48, 65 Maastrichtian 79, 131
Kimmeridgian 79 McCoy Brook Formation 55
Knollenmergel 56 Madagascar 82, 100, 101
Malawi 67
labiolingual compression 46 mammalian diet 67, 158, 164, 168–93
lacustrine 53, 55, 56 mammalian digestion 2–4, 10, 11, 42
lagomorphs 139, 190, 225 mammals 2–7, 12, 13, 42, 50, 53, 57, 146, 148,
land bridges 7 155, 157–79, 182–7, 190, 191–3, 202–22,
Late Bridgerian 211–12 223–7, 230–9
Late Carboniferous 5, 9, 14, 37 marginal dentition 4, 33, 34, 46, 48
Late Carnian 55, 60 marine specimens 79
Late Cenozoic 163, 164, 173 marsupial insectivores 145
Late Chadrionian 218 Mediterranean-type climate 228
Late Cretaceous 124, 131, 135, 137–40, 149, melanorosaurids 44, 60
171, 191 mesic environments 61
Late Early Arikareean 221–2 Mesozoic 10, 13, 35, 36, 72, 123–5, 130, 133,
Late Eocene 172–4, 178, 180, 182, 183, 184, 134, 135–9, 171, 173, 191
185, 186–93, 194, 195 metabolic rates 34, 35, 89, 168
Late Jurassic 94, 109, 111, 116, 131, 133, 134, ‘miacoid’ carnivorans 187
135, 139 micro-organisms 2, 4, 10–12, 34, 89, 168
Late Mesozoic 191 mid-Carboniferous 12
Late Miocene 234, 235, 236, 239–41 mid-Cretaceous 131, 133, 137
Late Norian 56 mid-Eocene 171–3, 189–188, 190–6
Late Oligocene 180, 182, 183–5, 189, 237 mid-Jurassic 61, 79, 94, 111, 116, 125, 126, 134
Late Paleocene 172, 178, 179–85, 190–2 mid-Paleocene 178, 179
Late Paleogene 187, 189, 192 mid-Pennsylvanian 13
Late Paleozoic 28, 33, 35, 72 mid-Triassic 20, 22, 31, 33
Late Pennsylvanian 9, 13, 14, 15, 17 Middle Bridgerian 211
Late Permian 5, 9, 20, 22–6, 28, 37, 68 Middle Chadronian 217–18
Late Pleistocene 230, 234, 241 Middle Miocene 230, 232, 233, 234
Late Tertiary 230, 240 Middle Pleistocene 240
Late Triassic 20, 22, 28, 31, 33, 43, 116, 128, Middle Tertiary 223, 228, 230, 239, 242
131, 135 Middle Wasatchian 208–9
Late Unitan 213, 214 migrant mammals 240
Late Wasatchian 209–10 Miocene 6, 7, 189, 192, 224, 225, 226, 230,
leptochoerids 189 231–7, 239–41
limb specialisation 19, 44, 53, 65, 91, 95, 100, mioclanenids 170
110–112, 113, 116 modern mammals 174, 175, 181, 187, 188, 190,
lineages 170, 171, 172, 177 191, 192, 193
Linnean Society 21 modernity 9, 175
litopterns 233 modifications 29, 64, 65, 71, 164, 165
lizards 3, 5, 10, 12, 18, 30, 34, 43, 46, 48, 49, molariform dentition 15, 28–9, 66–7, 72,
52, 62, 72 156–61, 175, 228, 229
Ilamas 178, 226 molluscivores 32, 65, 108
localities 55–9, 60–2, 110, 111, 174 Mongolia 103
logarithmic scale 177 monkeys 176
lophodont 172, 207–22 monophyletic clades 43, 68, 91, 124
Los Colorados Formation 56 Morocco 100
low browsers 94, 96, 105, 109–16, morphology 2–5, 33, 144, 169, 171, 174, 196, 223
125–8 Morrison Formation 110
Lower Elliot Formation 55, 60, 61 mouse deer 175
Lower Jurassic 53 musculature 4, 10, 20–6, 32–4, 65, 81, 82,
Lower Lufeng Formation 56–62, 90 93–8, 100, 102, 103, 106–15
254 Subject index
Nacimiento Formation 150, 151 osteological features 4, 6, 7, 10, 13, 14–24,

