LUCAS How Teeth Work 1st Ed 2004 PDF

DENTAL FUNCTIONAL MORPHOLOGY
How Teeth Work
Dental Functional Morphology offers an innovative alternative to the

received wisdom that teeth merely crush, cut, shear or grind food, and
shows how teeth adapt to diet. Providing an analysis of tooth action
based on an understanding of how food particles break, it shows how
tooth form from the earliest mammals to modern-day humans can be
understood using very basic considerations about fracture. It outlines
the theoretical basis step by step, explaining the factors governing
tooth shape and size, and provides an allometric analysis that will
revolutionize attitudes to the evolution of the human face and the
impact of cooked foods on our dentition. In addition, the basis of the
mechanics behind the fracture of different types of food, and methods
of measurement are given in an easy-to-use appendix. It will be an
important sourcebook for physical anthropologists, dental and food
scientists, palaeontologists, and those interested in feeding ecology.
Peter W. Lucas is a professor of anatomy at the University of Hong
Kong. He is particularly interested in the function and evolution of
mammalian teeth, but he also studies the factors involved in food
choice, particularly in primates.
DENTAL FUNCTIONAL
MORPHOLOGY
How Teeth Work
PETER W. LUCAS
University of Hong Kong
CAMBRIDGE UNIVERSITY PRESS

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, So 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/9780521562362
Peter W. Lucas 2004
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 2004
This digitally printed first paperback version, with corrections, 2006
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication data
Lucas, P. (Peter W.)
Dental functional morphology: how teeth work / by Peter W. Lucas.
p. cm.
Includes bibliographical references and index.
ISBN 0 521 56236 8
1. Teeth. 2. TeethEvolution. 3. TeethAnatomy.
4. Dental anthropology. I. Title.
QM311.L83 2004
611.314dc22 2003063885
ISBN-13 978-0-521-56236-2 hardback
ISBN-10 0-521-56236-8 hardback
ISBN-13 978-0-521-03540-8 paperback
ISBN-10 0-521-03540-6 paperback
The publisher has used its best endeavours to ensure that the URLs for external websites referred to in
this publication are correct and active at the time of going to press. However, the publisher has no
responsibility for the websites and can make no guarantee that a site will remain live or
that the content is or will remain appropriate.
To my wife Mariati and my daughters, Katherine and Diana,

with my everlasting love
Contents
Preface
Flickart
page ix
xv
1 How to get excited about teeth
2 The basic structure of the mammalian mouth
13
3 How the mouth operates
55
4 Tooth shape
87
5 Tooth size
133
6 Tooth wear
181
7 The evolution of the mammalian dentition
202
Appendix A
Mechanical properties and their measurement: material
properties made easy
257
Appendix B
Properties of teeth and potential foods
283
Notes
References
Index
292
305
346
vii
Preface
Teeth cause such dreadful problems in humans that interest in them by

non-dentists would seem both unlikely and unhealthy. Who could get
excited about tooth decay and gum disease? The physical reality of such
apparently moribund structures is paralleled in our cultural perception of
them. Diseased or not, the whole mouth is viewed as an unclean region
of the body in most parts of the world, especially when it is crammed full
of food. Parents, particularly in Western countries, often train children to
keep their lips sealed when they are eating even though this is difcult to
follow exactly and, indeed, little food seems to re-emerge if the instruction
is disobeyed. It is debatable if this training is necessary. While it is possible to
sit next to someone at a banquet and get sprayed with seafood, for example,
from his or her mouth, the nutritive loss to the diner, represented by the
sum of those ne particles, seems negligible compared to what is obviously
going down their throat. This is a clear sign of the efciency of the chewing
process. The main reason, in fact, that food particles are expelled is that
the person is talking while chewing. Talking involves the expiration of air
and that is what pushes food particles forwards. This may seem a strange
example but it makes a strong point: the thought of even catching sight
of food that was, a moment previously, decorating a plate evokes visceral
feelings (of a somewhat inside-out kind) rather than artistic ones. The plate,
too, seems to lose its appeal after most food has disappeared and may be
quickly consigned to the wash. In short, we appear often to be embarrassed,
if not disgusted, by the major biological function of feeding and the need
that underlies it, although we dont go as far in hiding it from social view
as we do activities at the other end of the gut. Presumably, an intuitive
understanding of hygiene explains that disparity, but restrictive practices
about eating pervade most human societies and, according to an intriguing
account by Visser (1991), have many cross-cultural features that are not
easily explained on grounds of hygiene (the latter, of course, preoccupies
developed societies: Lacey, 1994).
ix
Preface
So why write a book on how teeth work? I answer this, not by seeing any
disguised elegance in the appearance of a good feed, or in the avocation of
new table manners, but by admitting and attempting to transmit through
this book a personal fascination with the fundamental role that teeth have
had in our evolution and my dissatisfaction with current explanations on
both how they work and why they evolved.
How do teeth work? One relatively uniform answer to this is already
provided in numerous accounts in top journals, encyclopaedias and even
school texts. It is that teeth variously crush, cut, shear or grind food. And
with slight complications, that is more or less the prevailing wisdom in
dentistry, zoology, palaeontology, anthropology and many other biological
elds (with the eminent exception of food science). Despite apparent unanimity, these accounts are completely wrong. A genuine analysis of tooth
action, one that could possess explanatory power rather than glib description, starts in the still somewhat obscure world of fracture mechanics in
the understanding of how food particles break. Such an analysis is not likely
to be as edible as slogans like shearing or grinding: inevitably, analytical
depth requires more than the coining of facile words and phrases.
The action of the teeth cannot be separated from that of the mouth, so
an attempt is made consistently to understand oral processing as whole.
However, this is the oral processing of solids, not liquids. The ingestion
of liquids like nectar and honey is all tongue and no teeth, and drinking
is actually not that common an activity in mammals certainly not in
primates once they grow up. The process of growing up, of development,
is not discussed, so those interested in suckling and weaning and how the
young cope will not nd anything on it. The largest body of information
to be excluded here though is neurobiology: there is no space to include
much of it here and some of its alleys seem currently to be very dark.
I am sensitive to the knowledge that the further the book sinks into a
world requiring the learning of new terms, the more potential readers will
be lost. Accordingly, I have tried to present my viewpoint in as simple a
way as possible, deliberately seeking light generalization rather than long
and dark specics. One of the worst aspects of biology is the plethora of
terms that it employs. If I added a full suite of terms from mechanics, then
the book would be a slow read. Sensitive to this, I have made a deliberate
effort to reduce the number of terms to a minimum. The overall intention
is to provide a fundamental analysis of the feeding apparatus of mammals,
based on the interface between outside and inside, i.e. the contact between
foods and teeth. If this interface is properly understood, then I contend
that the optimal design of the working surface of teeth, the organization of
Preface
xi
the structural support for this surface and the production of bite forces by
muscles that move it all should follow in a predictable way. If this analysis
succeeds, then it should open the way to a fundamental understanding of
the evolution of feeding adaptations in mammals. On the plus side, I hope
that this book will be of value to anyone in bioscience with an interest in
feeding. On the debit side, I will undoubtedly have made some dreadful
mistakes and may sometimes appear uncharitable to those with other views.
However, the book is meant to be constructive and, in this sense, research
life has some resemblance to a game of chess: no one ever excels by just
making moves that have been seen before. Unless I am mistaken, a lot of
the moves recorded here are new.
I hope that this book offers a cohesive framework on the function of
teeth. For dentists and those basic scientists whose work might be covered
by the term oral biology, this is an account of how teeth break foods
down, untainted by the modern reverse trend. For food scientists and those
concerned with food texture, it is about oral physics, on which psychophysical investigations of food texture can be superimposed (see Chapter 7).
Common to both dentistry and food science has been a strong interest in
applied science. Whereas dentists look at patients, food scientists look at
consumers. However, dental surgery exists as a discipline entirely separated
from medicine because the dentition is the one area of the body requiring
regular surgery. It can be argued then that in many parts of the world,
the patient is also the consumer. Both dentists and food scientists might
benet from basic models of oral processing in order, on the one hand, to
predict the outcome of surgery or, on the other, to provide a foundation
for psychophysical investigations of food texture. Until recently, there was
very little cross-talk between these disciplines in most countries (I exclude
Japan from this), something possibly caused by lack of a sufciently overarching viewpoint. For ecologists, I describe the actual mechanical properties of foods that could inuence dietary niches and feeding rates. For
materials scientists, it may provide some information on a restricted group
of biomaterials foods. For palaeontologists and those concerned with the
evolution of the feeding apparatus in a wide group of organisms, the hope
must be that enough is explained here to help in the generation of general
theories for evolutionary change.
It seems a book tradition to tell people where they can nd articles on its
subject matter. Each of the above elds has scientic journals responsible
for the vast literature on the structure and function of mammalian teeth,
but if the denition of this books scope is taken to encompass feeding,
then the answer to this is really just about anywhere, even in physical
xii
Preface
sciences. Well-cited journals here include the American Journal of Physical

Anthropology, Archives of Oral Biology, Journal of Dental Research, Journal of
Human Evolution and Journal of Texture Studies. Some, such as the Journal
of Prosthetic Dentistry, are underrepresented because of their clinical slant.
Unfortunately, in recent years, there is another candidate for this roll of
honour, one that likes to bestow its own credits. The Annals of Improbable
Research has devoted much of its space to feeding research and several papers
have won IgNobel awards. Sometimes, this recognition is truly deserved,
but the frequency with which food research gets treated this way in the
journal seems to suggest that its contributors are actually obsessed with the
eld too, although clearly from a different perspective.
The problem with being sure about scientic novelty is the need to wade
through the mounds of information that modern biology accumulates so
rapidly. Really, there is too little time to sit down with the enormous body
of relevant literature to see if a theory ts well with the evidence or, alternatively, casts it into serious doubt. The purpose of this book is overtly to
try to bring together some of these piles according to the overarching theoretical model that dominates the book. Anyone who reads it will probably
know some of the areas that I have explored here, and in particular areas
of expertise, may well know more than I do, but what I am banking on is
that few will be acquainted with the full scope of this book.
One of the great benets of writing this, afforded for the only time in my
career, is the space that it provides in the preface to thank those who have
been seminal inuences. I am very grateful to some of the great men of
materials science and biomechanics who sat down and just had a word (or
several) with me. To give their names would be to suggest that they might
support some of what is written here or even know about it at all. They dont
necessarily know in either sense, so the temptation to thank them by name
is resisted. Early inuences are always the strongest. Of these, Bob Martin
was extremely important in directing my thoughts as an undergraduate at
University College London. As a postgraduate, the seminal inuence was
Jeffrey W. Osborn. Although we overlapped at Guys Hospital for only
18 months, his inuence was so powerful that it pervades the book. He
was, and still is, the greatest of teachers and the most powerful thinker that
I have ever met: I am grateful for the chance to thank him here. I should
have learnt from Jeff just how hard it is to write a coherent book in uniform
style, but I didnt. Although his research papers and a (now out of print) coauthored book called Advanced Dental Histology (many editions, published
by Wright, Bristol) have probably been most inuential, a much underrated
(also out of print) book for dental undergraduates that he edited, entitled
Preface
xiii
Dental Anatomy and Embryology (Volume 2 of A Companion to Dental

Studies, edited by A. H. R. Rowe & R. B. Johns, Blackwell, Oxford, 1981)
has provided a magnicent general guide while writing this book. I would
also like to thank Karen Hiiemae for getting me started in research, for
persevering with me and for offering general advice on research directions.
The subtitle of the book is adapted from an article by A. W. Crompton
and K. Hiiemae, entitled How mammalian molar teeth work, published
in Discovery (Yale University) 5: 2334 (1969). Douglas Luke helped me
survive and turned me from a neophyte into, well . . . into something else.
He also kept me in the business by pushing me to apply for a postdoctoral
fellowship, which led to several years of joint publication.
Despite all this, if I had to pinpoint any particular period when I began
to feel comfortable and capable in research, then it would have to be nine
years spent in the Department of Anatomy at the National University of
Singapore. I am eternally grateful to all the members of that department and
to Professor Wong Wai Chow, its then head, for supporting me while I grew
up. I miss that department and its staff very much and would like to take
this opportunity to offer my greetings to them. In particular, I recall with
gratitude, the advice offered in many conversations by Samuel Tay, Gurmit
Singh and K. Rajendran (acknowledged for Fig. 2.17). To the current head,
Professor E. A. Ling, I wish you all the best in your energetic leadership of
the department. Mark Teaford was for a long time a co-author of this book.
His sense of perspective and great friendship has helped sustain me through
many crises of condence. Mark very kindly read through much of the text
and helped correct and clarify it in many areas. Through both Singapore
and Hong Kong, I have collaborated at length with Richard Corlett, who
taught me ecology. Brian Weatherhead brought me to the University of
Hong Kong: thanks a million to you, Brian, and to the Anatomy staff here.
I have had a warm working relationship with Brian Darvell in Dental Materials Science for most of my 10 years here, and also with his technicians,
Paul Lee and Tony Yuen. Together, they designed the HKU Darvell eld
tester, described in Appendix A. A long-term collaboration with Iain Bruce
in Physiology has also been very stimulating and fruitful. In recent years
at the University of Hong Kong, I have been lucky enough to associate
with several PhD students (Choong Mei Fun, Nicola Parillon, Jon Prinz,
Kalpana Agrawal and Nathaniel Dominy) and postdoctoral fellows (David
Hill and Nayuta Yamashita), all of whom have had great inuence on
me. Of these, Jon and Nate have been towering inuences. P. Y. Cheng
has worked with me for 11 years: I am deeply grateful for all his
help. Recently, a medical student with strong palaeontological leanings,
xiv
Preface
Sham Wing Hang, has done a lot to focus my thoughts clearly. I am deeply
grateful. In addition, I have beneted greatly from interaction with Kathryn
Stoner and Pablo Riba during the Pantropical Primate Project. For help
during production of this book, I thank Eastman Ting, for several pieces
of artwork that set the style, and Johnny Leung for photographs. Henrique
Bernardo ([email protected]) did the cover and offered much
helpful advice. Gavin Coates ([email protected]) created the ick
art. In addition, I would like to thank for either direct help or inspiration:
Holger Preuschoft, Charles Peters, Roland Ennos, Walter Greaves, Patricia
(Trish) Freeman, Mikael Fortelius (for lengthy correspondence as well as his
papers, the inuence of both of which permeates the book in many places),
Josena Diaz-Tay, Robin Heath, Jukka Jernvall, Michael LaBarbera, Mark
Spencer, Peter Ungar and Chris Vinyard. I am also extremely grateful to
Dr Rob Hamer (Wageningen Centre of Food Sciences) for inviting me to
a food summit in Wageningen in November 1999, without the inuence
of which this book would not have been nished. I had the great privilege
there of meeting many of the greats of food science, such as Drs Alina
Scszesniak and Malcolm Bourne.
Specic acknowledgements for permissions to use gures go to the Journal of Anatomy (Fig. 3.8), Food Quality and Preference (Figs. 3.4, 3.5, 3.13, 3.14,
4.7, 4.8 and 4.10), Archives of Oral Biology (Fig. 3.12), Cambridge University
Press (Fig. 4.12) and the British Dental Journal (Fig. 5.10).
Finally, to Tracey Sanderson and Cambridge University Press, thank
you very much for sticking with this. To my mother and my extended
family back in England, thank you for all your support. Lastly, and most
importantly, to members of my immediate family, my wife Mariati and my
daughters, Katherine and Diana: you have supported me loyally throughout
my career and I simply do not have the words . . . emotions are better and
mean much more.
Flickart
A ick-page animation starts here, illustrating the evolution of the human

lower molar from the single-cusped tooth of an early synapsid that lived
about 300 million years ago. The sequence is designed to morph between
existing fossils and may, inadvertently, involve variation off the main lineage. It is hardly possible to know for sure, but only basic features are shown
anyway. Some partial restoration of cusp form has been necessary for fossils
with worn teeth. The timescale is kept relatively even, so some changes are
more rapid than others. Only the evolution of the lower molars is gured
because that of the upper molars is more complex. The original names
for their cusps, given in the nineteenth century (Osborn, 1888), make no
evolutionary sense and tend to confuse.
Each diagram shows the jaw bone in grey with tooth crown evolution
displayed above it. The tooth is viewed from the lingual aspect. The mesial
side of the tooth is always to the left. Root development is indicated only
by the bifurcation into two roots a little further on. Note though that the
earliest teeth in the sequence did not have roots.
The trends in the animation can be noted as follows:
(1) A single-cusped reptiliomorph tooth, equivalent to the protoconid cusp
of a mammalian molar, is the sole initial cusp. It always forms rst, even
in living mammals, and is generally very large.
(2) Two separate cusps form on either side of the protoconid. These are
the mesial paraconid and the distal metaconid. Early mammaliaforms
had this three-in-a-line molar form.
(3) A small shelf develops low down on the crown. This is the cingulum.
(4) The three cusps then triangulate by relative movement of the paraconid and metaconid to form a trigonid. This tooth form is called a
tritubercular molar.
(5) The cingulum extends distally to start forming the talonid, a shelf
that starts to develop three cusps. The resulting six-cusped tooth is a
xv
xvi
Flickart
tribosphenic molar from which those of all living therian mammals

have adapted.
(6) On the primate lineage towards humans, the cingulum and
paraconid are lost. The talonid evens up in height to match the remnants of the trigonid. Late on, all the cusps become blunter and reduce
in height.
How to get excited about teeth
introduction
Animals are destructive by nature. They do not build as plants do, by taking simple molecules and making themselves from them. Instead, animals
operate by taking complex ready-made structures and breaking them down
in their guts. Vertebrates are distinctive in that they are prone to attacking
the largest structures. They are, and probably always were, very active organisms, with consequent big energy expenditures, and they evolved teeth
and jaws early on to increase the rate of acquisition of these food items.
Mammals descended from this line of biological warriors, but they evolved
mechanical comminution (chewing) of food particles to precede the chemical comminution in their guts with so as to increase the rate of energy
ow. They needed to turn an energy stream into a river so as to fund their
all-weather activity cycle, attaining the latter at the immense cost of maintaining an internal body temperature well above their surroundings. We
(making the assumption that the reader is human) are super-mammals,
having extended this characteristic mechanical and chemical destruction
beyond the realm of our bodies to our environment. Probably none of
their devastating nature gives animals a good reputation among plants,
which have developed an enormous array of defences to try to stop being
eaten, and were any readers to be organisms other than human, then they
would surely vouch for the exceptionally poor reputation that we currently
have with every other species.
It is my set task in this book to try to dissect out facets of this general
picture, to make them glint in the light and then claim that these features
explain it all. Of course, this is a ridiculous remit, so I will go for a smaller
assignment and just try to make something glint. That something is the
mammalian dentition, set in the mammalian mouth. As with everything
else in biology though, even an apparently manageable and limited undertaking can start to become awesome once you get into it. There is a lot
1
to explain about just about anything and to be credible that explanation