narial retraction 85 25–31, 32–4, 64, 65, 102–3, 129
Navajo Sandstone 55 outgroups 68, 71
Nebraska 235, 237 oviraptorosaurs 65
Neogene 6, 7, 178, 189, 194 Oxfordshire 79
nest builders 12
New World 189, 223, 225 paca 189
Newark Supergroup 55 pachycephalosaurs 64–5, 127, 128, 129, 135
niche partitioning 61, 110, 111, 144 Pakistan 236
nodes 111–16 paleobiological interpretation 42, 72, 168,
nodosaurid ankylosaur 126 169, 195, 196
non-amniotes 12 paleobotanical data 195
non-avian dinosaurs 42, 144 Paleocene 6, 144, 149, 150, 151, 155, 156–79,
non-dinosaurs 28, 35, 62, 66 180–94, 202–8, 233
non-eutherian 187 paleoecology 61, 80, 168–9, 173, 190–6, 234,
non-mammals 4, 11, 28, 35 235, 239
Norian Age 55, 56, 60, 131 paleoenvironments 55, 56, 57, 60, 61, 80,
North America 6–9, 15–19, 25, 105, 110, 168–9, 188–96, 202–22, 234, 235, 239
125–6, 169, 172–6, 187–96, 230, 232–7, Paleogene 6, 168–70, 173, 176, 178–96,
238–41 202–22
Northern Hemisphere 127, 196, 236, 238 paleogeographic configuration 196
notoungulates 7, 225, 239 paleotemperatures 170, 173, 178–9, 192–6,
Nova Scotia 13, 53 235
nutrition 2, 3, 108, 110, 116, 144, 153, 158, 159, paleontology 3, 4, 5, 168, 196, 226
168 paleosol 236, 239
Paleozoic 5, 9–11, 12–14, 23, 28, 33–5, 36, 72
obligate arboreal mammals 193 palinal movement 34, 106, 109
obligate herbivores 3, 10, 71, 193 pampas environments 57
occlusal dentition 3–4, 28–9, 147, 148, 150, Panamanian land bridge 7, 241
151–63, 225, 232 pandas 3
okapi 178 pantodonts 170, 172, 176, 182
Oklahoma 19 paraphyletic 43, 86, 90, 94
Old World 223, 230, 240 PAUP 135
Oligocene 6, 160, 168–70, 180–9, 193, 194, peaks 79, 131, 134, 136, 137
195, 219–22, 233, 237 peccaries 178, 182
omnivores 5, 6, 12, 18, 22, 29, 30, 44–6, 49, Pennsylvanian epoch 9, 13
50, 66–72, 171–5, 190, 191 periptychids 149, 150, 155, 157, 170
omnivory 12, 16, 64–8, 70–1, 172, 178 perissodactyls 6, 7, 172, 175, 181–2, 186, 188,
omomyids 187 196
ontogeny 50, 53, 57, 58, 60, 117 Permian 2, 5, 12, 14, 16, 18, 19, 20, 22–3, 25–6,
opportunistic herbivores 12, 67–8 28, 36, 37, 68
opportunistic omnivores 50, 52, 68 phenacodontids 170
optimization 135 photosynthesis 227
oral features 3, 18, 23, 28–9, 34, 42, 43–8, phylogenetic analysis 1, 6, 11, 15, 18, 23, 26,
50–9, 60–8, 71–2, 79–82, 88, 90–9, 33, 36, 42, 44, 51, 68, 69, 99, 113, 124, 125,
102–6, 114, 116, 125–9, 135, 145–65, 171–5, 130, 131, 137, 177, 196, 240–1
189–93, 226 phylogeny 69, 70, 72, 113, 115, 124, 130, 131
Orellan Age 180, 183, 184, 219 physiology 2, 3, 5, 6, 11, 71, 88, 171
oredont 186 physiognomy 124, 139
ornithischian dinosaurs 1, 34, 46, 62–8, 70, phytoliths 106, 224, 225, 230
110, 114, 124, 127, 129–39 pigs 174, 178, 186, 187
ornithomimids 65 piscivory 44
orthal jaw activity 29, 102, 112, 115, 126–9, placental herbivores 155, 223–7
135, 148–51, 163 plant biomass 223, 227, 228, 229
orthal pulpers 135, 136, 137, 138–9 plant evolution 124, 158, 159
orthal slicers 136, 137, 138, 139 plant physiology 2, 227
Subject index 255
plant eaters 1–5, 12, 14, 18, 19, 26, 28, 33–7, savanna grasslands 57, 192, 230
43, 46, 50, 51, 56–8, 66–8, 71, 72, 79, 105, scansorial habitats 37
108, 110, 123–31, 137, 155, 159, 165–71, scavengers 45, 48
227–9, 238–41 sedimentary deposits 53, 55, 56, 60, 61
plant fossils 196 selective browsers 172, 190, 193
playa lakes 53, 61 selenodont artiodactyls 163, 181, 188
pleisomorphs 15, 24, 33, 90, 94, 113 selenodonts 175
Pleistocene 230, 240, 241 semi-arid conditions 53
plesiadapiforms 187 semi-lophed 185, 188–9, 193–5, 203–22
pleurokinetic hinges 64 shearing mechanisms 29, 33, 34, 72, 145–8,
Pliensbachian-Toarcian Age 55 151, 152, 153, 164, 172
Pliocene 230, 231, 232, 239, 240 ‘shoe-horn’ 44
pluvial periods 56, 57 Sichuan Province 60, 111
polymers 2 Sinemurian Age 131
poposaurid archosaur 44 Sinemurian-Pliensbachian Age 55
Portland Formation 55 size distribution 177, 178
Portugal 89 skeletal mobility 88, 94–5, 102–3, 115, 116,
postcrania 18, 26, 27, 94–8, 100, 102, 103, 104, 117, 127
106, 115, 116 skeletal structures 4–6, 10, 15–20, 21, 22–4,
‘precocious hypsodonty’ 238 34, 35, 42–8, 53, 58, 66–8, 71, 80, 83, 84,
predators 26, 45, 48, 65, 68 85, 87, 88–99, 102–6, 115, 126–7, 154–5,
primates 172–9, 187–9, 191–5 177, 237
prism decussation 154–7, 161–5 South Africa 9, 20, 21, 22, 33, 44, 60
propaliny 67, 68, 96, 97, 103–9 South America 7, 188, 225, 230, 233, 236, 237,
prosauropods 5, 34, 43, 44, 45, 46, 47, 52–9, 238, 239, 240, 241
60–6, 70, 72, 85, 89, 91, 113, 114, 117, 128, 134 Spain 65
‘proteutherians’ 187 species extinction 196
protists 2 species origination 196
protoconids 149 sphenodontids 53
protohippus 231 star-burst cladograms 131
pteridophytes 123 statistics 136, 137, 137, 139, 159, 170, 177–80,
pteridosperms 68 183–94, 223–38
Puercan 150–8, 160–1, 180–4, 202 Stephanian B. 13
pygmy hippo 176, 178 stratigraphic records 1, 130, 202–22
stress 152, 153, 154, 155
quadritubercular forms 174 subterranean habits 22
succulents 227
‘rabbit-sized’ 177, 183 suiform artiodactyls 172, 182, 188
rabbits 230 surgeonfish 49
rainfall patterns 192, 193 symbiotic support 2, 4, 10–12, 37, 169
reptiles 3, 5, 10, 12, 18, 23–7, 30, 32–7, 43, 48, sympatric patterns 130
53, 62, 68 synapomorphy 89, 114
retroarticular processes 106
‘rhino-sized’ 178, 183 tables 49, 55, 69, 81, 82, 83, 86, 154, 202–22
rhinocerotoids 186–8, 196 Tanzania 111
rhinos 175, 186, 225, 235, 237 taphonomic bias 124, 168
rhizomes 33 tapinocephalid dinocephalians 20
rhynocosaurs 31, 32, 34 tapirid 188
rodents 7, 168, 188–90, 195, 225, 230, 233 tapiroids 182, 196
roe deer 178 tapirs 85, 178
ruminant artiodactyls 175, 186, 188, 189, 193, taxa 3–22, 26–45, 55, 61–8, 70, 72, 102, 117,
225 124–31, 149, 159, 161, 171, 172–8, 180–95,
Russia 20, 23, 25 233, 236, 238
taxon 1, 11, 23, 89
Salla 237, 238 taxonomic diversity 1, 3, 5, 15, 17, 28, 36, 43,
sauropod crania 81, 116 61, 62, 81–84, 89, 124, 128, 135, 171–96, 233
256 Subject index
temperate vegetation 192, 193, 227 turtles 12, 22, 48, 66

temperature changes 170, 173, 178, 191–6 tylopods 186, 188, 193
Tendaguru beds 111
Tennessee 230 Uintan Land Mammal Age 180, 181, 183, 184,
terrestrial grazing guilds 223, 227, 235, 239, 187, 188
241, 242 uintatheres 170, 172, 176, 178
terrestrial mammals 176, 177, 190–3, 234, ungual structures 33
235 ungulates 1, 3, 6, 7, 33, 145, 149, 150–69,
terrestrials 15, 28, 35, 37, 72, 79, 80, 117, 170–84, 186–96
171–8, 182, 186–92, 223, 236 United States 79, 171
territorial behavior 64 Upper Cretaceous 99, 103
Tertiary 6, 145, 168–71, 191, 223–42 Upper Elliot Formation 55, 60, 61
tetrapods 1–15, 20, 23, 28, 33–7, 124 Upper Jurassic 89, 105, 110
Texas 19, 33 USA 55, 79
theridomyids 190 Uzbekistan 71
Tiffanian 180, 183, 184, 204–7
time charts 233 varanids 96
time intervals 174 vegetation 56, 57–61, 239
titanosaurids 92, 102, 103, 113, 128 vertebrates 2–5, 9, 10, 28, 35, 55, 223
Toarcian Age 55 voles 168
‘tooth-comb’ 79, 102, 110, 114, 115 vulcanodontids 91, 94, 113, 114, 117
tooth types 178, 179
toothless groups 65, 66 Wales 56
Torrejonian 180, 183, 184, 203–4 wapiti 178
tortoises 50 Wasatchian Age 162, 164, 180, 181, 183, 184,
trackways 80 208–10
tragulids 176–7, 189, 193 water chevrotain 176
traguloid 186, 188 Westphalian B. 13
Transformation 230 Whitneyan Age 180, 183, 184, 219–20
transition diets 44 woodlands 192, 193
transition dinosaurs 66 Wuding Basin 60, 90
transverse grinders 135, 136, 137, 138, 139
traversodonts 28, 29 xerophytes 57, 60
tree hyrax 175, 176, 177
Triassic 5, 10, 11, 20–44, 116, 128, 131, 134–5 Young’s Modulus test 153, 154
tribosphenic dentition 145, 148, 152 Yunnan 56, 58, 60, 90
tritylodontid synapsids 66
trophic dynamics 72, 124, 139, 172, 173, 193 zapodids 190
trophic interaction 9, 37, 72 zhelestids 171
tropical environments 193, 227 zig-zag dentition 33
tropical forests 188, 189, 191, 192 Zimbabwe 55
tubers 33 zygapophyses 35, 88