needs to start with the basics. Most scientists seem to have found the same
thing in any eld of investigation and it is as though they jump ship when
they realize it, becoming honorary fellows of the plant kingdom, because
explanations have to be built.
the origins of teeth
Vertebrates have over 500 million years of ancestry, dental tissues being
present in the very rst forms. Until recently, it was thought that these
tissues developed not in the mouth, because the rst vertebrates were jawless and fed on very small organisms, but on the surface of the body as
exoskeletal protection dermal armour as it is called (Smith & Sansom,
2000). The inclusion of dermal armour as teeth denitely takes the very
respectable ancestry of these organs back to the Silurian period (Janvier,
1996), and possibly to late Cambrian times, over 500 million years ago
(Young et al., 1996).1 This view has teeth as the last remnants in a mammal
of a vertebrate exoskeleton (an external rigid coat that not only provides
support but also contributes protection as a form of armour). A mineralized exoskeleton (mineralized tissue being that which contains crystals of
an inorganic compound in vertebrates, always a compound containing
calcium and phosphorus called hydroxyapatite) seems to have preceded
the evolution of the bony endoskeleton that vertebrates now have. While
an exoskeleton has great advantages in a small animal (that is why insects
possess them), providing optimal stiffening and direct protection, a large
animal would have to develop an extremely thick exoskeleton in order for
it not to buckle (Currey, 1967). Such heavy, and probably very insensitive,
animals would clearly be at a competitive disadvantage to those with endoskeletons, which is presumably why the exoskeleton of vertebrates has
been lost.
Enamel, the outermost tooth tissue (that which may contact a readers
ngernails in some circumstances), is unusual in that it is formed from
the outermost layer of the body, the epidermis, in a process involving a
very distinct set of proteins. Indirect evidence from molecular analysis has
recently been interpreted as suggesting that some of these proteins might
even have evolved back even in pre-Cambrian times (Delgado et al., 2001).
Dentine is the foundation of modern vertebrate teeth and derived from
underlying connective tissue. It may have evolved alongside a bone-like
tissue at a later date (Janvier, 1996; Smith & Coates, 2000), but the issue
of what evolved rst will probably only be resolved by examination of the
The origins of teeth
genes that make these proteins (Kawasaki & Weiss, 2003). There is no need
to provide a xed viewpoint on this debate here: the science itself will evolve
rapidly.
Whether teeth evolved rst on the surface of the body or actually in
the mouth, it seems probable that they evolved in jawless forms. These
early forms were very likely to be active feeders that sought large prey
items. We know that conodonts (vertebrates off our direct line of descent
and that lived over 440 million years ago) had muscle tissue somewhat
like that of modern vertebrates (Gabbott et al., 1995), which indicates that
they were highly active organisms. They also had teeth set in a jawless
mouth that show evidence of microscopic wear, indicating they were used
in food acquisition (Purnell, 1995). Microwear, as this microscopic wear is
called, is a clear indicator of the manner in which teeth are used and is
strongly linked to diet, as shown in Chapter 6. The rst jawed vertebrates
were the placoderms. Curiously, the earliest members of this group did not
have teeth. Later ones did, but may have evolved their teeth independently
(Smith & Johanson, 2003).
Most vertebrates have teeth just for ingestion, the term given to the
process by which food is taken into the mouth. The real breakdown of
food is chemical. Although all animals are distinguished by the chemical
breakdown that takes place within their guts, the most distinctive feature of
a mammal is the mechanical breakdown that takes place in its mouth prior
to this chemical activity. Although we tend to think of teeth as a chewing
instrument, it is only mammals that really chew.2 The earliest mammals
added chewing (or mastication) to ingestion. Rather than just break food
down chemically with their gut, they also comminuted food with their
teeth. Why?
It is generally agreed that the need for mastication is related to the rate of
energy requirement by mammals. Among vertebrates, only birds and mammals have exact control over their body temperatures. Other vertebrates
have restricted activity cycles and limited ecological niches. Adaptation to
cooler climes requires the development of a locomotor stamina that coldblooded reptiles do not possess. The elevation of body temperature to a
standard that is usually well above the ambient is very costly and demands a
large increase in basal metabolic rate (which is the rate of energy consumption required to keep the body functioning). There are two basic methods
by which such energy demands could be met in mammals. One involves
simply ingesting much more food and letting a very large gut extract more
nutrients per unit of time. This, however, works against mobility because
the mammal would have a heavy weight of inert material in its gut that
would simply slow it down. The alternative solution, to which mammals

have resorted, is to reduce food particle size mechanically prior to chemical reduction in the gut. This adds new surface area on which enzymes in
the gut can act more quickly, so increasing the rate of chemical breakdown
(digestion). It is mastication, veiled by the lips and concealed by the cheeks,
that provides this key advantage to mammals.
Birds are the other surviving group of vertebrates to possess high
metabolic rates and it will not have escaped the attention of observant
readers that they do not chew. It is unclear why this is and it should not
be assumed that mammals are somehow superior: early birds appear to
have had (simple) teeth but lost them. Weight was probably an important
consideration for longevity of ight. Beaks are much lighter than teeth, but
have disadvantages in being much more limited in their ability to break
down foods and more importantly, are relatively insensitive compared to
a dentition. The sensitivity of teeth is provided by receptors housed both
inside them and in their sockets, which offer ne-scale detection of forces
and also modulate salivary secretions. Both features signify a key asset of a
complex mammalian mouth the ability to discriminate foods by texture
and taste at the front end of the gut so as to assess food quality prior to
decisions about swallowing. Non-mammalian vertebrates have much less
ability to do this because they generally only perforate large food particles
and cannot make any detailed assessment of their contents.
Thus, birds may lack the ability to make many of the ne judgements that
mammals can achieve with the plethora of sensory receptors in and around
their mouths. Instead, they probably have to rely more on regurgitation
or vomiting if they make errors during feeding. These reverse movements
of the gut are part of the normal behaviour patterns of many birds when
feeding their young. They are also seen in owls when they expel bones and
fur through the mouth and in fruit-eating hornbills that regurgitate large
seeds of the nutmeg family (Kemp, 1995).3 The reliance of a mammal on
oral sensitivity is made doubly clear by the rudimentary development of the
sensory capabilities of the rest of the gut. In stark contrast to the mouth, the
abdominal organs send very unclear messages to the brain about problems
with their contents. It is extremely difcult, for example, to get a clear idea
about where visceral pain is coming from and it often gets referred to the
skin. There is no reason to suppose that these human sensory dilemmas
differ from the situation in other mammals. Clearly, all the major feeding
decisions in a mammal are being made up front. This provides them
with a dietary diversity that would make it difcult to interpret signals
from the gut very precisely anyway. For example, a monkey might feed
Are teeth just hangers on?
on several food items within a 2-hour period. How could an individual

animal with a diet of that breadth connect subsequent abdominal pain or
diarrhoea to any one of these particular foods? It is thus possible to argue
that a limitation on sensory discrimination within the mouth forces birds
to adopt a more stereotyped diet than terrestrial mammals. They (and
other non-mammalian vertebrates) probably use longer-distance cues to
recognize food sources than mammals do. There is clearly a linkage here
to the general diurnal (daytime) niche of birds versus the usual nocturnal
niche of mammals. Lacking the visual ability to forage at distance due to
nocturnal activity, early mammals probably had to develop shorter-range
cues that included the ability to sense food texture very accurately in the
mouth. This theme is developed further in Chapter 7 in relation to food
selection in primates including humans.
Considerations like the above show the enormous timeframe over which
teeth have evolved. Their beginning was a very long time ago. Given the
wonders of modern dentistry and medicine, we sometimes forget the evolutionary basis for much of what goes on in the mouth. What is the connection between such problematic, easily infected, structures as our teeth
with a noble lineage of 500 million years? The answer lies in our very recent
history.
are teeth just hangers on?
Excepting humans, there is circumstantial evidence that mammals whose
dental function is sufciently impaired die and that the lifespan of the dentition could be the operative factor limiting the lifespan of many mammals
(Chapter 6). In contrast, humans have in the last 10 000 years tried to control their food supply. No longer content with merely understanding where
plants grew and when their parts were potentially edible or not, nor with
simply tracking animals, our ancestors started to domesticate both plants
and animals to produce reliable proximate food sources. And that is when
problems started. Most of the dental decay (dental caries) and periodontal
(gum) disease that we see started with the high consumption of cultivated
(seed) grain products just less than 10 000 years ago. The nutritional content of these food items comes from contained starches that are usually
bound in seeds into relatively insoluble, and thus indigestible, granules of
micrometre size. Allied to cooking, a procedure recently supposed to dene
humanity better than tool use (Wrangham et al., 1999), the starch is solubilized and its mechanical properties radically altered, such that it breaks
up into very small particles (probably fragmented granules), which do not
get cleared from the mouth very easily. An enzyme in the saliva starts to
break the starch down into sugars and bacteria living in the mouth then
convert those sugars into acid. Unfortunately, once acidity drops to below
a certain level, tooth tissues start to dissolve and decay sets in. The major
problem with the modern (Western) diet is not with these starches, but
with sucrose. This forms an exceptionally adhesive layer to the teeth, more
so than other common sugars. It may be thought that this effect should
have some natural defence, but sucrose is not a common sugar in plants.
It is the transport medium for energy but is rarely concentrated, being
rapidly converted to other sugars within plant cells e.g. for building up
the cell wall. Many fruits do not contain sucrose and, indeed, a large group
of fruit-eating birds lacks the enzyme sucrase needed to break this sugar
down, developing diarrhoea if they consume it (Martinez del Rio, 1990).
Our mammalian ancestors could process sucrose, but the effects of high
concentrations of this sugar on the teeth in recent human evolution have
been profound.
So teeth are not intrinsically dirty structures in an unclean cavity. Far
from being unclean, the mouth of mammals is a very efcient self-cleaning,
self-clearing, system for natural diets. It has to be because food is certainly
not sterile and many microorganisms nd the mouth an acceptable environment in which to survive, if not thrive. The epithelial lining of the
body provides protection, of course, but only so long as there is no break
in it. The teeth, however, provide that break. Besides the risk of infection
around their roots, the vulnerability of tooth material itself is great. The
greatest danger facing dental tissues is acid. Any drop in the oral pH below
about 5.5 (which is really only very mildly acidic) and tooth tissues start
to dissolve. What prevents this? The major factor is oral uid. Saliva jets
out into the mouth from four major orices, and many smaller ones, providing a mildly alkaline bicarbonate spray to prevent dissolution (Edgar &
OMullane, 1996). For further detail, see Chapter 3.
I believe that that the presence of teeth in every major vertebrate lineage
except birds makes it obvious that there was very high selective pressure
for the development of sturdy teeth in strong jaws. This pressure is also
responsible for the diversity of tooth form in mammals in relation to diet.
The strongest evidence for this comes from the frequency of convergent (independent) evolution of dental features in mammals. The most striking of
these is the suggestion that the basal form of the mammalian cheek tooth,
the tribosphenic molar (the evolution of which can be icked between
pages 1 and 159), may have evolved twice, quite independently, about 200
150 million years ago (Luo et al., 2001).4 More recently, a prominent feature
Basic functional considerations
of upper molars called the hypocone, not present in the tribosphenic

form, has arisen independently (convergently) in many mammalian lineages (Hunter & Jernvall, 1995). In fact, the adaptability of hard tissues like
the teeth has recently been termed too great to rely on for characters by
which to judge the evolutionary separation of mammalian groups: some
aspects of soft-tissue arrangements may be more conservative (Gibbs et al.,
2000). This is a complete reversal of the predominant view just 20 or so
years ago. This pliability of the teeth is despite these structures being the
complex product of the action of more than 200 genes (Jernvall & Thesleff,
2000), and the fact that a dental tissue like enamel (Chapter 2) possesses a
sophistication far beyond that which can be produced articially.
basic functional considerations
The teeth are set as close as possible to the sense organs (the ears, eyes, nose
and tongue) so that they can aimed accurately, and set close to their central
processing unit (brain) so that signals can be sent to and fro with minimal
delay.5 All this apparatus is housed in a light appendage (the head), set on
an agile muscular stalk (the neck). The following is an attempt to sketch
the arrangement of the functions of the mammalian head and neck and the
compartments that they occupy (Fig. 1.1), emphasizing just how distinct
mammals are from their reptiliomorph ancestors and other vertebrates.
Chewing and swallowing
The major function of the head is to ingest (take in) food. Much of the
anterior (front) part of the head contains the jaws and teeth that ingest
food particles. Mammals chew (masticate) food whereas virtually all other
vertebrates do not. The evolution of mastication has required many evolutionary modications to the head from the reptilian state. Moving the
jaws so as to break food particles requires large muscles. These masticatory muscles, the temporalis, masseter and pterygoids (Chapter 2), have
differentiated in mammals so as to bite food without sending this force
pointlessly through their jaw joint. This keeps head weight down because
the food should absorb a considerable amount of the energy imparted by
muscular work, the optimum being to channel all this work eventually into
its fracture.
In most vertebrates, ingestion is followed immediately by swallowing,
with a very sensitive muscular bag, the tongue, organized to direct food
particles backwards to the next part of the gut. In mammals, the tongue
Fig. 1.1 The main diagram shows functional compartments of the anterior part of the head
and neck in the human. The diagram above left shows how small the oral cavity really is,
being the unshaded space lying between the tongue and hard palate. The rigid hard palate
extends backwards into the pharynx as the soft palate, which is a mobile ap.
has developed a capacity to throw food sideways onto the teeth for chewing
as well as backwards for swallowing. Assessment of whether to swallow or
to chew further requires the tongue to have a surface positioned directly
above it against which it can manipulate the food. The palate, another
mammalian innovation, acts as this template.
Throwing food sideways requires a wall on the other side of the dentition
to prevent food from escaping. Muscular cheeks provide this. When food
particles have been chewed sufciently, they tend to aggregate into a sticky
mass called a bolus. The tongue can then propel this bolus backwards for
swallowing. The tongue, palate and pharynx (the muscular tube directly
behind the mouth), all act together during swallowing. The pharynx is
Basic functional considerations
uniquely mammalian possessing a musculature that appears to have no

equivalent in other vertebrates (Smith, 1992).
Sensing food
The senses are mounted largely as paired organs on the head and are generally separated widely so as to allow the location of food to be xed. The
ears sit on the sides of the head while the eyes sit in the orbits just above
the jaws. Both are admirably equipped to localize food direction. However,
the nostrils (not the eyes) are too close together. Thus, the nose probably
has little directional ability to recognize smells, which have to be determined by the animal following the scent of a particular
chemical until it reaches the source (or the source reaches it). Airow is
needed for smell, so this sense is built into the air intake for the lungs. The
nostrils lead to large nasal cavities lying above the mouth and separated
from it by the hard palate. However, smell is not just important for leading
mammals to food. When food is being chewed, the soft palate does not
seal off the foodway from the airway completely. It is probable that food
vapours pass from the mouth back around the soft palate, and up into
the nasal cavity (Fig. 1.1). In the part of the pharynx behind the mouth,
the so-called oropharynx, the airway and foodway are a single structure to
allow this. Lower down, the airway and foodway separate with the larynx
and trachea lying in front of the oesophagus.
The last sense is that of taste, which is largely embedded in the surface
of the tongue.6 Potentially, anything in solution in the mouth can be tasted, but it appears that many vital substances are not, key among these
being proteins, calcium and phosphate ions. Recent developments in taste
research make pronouncements uncertain though, and I postpone more
consideration until Chapter 3. Taste and salivation combine as detection
mechanisms.
Much of the head is dedicated to housing and protecting the central
processing unit of this sensory information, the brain. This needs to be set
close to the jaws because timing losses due to the speed of nervous impulses
are then minimized.
Communication
The crossing of the foodway and airway not only allows the ability to smell
food in the mouth, but also the possibility of vocalizing through it. Most
mammals make noises. These noises can seem rather limited to us because
10
these animals seem to communicate better by smell and body language

(perceived by sight). Wilson (1975) sets communication via sound on the
highest grade. In order to speak, humans use the larynx to break up expired
air into bursts by opening and closing the airway. Our larynx lies very low
so that these bursts of air pass through the oral apparatus, with the tongue
moving against the palate and teeth, not to chew, but to produce speech.
The consequence is a potential for choking. However, unless it can be
demonstrated that this need to communicate has inuenced the design of
the oral cavity, this is all really beyond the scope of this book. There are other
candidates for evolutionary modication to the chewing apparatus than
sound. Higher primates (monkeys and apes, including ourselves) can make
faces, adding these expressions to behavioural gestures and vocalizations.
These movements are mostly a function of our facial muscles. However,
some of these have important actions during food ingestion and chewing,
so there is the potential for conict and compromise of function here too.
the basic model
This book considers the function of teeth in relation to the ingestion,
chewing and swallowing of food. It attempts to elucidate the principles that
underlie the evolution of tooth shape and size and those mechanisms by
which the dentition can be maintained. To do this, I need to understand the
properties of foods that drive evolutionary change. The following model has
its antecedents in food science, most particularly in a paper by Szczesniak
(1963), but it differs slightly in emphasizing a food surface versus nonsurface dichotomy and by viewing the action of the teeth as to change the
boundary between the two by the process of fracture.
When a mammal chews food, it usually breaks particles into fragments,
producing new surface area by cracking. This extra area increases the rate
at which digestive enzymes in the gut act. This is the only general explanation advanced for the evolution of mastication in mammals (though more
specic explanations will be offered in Chapter 3). I assume in this book
that the rate of food breakdown is optimized to an animals needs and,
therefore, that features of its anatomy and physiology can be interpreted in
this light.
The effectiveness of the forces that are produced when teeth contact
food depends on the characteristics of the food surface that they act on.
The form and extent of the external food surface is referred to as the external
physical attributes of the food. These characteristics include food particle
size and shape, the total volume of particles in the mouth and attributes
The basic model
11
of surface quality such as surface roughness and stickiness. These surface

attributes can be contrasted with the resistance of food to the formation of
new surface, this being the object of the chewing process. This resistance
is embodied in the internal mechanical properties of the food, including the
Youngs modulus, yield strength and toughness.
The rate at which food particles break down can be subdivided into
many anatomical and physiological variables. However, at a conceptual
level, it can be subdivided simply into the product of two discrete events:
the chance of hitting a food particle on the one hand and the fragmentation
that is produced when any particle is hit on the other. From the food side
of the problem, the chance of hitting a food particle depends only on its
external physical attributes. For example, small particles are less likely to be
hit than large ones. A smooth particle may slip from under the load whereas
a rough one might be trapped and so on. I will deal with these attributes
in Chapter 3. Once the particle is hit, then its resistance to fragmenting
is dened by its internal mechanical characteristics. These characteristics
include major mechanical properties such as toughness, Youngs modulus
and strength (dened in Appendix A). From the mammals side, the chance
Fig. 1.2 A binary model of food properties developed in this book stressing the difference
between surface attributes (right column) and those properties that act variably to prevent
the surface from being extended by food breakdown (left column).
12
of hitting a food particle with the teeth is enhanced by making the tooth
bigger i.e. by changing tooth size. In contrast, the effect of the force that
the tooth exerts on that particle depends on the contours of its working
surface i.e. on its tooth shape.
Figure 1.2 shows how this binary model is developing into two separate
streams of reasoning. A fundamental aspect of this book will be pushing
this model to the limit, progressively examining this classication of food
properties so as to understand how teeth work. To do this, I rely a lot
on optimization theory, even when the relevant numerate models are not
yet constructed. For those averse to this method of investigation, I refer
them to a book on this subject by R. M. Alexander (1996) and his recent
short defence of it (Alexander, 2001). Achieving my objective requires some
descriptive detail and literature citation. Vital though it is, citation interferes
with textual rhythm and seems somehow to make it all rather sombre.
Heraclitus said that everything ows (Barnes, 1987), but perhaps he might
have exempted much of this book.
The basic structure of the mammalian mouth
overview
Before processes of ingestion, mastication and swallowing can be considered, the structure of teeth and their general arrangement in the face have
to be sketched. Function creeps in, but only so that the sense in certain
structural arrangements is claried.
what are teeth?
The teeth of a mammal are stones, anchored in tight-tting holes in the
bones of the upper and lower jaws and projecting through the lining tissue
of the jaw into the mouth (Fig. 2.1). The part of the tooth that projects into
the mouth is called the tooth crown while that part set into the jaw is called
its root. The working surface of the crown is described by its most prominent features: pointed elevations are called cusps if they are large, tubercles
if they are small. Roughly circular depressions are called fossae (singular,
fossa). Raised folds can be called ridges or crests (sometimes written as
cristae; singular, crista), but I will sometimes refer to these features later as
blades. The creases that run between the bases of cusps are called ssures.
Figure 2.2 shows some of these features diagrammatically on a typical mammalian molar and gives compass bearings needed to identify them unequivocally in a skull. Many features, such as ssures (one is indicated on
Fig. 2.1), are depicted best in photographs and reect the fact that the cusps
of most molars rise like mountains alongside V-shaped valleys. Fissures are
like the rivers running at the base of those valleys.
A typical mammalian tooth contains three mineralized tissue layers that
are rmly bonded to each other in spite of the fact that few structures
traverse their junctions. The outermost layer of the tooth crown is the
enamel (Fig. 2.1). This tissue is developed from the epithelium of the mouth
and is very heavily mineralized. It is backed by dentine, a tissue that forms
13
14
2 The structure of the mammalian mouth
Fig. 2.1 The basic form of a mammalian tooth, illustrated by a human molar. The crown is
covered by enamel and usually projects entirely into the oral cavity, while the enamel-less
roots are buried in a socket in the jaw bone. The junction between the two is called the
cervical margin (the neck) of the tooth. The gingiva (gum) forms a roof for the periodontal
ligament, the soft tissue of the tooth socket. The ligament (much narrower than shown)
is crossed by collagen bres that are anchored in the cement of the tooth and in the bone
of the socket wall. The tooth is kept alive by blood vessels that enter through holes in the
roots and which feed the cells that form the dentine (odontoblasts cell bodies shown as
circles). Nerves that supply the periodontal ligament also pass into the tooth.
the foundation of all teeth, and which derives from underlying mesodermal
(or, more strictly, ectomesodermal) tissue. Inside the dentine is a pulp cavity
containing living tissue, which can repair dentine as a response to tooth
wear. A thin tissue called cement (or cementum) lines the outer surface
of the dentine in the tooth root. The cement and dentine, together with
the soft tissue of the tooth socket, the periodontal ligament, and also the
Caption for Fig. 2.2 (cont.) The equivalent posterior extension in the upper molars, the
talon, was a late evolutionary development (evolving independently in many mammalian
lineages) and usually has one major cusp. Additional smaller cusps can be present and any
or all of these can be connected by ridges/crests. The named basins (fossae) lie below cusps:
the protocone is aligned over the talonid, while the hypoconid is lined up under the trigon.
There is a lot of variability on this basic form (for example, the ickart on pages 1159
shows how the paraconid was lost in the evolution of higher primates). Everything possible
on a tooth crown is named and the wealth of terminology is truly awesome. In line with
the advice of Butler (1978), I have strived to keep things simple.
What are teeth?
(A)
15
Anterior
Posterior
BUCCAL (B)
MESIAL
(M)
Medial
LINGUAL (L) LABIAL
(Lab)
DISTAL (D)
BUCCAL (B)
Lateral
(B)
M
B
protoconid
metacone
paracone
hypoconid
hypoconulid
entoconid
protocone
hypocone
paraconid
UPPER
trigon
metaconid
LOWER
trigonid
talon
talonid
talonid
basin
trigon
basin
L
Fig. 2.2 (a) Anatomists use a straight compass to describe directions, but the mouth has
its own bent version. Anatomical directions are shown above, where anterior (to the front)
is the opposite of posterior, and medial (towards to the midline) is contrasted with lateral
(away from it). In the mouth, mesial is the opposite of distal, while buccal is opposed to
lingual. Buccal (meaning towards the cheek) is replaced by labial (towards to the lips) for
the anterior teeth. (b) The basic mammalian molar form. Cusps are indicated by circles,
while ridges or crests are marked by lines. Both upper and lower molars have a triangular
region. This is called the trigon in the upper molar (the sufx -id being added for lowers)
and is bounded by three major cusps. The lower molars have a platform (posterior to the
trigonid in all living forms) called the talonid, which is bounded typically by three cusps.
(cont. on previous page)
16
surrounding jaw bone all seem to have a common developmental origin