Origen Herviboria

Uploaded by

Copyright:

Available Formats

Origen Herviboria

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Origen Herviboria

Uploaded by

Copyright:

Available Formats

Evolution of Herbivory in Terrestrial Vertebrates

Perspectives from the fossil record

Although herbivory probably ﬁrst appeared over 300 million

h a n s - d i e t e r s u e s is Vice-President for Collections and

Cambridge University Press

© Cambridge University Press 2000

This publication is in copyright. Subject to statutory exception

First published 2000

Library of Congress Cataloguing in Publication data

ISBN-13 978-0-521-59449-3 hardback

ISBN-13 978-0-521-02119-7 paperback

List of contributors page vii

1 Herbivory in terrestrial vertebrates: an introduction 1

Taxonomic index 245

This book presents a collection of paleontological perspectives on the

thanks are due to the co-organizer of this symposium, Conrad C. Laban-

Herbivory in terrestrial vertebrates: an

Understanding the ecological attributes of extinct organisms has

numbers of herbivores supporting a relatively small number of carni-

Herbivory in extinct vertebrates

Unequivocal evidence for trophic interactions between extinct animals

Luke 1984) and, in many instances, can be tested experimentally in extant

than those of their closest faunivorous relatives, or the trunk region is

MacFadden (Chapter 8) presents an overview of recent work by him and

Bjorndal, K. A., Bolten, A. B., and Moore, J. E. (1990). Digestive fermentation in

Herbivory in late Paleozoic and Triassic terrestrial

The exploitation of land plants as a major food resource by amniote

Morphological and physiological correlates of herbivory

The effective utilization of high-ﬁber plant material by vertebrates

up by oral processing or by comminution in a muscular gizzard using

symbiotic micro-organisms for fermentative breakdown of the cellulose

Acquisition of cellulytic endosymbionts by herbivorous

The aforementioned anatomical adaptations for feeding on high-ﬁber

Carboniferous Diadectidae present an interesting conundrum. They pre-

Diversity of late Paleozoic and Triassic herbivorous

The known fossil record of amniote tetrapods begins in the Mid-Pennsyl-

Late Paleozoic taxa

that diadectids formed a common element of the terrestrial tetrapod

of unworn teeth bear ﬁne, oblique serrations, which subsequently

ranging from the primitive pattern seen in Cotylorhynchus (with trans-

Synapsida: Therapsida: Dinocephalia

Synapsida: Therapsida: Anomodontia

Figure 2.5. Skulls of three representatives of Anomodontia ((A) Patranomodon,

dentition is still poorly known, but it apparently lacks any specializations

Reptilia: Parareptilia: Pareiasauria

somewhat controversial, the Procolophonoidea are the most likely sister-

mandible, which is consistent with a greater anteroposterior length of the

Figure 2.8. Skulls of three representatives of Captorhinidae (A, Captorhinus;

related to feeding on plants. These include some degree of elongation of

Synapsida: Cynodontia: Gomphodontia

Figure 2.9. Skulls of two representatives of Cynodontia (A, Thrinaxodon and

massive, with deep zygomatic arches and often pronounced ectocranial

Figure 2.10. Skulls of three representatives of Procolophonoidea (A, Owenetta;

high-ﬁber plant material evolved within the clade Gomphodontia. This

Reptilia: Parareptilia: Procolophonidae

Reptilia: Diapsida: Archosauromorpha: Rhynchosauria

Figure 2.11. Skulls of two representatives of Rhynchosauria (A, Mesosuchus and

In the most derived taxa such as Hyperodapedon, the transverse width of

the presence of a voluminous digestive tract (Benton 1983). Some taxa