and possess structural similarities, being built on a collagenous framework.
Cement forms an anchorage for the thick bundles of collagen bres that
pass through the living tissue of the periodontal ligament to the densely
mineralized bone of the socket wall. However, it can extend over the crown
of a tooth as well as the root. The pulp communicates with the periodontal
ligament via a hole or holes in the tip (apex) of the tooth root. The teeth
at the back of the mouth tend to have more than one root (a two-rooted
tooth is shown in Fig. 2.1) for greater stability under load. The pulp and
periodontal ligament are well supplied with blood and there are also nerves
that supply the walls of the blood vessels and conduct impulses from sensory receptors. The collagen bres of the periodontal ligament have various
orientations and are separated by cells of both epithelial and ectomesenchymal origin. There are also bres in the elastin family called oxytalan bres
and much extracellular gel-like material. The ligament cushions the tooth
and controls small movements within the socket, as will be discussed a little
later.
types of teeth and their denitions
There are three classes of teeth in a mammal incisors, canines and postcanines. The incisors and canines will often be referred to as the anterior
teeth, while those behind the canine will be called posterior teeth. All develop inside the jaws and erupt through the jaw sequentially so as to appear
in the mouth. Each tooth class develops in its own series over two distinct
tooth generations. Figure 2.3 shows a sketch of the permanent dentition of
the lower jaw of two Old World higher primates: a male adult long-tailed
macaque (the monkey of much medical literature) and a male gibbon (an
ape). The normal dentition has bilateral symmetry, and the lower teeth are
generally aligned fairly precisely against the upper teeth, which are embedded in the upper jaw.
Set at the front of the jaws, the incisors generally have a simple shape. In
most mammals, these are short pointed teeth. However, in higher primates
(New World monkeys, Old World Monkeys, apes and humans), incisors
have a bladed working surface. In rodents, lagomorphs (rabbits, hares, etc.)
and one primate (the aye-aye), the central incisors are wide bladed teeth
that grow throughout life. The canine teeth, when they are present, are
set behind the incisors. Canines nearly always have pointed crowns and
generally project beyond the rest of the dentition. Behind these are the
postcanines, which are responsible for mastication. There are two names
Types of teeth and their denitions
17
head of
condylar
process
molars
premolars
canines
incisors
GIBBON
MACAQUE
Fig. 2.3 The lower jaws of two male primates, an ape (the Bornean gibbon, Hylobates
muelleri) and an Old World monkey (the long-tailed macaque, Macaca fascicularis), to
illustrate basic features of the arrangement of the jaw and teeth of mammals. The broad
lower incisors are called spatulate in the literature.
generally given to postcanine teeth: molars and premolars. Molars develop

in the rst generation, but the most posterior members of the series are
retained throughout life. However, most of the more anterior molars are
replaced by premolars. These are the descendants of molars: by denition,
they are postcanines of the second generation. Deciduous molars are those
that are shed (fall out) so as to provide space for their descendent premolars,
whereas permanent molars are not succeeded and remain in function until
death.
It is thought probable that early mammals had at least 1213 permanent
teeth in each half of each upper and lower jaw: up to ve incisors, a canine,
four premolars and three molars. In some mammalian lineages, there has
been a marked tendency to reduce this number (e.g. in primates: Martin,
1990), while in others like the cetaceans (whales/dolphins), there has been
an increase. Either an increase or a decrease creates a problem: in order to
describe teeth, we need to identify them.
When a common name is given to structures in different mammalian
species, we really need to establish their genetic identity. Structures in
different species of organisms formed by the same set of genes are called
homologous structures. The problem arises because, for most mammalian
structures beyond the simplest, the developmental mechanisms that create
18
these structures have not been tracked back to the complicated interaction
of gene products on which their existence depends. Far too many genes act
together to produce the dentition to be clear about what creates the identity
of a tooth class. There are clues (Yucker et al., 1998), but no certainty. So
until the homology of structures is known for certain, arbitrary indicators
must be used.
The homologies of the teeth of mammals are currently determined in a
very arbitrary way, being identied by their position with respect to a bony
suture lying between two bones of the upper jaw, the premaxillary bone
in front and the maxilla behind it. The rst tooth behind this suture is
named the upper canine. There is, for some obscure reason, only allowed
to be one such canine on each side of the upper jaw. Incisors lie in front of
this canine, while postcanines lie behind it. The incisors and postcanines
(deciduous molars, premolars and molars) are each numbered from mesial
to distal. However, this procedure only denes upper teeth. Names for the
lowers are organized according to a single criterion: the tooth tting just
in front of the upper canine is the lower canine. Everything follows from
this.1
Simple though this nomenclature appears to be, it is just deeply entrenched convention. There have been surprisingly few attacks on it, partly
because of the tenacity of the conservative response. Yet a logical argument has been advanced which casts great doubt on the relationship of
this convention to actual homology (Osborn, 1978). Osborns theory, the
developmental details of which are not of concern here, suggests that tooth
classes can be recognized by a commonality of shape. Within any toothclass series, there will be a clear gradient of size and shape, the form of any
individual tooth depending on its position within one of these series. The
boundaries between tooth classes can then be recognized by non-sequiturs
of form, i.e. by recognizing abrupt breaks in shape and size. Osborn further
suggested that the numbering of teeth should follow the order in which
teeth develop rather than their position in the jaw. This is based on the
assumption that teeth tend to be lost at the ends of series (where the end of
any series can be marked by identifying the last tooth to develop in either
the rst or second generation). Thus, the rst teeth to develop in a particular tooth class in any mammal would be homologous across all mammals
with the others numbered in order of their subsequent appearance. Using these simple rules, Osborn showed that the application of this system
sometimes leads to very different tooth identities to those of conventional
wisdom. For example, his system has no need to postulate that there is only
one canine on one side of an upper or lower jaw and there is no reason to
Alignment of teeth
19
identify canines by their position in that jaw, either in relation to a bony

suture or to a tooth lying above and behind it. Just a little thought by
anyone who has not been taught the prevailing dogma is perhaps enough
to see that the evolution of the dentition is unlikely to have been constrained by a xed relationship to bone. Without dwelling on this further
(because it will only be returned to briey in Chapter 7), it seems quite
likely that this novel system of homology will be shown to be correct. The
ramications could be considerable. The commonest reason for ignoring
it that it would destabilize long-held nomenclature for the dentition
should not be a barrier to its investigation. Even Osborn (1978) felt that the
two issues, taxonomic nomenclature and biological identity, should best be
separated.
The development of the dentition is the subject of a lot of current
research, but it is far from generating clear answers. Among the general
facts that developmental biologists have to contend with is that teeth can
develop perfectly normally outside the mouth, e.g. the ovarian cyst pictured
in Sharpe (2000) that looks to possess most of the tooth types in the lower
jaw. The crowns of these teeth look completely normal but are located in
an utterly bizarre setting. (Such a structure has, on occasion, been reported
to have broken through into the vagina of a patient.) Yet location must be
important in dental development or else upper and lower teeth could not
develop in approximate alignment, a matter for the next section.
alignment of teeth
The upper and lower teeth of most vertebrates other than mammals never
contact and wear only because they contact food. When the jaw is closed,
the simple conical crowns of the upper and lower teeth of a typical reptile, like a lizard, interdigitate with each other, but are spaced such that
there cannot be contact. Mostly, this is achieved by having the lowers
lie inside the uppers forming a zigzag arrangement (Fig. 2.4), although in
some living crocodilians, the upper and lower teeth lie along a line (Brochu,
1999).
In contrast, the need of mammals to fragment food particles requires
complex postcanine teeth that have to touch when the jaw is closed. There
is obviously great selective pressure against adding tooth-on-tooth contacts
to those produced by food, because this is potentially very dangerous.
The principal mode of wear of teeth is abrasive wear, in which hardness
is a major factor (Chapter 6). Teeth are generally the hardest items in the
mouth and are therefore likely to do much more rapid damage to each other
20
Fig. 2.4 The arrangement of the upper (unlled) and lower (black) single-cusped teeth of a
reptile. (a) A view from below, the outline being that of the maxilla. The lower dental arch
ts inside the upper and the teeth are spaced so as to avoid toothtooth contact and wear.
(b) A side view with interdigitating cusps. In some living crocodilians, the upper and lower
teeth lie in line rather than one inside the other.
than food is capable of. Forceful contacts are made during chewing (just
after breaking food particles) and swallowing. Wear potential from these
contacts is minimized by various anatomical and physiological variables.
The rst is precise alignment of upper and lower teeth (called occlusion
by dentists). Opposing teeth have surfaces that match very well and the
contacts that are made as they glide together are used as a physiological
guide to bring the teeth together in a standardized fashion. The actual
precision of this alignment depends on the type of tooth. There are fewer
misalignments (called malocclusions by dentists) in mammalian species
where the postcanines have tall cusps with more prominent ridges than in
those with low cusps where malocclusions are common (Mills, 1955).
The alignment of teeth cannot just depend on the position of a tooth
after it erupts because this is unlikely to be accurate enough. It depends
also on the ability of mammals to move teeth very subtly in the jaw so
as to adjust their position. This movement is possible because the bone
surrounding the periodontal ligament can remodel under stress.
Individual tooth tissues
21
Spreading the force of tooth contacts across the dentition by precisely

aligning contacts is not necessarily sufcient to avert wear at the end of
chewing strokes. It is also necessary to decrease the time of contacts. This is
achieved by switching off the activity of the key jaw closing muscles about
12 ms after toothtooth contact (Gibbs et al., 1981).
Just as in other vertebrates, there is a tendency for the mammalian lower
dentition to lie medial to the upper dentition. It follows that, in order for the
teeth to be aligned for conjoint action, the lower jaw must be moved laterally
at some point. In most mammals, this is the only other movement besides
simple open and close, but some mammals, like rodents and elephants, have
an anteroposterior component to jaw movement during chewing (Maglio,
1973; Hiiemae, 1978).
individual tooth tissues
Each hard tissue is a composite, i.e. an organized mixture, in which minute
ceramic particles (inorganic crystals) are set in organic glue containing
plenty of water. The protein in the glue acts like a scaffolding while the
tissue is being built. This protein may be brous, as in dentine, or mostly
globular as in enamel (Fincham et al., 2000). Normally, this scaffolding
remains after formation, but to get the very high mineral content of enamel,
it is largely removed during a process called maturation. The orientation
of the protein framework (called a protein matrix by biologists) seems to
determine how the mineral is laid down.
The inorganic crystals in each tissue are made of hydroxyapatite, a calcium compound with the formula Ca10 (PO4 )6 (OH)2 . In large blocks, this
stony material is stiff and hard, but it lacks fracture resistance in tension
or bending due to its low toughness. However, as outlined in detail in
Chapter 5, the smaller a solid object is, the stronger it usually is. The
crystals in hard tissues are nanometre-sized and are actually so strong that,
when these tissues crack, the cracks run between crystals in the surrounding
protein matrix and not through them.
The protein matrix surrounds the crystals and appears to contain narrow channels through which water can ow. When dental tissues are compressed, they reduce their volume slightly, densifying because their Poissons
ratios are less than 0.5 (see Appendix A). Assuming these tissues are saturated with water (and the tooth crown is always coated with saliva), this
loading must push water out of the tissue because water, like other uids,
has a low compressibility. The results of Haines et al. (1963) and Fox (1980)
are both consistent with water owing through pores in dental tissue. In
22
part, this ow must be through enamel, but Paphangkorakit & Osborn

(2000) have actually demonstrated this uid ow for dentine under load.
The two-phase liquidsolid structural organization of dental hard tissues gives them very different properties to a slab of plain ceramic, as
in an unglazed teacup, a slab of cement or a block of limestone, which
they might supercially appear to resemble. The mechanical advantages
of their intricate make-up are clear-cut: at the cost of a lower stiffness
and the loss of some compressive strength, dental tissues have much greater
toughness than plain ceramics. As Gordon (1991) pointed out, the greatest sin in engineering materials is a lack of toughness. Selective pressures
on structures like dental tissues prove the point: teeth do not break down
themselves, even after breaking down food particles tens of thousands of
times.
Enamel
Enamel forms a hypermineralized cap for the tooth, never more than a
few millimetres in thickness, that coats at least part of the dentine so as
to provide a hard working surface to a tooth crown.2 It is the most highly
mineralized of all vertebrate tissues, being between 85% and 92% mineral by
volume in the permanent teeth of humans. Its other constituents are mostly
water, with some organic matter that is largely protein. The mineral crystals
of enamel have the same constitution as in bone or dentine but grow much
bigger during the nal stage of enamel formation, called enamel maturation.
At nal size, these crystals, which resemble long bars with a hexagonal crosssection, measure about 3040 nm across. They are much longer than these
cross-sectional values, but the actual length seems contentious.
The rst-formed enamel, right at the enameldentine junction, has not
been much studied, but appears structureless under the light microscope
and its crystals appear to have a random orientation. The rest of the enamel
is highly organized with a very denite pattern of crystal orientation reecting that of the pre-existing protein matrix (Osborn, 1981). Neighbouring
crystals are usually roughly parallel, but there are tiny areas where crystals are strongly angled to each other. The crystals pack less well in these
areas, resulting in a 0.1 m wide strip of tissue that is relatively decient
in mineral. This crystal-free region is circular in form (Fig. 2.5). In some
species the circle is complete, whereas in others, such as in human enamel,
it is horseshoe-shaped. Complete or not, this crystal-free region bounds a
roughly circular patch of enamel, usually around 5 m in diameter, called
an enamel rod or enamel prism (Fig. 2.5). These rods are visible under
23
Fig. 2.5 Scanning electron micrograph of acid-etched enamel showing the form of enamel
rods in human tissue. The multi-crystal rods are viewed end on, each being dened by a
semicircular hood where crystals are decient. This hood, sometimes called a rod sheath,
lies at the boundary between the territories of secreting cells. At these points, crystals lie at
large angles to each other. This is no developmental blunder, but central to the mechanical
properties of enamel. Rods have a variety of shapes in different mammalian species. In
human tissue, the rod is bounded on three sides by this crystal-free sheath but, at the lower
border of the rod, there is no boundary because cellular territory is continuous and crystal
disjunction does not result. Scale bar, 5 m.
a light microscope (Fig. 2.6) and extend through the enamel from close
to the dentine border nearly to the enamel surface. Very often, the rods
curve.
Human enamel has been particularly well studied. The path of each
rod is often slightly out of phase with that of its neighbours above (i.e.
towards the cusps) and below (towards the root). This phase change is
regular in the inner enamel and sufcient in magnitude to make rods
that are 10 units apart appear to be running at a large angle to each other
(Osborn, 1974). These periodic shifts, called decussation, give rise to optical
effects called HunterSchreger bands. Many mammals have these bands
(Fortelius, 1985; Boyde & Fortelius, 1986), which probably help to prevent
a straight crack from cleaving the entire thickness of enamel (Rensberger,
2000). On exposure at the working face of the tooth, they also roughen
its surface, so helping grip. At the cusp tips, rods twirl around each other
at very tight angles and the crack-stopping effect of this arrangement has
been demonstrated experimentally in the cusps of human teeth (Popovics
et al., 2002).
24
Fig. 2.6 A low-power light micrograph of inner human enamel. The rectangular box on
the right indicates where the section belongs within a tooth. In the centre of the picture,
enamel rods can be seen end-on with their characteristic semicircular hoods. To either
side of this band, rods are seen in side view. This banded appearance is due to the wavy
arrangement of enamel rods, with the path of each rod being slightly out of phase with
that of its neighbours. This phase change is sufcient to produce the changes in orientation
seen in the picture. Viewed from a distance, rod directions appear to change abruptly, but
closer scrutiny (e.g. to the upper right side of the central band) shows that the change is
gradual. These side view/end-on/side view bands have been known for about 200 years and
are called HunterSchreger bands. The dark colour of the central band is an optical effect
and has no signicance. Ground longitudinal section stained with Bohmers Haematoxylin.
Scale bar, about 50 m.
Enamel formation by ameloblasts is continuous, but not completely

regular. The quantity of matrix being secreted varies on a daily basis. Daily
increments can be visualized as cross-striations (Fig. 2.7a), the measurement
of which in apes and humans shows that 2.57 m of this tissue forms each
day (Dean, 2000). Somewhat staggeringly, there is also a periodic change
in formation on a 410-day schedule (Dean, 2000) when, for no known
reason, it appears that the entire population of ameloblasts suddenly shifts
slightly in the tranverse plane (Osborn, 1971a). This shift leads to lines called
striae of Retzius (Fig. 2.7b). Integrated measurement of these incremental
lines gives accurate information about the timing of tooth formation in
living and fossil mammals (Dean, 2000).
25
Fig. 2.7 Optical effects called incremental lines that form in human enamel on a regular
basis. In both (a) and (b), the white arrows are directed at the striae of Retzius, a fancy
name for the results of an event without parallel in any other mammalian tissue. At an
interval that hovers around a mean of 1 week (depending on the species), ameloblasts
suddenly move transversely, causing a wriggle in the subsequently formed enamel rod,
which when viewed at sufciently low magnication (as in (b)), looks like a dark line.
(a) also shows rod direction (parallel to the thick black lines) and cross-striations (perpendicular to rod direction, indicated by the thin black lines). The latter represent daily
increments of enamel formation.
26
Unlike the mineral, the protein matrix of enamel has been difcult
to study because the soft tissue tends to collapse after demineralization
because there is so little of it. Some proteins have now been identied, but
detailed understanding is still not there. The amelogenins, of which there
are at least seven, initially make up about 90% of the protein and are most
important. They are hydrophobic and probably control crystal size and
orientation by laying down the scaffolding on which crystals develop. This
scaffolding appears to consist, rather bizarrely, of a mass of macromolecular
spheres, each about 20 nm in diameter (Diekwisch, 1998; Fincham et al.,
2000). Some of the remaining 10% of proteins, called enamelins, are acidic,
proline-rich, and may nucleate crystals. They are concentrated in the region
around the enameldentine junction creating a mineralizing front. There
are also tuftelins (Deutsch et al., 1995), so called because they were identied
in hypomineralized areas called enamel tufts. Amelogenins are removed
during maturation by enzymes released from ameloblasts, which then suck
out vast amounts of protein (degraded with proteases) and water. The other
proteins remain in some form (Robinson et al., 1997), particularly close
to the enameldentine junction. However, unlike mesodermal (or, more
accurately, ectomesodermal) mineralized tissues such as dentine, cement
or bone, these proteins do not appear to be brous. A better description
might be brillar and perhaps, in combination with the water, they form
a thick gel.3
The maturation process must clearly rely on ow channels in the matrix.
The crystals of enamel, which are initially small, grow to ll much of the
space left by the withdrawn matrix. However, channels remain in mature
enamel and appear to have a bimodal size distribution, being very narrow
except in the crystal-free edges of rods (Zahradnik & Moreno, 1975; Shellis
& Dibden, 2000). Fox (1980) hypothesized that the narrowest of these
channels are extremely important in absorbing energy under load. He supposed that uid is squeezed slowly out of the gel-like matrix when the tooth
is loaded, but it returns afterwards because this matrix attracts water, being
hydrophilic. In itself, this model does not provide much resistance to ow,
but when channels are as narrow as in enamel (in the low nanometre range),
the electrical layers on the walls of the channels interact signicantly with
ions in the water to raise overall resistance to their ow. Fox predicted, and
found, that soaking enamel in solutions of greater ionic strength (such as
stannous uoride) would increase resistance to ow, effectively toughening
the enamel because the energy absorption in viscous ow cannot be fed
into fracture in the way that elastic energy absorption can. In some ways
then, enamel may resemble the stiffest of sponges.4
27
Dentine and pulp

Dentine actually forms most of the tooth. It is rather like bone and cement, with about 48% of its volume being mineral and an organic matrix
based on collagen bres. The cells that form it are called odontoblasts.
Unlike in bone, where the formative cells become embedded in the tissue, the odontoblasts retreat in a row away from the tissue as they develop
it towards the pulp. They take their cell bodies (their nuclei and most
cellular components) with them but lay out long cellular processes behind them as they go. These processes extend initially through the entire
thickness of the dentine, remaining in unmineralized spaces called dentinal tubules (Fig. 2.8). The cell bodies of the odontoblasts end up lining
the pulp cavity. The very rst formed dentine, in humans about 20 m
or so, is called mantle dentine and is less mineralized than that which
follows.
The hydroxyapatite crystals are plate-like in dentine, being very long as
in enamel, but only 23 nm in thickness. The organic matrix of dentine
is based on collagen bres, whose orientation seems to determine that of
the crystals that initiate within and around them. Scattered throughout
the dentine, there are spherical regions of crystals called calcospherites.
Calcospherites are obvious structures when dentine is forming, but they
generally fuse together as mineralization proceeds. However, their margins
can be recognized in several sites after formation where mineralization
remains incomplete (Fig. 2.8a).
The pulp cavity is the living centre of the tooth, existing just to service the nutritional and functional requirements of odontoblasts on its
periphery. Odontoblasts live on for most of the life of a tooth, regularly
forming dentine at a slow pace after tooth eruption. As a result, the size of
the pulp cavity gradually reduces as the dentine is thickened. The inside
of the tubules gradually mineralize (Fig. 2.8b) and are slowly obliterated,
particularly in root dentine. All this provides a physiological adjustment
for gradual tooth wear. However, this leisurely secretion is not likely to be
the main reason why teeth have a vulnerable soft centre. A much more
likely reason is the potential that it gives for a local response to more rapid
wear of the tooth crown. The mechanism for this response is not properly established, but the long processes of odontoblasts appear to detect
the movement of uid inside the dentinal tubules following tooth loading
(Paphangkorakit & Osborn, 2000). The pulp is well supplied with nerves,
most of which head towards its periphery where they form a plexus near
the odontoblasts. Some nerve bres even pass for a short distance into the
28
Fig. 2.8 The structure of mammalian dentine, again exemplied by the human.
(a) Dentine contains cylindrical cavities called dentinal tubules that wind from the junction with the enamel back to the pulp cavity. When the tooth is rst formed, they contain
elongated processes of the odontoblasts and also uid. Initially, the odontoblastic processes
extend right to the enamel and can even be squeezed into it, but they later retract away.
Note the hypocalcied region in the dentine, marked by the semicircular contours of calcospherites that have failed to fuse in this region (see Fig. 2.10). For some reason, this appears
to have no functional consequences. The hydroxyapatite crystals in dentine are completely
separate to those in enamel (Diekwisch et al., 1995), again without consequence. (b) The
tubules gradually inll with peritubular dentine, which is much more heavily mineralized
than the dentine between tubules (called intertubular dentine).
29
Fig. 2.9 The pulp exists to service odontoblasts, cells whose nuclei can be seen as dark oval
masses next to the dentine. Most of the nerve bres in the pulp form a plexus near the odontoblasts and some, including that which is arrowed in the gure, pass between them to enter
a dentinal tubule. One way or another, this makes dentine sensitive to load, triggering the
commissioning of extra tissue so as to thicken it.
tubules (Fig. 2.9), but none appears to communicate with the odontoblasts
via synapses (Holland, 1994).
It is thought that the cell body of the odontoblast in the pulp may relay
this signal to nearby nerves. Quite how this signal is organized is unclear,
but odontoblasts start to produce dentine at an increased rate as soon as a
sufciently intense signal is received. There appears to be no need for the
dentine itself to be exposed in the mouth for this response to begin.
As with other highly structured mineralized tissues (e.g. mollusc shell:
Currey, 1980), the organization of dentine suffers when it is laid down
rapidly. Reparative dentine does not contain ordered tubules and has a
lot less mineral than the dentine that is formed before a tooth is in
function. This difference in mineral content between regular and reparative dentine is important in understanding the wear patterns of exposed
dentine.
30

The properties of enamel and dentine
Appendix B lists the principal mechanical properties of these crucial tissues.

Unsurprisingly, enamel is generally much stiffer and harder than dentine.
However, it is becoming clearer that there is some overlap. Peritubular
dentine, for example, has close to the hardness of enamel (Kinney et al.,
1996) and there is probably nothing difcult in developing very high mineralization levels in vertebrates. Lungsh, for example, long separate on their
own evolutionary path, have a form of dentine (petrodentine) within their
massive tooth plates that has the hardness of mammalian enamel (Bemis,
1984). What seems special about mammalian enamel, however, is its mechanical anisotropy (difference in property values in different directions).
Microstructurally, enamel and dentine are both very heterogeneous tissues and it might be thought likely that this is reected in their properties.
The surprise though is that enamel seems to show more anisotropy than
dentine, e.g. in toughness (Rasmussen et al., 1976). The reason for this
seems to lie in the avoidance of crown fracture. The theoretical work of
Spears et al. (1993) suggests also that some anisotropy of the Youngs modulus of enamel is essential if stresses are to be passed from the enamel
through to the dentine. Their specic prediction, that enamel should be
considerably stiffer along rods than across them, remains unproven (White
et al., 2001), yet it makes much sense out of tooth structure. Many prior
studies have shown that if enamel were isotropic, then stresses would be
taken mainly by the stiff enamel cap and only transferred to the dentine around the cervical region, the neck of a tooth. Spears et al.s suggestions would result in the dentine being loaded much more evenly and
make a lot of structural sense. How, for example, could odontoblasts
in the pulp respond to wear if the dentine were not under signicant
load?
Inner enamel behaves differently to the bulk of the tissue. The rstformed enamel, that closest to the dentine and which appears the least
organized region, resists cracking under indentational loads much better
than the rest and seems to protect the enameldentine junction (Habelitz
et al., 2001), something that also only makes sense if this region is heavily
loaded. One of the best-documented features of enamel, and one that is
clinically important, is the effect that enamel structure has on controlling
cracks. These fracture paths mirror toughness values in that cracks avoid
crossing the rods and tend to travel only in the rod margins. Thus, variation
in rod directions and in the amount of decussation of the enamels of
different mammals could and should have a very strong effect on wear
31
resistance and on the maintenance of the sharpness of enamel ridges. Work

on this is now progressing (Rensberger, 2000).
Dentine is structurally just as heterogeneous as enamel, but it has
nothing like this mechanical anisotropy. Fracture paths show no denite
relationship to tubule directions (Andreasen, 1972) and, despite differences in toughness, there is little evidence for any crack deection at the
peritubularintertubular dentine boundary. Yet, there are clearly important
regional variations in dentine within a tooth that have important consequences for wear when this tissue is exposed on the crown surface. The
dentine that lies just under the enamel, called mantle dentine in humans, is
at least 10% less hard than the rest of the rst-formed (or primary) dentine
(beavers: Osborn, 1969; gibbons: pers. obs.; humans: Renson & Braden,
1971) and overall, the microhardness of primary dentine is generally twice
that of secondary dentine (Baker et al., 1959; Habelitz et al., 2001). A careful study of the wear patterns of the dentine of the incisors of the beaver
(Osborn, 1969) shows a close correspondence with the microhardness of
the dentine and there is circumstantial evidence for this in virtually all
mammals where both primary and secondary dentine are exposed.
Yet, it should not be thought that enamel and dentine are faultless and
impeccably nano-engineered. In fact, aws are obvious in both tissues by
observing them at low magnication in a light microscope. Figure 2.10
shows one typical aw in both tissues. Relatively massive spaces (massive
compared to processes involved in crack formation anyway) called interglobular dentine are commonly found and represent regions of very low
mineral density. Why do these not just collapse under load and send cracks
spraying out to the enameldentine junction? The answer does not seem obvious. Hypomineralized regions called tufts, attened irregular sheets, also
characterize the inner enamel (Osborn, 1981). The aws may run considerable distances towards the crown surface, without apparent consequence.
More rarely, normal enamel contains what look to all intents and purposes
to be cracks. These lamellae are thought to represent cracks that developed
early in enamel formation and which are lled by protein.
Much has been made of the hierarchical complexity of mineralized tissues like enamel and dentine and of the contribution that an understanding
of their structure could make to improving the properties of articial composite materials. The aim of some is to produce similar materials that, like
enamel and dentine, are structured as a nanometre scale. However, perhaps
when these articial nanomaterials can be designed to survive the scale of
aws that real dental materials possess without apparent consequence, this
will signal true understanding of these amazingly complex dental tissues.
32
Fig. 2.10 Flaws in human enamel (E) and dentine (D) are obvious in this micrograph. The
large spaces (arrowed in D) are called interglobular dentine, while the lines in enamel
extending from the tissue junction are called tufts (arrowed in E). Both are very low
in mineral density. The multiple semicircular structures bounding interglobular dentine
are called calcospherites and represent the manner in which dentine mineralizes. Scale
bar, 1 mm.
jaw bones
The upper teeth are housed on the maxilla and premaxilla, the lowers on
the mandible. The teeth in both jaws form arches that follow the margins
of the tongue. Normally, the lower arch is narrower than the upper arch,
which means that lateral movements of the mandible are necessary in order
to align upper and lower postcanine teeth properly for chewing.
In most mammals, dependent on their age, there is a pattern of sutures
(stiff brous joints) between the bones in the upper part of the face allowing
for very limited movement between them. The lower teeth are mounted in
the mandible that may or may not be a single bone. Anteriorly, at the midline, the two dentaries (as the bones of the mandible are called) are often
connected much less tightly and there can be a substantial amount of movement between them (e.g. in the dentally primitive Madagascan mammal,
the tenrec: Oron & Crompton, 1985). Scapino (1965) even referred to this
anterior symphyseal region5 in carnivores as akin to an extra jaw joint. Studies of the anatomy of the tissues in primates with unfused symphyses suggest
Jaw bones
33
that its exibility is variable, with some species having very tight connections
and others much looser ones (Beecher, 1977, 1979). Quite independently, a
number of mammals, including the higher primates and the giant panda,
have fused their symphysis (i.e. converted it to bone), so that the whole lower
jaw is just one unit. A large amount of work by Hylander and colleagues
(Hylander et al., 2000) suggests that symphyseal fusion can be explained
by the need of some mammals to recruit jaw-closing muscles of both sides
of the body to elevate the force in a bite (although this is not without
counter suggestion; Lieberman & Crompton, 2000).
Bone differs from tooth tissue in being alive. It contains blood vessels
that run in a network of channels throughout the tissue, feeding the boneforming cells and also transporting bone-eating cells rapidly to any location.
The result is that bone is constantly in ux both internally and externally,
allowing it to respond homeostatically both to changes in mechanical strain
within the bone itself and to whole-body equilibrium. An example of this
is provided by tooth loss in humans. The mandible and maxilla have a basal
part, which forms the basic support of the face and which remains even
if the dentition degenerates. The roots of the teeth are housed in what is
called the alveolar process of these bones. If the teeth are lost, this bone is
gradually resorbed (eaten away) and over a period of 10 years or more can be
completely lost. Heath (1982) studied human denture wearers and found a
correlation between maximum bite force and the height of the mandibular
body.6
The local signal for bone remodelling has been thought for more than
100 years to be related to its mechanical environment and there is a quite
massive body of literature that has attempted to dene the biological rules
that control this. Whenever food is loaded between the teeth, the teeth and
surrounding bone are deformed. It is thought that bone tissue remodels in
response to the strain levels that it sustains. Biewener (1992) describes how
this has been investigated, but the strain threshold level for either bone loss
(resorption) or deposition has not been established clearly and it has not
even been proved that strain provides the trigger for tissue changes. Having
a cellular response purely based on a strain criterion will not guarantee
freedom from fracture and microcracks in bone are in fact quite common
(Lanyon & Rubin, 1985). Local compressive strain level within bone has
been linked with the onset of bone formation, but the rhythm of loading
patterns is also implicated (Rubin et al., 2001; Mao, 2002). It is entirely
possible that local strain energy density, i.e. the energy absorbed per unit
volume of tissue, plays an important role. The skull may behave differently
from the limbs and one study found that no local stimulus was necessary
for skull bone deposition: that of the armadillo thickens substantially in
34
response to general exercise, apparently quite independently of workouts

of its jaws during feeding (Lieberman, 1996).
The stress that corresponds to a particular strain will depend on the
stiffness (Youngs modulus) of bone (Appendix A). The outer bone appears
completely dense to the naked eye and is called compact bone. Its stiffness
varies subtly with location (Dechow & Hylander, 2000), possibly related to
the direction of muscular pull. Inside this is a network of bony struts getting
the name trabecular bone. The stiffness of trabecular bone depends not
only on the density of the struts, but also on their geometrical arrangement
and orientation (Gibson, 1985).
Sadly, I conclude this section by confessing that my efforts to establish
overall generalities relevant to the control of bone formation or loss in the
face have failed.
the tooth socket

Non-mammalian vertebrates generally have their teeth xed to the skeleton
of the face either by brous tissue or direct to bone (this bone being termed
bone of attachment). In contrast, mammalian teeth nestle in sockets: a
small fact with great implications. It is the socket that allows a tooth to
drift slowly in the jaw and come to rest in a position where it is accurately
aligned to its counterparts in the other jaw. The movement involved is
not rapid, depending on the rate at which the body can remodel bone.
As with so many aspects of dental function, the mechanisms involved are
not well understood, but without this ability to move, mastication would
be impossible because the teeth would not match up. The structure of
the lining tissue of the tooth socket, the periodontal ligament, is therefore
worthy of some attention.
Most of the ligament, three-quarters of it or so, is brous tissue in the
form of long bundles passing from the cement to the bone lining the socket
(Fig. 2.11).7 These bre bundles form an open network in the periodontal
ligament, running in a variety of directions that enclose cells, interstitial
uid, blood vessels, lymphatics and nerves. There are also elastin bres,
although their role is not yet very clear (Kirkham & Robinson, 1995). Nerves
and blood vessels cross the ligament to enter the apices of the tooth roots
but the ligament itself is also very richly supplied. The collagen bres of
the ligament have a rapid metabolic turnover, which requires a good blood
supply. Thus, copious blood vessels enter the ligament through numerous
small holes in the bone of the socket. The ligament is also well innervated
The tooth socket
35
Fig. 2.11 The periodontal ligament (PL), lying between the cement (C) and bone of the
socket (B), is largely composed of collagen bres set in a gel-like matrix. Both are important
in explaining why teeth do not abrade against the socket walls, given that the ligament is
so narrow. The dentine (D) underlies the cement. The periodontal ligament is generally
between 0.1 and 0.3 mm thick.
and contains mechanoreceptors that relay information about the direction

of loading on a tooth (Linden, 1990).
Cement starts as mineral glue that forms around the incoming collagen
bres of the periodontal ligament. However, unlike enamel, it can be added
to throughout the life of a tooth so as to compensate for the loss of tooth
height. Cement that is laid down later, after tooth eruption, develops its
own, much ner, collagen framework (Jones, 1981). If it is formed slowly,
then the cementoblasts that form it can retreat away, but if it is formed
rapidly, cells get trapped in the matrix and eventually die. The rate at
which cement has formed is evident from its cellularity. In humans, rapid
cement formation around the apices of the roots is a response to rapid wear
of the crown. In some herbivores, the crown of the tooth is also coated in
cement (Jones & Boyde, 1974).
The periodontal ligament is generally very narrow, generally being
0.10.3 mm in width in the human. Large transverse (lateral) forces will
widen the ligament. For example, a very wide socket is normal for the upper
canines of Chinese water deer (Hydropotes inermis), which are presumably
continually being knocked by lateral movements of the jaw (Aitchison,
1946). However, the teeth of most mammalian species tend to t very
36
tightly into their sockets. Synge (1933) suggested that the reason for this
tight t is that the displacement of the tooth crown in a lateral direction at
any point on its surface is proportional to the cube of the width of the ligament. In other words, doubling the width of periodontal ligament would
increase tooth movement eight-fold. A snug t of a tooth in its socket is
thus absolutely essential or it will waggle like a water deers.
The assumptions that were necessary for Synges result do not sit very
well with its biology. He assumed that the ligament was elastic and incompressible (rather like a rubber) and sandwiched between an innitely rigid
wedge-shaped tooth and bony socket. However, the critical assumption is
probably that of incompressibility, which is likely to be provided by the large
quantity of intra- and extracellular uid within the ligament, much of it
bound in a gel-like matrix. Incompressibility is supported by the observation that when a tooth is pressed into its socket, the bony crests standing just
beneath the gum move away from the tooth (Picton, 1965). Whatever, the
ligament is far from ideally elastic, taking time to recover after being pushed
into its socket. Synge knew this from investigations done with Dyment
(Dyment & Synge, 1935) and he made a number of revisions to his theory,
including the effect of a small compressibility. However, his theory did
not offer a specic role for the collagen bres in the periodontal ligament,
stating that if these alone were responsible for resisting tooth movement,
then the displacement of the tooth crown in a lateral direction at any point
on its surface would be directly proportional to the width of the ligament.
Without the cube relationship, an argument for the tightness of the socket
is lost. Rather like tendon bres, the collagen bres of the ligament possess
an undulating form (called a planar crimp). This probably pulls straight at
a 4% elongation (Kastelic & Baer, 1980), but prior to that, these bres
offer no resistance to tooth movement. Once straightened, they are predominantly elastic and must be very important in preventing a tooth from
bottoming out in its socket. There is experimental evidence in goats that
collagen bre direction adapts to the degree of tooth depression (Lieberman,
1993).
If a tooth goes out of function, then the ligament narrows. It is accordingly likely that most teeth are loaded mostly in a direction close to
their long axis. If they were not, then they would realign until this were so
or else quickly work very loose. Resistance of the ligament to axial forces
that press teeth directly into their socket is obviously extremely important. Synge (1933) also gave an estimate of this for a human upper central
incisor with a conical root. The pressure, P, in the ligament is highest
at the apex where
Movement of teeth within the jaws

P = PA (2F /A)
37
(2.1)
where PA is atmospheric pressure, F is the axial force and A, the area of the
cross-section of the tooth root at the top of the socket.8 If the ligament is
assumed to possess the properties of rubber, then Synge worked out that
an axial force of 1.7 N moves the tooth only 7 m into its socket, while a
transverse force of 0.85 N applied to the incisal edge moves it 226 m. Real
data somewhat exceed this for axial forces but not for transverse ones. For
example, Partt (1960) found experimentally that an axial force of 1.7 N
actually pushed an upper incisor 25 m into its socket. (Anderson (1976)
states that human teeth probably only depress about 50 m at most into
their sockets.) This is much greater than Synges predictions, but does not
(in my opinion) detract from the value of his work. Muhlemann (1951)
found that the same horizontal force as Partts would move the tooth
150 m. However, the ligament is less stiff at low forces than higher ones,
which suggests collagen bre involvement that Synge would not have anticipated. Much experimental work has followed these early studies and
is summarized in Berkovitz et al. (1995), but fundamental explanations of
their results seem far away.
movement of teeth within the jaws

Teeth develop inside the jaw, emerging as and when needed by a process
called tooth eruption.9 A soft lining to the tooth socket allows the erupted
teeth of mammals to drift in the jaws so as to make minor adjustments in
alignment. This drift can be either mesial or distal in direction (Berkovitz
et al., 1995). Very likely in ungulates that have a diastema (shown in Fig. 3.10)
between the cheek teeth and incisors, the anterior cheek teeth drift distally
while the posteriors drift mesially, thus keeping the cheek tooth battery
together in an unchanged position. Unless the tooth is completely stable,
areas of both bone formation and resorption can be found in and around
the socket so as to facilitate this tooth movement.
The principal reason for this drift must be the direction of the muscular
force. This will determine the effective direction of the bite force on all
food particles except the most rigid. While the resultant of the force must
be directed along the long axis of the tooth in order for the position of the
tooth not to shift, the direction of the force depends on gape. In humans,
the muscle closing force has an anterior component when the jaws are
almost closed, and a slightly posterior component when the jaws are wide
38
Fig. 2.12 The structure of the human temporomandibular joint. This has been researched
in great detail, but some aspects of its structure and the explanation of these remain
controversial. In mammals, a brocartilaginous disc separates bony surfaces lined by
brous tissue, effectively forming two joint spaces, one above and one below this disc.
The temporomandibular joint
39
apart (Paphangkorakit & Osborn, 1997). The force direction at jaw closure
is consistent with the results of Osborn (1961) who found that teeth push
forwards (mesially) in the jaw during an empty clench. The approximal
wear that develops from this is described in Chapter 6.
the temporomandibular joint
The upper and lower jaws are connected together by a synovial joint, called
the temporomandibular (or jaw) joint, which ensures their stable alignment. The joint lies between the head of the mandibular condyle and the
squamous part of the temporal bone of the skull and is unique to mammals.
The joint has been investigated in great detail because of its complexity and
importance (Rees, 1954; Hylander, 1992). Like all synovial joints, it has a
brous capsule enclosing the joint surfaces. Viscoelastic synovial uid
bathes the joint surfaces, so reducing friction to very low levels (the coefcient of friction in joints varies, but is generally <0.04; Swanson, 1980).
In most mammals, the joint is basically a hinge (capable of rotatory
motion), but it can also glide (translate) at wide openings. In some mammals
with fused symphyses, any movement of the lower jaw necessarily involves
movements of both the left and right temporomandibular joints because
the mandible is a single bone. The limits of jaw movement in humans are
mainly due to two prominent ligaments, the lateral ligament that lies on
the capsule and the sphenomandibular ligament, which is separate from it
(Fig. 2.12) (Osborn, 1993). In humans, the mandible appears to translate
under the control of the upper end of the lateral ligament, but rotate around
its lower end, rather like an upside-down childs swing (Osborn, 1995a).
In humans, the articular surface of the mandible sits proud as a condyle,
which nestles behind a projection of bone on the posterior aspect of the
zygomatic part of the squamous temporal bone of the skull called the articular eminence (Fig. 2.12). These bony surfaces are covered in brocartilage
rather than the hyaline cartilage that covers the surfaces of bones in most
other synovial joints. This is partially a consequence of these bones developing in brous membrane rather than from cartilaginous precursors
Caption for Fig. 2.12 (cont.) This is a unique arrangement for a synovial joint. According
to Osborn (1985), this is to allow the head of the mandibular condyle to bed into the disc
at low stresses, due to the J-shaped stressstrain curve of this soft tissue (see Appendix A
and Fig. 5.7). The two important ligaments that help to control jaw movement, the lateral
and sphenomandibular ligaments, are shown.
40
(Hylander, 1992), but the distinction is important mechanically too. Unlike hyaline cartilage, brocartilage has a J-shaped stressstrain curve (Fig.
4.13) (Tanne et al., 1991; Teng et al., 1991). At low stresses, it is very pliant
which means that the head of the condyle tends to sink into the brous tissue under the skull, allowing the condyle to bed down (Osborn, 1985). At
high stress, this tissue is much rmer, so allowing it to slide (see Fig. 5.7).
The joint surfaces do not contact because a brocartilaginous disc is
placed between them. It is attached to the capsule all round, so dividing
the joint cavity into separate upper and lower spaces. Anteriorly, the disc
attaches both to the squamous temporal bone and the neck of the condyle.
It also appears to receive bres of the lateral pterygoid muscle, although
whether these bres embed only in the capsule of the joint or only pass
into the front of the disc instead is not clear. Essentially, the disc moves
with the mandibular condyle, but not in complete synchrony, something
that is a cause of trouble with the joint in humans.
soft tissues
The mammalian mouth (or oral cavity) is designed around the need for
mastication. It is bounded above by the hard palate, which separates it
from the upper respiratory tract lying above. Reptiles, other than some
crocodilians (Smith, 1992), do not have a hard palate because they do not
chew. They swallow large particles, which will not t into the airway. In
contrast, the small (sometimes minute) fragments that can be produced
by chewing can be inhaled if the airway is not protected, so particle size
reduction of foods by mastication requires the airway be as disconnected
as possible from the foodway.
Food that enters the mouth encounters a cavity lined with a tissue called
oral mucosa that is somewhat like thin skin (Fig. 2.13). As with skin, the
lining of the mouth is keratinized in all places where there is considerable
friction. Contact with food particles produces this friction and the high
toughness of keratin (e.g. in horse hoof: Appendix B) prevents any subsequent wear. The frictional coefcient of dry oral mucosa against dry food is
high (see below). To avoid this, the mouth needs to be constantly lubricated
by saliva.
The oor of the mouth houses the tongue, which is a large muscular bag
capable of a wide range of movement. The oral cavity is actually mostly
tongue. It occupies virtually all the space inside the dental arches, leaving
only 1020% of cavity volume vacant for food particles. Tongue movements
do not impinge on this space because the tongue is thought to have a
Soft tissues
41
Fig. 2.13 Above, features of the musculature of the tongue, soft palate and pharynx in the
human, exposed from the side. Below, a cut across the face in the coronal plane to show
details of the oor of the mouth.
42
pterygomandibular
raph
orbicularis
oris
middle constrictor
of pharynx
buccinator
inferior constrictor
of pharynx
superior
constrictor
of pharynx
stylopharyngeus
Fig. 2.14 The major muscles of the pharynx, the constrictor muscles, are a mammalian
novelty. They form a semicircle of muscle around the back of the oral cavity and larynx
and contract in a peristaltic wave during swallowing. Curiously, the superior constrictor
is attached to the cheek muscle, the buccinator, another novelty, even though these have
different functions, developmental origins and nerve supplies. The vertical line at which
the buccinator and superior constrictor bres interdigitate is called the pterygomandibular
raphe. Anteriorly, the buccinator appears to send many of its bres into the orbicularis oris,
the muscle of the lips. These bres decussate (cross) at the corner of the mouth.
constant volume whatever its position (Kier & Smith, 1985). The oral surface of the tongue is purposefully rough. Its upper surface is roughened
with keratinous spikes called liform papillae in order to grip food particles. Taste buds on the tongue surface are associated with other types of
papillae. All taste buds are anatomically similar, but their component cells
are sensitive to different chemicals depending on their membrane receptors
(see Chapter 3).
Salivary glands
43
Below, the mouth is separated from the neck by a thin diaphragm called
the mylohyoid muscle (Fig. 2.13). This is connected at the front and sides
to the lower jaw. Behind, it is attached to the hyoid bone, a small bone
that oats below the tongue suspended by ligaments and the tone of
its muscles. Normally, the mylohyoid muscle hangs loosely downwards,
but it contracts strongly during swallowing (Hrycyshyn & Basmajian,
1972).
Posterior to the mouth lies the opening to the pharynx. The tongue
bridges the two spaces with its posterior one-third lying in the pharynx. In
contrast to the roughened oral surface of the tongue, the posterior part is
smooth so as to allow food particles to slip over it easily. In the pharynx, the
airway crosses the foodway (Fig. 1.1). The mammalian pharynx possesses a
series of semicircular constrictor muscles attached to the back of the mouth
and larynx, which reptiles do not possess (Smith, 1992). Rapid peristaltic
contraction of these muscles assists in pushing the food bolus down the
pharynx to the oesophagus as rapidly as possible (Fig. 2.14).
salivary glands
Lubrication of the mouth is achieved by the output of an oral uid called
saliva secreted from a large number of glands located around the mouth. The
largest of these glands in most mammals are the parotid, submandibular
(sometimes called the submaxillary) and sublingual glands (Fig. 2.15). The
rst two are much more important than the last, which is actually a mass of
very small glands lying close together. Saliva jets from four major orices,
two on either side of the midline, and from many small openings. The
largest glands are the parotid glands, which are housed just behind the
lower jaw and whose ducts open into the vestibule of the mouth opposite
the molar teeth, the most important postcanines (Fig. 2.15). The ducts spray
these teeth directly. The submandibular glands lie partly in the oor of the
mouth and partly in the neck, wrapped around the posterior edge of the
mylohyoid muscle (Fig. 2.15). Their ducts open close together in the oor
of the mouth just behind the lower incisor teeth.
Saliva itself is a dilute solution of glycoproteins with four very important
functions from the viewpoint of this book:
(1) It reduces friction between food and mucosa to low levels. This requires
fairly high glycoprotein concentrations.
(2) It wets the new surface of the food produced by particle fragmentation
so that these will be encouraged to bind together to form a bolus for
swallowing.
44
sublingual
gland
parotid gland
and duct
mylohyoid
muscle
submandibular
gland and duct
Fig. 2.15 A sketch of the major salivary glands in the human. Both the parotid and submandibular glands have grown so large that their bulk, or part of it, lies outside the
oral cavity.
(3) It helps to solubilize potential taste compounds in foods such that they
can be sensed on the tongue.
(4) It buffers food acidity, so preventing acid erosion of the mineralized
tooth tissues.
The composition of saliva depends substantially depending on the gland
that produces it. Minor glands, including the sublingual, mostly produce
viscous secretions that cover the mucous membrane. This thin covering is
not lost at swallowing and is needed to keep intra-oral friction low. J. F.
Prinz (pers. comm.) shows that surface-dry mucosal surfaces can have a
coefcient of friction between them as high as 1.5, while saliva brings this
The muscles
45
down 10-fold or more. The secretions of the parotid gland are much less
viscous, with submandibular saliva being intermediate.
Physically, saliva has many functions. It is the anti-stick lining of the
mouth only fatty lms from foods usually evade the clearance of food
by swallowing. Even chewing gum does not stick either to the teeth or
mucosa. Nevertheless, saliva (particularly parotid saliva) is good at wetting
food surfaces, due to a surface tension well below that of water (Glantz,
1970), and good at sticking food particles together into a bolus by virtue
of a viscosity that is well above it (Prinz & Lucas, 2000).
The viscosity of saliva has always been measured by conventional viscometry, i.e. by shear. Currently accepted values, which vary with the rate
of shear, are probably underestimates. There is circumstantial evidence
during critical events in the mouth, such as during swallowing or food bolus
evaluation, when the tongue thrusts up against the hard palate, that saliva
is subject to extensional ow, probably biaxially (de Bruijne et al., 1993;
van Vliet, 2002). This type of ow changes the volume of saliva slightly
by stretching it. Volume change may seem counter-intuitive to those who
think in terms of the action of hydraulic pistons, but it is perfectly possible
to change the volume of liquids in some circumstances, and this is the only
way in which a lm of oil in an immiscible uid can be emulsied (i.e. can
break down into small droplets). Such emulsication does happen in the
mouth, as also the production of many small bubbles in the saliva, which
is somewhat analogous (de Bruijne et al., 1993). The point is that resistance
to extensional ow appears much greater than in normal shear, requiring
a rethink of salivary viscosity.
the muscles
There are a very large number of muscles involved in oral function and
their general morphology and bre direction varies substantially between
mammalian species. The muscles can be grouped into (1) jaw movers,
(2) tongue muscles, (3) facial muscles that move or x the mouth slit and
(4) neck muscles that adjust the position of the cranium.
Jaw movers
The major work input for fracturing food particles comes from the large
muscles that elevate the lower jaw. These are the masseter, temporalis and
medial pterygoid muscles (Fig. 2.16). The temporalis is a fan-shaped muscle,
attached to the side of the vault of the skull and fascia overlying it. Its bres
46
(a)
anterior
fibres
(b)
posterior
fibres
deep
fibres
superficial
fibres
(c)
upper head
and
lower head
of lateral
pteygoid
muscle
(d)
posterior belly
of digastric
condylar
process
of mandible
medial
pteygoid
muscle
anterior belly
of digastric
Fig. 2.16 The orientation of (a) masseter, (b) temporalis and (c) the medial and lateral
pterygoid muscles in humans. All except the last muscle provide power to the bite. The
lateral pterygoid is also shown in (d) assisting the digastric muscle in opening the jaw.
descend medial to the zygomatic arch and attach to the coronoid process
of the mandible. The masseter muscles run from the whole length of the
zygomatic arch, both from the undersurface and medial side, down to the
lateral side of the ramus of the mandible. In most mammals, it is a complex
multipennate muscle (Anapol & Herring, 2000). The medial pterygoid,
which has bres that are roughly parallel to those of the masseter, but which
is on the inside of the mandible rather than the outside, adds weight to the
bite.
The jaw-closing muscles (the elevator or adductor muscles) of lower vertebrates are critical for feeding, but they are relatively undifferentiated,
The muscles
47
forming a common muscle block roughly where the masseter and medial
pterygoid lie in mammals. When this jaw-closing muscle acts, it probably
presses the jaw joint into the base of the cranium. This does not matter too
much if the forces are small, but mastication can involve very large forces.
Crompton (1963) was the rst to work out a potentially optimal solution,
showing how the development of a discrete masseter, medial pterygoid and
a posterior temporalis musculature considerably eases the elevation of a
jaw. If the masseters acted alone, they could only close the jaw by exerting a very high compressive load on the jaw joint. The posterior temporalis muscle is important in continuing the movement without driving the
condylar head into the skull. Figure 2.17 sketches the logic of the arrangement of jaw-closing muscles in mammals, but, since it can be difcult to
grasp geometrical ideas like this from paper descriptions, Fig. 2.18 suggests a
demonstration with a human mandible. Despite the logic of this, Osborn &
Baragar (1985) showed clearly that it is impossible to unload the jaw joint by
muscle arrangements such as this. Osborn (unpubl. data) suggests that
it is far more probable that evolutionary changes favoured muscular
efciency.
The jaw-opening muscles (also called jaw depressors or abductors) include
the lateral pterygoid, the digastric (Fig. 2.16d ) and other suprahyoid muscles
(such as mylohyoid, geniohyoid and sternohyoid Fig. 2.19). The anterior
digastric and the geniohyoid are probably the most important. In humans,
the anterior digastric and the lateral pterygoid muscles act as a couple to
open the jaw by both rotation of the joint (at any angle of opening) and
anterior translation of the condyle (pronounced at wider gapes), but this
is not true in many other mammals. The digastric muscle, for example, is
variable even in closely related primates like humans and great apes (Aiello
& Dean, 1990).
The mammalian mandible can be moved from side to side and forwards
and backwards. Side-to-side movements are more pronounced in most
mammals. In some ungulates, this is extensive, while in carnivores it hardly
exists at all (Hiiemae, 1978). The lateral pterygoid muscle in particular,
acting just on one side of the jaw, drags the lower jaw towards the opposite
side, but other jaw muscles are potentially capable of assisting it depending
on whether they are situated so as to produce mediolateral pull. There can
be considerable variation in lateral movement even within a masticatory
sequence, when processing one mouthful of food. In contrast, when jaw
movements are viewed from the side, they appear stereotyped, varying only
in amplitude. This pattern is reversed in rodents where the typical chewing
orbit (see Fig. 3.6 for this in humans), with different opening and closing jaw
48
centre
of jaw joint
jaw adductors
(a)
(b)
(c)
M
M
T
T
(d)
small gape feeder
large gape feeder
Fig. 2.17 Schematic diagrams to explain the logic behind various arrangements of the
jaw joint and muscles in mammals. (a) A pseudo-mammalian jaw given one jaw-closing
muscle the equivalent of the pterygoideus part of the jaw adductor mass (called M) of a
reptile moving the jaw through a wide gape. At right, this jaw is modelled as a disc with
a central joint. The jaw muscle can rotate the disc through a given angle, close to where
it forms a perfect tangent, but at other angles will tend to pull on the joint itself. (b) This
mammal has now developed a process in front of the joint (the coronoid process) and differentiated a muscle mass (T, standing for temporalis) that pulls at a very different angle to M
The muscles
49
Fig. 2.18 A game for two players to demonstrate Cromptons (1963) explanation for the
organization of jaw muscles in mammals. If you have access to skeletal material or a plastic
model, you could try the following on a mandible with a friend. (a) Grip the outside of
the mandibular condyles with the thumb and index nger of one hand and attempt to
swing the jaw so as to close it by pulling upwards on the angles of the jaw with the other
hand. This will simulate the action of the masseter and medial pterygoid muscles. The
jaw will move but with difculty. You will feel the force through the mandibular condyles
that would, in life, push them through the oor of the skull. (b) Now ask a friend to pull
back lightly on the coronoid processes at the same time. This simulates the action of the
posterior temporalis muscle and causes the jaw to elevate with ease. The point is that, in
order to rotate the jaw upwards, no one pair of muscles with a xed line of action can
produce an even bite force at all angles of jaw opening. (Courtesy of Dr K. Rajendran.
Note that a human jaw is depicted: all human material should be treated with very
great respect.)
trajectories, is seen not from in front, but from the side (Hiiemae & Ardran,
1968; Weijs, 1975).10 The explanation for this anteriorposterior movement seems to lie in the mastication of very large mouthfuls of food, such
that they can be processed bilaterally (i.e. on both cheek tooth rows at
once).
Caption for Fig. 2.17 (cont.) (drawn for the medial pterygoid muscle, but which could also
represent the masseter). This muscle arrangement closes the jaw through a far wider angle
without loading the joint, as shown by the disc. (c) Many models of jaw-muscle action (too
complex to describe here) emphasize the optimization of muscle activity so to produce the
largest bite force for either the least effort or the least loading of the joint. A mammal that
feeds at a variety of gapes may show one of two solutions to this. At left, variation in muscle
bre direction, coupled with sophisticated neural wiring, activates only those bres with
optimal orientation. Many mammals show some variegation in the temporalis and masseter
muscles like this (McMillan & Hannam, 1992; Anapol & Herring, 2000). At right, instead
of this arrangement, the jaw joint no longer just rotates, but slides its centre of rotation
forwards towards M. Primates do this to reduce the stretch in M (Carlson, 1977; Hylander,
1978). (d) The jaw of small- and large-gape feeders. At small gapes, M is favoured, while at
larger gapes, M is stretched too much and T (temporalis) is more efcient. Thus, small-gape
feeders like many ungulates have large M, while carnivores have large T.
50
stylohyoid
mylohyoid
omohyoid
sternohyoid
Fig. 2.19 Some of the prominent muscles attached to the hyoid bone.
Tongue muscles
The tongue is very mobile, obtaining this ability from three sources. The
rst derives from the anchorage of the tongue to the hyoid bone. In reptiles, movements of the hyoid basically determine the movements of the
tongue (Smith, 1992). The most prominent tongue movement in lower
vertebrates is an anteroposterior motion, which when receiving ingested
solids or liquids sometimes requires that the tongue extend out of the
mouth for very long distances. In mammals, this anteroposterior movement is also important, being produced largely by muscles that move the
hyoid bone and muscles that link this bone to the tongue (Fig. 2.19).
They produce anterior (genioglossus muscle) or posterior (styloglossus and
hyoglossus muscles) movement, depression (genioglossus and hyoglossus)
The muscles
51
and elevation (styloglossus and palatoglossus), changing both tongue shape

as well as its position. However, the capacity for lateral movement of the
tongue, involving unilateral contraction, is extremely important for food
manipulation (Fig. 3.7). Other muscles that lie just below the oral surface
of the tongue, the intrinsic muscles, have no bony attachments and cannot,
therefore, move the tongue around as a body. Nevertheless, these muscles
are also extremely important in the manipulation of food particles. There
are three sets of intrinsic muscle bres arranged perpendicular to each other:
vertical, transverse and longitudinal (Fig. 2.13).
Facial muscles
The lateral boundary of the vestibule of the mouth is the muscular cheek.
The cheek muscle, called the buccinator, is the essential facial muscle for
intra-oral control of food particles. It has bres that run anteroposteriorly
in humans, with rm attachments above and below to the maxilla and
mandible. Anteriorly, it runs into a brous knot of tissue called the modiolus
(Fig. 2.20), while posteriorly, it is a continuation of the circular superior
constrictor muscle of the pharynx. The interdigitation of the bres between
the two muscles is called the pterygomandibular raphe (Fig. 2.14). When
a human subject yawns, this raphe bulges inwards noticeably and can be
used as a landmark by dentists seeking to block the inferior alveolar nerve.
The buccinator also has strong attachments to the maxilla and mandible.
This is because the cheek is an essential device for ensuring that particles
thrown laterally by the tongue in the early opening phase of the chewing
cycle stay on the working surfaces of the molars and do not fall into the
sulci on the lateral side of the teeth. Effective mastication depends on the
buccinator contracting during the late opening phase of a chew (Blanton
et al., 1970).
The anterior limit of the cheeks is the anterior limit of the postcanines.
Anterior to the mouth are the lips, which control the width of the mouthslit
and guard the entrance to the mouth. This control is achieved through a
large number of muscles on the face that converge onto the corners of
the lips. Just lateral to the angle of the mouthslit, these muscles meet at a
brous knot called the modiolus (Fig. 2.20).11 The details of the modiolus
are complex. Some accounts describe it as a brous knot while others have
muscle bres passing through it (Lightoller, 1925). Whatever, there is a
circular muscle surrounding the mouth called the orbicularis oris. This
muscle, which has no bony attachments, acts as a sphincter. The muscles
that converge on the modiolus from above and below are dilators or openers
of the mouthslit. The modiolus serves another role. The action of some
52
Fig. 2.20 The facial muscles in the human originally evolved as sphincters and dilators
of orices (mouth, nostrils, etc.). Many muscles converge on the mouthslit, particularly
laterally towards its corners. Most are vital for the control of food at the front and sides of
the mouth. Many of these muscles double up for the purposes of a facial body language.
upper and lower muscles can x its position. The buccinator depends for
its action on this xation (Figs. 2.14 and 2.20). Though its bres run from
front to back, its main action is to press against the lateral side of the
postcanine teeth. It can only achieve this if it is xed in some way. By xing
it anteriorly using the modiolus, its contraction tends to move the whole
cheek medially, which is the required movement in chewing.
In some rodents (Ryan, 1986), cercopithecine Old World monkeys
(Murray, 1975) and even the platypus, the cheek contains pouches that act
The muscles
53
Sternocleidomastoid
Fig. 2.21 The sternocleidomastoid in humans is a major feature of the neck, having
reoriented itself to assist in the head posture of a newly upright mammal.
as food stores, allowing small ingested food particles (e.g. fruits or seeds)
to be moved from their source prior to oral processing.
Neck muscles
These are generally set around the neck attaching to the undersurface of the
cranium and the uppermost vertebrae. Small delicate deeply set muscles,
54
particularly those located between the upper cervical vertebrae and the
back of the head, allow for small movements that align the visual axes.
Larger, more supercial muscles, such as the sternocleidomastoid (which
is particularly prominent in humans Fig. 2.21), are responsible for the
rapid powerful movements of the head required in feeding. Deeper muscles, such as the longus colli, longissimus cervicis and longissimus capitis,
also act to ex the vertebral column in the neck though with a lower mechanical advantage (Basmajian & De Luca, 1985). All these muscles are also
responsible for balancing the head and their size reects the position in
which the head is held. In most mammals, support is essentially by a long
elastic ligament that runs along the posterior side of the vertebral column
called the ligamentum nuchae. However, humans have an habitual upright
bipedal stance in which the head is balanced only by continuous active
contraction of posterior neck muscles such as the longissimi. Without this
action, the head falls forward such as when dozing over a book like this.
That is enough then about structure.
How the mouth operates
overview
This chapter describes how the mouth operates, placing this in the context
of a general physiological model. The chapter draws extensively on data
from humans because the need for cooperation in many chewing experiments exceeds that which can be obtained from trained animals.
introduction
The literature on mastication and swallowing is vast and it is impossible
to cite more than a small fraction of the papers that comprise it. Emphasis
has been placed here on those papers that describe facts relevant to physiological modelling of the process, however primitive such models might
be at this point. Central to unravelling what happens to food particles
in the mammalian mouth have been cine- and video-uoroscopic studies.
These date back to Ardran et al. (1958) on the rabbit, through Crompton &
Hiiemae (1970) on the American opossum, up to the present. These X-ray
movies have permitted views of the intra-oral processing of food otherwise
obscured by the cheeks. Excellent reviews are available (e.g. Hiiemae &
Crompton, 1985; Orchardson & Cadden, 1998; Thexton & Crompton,
1998) as well as studies on particular species. It has required the combination of a wide range of experiments and detailed observation (e.g. Weijs
& Dantuma (1981) on the rabbit) to establish the major features of oral
processing. However, I have yet to see some synthesis that attempts to weld
all this information on mastication and swallowing together. This chapter
intends not so much to review what is regularly reviewed in journals and
edited volumes, but to show how mastication and swallowing can be related
to features of the food input. Many will view the account as speculative in
places. However, the aim here is not to write a novel, but to drive towards
55
56
an understanding of how the description of the process can be tied to its

biological role.
why did mammals start chewing?
Mammalian mastication may have evolved for three reasons. The rst and
most general reason is that particle size reduction increases the specic
surface (the surface area per unit volume) of a solid. The rate of action
of enzymes in the gut is proportional to the surface area that they act
upon. Mammals consume energy at much higher high rates than reptiles
(Schmidt-Neilsen, 1972). Thus, it seems likely that the high metabolic rates
of mammals favoured the evolution of mechanical breakdown of solids,
prior to any chemical breakdown, because this can expose new surface very
quickly. Thus, the evolution of mastication in early mammals may be linked
to an increase in metabolic rate.1
Some mammals ingest foods that already have a high specic surface.
Yet, they still chew them. Familiar examples are ungulates that eat leaves
and grasses, which are so thin that they yield little extra surface when they
are fractured. The major reason for mastication in these circumstances is
likely to be the need to gain enzymatic access to food because it is otherwise
indigestible. Seed eating comes into the same bracket because seed casings
also obstruct digestion. Mammals that swallow whole mature seeds always
defaecate them in an unaltered state this is how they disperse seeds away
from the parent plants. A second reason for mastication in some mammals
then is to gain chemical access to foods that are otherwise impervious to
the action of their guts. This explanation is not linked to metabolic rate
per se and there is good evidence that the pre-Mesozoic reptile Suminia
getmanovi, dating from 260 million years ago, chewed plant material for
just this reason (Rybczynski & Reisz, 2001).
Most mammalian carnivores do very little processing of food in their
mouths and swallow foods in very large particle sizes (Savage, 1977). The
extra surface area produced by one or two chews is small and vertebrates
contain all the enzymes necessary for the digestion of the animal kingdom.
However, many mammalian carnivores attack prey that is as big as or bigger
than they themselves (Peters, 1983). These predators cannot ingest prey like
that in one lump and are therefore forced to reduce the size of their prey
simply to accommodate it in their gut. The argument cannot be pressed too
far because some vertebrates such as snakes can dislocate their jaw joints
to accommodate prey of colossal size. That carnivores do not do this is
probably explained by the fact that they cannot afford to store complete
Particle size distributions
57
carcasses in their guts without compromising their activity levels. This leads
to a third reason for evolving food breakdown in the mouth: it could have
evolved simply to reduce the particles to a swallowable size. Exactly what is
capable of being swallowed by mammals is considered a little later in this
chapter.
I start with the rst and most general of these reasons that mastication
is essential to the rate of food processing of the mammalian gut and treat
mastication strictly as a comminution process designed to expose new food
surface as rapidly as possible.
particle size distributions
Mammals usually ingest food in batches (mouthfuls). When particles of
any solid batch are broken down repetitively, they fragment into smaller
pieces that range often across several orders of magnitude of size. This comminution process expends about 5% of the planetary energy generated by
human industry (Lowrison, 1974). Despite the magnitude of this effort,
and the scale of its alleged inefciencies, the process apparently remains
incompletely understood. Yet what has been discovered provides myriad
observations of great importance for understanding mastication. By denition, it is mammalian to possess a dentition with a natural comminuting
capability, but it is distinctly human to comminute with tools; in fact, this
is probably the evolutionary basis of the development of a tool industry
(Chapter 7).
The particle size distributions formed during chewing are virtually always
skewed, meaning that instead of looking like the symmetrical bell shape of
a normal distribution, as tted by a Gaussian curve, they appear variably
distorted. The degree of this skew is initially dependent on what has been
ingested, but if the ingested particles are a group of similar-sized particles of
a relatively homogeneous food like a bunch of peanuts, then distributions
appear to develop as shown in Fig. 3.1. After a few chews, the long tail of
the distribution is initially stretched out to the left, but later it moves to
the right as particles get smaller with further chewing and as both the feed
size and its characteristic inuence disappear.
It is difcult to be entirely sure of this pattern because of the limitations
imposed by the sizing methods that have, until recently, been available.
Most studies on humans over the last century have used sieves, organized
so that their size intervals form something like a geometric progression
(Fig. 3.2).2 This is logical because any kind of comminution process, such
as the cutting up of vegetables in a kitchen for example, appears to have
58
Fig. 3.1 Schematic representation of the development of food particle size distributions
produced in chewing. The feed particles are a group of ingested particles with very similar
sizes. After just a few chews, several particles are broken, which sends a tail of fragments out
to the left-hand side. After several more chews, as more of the feed particles are broken, the
distribution begins to look like a normal bell curve, but as comminution proceeds further,
the tail drifts to the right-hand side and stays that way. The rate of these events depends on
the rate of food breakdown, but this normally takes only about 10 chews in humans (Lucas
& Luke, 1983; Olthoff et al., 1984).
hew
15
ws
e
ch
10 c
ew
s
ew s
ew
ch
25
ch
20
ch
s
ew
ch
80
30
% of undersize particles
100
60
median
particle size
40
20
0
0
10
15
Food particle size (mm)

Fig. 3.2 Real particle size distributions resulting from sieving spat samples of raw carrot
chewed by humans. The spacing of the mesh sizes is not even, but follows a geometric progression. The plots are cumulative frequency plots in which the y-axis reports the percentage
of particles in a distribution that are less than a given (sieve) size. Due to the variable skew of
the particle size distribution (shown in Fig. 3.1), the median particle size is the best indicator
of the degree of particle size breakdown.
Particle size distributions
59
some geometric simplicity to it. For example, we could imagine continually

halving the sizes of vegetable particles with a knife. Even though we could
not cut all of them with each chop, we might expect an overall geometric
progression in size to develop.
Epstein (1947) claimed that, if particles were comminuted for long
enough, then the eventual result would be a distribution that would be
normal if the logarithms of particle sizes were taken. In fact, such lognormal distributions do not materialize. It is more difcult to break smaller
particles than larger ones and limits can be reached at which the fracture
behaviour of foods changes. Provided that no such fracture limit is reached
though (for an explanation of which, see Chapter 4), then a large body of
data obtained from humans shows that distributions appear to drift towards
square-root normality (Fig. 3.3); (Voon et al., 1986). The same should be
expected for any mammal that reduces particle sizes far enough, such as
other primates (Walker & Murray, 1975; Sheine & Kay, 1979) or ungulates
(Rensberger, 1973).
It is useful to be able to characterize particle size distributions by a mathematical formula. The RosinRammler equation does this, providing a
BRAZIL NUT
1
RAW CARROT
median
particle size
0
BRAZIL
NUT
-1
RAW
CARROT
Frequency
Cumulative frequency (probit transformation)
Food particle size distributions after 30 chews

2
-2
Food particle size
-3
0.3
0.5
1.0
2.0
4.0
Food particle size in mm (square root of particle diameter in mm)
Fig. 3.3 Particle size distributions produced by one subject with raw carrot and brazil nuts.
After about 30 chews, a straight line is obtained when the square roots of particle sizes are
taken and when cumulative frequencies are converted to probit units. The latter convert
the sigmoid shape of normal cumulative frequency plots to straight lines (Finney, 1971).
60
reliable estimate of both the median particle size, the centre of the distribution, and its spread (Olthoff et al., 1984). Generally, as shown in Fig. 3.3
for brazil nuts and raw carrots, the faster a food breaks down per chew, the
broader its particle size distribution.
analysis of particle size reduction

Epstein (1947) developed a binary method of analysis of comminution
processes which has been of great help in understanding mastication because
it can help factor out the oral variables that inuence the rate of food
breakdown. The following denitions stem from Gardner & Austin (1962),
the only amendment being to replace per unit time by the concept of a
breakage step (per chew). Once you get the hang of these concepts, they
can be extremely useful. The rate of mastication depends on two factors:
(1) The chance that food particles have of being fractured by the teeth
during any chew. This is termed the selection function, S(x), where
S(x) x, for sufciently small x, is that proportion of particles of size
range x to x + x that are broken per chew.
(2) The size distribution of fragments produced by any particle that fractures. This is the breakage function, B(y, x), where B(y, x) x is that
proportion of the fragments by volume of size range x to x + x that
break to below size y per chew (where y x).
To illustrate what the selection function means, look at Fig. 3.4 where four
discrete particle sizes are pictured. Particles with crosses have been broken
(i.e. selected) in that chew. The number beside each particle grouping gives
the proportion of particles for each particle size that have been selected.
This is the selection function. Note that the smaller the particles, the lower
the value of this function. This unsurprising result has been found true for
every type of comminuting process including mastication (Lucas & Luke,
1983; van der Glas et al., 1987) where, for any given mouthful of food, the
selection function appears to be related to particle size by a power law:
S(x) = cx a where a is an exponent of particle size and c is a constant that
depends on the unit of measurement (Lucas & Luke, 1983; van der Bilt
et al., 1987; van der Glas et al., 1987). The best estimate of the exponent
from experiments on humans is a = 2.0, i.e. the selection function varies
roughly as the square of particle size. This relationship is indicated in the
graph at the bottom of Fig. 3.4. It should be emphasized though that this
is a statistical average and experiments need to be run several times before
this result is approximated.3
Analysis of particle size reduction
61
Fig. 3.4 The selection function a measurement of the probability of particles being fractured. See text for explanation. (Reprinted from Food Quality and Preference vol. 13, Lucas,
P. W., Prinz, J. F., Agrawal, K. R. & Bruce, I. M., pages 203213, Copyright (2002), with
permission from Elsevier.)
The breakage function is illustrated in Fig. 3.5 where, for simplicity, it is

assumed that a mouthful of particles, all of particle size x, has been ingested.
Twenty per cent of these particles have been selected in the rst chew (i.e.
S(x) = 0.2). These selected particles and their fragments are identied in
the gure by being crossed out. We can grade these fragments by assessing
their sizes relative to the size of the parent particles from which they broke.
This is commonly done on a cumulative basis. The gure shows a grading
level where a particle size y is half the size of x (i.e. y = 0.5x). Ten per
cent of the fragments that broke from x lie below y, i.e. B(y, x) = 0.1
for y/x = 0.5. If this procedure is repeated for other values of y, we can
build up the distribution of fragment sizes relative to x, the size of their
parents.
Measurement of the breakage function means sorting out which fragments break from which parent particle. The only direct method is to take
one particle, get a person to chew it just once and spit the fragments out.
This has to be repeated many times and the products sized. The results of
62
Fig. 3.5 The breakage function the description of the size distribution of particle fragments produced by a fracture event. The graph shows the t of one suggestion distribution
(GaudinMeloy) to fragmentation data from a human subject. Accurate measurement of
fragmentation requires a lot of repeats in an experiment to get stable results. (Redrawn from
Food Quality and Preference vol. 13, Lucas, P. W., Prinz, J. F., Agrawal, K. R. & Bruce, I. M.,
pages 203213, Copyright (2002), with permission from Elsevier.)
such tests often resemble Fig. 3.5 and can be characterized by two equations. The rst is by Gaudin & Meloy (1962), who claimed a fundamental
derivation relating the size of the parent particle and that of its fragmented
offspring, when the latter is from a single fracture. This relationship has
the form: 1 B(y, x) (1 x/y)m , where m is an exponent. The other is
Analysis of particle size reduction
63
an empirical suggestion deriving from Austin (Lowrison, 1974), whereby

B(y, x) xb , where b is an exponent.
The selection and breakage functions can be combined. In the example
taken above, the fraction of the total volume of fragments of this size range
x to x + x that break to less than their half their original size, i.e. to a size
less than y per chew, is S(x)B(y, x); in the example, this is (0.2 0.1) =
0.02 for y/x = 0.5. So 2% of the particles break into fragments that are
half the original size per chew. Since the rate of particle size breakdown
depends only on selection and breakage functions, then repeating this for
all size fractions computes the rate of particle size breakdown without all
that chewing, spitting and sieving after different numbers of chews.
Doing this by hand for all size fractions and repeating this for different
numbers of chews is not practical, so formal mathematics are necessary. If
the percentage of the total volume of particles of size range x to x + x
before the nth chew is Pn1 (x)x, then the percentage of particles below
size y after the nth chew is

y
Q n (y ) = Pn1 (x )B(y , x )S(x )d x + Pn1 (x )d x
y
(3.1)
where Pn1 (x )d x per cent of particles exist below size y before the
0
nth chew. The percentage of particles of size range x to x + x before
the (n + 1)th chew, i.e. Pn (x)x, can be obtained from
Pn (x ) =
dQ n (x )
.
dx
(3.2)
A solution to these equations is possible if the behaviour of the selection

and breakage functions with respect to particle size can be specied and
if these are not affected by the time that particles have been resident in
the mouth. This analysis has been widely adopted in industrial processes,
solving for the selection and breakage parameters using computers (Gardner
& Austin, 1962). However, this can now end because the problem has been
solved analytically by F. A. Baragar (University of Alberta), specically
for understanding mastication (Baragar et al., 1996). The t to real data
looks amazingly good.4 When the median particle size is taken as the
appropriate marker for the degree of size reduction of a batch of food in the
mouth, the rate of breakdown in human mastication after about 10 chews
follows an exponential curve, such that a loglog plot produces a straight
line.
64
Even if this method of analysis is valid, it still has to be shown that

it has some use. This can only be achieved by relating these functions to
characteristics of the food intake. In Chapter 1, I linked the surface attributes
of food particles with their probability of fracture (the selection function),
while the mechanical properties of foods that act to resist the formation
of new surface area, were tied to the degree of particle fragmentation (the
breakage function). I went further and linked tooth shape to the breakage
function and tooth size to the selection function, though I did not use
this terminology at that time. The logic of this is not difcult to grasp.
The mouth has a large surface area, not much of which is given over to
the dentition. Any increase in the size (surface area) of the teeth must
enhance the chances of a particle being fractured. This chance can only
depend on food particle characteristics that are exposed to the surface of
the teeth. Thus, evolutionary changes in tooth size in mammals are linked
to change in the surface characteristics of foods in the diet. In contrast,
evolutionary changes in tooth shape are likely to be a response to those
food characteristics that obstruct the formation of new surface area in food
particles by resisting their fracture. These characteristics are the internal
mechanical properties discussed in Appendix A.
the masticatory sequence
Any masticatory sequence consists of a series of chewing cycles in which the
lower jaw moves in orbits. In each cycle, the jaw rst opens to allow space for
food particles to be collected by the tongue. The jaw then turns and starts
to close. Before this happens, however, the tongue throws a set of these food
particles towards the teeth, usually (but not always) just to one side of the
mouth. The jaw then continues to close in order to fracture these particles.
After the jaw opens again, the fractured particles spill out, usually towards
the tongue. The postcanine teeth of mammals generally appear designed to
encourage this movement. If this spillage did not happen, or if food stuck
to the teeth, then because only a fraction of the particles in the mouth
get fractured in a chew, bimodal particle size distributions would quickly
become evident (one modal size representing unbroken particles, the other
the average size of comminuted fragments). None of the extensive data
from humans or the limited data for other mammals shows any evidence
of such bimodal distributions.5
Figure 3.6 shows the orbital path of the lower jaw, viewed from in front,
in the human. This movement pattern is not species-specic and could also
be found in a very large number of other mammals, including herbivores
The masticatory sequence
65
Fig. 3.6 The trajectory of any convenient point of the mandible (like a lower incisor) during
a chewing cycle. The path is highly variable, but the essential features are (1) the differing
path taken by the jaw in opening and closing and (2) the characteristic changes in jaw
velocity at different points in the cycle.
(Hiiemae, 1978) and insectivores (Fish & Mendel, 1982). However, it would
be an impossible movement for carnivores, which have a movement path
that is almost entirely vertical, and it is not typical of rodents and elephants either, which tend to have a distinct forward component to their
jaw trajectory instead.
In most mammals, the opening path is mainly vertical (e.g. in humans:
Mongini et al., 1986), but early in closing, the jaw turns laterally towards the
side on which food particles are to be fractured. During further closing, the
mandible starts to turn again, directing its thrust upwards and medially.
The mandible may then pause with the teeth together, or nearly so, or it
may slide a small distance rst across the upper teeth, before setting out on
another opening journey. The cycle is rapid and takes less than a second in
the human. However, it can be up to four times faster than this in other
mammals (Table 5.2).
The timing of chewing cycles is relatively inexible, being controlled,
just like locomotion is, by neural circuitry called central pattern generators
(Lund, 1991). These generators are circuits of neurons within the brain stem
that set body rhythms (Luschei & Goldberg, 1981). Everything other than
the timing though is very variable. Indeed, Fig. 3.6 does not have to be drawn
very exactly, because it is based on an idealized teardrop shape of chewing
cycle that is actually rarely seen. This highlights a major difference between
66
mastication and locomotion, the two major motor activities of mammals.

Terrestrial locomotion is more stereotyped in form. For example, the gait of
a terrestrial tetrapod varies little whether it is walking over hard rock or soft
sand. If it did vary greatly between these substrates, then fossil footprints
would provide a biased record of the movements that these animals made;
instead they supply crucial raw data for analysing their pace of life.6
In contrast, movements in every masticatory cycle vary between chews.
This is not likely to be due to the mouth having inadequate mechanisms
to know where food particles are, or to some jerky motor program driving
the muscles, but to the fact that the substrate that the mouth works on
is changing as it is (deliberately) being broken down and subject to the
statistical patterns described by selection and breakage functions.
For the purposes of description, each chewing cycle is usually broken
down into phases. Jaw opening and closing are obvious subdivisions, but
each opening and closing phase can be further divided up by differences
in jaw velocity. Not every mammal shows this to an equally clear extent,
but a slow open fast open fast close slow close pattern of movement
in a chewing cycle has been described now in many mammals (Hiiemae &
Crompton, 1985).
During early jaw opening (slow open), the tongue moves forward
rapidly so as to catch food that tends to fall from the lingual sides of
the teeth after jaw closure. The tongue may dip its tip into the pool of
submandibular saliva that has accumulated behind the lower incisor teeth.
Picking this saliva up, somewhat in the manner of an intra-oral lap, the
tongue then mixes the food particles with saliva by pressing them against
the hard palate. The considerable amount of sensory processing necessary
for this probably dictates the slow speed of this part of the cycle. The tongue
then releases its pressure on the particles and starts to move backwards. As
the food falls from the hard palate, the tongue senses whether particles stay
clumped together or else fall apart. If the particles all fall apart, the tongue
throws the food laterally into the space between the upper and lower postcanine teeth. The positions of the tongue during jaw opening are depicted
in Fig. 3.7.
Foods that fall too far laterally seem to bounce off the cheek as the buccinator contracts (Blanton et al., 1970), so reecting them, in a damped
fashion, back onto the teeth (Prinz & Lucas, 2001).7 Before the jaw turns
and starts to close, of course, the cheeks must start to relax. Then, with
both the tongue and cheeks out of the way, the teeth close rapidly towards the food particles. Close to or at toothfood contact, the mandible
slows, probably allowing perception of food fracture properties. As the food
Sensory feedback
67
Fig. 3.7 Probable positions of the tongue during early and late jaw opening. In (a), the
tongue is collecting food particles in early opening, while in (b), it is throwing them laterally towards the teeth. (After Abd-el-Malek, S. (1955) The part played by the tongue
in mastication and deglutition. Journal of Anatomy 89: 250254 (Oxford: Blackwell).
With permission.)
fractures, the nearby parotid duct starts to spray saliva over the newly formed
fragments, helping to wet their surfaces (Anderson et al., 1985). Then as the
whole cycle starts again, these fragments fall towards the tongue.
sensory feedback and the control of mastication

This complicated sequence of events cannot have an inexible neural program because the process depends critically on sensing where food particles
(and soft tissues like the tongue and cheeks) are (Fig. 3.8). Such knowledge
should not be a problem. Animals are exquisitely well organized for the
detection of information about their external environment, so much so
that some mechanisms, such as those for perceiving light or chemicals, operate at the level of individual photons and molecules respectively (Bialek,
1987). However, current knowledge indicates that mechanisms for detecting the state of the oral environment fall far short of those dictated by the
laws of physics. Instead, natural selection seems to have acted to assign the
greatest sensitivity to two critical areas: the detection of swallowable particles so as to clear the mouth and to the avoidance of damage to the teeth
by wear.
There are many sensory receptors in and around the mouth, as will be
evident below, but they can be categorized into those that supply information about jaw displacement (and the rate of its displacement) and those
that monitor stress (Fig. 3.8). The overall pattern of control appears to
be displacement dominated, with the aim being to keep the jaw moving.
However, when there is contact, or anticipated contact (Ottenhoff et al.,
68
Fig. 3.8 The sources of sensory information available in and around the mouth considered
in relation to food properties and the direction of motor responses.
1996), a stress control creeps in. The latter part of a cycle is associated
with a slowing of movement, even with pauses (Hiiemae & Crompton,
1985). A stress control seems to pervade movements during the entire
chewing cycle when foods of high Youngs modulus are chewed (Bourne,
1976).
The events during jaw opening are critical for the success of a chew.
Early opening is slow and there is strong evidence for this being mediated
by sensory receptors (Ostry et al., 1997). It is not clear where these receptors
are neuromuscular spindles (see below) are almost absent in jaw-opening
muscles. Yet this slow speed is a sure sign that food properties are being
evaluated. Since there is no fracture involved in opening, it must be surface
attributes that are being evaluated on the oral mucosa.
Sensory feedback
69
The tongue manipulates food particles in this phase of a chewing cycle

and touch sensors on both hard palate (Dixon, 1963) and tongue are probably vital in this evaluation. The two-point discrimination (a test for the
stereo sense of touch) of the most sensitive areas of the oral cavity of the
human, the anterior hard palate and tongue tip, is about 12 mm (Ringel
& Ewanowski, 1965; Laine & Siirila, 1971). In other words, particles smaller
than this that rest on the mucosa cannot be discriminated in size. This is
very likely related to the general propensity of humans to swallow particles
in the low millimetre size range (Lillford, 1991). A swallowing decision that
is related to particle size makes sense of the sensitivity range. If 12 mm
particles are those desired for digestion, then there is no advantage in detecting the exact size of smaller particles since a swallowing decision will
already have been made. There is considerable circumstantial evidence for
the operation of a low-millimetre oral threshold in a number of mammals,
including rodents (Magnusson & Saniotti, 1987; Corlett, 1996), tree shrews
(Emmons, 1991; Emmons et al., 1991), cercopithecine Old World monkeys
(Corlett & Lucas, 1990; Lucas & Corlett, 1998; Otani & Shibata, 2000;
Su & Lee, 2001) a fruit bat (Phua & Corlett, 1989) and many herbivores,
such as cattle, sheep, goats and horses (Uden & Van Soest, 1982; Van Soest,
1996). There is also experimental evidence for this in humans (Prinz &
Lucas, 1995).8
As the jaw turns to close, the displacement sensors that are likely to be
most important are those around the capsule of the jaw joint (Klineberg
& Wyke, 1973; Klineberg, 1980). At wide gapes, these receptors seem to
gauge movement patterns accurately, but at smaller gapes they fail to do
so (Owall,
1978). This is understandable because these sensors are set very
close to the pivot. Accordingly, as the jaw closes further, other sensors must
monitor its pattern of movement. The neuromuscular spindles in muscles
are undoubtedly important and organized in mammals such that they can
sense the state of muscular contraction at a wide range of gapes (Matthews,
1972).
Closer to contact with food particles, and to upper and lower tooth
contacts, force sensors are also required in order to monitor fracture
events. The evidence is that mechanoreceptors in the periodontal ligament
(Linden, 1990) initially act in this way. An additional mechanism is undoubtedly provided by dentine sensitivity (Paphangkorakit & Osborn,
2000). As pointed out by these authors, the periodontal ligament is rather
pliant, with the result that its sensitivity must be limited with stiffer foods.
At higher stresses, these authors provide strong circumstantial evidence
for an intra-dental (pulpal) sensor. As there are no specialized receptors
70
in the pulp (Holland, 1994), it is probable that uid movement in dentinal tubules under the stressed region stimulates the nerve endings. (This
reasoning requires that the modelling of Spears et al. (1993) of force transmission from enamel to dentine be correct; see Chapter 2.) However, wherever the mechanisms are located, the ability to sense a particle between the
teeth is very ne. Depending on method and on the particular teeth being
tested, particles of only 815 m in thickness or diameter can be detected
(Utz, 1986). This distance is only a little above the sensitivity of individual
neurons to displacement (23 m: Anderson et al., 1970). Undoubtedly,
this ne sensitivity is related to the need to avoid dental wear. Most of
the wear is produced at this micro level, either by extraneous grit ingested with food (Teaford, 1994) or by the opaline silica contained in many
plant species (Baker et al., 1959; Walker et al., 1978; Lucas & Teaford,
1995).
Slowing of movement during late closing is not likely to be connected
just with perception of food fracture properties, but with avoiding damage
to the teeth. This is generally more difcult when a device, be it articial or
organismal, is functioning under stress control.9 Tooth contacts in human
mastication are well known in both chewing and swallowing (Anderson,
1976). In order to avoid tooth wear, electrical activity in the jaw closing
muscles has to be switched off as soon as jaw movement is detected as being
impeded. A reex (which on the sensory side seems uniquely to contain only
one neuron running all the way back to the midbrain without a synapse)
seems to do this about 12 ms after toothtooth contact (or a similar period
after contact with very stiff foods), but due to a forcetime lag, the force
invariably builds up for an average of 40 ms beyond this (Gibbs et al., 1981).
So there is at least 50 ms in which tooth wear can take place.
Figure 3.9 shows how the basic events in a chewing cycle are related
to food properties. There is no fracture during early jaw opening and
so palpation of food with the tongue can only detect external physical
attributes. Sensations during opening regulate the vertical amplitude of a
chew. The internal mechanical properties of foods can only be detected
during late jaw closing during fracture. These may affect the degree of
lateral movement during closing.
There is one extra source of feedback (mentioned in Fig. 3.8) that has
been neglected by oral physiologists, though not by food scientists, and that
is sound. Fast fracture in food produces noise (Drake, 1963, 1965; Vickers,
1981) that can be interpreted by subjects in terms of specic food properties,
as shown by Vincent et al. (2002). The effect of this on jaw movements
though appears unknown.
Jaw mechanics
71
Fig. 3.9 The relationship between food properties and the movement sequence during a
chewing cycle. (Derived from Lucas et al. (1986e) and Agrawal et al. (2000).)
jaw mechanics and the control of jaw muscles

Of all the sensory receptors discussed above, those with the smallest role
appear to be the joint receptors. Yet they seem to have substantial innervation. What are they there for? Well, there is another role that could be
accorded to these receptors that relates to a general problem called jaw
mechanics and which was alluded to in Chapter 2 (Fig. 2.16). The issue
can be stated succinctly:
The jaw is a lever, anchored at a joint and moved by muscular efforts
acting at xed locations. Yet the jaw is also loaded by contact with food at
various positions along its length and at a range of opening angles. How
can the force of loading, the bite force, be organized so as to minimize the
effort or without being transmitted through the joint?10
The bite force that the food endures has to be distinguished from the
resultant of the muscular forces moving the jaw. However, except when
72
chewing very stiff foods (Rensberger, 2000), the distinction between the
two is unimportant. From now on, I will mainly refer to bite forces, but the
directions of both are almost certainly going to be approximately parallel to
the long axis of the teeth, otherwise they would start to tip. Some imbalance
is inevitable due to change in the muscle force vector with gape, but any
inconsistency needs to be harnessed so as to produce a protable result, e.g.
by keeping teeth in a single functional group together.
A compressive force through the joint can be countered by building up
the bones that form the joint so as to prevent total catastrophe: the head
of the mandibular condyle breaking through into the cranial cavity. The
joints of some mammals, e.g. those of carnivores, are hinges, so limiting
any possibility for anterior dislocation, but others are capable of translation and rotation as in humans. Whatever their form, none appears to
be built so as to sustain high compressive forces. A greater problem than
compression though is the possibility of a force that could separate the
bones, a distracting force that moves the mandibular condyle downwards,
so rupturing the joint capsule. Such a force could be detected by stretch
receptors in the capsule, which may well provide feedback to the muscles
affecting their ring patterns. Some of these receptors may be able to sense
compression at the joint as well, e.g. those in the lateral ligament of the
capsule in humans, a structure that seems to be taut when the teeth are
together (Baragar & Osborn, 1984). In this regard, joint receptors could
be analogous to mechanoreceptors in the periodontal ligament: they need
to sense compression, but actually respond themselves to being stretched
(Linden, 1990).
The ring patterns of muscles are not stereotyped, but it is unclear what
the efciency limits for muscles are and what their neural control might be
optimizing. The smallest order of physiological organization is the motor
unit, which is the number of muscle bres controlled by a single nerve
bre. However, even if muscle recruitment could be organized that nely,
then there would still be light compressive loads at the jaw joint (Osborn &
Baragar, 1985).11 So what is behind the variability of muscular contraction
seen physiologically? Osborn & Baragar suggest that the muscles may be
acting to produce the maximum bite force at minimal cost in terms of
bre contraction. This approach has produced some surprises: the pull of
the lateral pterygoid, probably the superior head of this muscle for the
most part, seems important in generating an efcient bite (Osborn, 1995b).
Spencer (1998) has recently reviewed the subject as a whole, conducting
experiments on humans that test one of the most original models of jaw
Saliva and so many senses
73
function, that of Greaves (1978), but there appears to be no current model

that can account for all observed patterns of muscle activity.
The internal architecture of the jaw muscles is variable. In some muscles,
the bres lie at an angle to the bite force. This packs more bres into a given
volume than parallel bre arrangements, but reduces the gape at which the
muscle can act. The masseter is a key example of a muscle like this (Anapol &
Herring, 2000). Also, the ring pattern of the bres themselves is variable.
They may contract rapidly, but fatigue quickly or vice versa. An important
review by Mao et al. (1992) has reviewed the distribution of these bres
in different jaw muscles and across mammals in different dietary groups.
The muscles of carnivores and rodents tend to be fast acting but quick to
fatigue, while those of herbivores are the opposite.
A nal aspect of this is that muscle tissue does not generate force
equally well at different amounts of stretch (Herring & Herring, 1974).
Paphangkorakit & Osborn (1997) show in humans that this has a sizeable
effect on the jaw-closing muscle that acts most efciently to produce a bite.
At smaller gapes, it is the masseters, while at wider gapes it is the temporalis muscles. Mammals that use wide gapes, like carnivores, have heavily
built temporalis muscles, while herbivores use much smaller gapes and have
relatively larger masseters (Fig. 3.10).
saliva and so many senses
This section discusses the range of stimuli, including tastes, smells and
abrasives that inuence the rate of salivation and the properties of saliva
itself.
The best time to monitor the quality of the potential food intake is as
soon as possible. For this reason, all the special senses of the body operate
before, and during the oral processing of food. Smell, taste and texture
are all enhanced and/or facilitated by the exposure of new food surface by
fracture and fragmentation during mastication. While odours follow the
local airow to the nasal cavities (probably passing backwards and around
the soft palate), the most important chemical sense in the mouth is surely
that of taste. Taste requires the presence of saliva because only dissolved
compounds trigger it. Consequently, it is unsurprising that as food particles
are broken down, several sensory mechanisms have developed to pour saliva
over them. Saliva is stimulated by the physical presence of foods against
the mucosa, by food particles being loaded by the teeth (Anderson et al.,
1985), by taste (Spielman, 1990), by odour (Lee & Linden, 1992) and by
74
Fig. 3.10 The jaws of a herbivore (above) and a carnivore (below) to show the relative sizes
of the masseter (M) and temporalis (T). The sizes are related to the gape angle at which
muscles act.
abrasives (Prinz, in press). A good review of much of this is by Hector &

Linden (1999).
Tastes have been categorized traditionally as sweet, sour, bitter and salty.
However, this view is breaking down due to a revolution in research on this
topic. I have no option but to review this here and so need the following
short, but necessary, interlude to the text.
The sensation of saltiness is triggered by sodium and potassium ions
(Boughter & Gilbertson, 1999). It is difcult to relate to any other aspect of
this book, except with respect to geophagy the consumption of special soils
and minerals by mammals that can cause considerable dental wear. There
is no evidence that calcium or phosphate ions in foods can be detected
by any specic mechanism, despite the bodys requirement for these in
mineralized tissue. The sensation of sourness is due to low pH. Food acid
erodes enamel. Quite logically, therefore, it seems to be the most potent
known stimulant for saliva (Kawamura & Yamamoto, 1978).
75
Much of the recent work on taste though involves molecular discoveries

about sweet, bitter and free amino acid taste receptors. Most of it derives
from work on rodents, but there is now increasing evidence for it on human
taste cells too. Where both human and mouse tissues have been tested, the
response is not 1 : 1 because the relevant genes in both species share only
about 70% similarity. However, this is a lower gure than for comparisons
of most other human/mouse proteins, conrming that taste does adapt to
diet (Nelson et al., 2001).
Food sugars are important in the human because of their clinical significance in dental decay. They are detected by a receptor or receptors that
give rise to a sensation that we refer to as sweetness. While sweetness is
an attractant for mammals with a non-fermenting digestive system, it may
repel ruminants for which sugars jeopardize microorganisms in the fermentation chamber of their stomachs. The receptor is useful for both types of
mammal, but it probably evokes contrary psychological responses. Sweet
taste reception in mice and humans usually seems to involve the conjoint
action of two genes (Nelson et al., 2001). A wide range of sweet chemicals
can be perceived, including sucrose and fructose, both of which produce
a strong response. These sugars are major indicators of the carbohydrate
content of plant foods. Sucrose is the ubiquitous transport sugar of plants
while fructose is the commonest sugar in fruits. However, not just sugars
are sweet many d-amino acids are as well (Nelson et al., 2002). It is not
clear how important this is because the d-isomer is not involved in protein
formation. Some proteins themselves are also very sweet, but this is rare.
More importantly, not all sugars excite this receptor. Some sugars, including glucose, galactose and maltose, produce no response from this receptor
and yet are perceived as somewhat sweet by humans and other primates
(e.g. chimpanzees: Hellekant et al., 1996). Glucose is the transport sugar of
animals, but it is also common in fruits because the enzymatic breakdown of
one sucrose molecule produces one fructose and one glucose molecule. On
current evidence, it appears that glucose must be sensed by some other, as
yet unknown, mechanism.12 Recent evidence suggests something of the nature of this second receptor.13 Rats can taste oligosaccharides (longer chains
of saccharides) that are derived from starch and contained in commercial
corn starch syrups (Sclafani, 1991). Laska et al. (2001) have recently investigated the attraction of these corn syrups to four primates, one of which, the
pigtailed macaque (Macaca nemestrina), actually preferred these oligosaccharides and maltose to simple sugars. The most parsimonious suggestion
at present is that glucose, galactose, maltose and some oligosaccharides are
all detected by a second (unknown) sweetness receptor. Confusingly, both
76
Sclafani (1991) and Laska et al. (2001) employ the term starch receptor
but this seems unwarranted. Starch itself is not tasted and there seems little
point for most mammals to taste a small proportion of starch breakdown
products when their gut will not be able to digest the vast bulk of it. Only
ruminants are likely to manage this and rats and macaques are not ruminants. This point is explored further in relation to leaf eating and salivary
amylase in Chapter 7.
Free essential amino acids in foods can be assimilated without digestion
and one would think that mammals would crave them. Yet only recently
has it been discovered that they are sensed on the tongue by a specic
mechanism. Proteins are built from l-amino acids and only this optical form
triggers the taste receptor (Nelson et al., 2002).14 Umami, the taste produced
by monosodium glutamate, might be transducted by this mechanism, but
no one as yet knows. Most people seem to believe that the detection of amino
acids in the mouth acts as a proxy for the foods protein content. However,
some fruits, such as gs, have large free amino acid concentrations, but no
protein (N.J. Dominy et al., unpubl. data). Furthermore, of the hundreds
of amino acids known, only a few are essential. Many of the others may
actually be toxic, produced as a (probably infrequent) form of defence in
young leaves (Coley & Kursar, 1996). Nothing is known about how the
amino acid receptor responds to these compounds.
Aside from their normal metabolism, plants produce a wide variety of
compounds that appear to have no other specic role than to prevent
animals predating them. Some of these compounds are toxic in extremely
low doses. They are mostly detected via bitterness, which depends on a
large family (between 40 and 80) of receptor proteins, many of which can
be found on a single taste cell (Chandrashekar et al., 2000). A protein
called gustducin also apparently needs to be present for this taste pathway
to function. The range of receptors seems to parallel the range of so-called
qualitative plant secondary compounds. Qualititative here refers to the
devastatingly powerful effects of some of the compounds, e.g. the alkaloids,
some of which (e.g. ricin) can kill a cell with one molecule.
Some fatty acids are essential in the diet and there is now clear-cut
evidence in the rat for a fatty acid detector in the tongue. This is not a
receptor per se, but makes use of the ability of fatty acids to block an open
ion channel (e.g. Gilbertson, 1998). The response is quick and specic
and it has been shown using rats of different strains that preference for
dietary fat is related to sensitivity to free fatty acids (Gilbertson et al., 1998).
It is probable that humans have this detection mechanism. Tittelbach &
Mattes (2001) have shown from blood tests that exposure to fat in the
77
mouth boosts its absorption in the blood several hours later to levels far
above that of controls. The experiments are complex, but the only sensible
interpretation is that fats are detected in the mouth, very probably by
Gilbertsons mechanism, leading to an enhanced response later on. Fat
substitutes can have the same texture as fats, but without those fatty acids
being present there is no response (Mattes, 2001).
Lipids as a whole are the most energy dense of foods and can also
be detected texturally in humans as creaminess for example (Lermer &
Mattes, 1999). Rolls et al. (1999) have shown in an Old World monkey,
Macaca fascicularis, that a distinct region of the orbitofrontal cortex of the
brain responds to the texture of fats in the mouth. This neural response is
denitely textural because the cells respond to non-absorbable lubricants
in the same way. This is likely to be a tactile response to a lowering of
intra-oral friction (the opposite of that induced by tannins, which are discussed next).
Tannins are plant compounds that bind to proteins, generally forming
large molecules that may precipitate out of solution, so rendering the protein useless (Waterman & Mole, 1994). If these tannins react with digestive
enzymes, then the consequences for digestion can be disastrous because
the binding will put those enzymes out of commission. In low doses,
tannins inhibit digestion, but in higher doses, they are downright toxic.
Usually, tannins are viewed as defences against predation (Dominy et al.,
2001), but there is now some evidence that limited doses of them (usually
under the banner of polyphenols) in the diet improves human health.
For example, they can benet ruminants by speeding their digestion up
(Aerts et al., 1999). Yet large doses of this very diverse range of chemicals are
denitely toxic (Waterman & Mole, 1994) and levels of ingestion need to be
monitored.
About 70% of the proteins in human saliva, largely proteins found in
the output of the parotid and the submandibular glands, are rich in proline
(Edgar & OMullane, 1996). This amino acid is not usually well represented
in protein structure (if we discount those proteins involved in enamel formation, that is) and it is not entirely clear why it is there. It might be to
protect enamel structure because some of this protein adsorbs onto the
enamel surface (Jensen et al., 1992). Enamel has certainly been shown to
demineralize very rapidly in deionized water (Habelitz et al., 2002). However, it is (in my opinion) impossible to explain both the quantity and
the diversity of these proline-rich compounds simply by this human-based
explanation. Alternatively, as is well known, most bind strongly to tannins and some seem to have no other demonstrated role. The result of the
78
binding with tannins is an oral sensation often described as astringency

or a feeling of intra-oral dryness. This feeling is probably due to loss of
lubrication. While it may be felt as dryness, this is more likely to be a consequence of an increase in intra-oral friction (Green, 1993; Prinz & Lucas,
2000). At high tannin concentrations, this increase could simply be due
to glycoprotein being lost from the saliva. However, at low concentrations,
Prinz & Lucas (2000) and Dominy et al. (2001) have suggested that the
adsorbed protein coat on the enamel (the pellicle as it is called) could react
with them. This forms an ideal way of detecting tannins with the tongue,
since the roughness produced on the tooth surface following a tannin reaction is readily identiable, as when drinking red wine, for example. Why
then do some humans appear to like a (very modest) tannin intake? One
possibility is that people have always known what scientists have recently
begun trying to prove that polyphenols can be good for you. A more
credible explanation though is that of Prinz & Lucas (2000): food particles
or lipid lms attached to the oral mucosa by saliva might be dislodged
by a mouthwash of tannins binding to the salivary glycoproteins. This is
much like the use of tea in a Chinese banquet as a type of cleansing
agent.15
Increased friction, either toothfood or toothtooth, could increase the
rate of tooth wear. The addition of grit to chewing gum results in increased
salivation (Prinz, in press). This could be an evolved response to reduce the
potential wear threat from the grit, analogous to the way that ooding the
mouth with bicarbonate ions prevents loss of mineralized tooth tissues by
dissolution (erosion).
the decision to swallow
At some point during chewing, a batch needs to be swallowed so as to
allow a further batch to be ingested. Chewing for too long jeopardizes the
rate of energy acquisition: there can be no catching up later if further food
input is excessively delayed (Alexander, 1994). Consequently, it is logical
to presume that the duration of the ingestionmasticationswallowing
segment of the digestive sequence is optimized for the maximum digestive effect (Alexander, 1991). Accordingly, a physiological trigger should
be anticipated that could tip the central nervous system off so as to initiate a swallow at the best moment. There are some problems, however,
about generalizing from human swallowing experiments to the physiology
of other mammals. When humans swallow, we generally collect food in
our mouths until we initiate a swallow voluntarily.16 Food is then forced
Theories of swallowing
79
rapidly down the pharynx by a piston-like movement of the tongue. Before

the food reaches the pharynx though, the raising of the soft palate and the
larynx seals off upper and lower respiratory tracts respectively. If the process
is mistimed, such that either breathing is in progress or that food particles
do not move entirely as a group, then it is possible to send some particles
down the lower respiratory tract, so causing choking. In contrast, other
mammals have their larynx set much higher than in humans (Thexton &
Crompton, 1998). The change in humans appears to have been to facilitate
vocal communication. In other mammals, the epiglottis can connect with
the soft palate, such that it is possible for them to breathe while swallowing.
Their upper and lower respiratory tracts may connect in the midline, allowing food particles to slip down to the oesophagus around this connection
without jeopardizing respiration (Hiiemae & Crompton, 1985).
I take the view that, despite differences in gross anatomy, essential features
of the swallowing process, in particular the sensory trigger or triggers, would
be similar across mammals and that experiments in humans are relevant
to these features. Humans initiate swallowing voluntarily. Is there a trigger
for this or is it just habit? Studies in humans show that the point at which
swallowing starts is not random. For example, if one mouthful of food is
given repeatedly, the time to swallow in a given subject varies little between
trials. This could simply follow habit people may chew a certain number
of times and then swallow. However, if so, then this is a habit that can be
manipulated rather easily because simply changing the size of the mouthful
of any food (the food volume) changes the number of chews (or the time
taken) before swallowing and by a very substantial margin (Lucas & Luke,
1984a). Rather than habit then, it is more logical to suppose that something
about the state of the food is being detected.
theories of swallowing
Swallowing is a clearing process that is initiated voluntarily, but later
becomes reex once food particles have been cleared from the tongue
(Thexton & Crompton, 1998). For optimal efciency, the process should
move chewed food particles as a set from the mouth down to the stomach
in one batch. Food clearance lower down the gut is not a problem because
there is surplus uid resulting from glandular secretions and these ush
food particles along from one step to the next. In addition, processing in
the stomach turns food into slurry, which in itself facilitates transport: peristaltic movements of the gut move pliant material much more easily than
stiffer stuff (experiences with diarrhoea versus constipation are informative
80
in this regard). The mouth though is more of an earth closet than a water
closet: uid is in short supply.17
Particle size threshold model
A threshold in food particle size during chewing has been the most common suggestion for a trigger. Although there is some evidence that food
particles are generally reduced to a low-millimetre size range by the time
that swallowing occurs (Lillford, 1991; and see above), experiments and
surveys show that the decision to swallow is not triggered by particle size
directly. In particular, evidence from Yurkstas (1965) shows that there is an
inverse correlation in humans between particle size reduction rate and the
size of particles at swallowing. Humans with full dentures, who have very
slow particle size reduction rates, often swallow extremely large particles.
In experiments on subjects with natural dentitions, changing the mouthful
not only changes the number of chews to swallow, it also changes the particle sizes that are swallowed. A very small mouthful is always swallowed at
smaller particle sizes than larger ones (Lucas & Luke, 1984a). It appears that
the rate of food particle size reduction is probably not being sensed. It does
not have any obvious end point anyway and could continue indenitely
(were it not for deformation transitions Chapter 4).
Dual threshold model
Hutchings & Lillford (1988) made an important conceptual advance by postulating that swallowing might depend on a combination of two thresholds:
particle size and particle lubrication. Both of these could involve sensory
criteria that have to be satised in order for swallowing to take place. The
authors offered qualitative descriptions of what they refer to as the breakdown paths of various foods. For example, an ingested raw oyster satises
both thresholds on entry to the mouth and could be swallowed immediately. In contrast, a slice of a mandarin orange, while not being reduced
in size by chewing, needs sufcient juice expressed from it in order to be
swallowed. The slice satises the particle size criterion immediately after
ingestion, but not that of lubrication, so it is chewed until its surface is
lubricated by saliva and juice. A peanut appears to have to be chewed both
to reduce its size and to lubricate it.
These examples indicate that, if there is a particle size threshold, then this
depends on the food in some way: oysters and mandarin slices are much
larger than peanuts. Although Hutchings & Lillfords study was based on
81
SLOW
Exploratory chews
Two-threshold swallowing model
immediately
swallowable
FAST
Normal comminuting chews
Average chewing frequency
many years of practical experience, it was qualitative in nature and presents

considerable problems about how to detect these thresholds. Nevertheless,
Prinz & Lucas (1995) devised an experiment involving a varying liquid : solid
mix with just that intention: to try to make a putative food particle size
threshold move independently of a lubrication threshold. Subjects were
offered a spoonful of brazil nut particles suspended in yogurt. In these
solidliquid mixtures, both the percentage of brazil nuts by volume and
the size of the brazil nut particles were varied independently. The overall
volume of the mouthful was kept constant. The experiment also included
yogurt without any nuts in it as a control and the whole experiment was
performed blind both to the subjects and experimenter, with the mixtures
being coded and given in random order. The number of chews, if any, taken
by the subjects before swallowing and the time taken to make those chews
were both recorded.
The results of this experiment can be described by reference to the
schematic diagram in Fig. 3.11. When concentrations were high and particles
were large, the standardized chewing rhythm kicked in almost immediately.
particles
too big
too dry
A
B
C
D
Solid:liquid ratio
Fig. 3.11 Average chewing rates on nutyogurt mixtures. Variation in the speed of chewing
movements is related to difculty in detecting small particles. The solid lines describe
experimental data from human subjects, each referring to a single particle size where A
is smallest and D largest. When the ratio of nuts to yoghourt was high, subjects chewed
at average speeds, but when it was low, subjects began to explore the mix for particles,
so slowing their average chew speed. The dotted and dashed lines describe the suggested
particle size and particle lubrication thresholds. See text. (Reprinted from Archives of Oral
Biology vol. 40, Prinz, J. F. & Lucas, P. W., pages 401403, Copyright (1995), with permission
from Elsevier.)
82
However, when concentrations were low, and particularly when particles

were also very small, chewing motions were very slow. These are called
exploratory chews in Fig. 3.11. Ordinarily, chewing proceeds with a unchanging rhythm and it is only when exploratory manoeuvres of some sort
are required that this rate decreases. Three domains are indicated, controlled by two thresholds, one for particle size and the other for particle
lubrication. The solid lines indicate results for similar particle sizes in varying concentrations, where A is the smallest size and D, the largest. The
vertical axis describes the rate of chewing, where faster chews are closest to
the origin and slower chews higher up.
Superimposed on the results in Fig. 3.11 are boundaries that could demarcate HutchingsLillford thresholds. The experiment could be viewed
as support for these thresholds. That for particle size comes out very close
to the expected range (1.4 mm). However, there seems to be far too much
uid in the mix (4 parts liquid : 1 part solid) to explain how a ball of peanut
fragments, for example, can be swallowed safely.
Bolus model
The search for another model for swallowing is also due to Hutchings &
Lillford and concerns the nature of the food bolus and the forces that could
stick food fragments together. Lillford (1991) illustrates bolus formation
by showing a sequence of cooked meat samples that had been chewed for
different numbers of chews by one human subject and then spat out onto
a dish. Early in the sequence, the single beef particle starts to be broken
up into fragments. Later, these fragments begin to stick together to form
a bolus, even though the fragmentation process is continuing via further
chews. At the point at which the person wanted to swallow (but was instead
instructed to spit), all fragments were essentially stuck into a single ball.
When the subject was asked to chew on further, rather than swallow (in
this example, for 50 chews, when wanting to swallow after 30), the spat
sample then consisted of myriad separate fragments oating away in saliva.
The study, which was not quantied, seems to show that there might be
a unique point at which a bolus will form: if a swallow is too early or too
late, particles might be left behind.
Prinz & Lucas (1997) attempted to produce a quantied model of the
forces involved in bolus formation. As food particles are reduced in size and
simultaneously wetted, the forces in the mouth tend to make food particles
either aggregate or to stick to the oral cavity. Surface tension and viscosity
(Cottrell, 1964) were considered to provide the relevant forces and, in order
83
to keep the theory simple, the model was constructed in two dimensions.
It was also limited to food particles with spherical shapes, so that what
applied to one two-dimensional plane through a ball of particles would be
presumed to apply to any other. Also, the foods were assumed to release
negligible uid when they were fractured and, thus, that the only wetting
agent to consider would be the saliva.
The mucosal lining of the oral cavity is always lubricated by a mucus-rich,
salivary lm. This lm is ever-present and does not reduce in thickness after
swallowing (Edgar & OMullane, 1996). A spherical food particle entering
the mouth could easily be attracted to the lining by a surface tensional force
given by:
F A = 4r
(3.3)
where r is the radius of this food particle and is the surface tension of this
salivary lm (Fig. 3.12a). As can be seen immediately from this equation,
the larger the food particle is, the larger that the adhesive force will be.
After food particles are broken, a spray of saliva from the parotid gland of
the same side as that on which they are chewed wets them. The particles
then fall back onto the tongue, which has just collected submandibular
secretions. The tongue then presses the particles against the hard palate
during the early part of the opening phase of a chewing cycle. Imagine
that these spherical particles are pressed together into a ball. If the particles stick together at this point, it is likely to be because of a viscous
Fig. 3.12 (a). A spherical food particle sits on the oral mucosa, adhering to it initially by
virtue of a surface tensional force created by a mucus-rich salivary lm. (b) A section through
a spherical ball of food particle fragments, packed by the tongue, but separated by saliva that
inltrates the inter-particular space. This ball will be held together by a force deriving from
viscous adhesion. (Reprinted from Food Quality and Preference vol. 13, Lucas, P. W., Prinz,
J. F., Agrawal, K. R. & Bruce, I. M., pages 203213, Copyright (2002), with permission
from Elsevier.)
84
force that acts at very close range. The force required to separate these
discs is
F V = 3 D4 /64d 2 t
(3.4)
where is the viscosity of the oral uid lling the spaces between the
food particles, D is the radius of the ball, t is the time-span over which
the separation takes place and d is the average distance between particles
(Fig. 3.12b).
The prediction of the model is very simple. Bolus formation should
begin when the net cohesive force, FV FA , is greater than zero and the
optimal period to swallow is when FV FA is maximized because this would
be the safest time, with the least chance of leaving particles behind.
Testing the model requires particle size reduction rates for foods. Prinz
& Lucas chose brazil nut kernels and raw carrot because they appear to
represent the extremes of breakdown rates in the human mouth. Turgid
carrot particles break down slowly, while brazil nut breaks down quickly
(Lucas & Luke, 1983; Lucas et al., 1986d). Selection and breakage functions
for these foods were available and inserted in the comminution equations
given earlier to produce particle size distributions. The number of chews
taken to swallow and the particle sizes at swallowing were known for a large
number of subjects too (Lucas & Luke, 1986), allowing a full test of the
model. The last component of the model was the most difcult: how to
simulate the packing of particles by the tongue. A computer model was
built specially to do this (detailed in Prinz & Lucas, 1997).
The results are shown in Fig. 3.13 where the net bolus-creating force,
FV FA , is plotted against the number of chews. Note how just low these
forces are. Initially, the force is negative, which means that food particles
are more likely to stick to the oral cavity than to each other. However, after
just a few chews, the force rapidly becomes positive and peaks between
about 2030 chews. The peak is shallow with raw carrot particles because
these break down so slowly, but for both carrot and brazil nuts, the cohesive
force gradually declines from the peak as too much saliva is introduced,
letting particles oat away.
Predictions of the model
The model predicts several features of oral processing behaviour, some
of which have been observed in humans and some of which remain to
be tested (Prinz & Lucas, 1997; Lucas et al., 2002). The model predicts
Lillfords (1991) result with beef. Forced chewing beyond the normal point
of swallowing will start to swamp the particles causing them to oat apart.
85
Fig. 3.13 The bolus-making force, plotted on the vertical axis, as it varies during a masticatory
sequence. Early on, the forces that attract food particles to the walls of the oral cavity
are greater than that which would stick them together, so the overall bolus-making force
is negative. However, the viscous cohesive force predominates after just a few chews. It
declines only when excess saliva weakens it by separating food particles and allowing them
to oat away. (Reprinted from Food Quality and Preference vol. 13, Lucas, P. W., Prinz,
J. F., Agrawal, K. R. & Bruce, I. M., pages 203213, Copyright (2002), with permission
from Elsevier.)
The model also suggests why humans swallow fruit pieces very quickly: the
time to swallow should be promoted by the expression of uid from within
food. If uid expressed from the fruit were sugar-rich, then the adhesive
force would increase and swallowing would be promoted even further. Food
acidity should also promote early swallowing by stimulating extra saliva.
While fruit esh often forms an almost instant bolus, leaf fragments do
not. Their outer covering, the cuticle, is so hydrophobic as to be effectively
unwettable (Barthlott & Neinhuis, 1997; Neinhuis & Barthlott, 1997). This
is not just a function of the waxiness of the cuticle, but also of its microroughness, which makes the contact areas of the leaf surface with water,
grit particles, fungi and other microorganisms so small that these will fall
off the leaf.18 This keeps the lamina clean and free from infection for as
long as possible. The cuticles of leaves interfere in bolus formation because
leaves have such a high specic surface that the cuticle dominates exposed
surfaces of leaf fragments throughout mastication.
The general value of the model for explaining swallowing patterns in
other mammals can be doubted and not just because of anatomical differences. It cannot connect the efciency of oral processing up to later steps in
86
the digestive process in any convenient way, something that other models
can do (Alexander, 1991, 1998). However, the existence of a food bolus itself
though cannot be denied Lillfords (1991) pictures show it and articial
bolus makers can apparently demonstrate it in vitro. Unless there is some
neural evidence for a swallow trigger though, the model really remains
just a possibility.
There is a icker of hope for such a swallow sensor though in the textural
fat sensor of Rolls et al. (1999) that resides in reinterpretation of their data.
These authors actually show cortical neurons in an Old World monkey
that respond purely to lubricants such as oils (both vegetable and mineral)
and the like, i.e. to a stimulus that probably involves a drop in intraoral friction or, at minimum anyway, to a surface sensation in the mouth.
Unfortunately, the authors do not give a location for this stimulus in the
mouth, their concerns being to ascertain that the neurons in the cerebral
cortex did not re in response to either taste or odour. The most likely
location is the tongue where sensory nerves have been demonstrated that
are exquisitely sensitive to forces in the low millinewton range (Trulsson &
Essick, 1997). As stated previously, Rolls et al. have established that there is a
group of neurons with a specic frictional role, but it does not seem logical
that these cells exist to detect fat texture if there is already a chemosense
for these compounds. I suggest a possible alternative: could these neurons
relay information about the developing state of the food bolus and provide
the information critical for swallowing decisions?
Tooth shape
overview
Virtually the entire dental literature deals with the function of teeth from
a geometrical perspective. In this chapter, the rst of three forming the
core of this book, I advance the alternative hypothesis that the mechanical
behaviour of foods is the major inuence on tooth shape. The principal mechanical properties of foods involved are the Youngs modulus and
toughness. However, I will also be referring to more specic inuences such
as the level of stresses within foods, the yield strength, strain and Poissons
ratio. To understand these concepts clearly, the reader is advised to work
through Appendix A before reading further. While some of the inuences
on tooth shape can be quantied fairly precisely, others are too complex
for this. The analysis then has to borrow from simpler circumstances in
which the basic principles have been established. The chapter starts by assuming that foods are homogeneous solids, but the value of understanding
features of food structure such as their internal connectivity should quickly
become apparent. To keep things simple, all through the chapter, the basic
analysis assumes linear elastic behaviour in foods. Only general principles
of dentaldietary adaptation are emphasized because Chapter 7 deals with
real dentitions.
introduction
The toughness of foods plays a crucial role in shaping teeth. Unfortunately, toughness is an energetic concept that still seems alien to most
biologists. When most people think of fracture, they think in terms of fracture strength, whereas, in fact, there is no such thing as a unique fracture
stress for either a solid material or a biological tissue. I am not referring here
to the fracture stress in tension say, as compared to that in compression or
shear, or to the effects of varying the rate of loading and temperature, but to
87
88
4 Tooth shape
Fig. 4.1 A simple experiment to ascertain that cracking is controlled not by stress but by
energy. (a) A sheet of newspaper has been cut into the shape of a pair of trousers, one
trouser held by hand and the other attached to a bulldog clip from which hangs a plastic
bag. Ball bearings can be placed in the bag until the paper starts to fracture in the direction
indicated by the dotted line. (b) Changing the width, w, of the legs lowers the stress, but
will not vary the number of ball bearings needed to make the crack start, whereas doubling
the thickness by tearing two sheets at once, as in (c), will require a doubling of the number
of bearings.
the fact that there is no such thing as a unique strength for a material under any given loading circumstances. Grifth (1920) established this about
80 years ago. However, with the exception of two popular books by Gordon
(1978, 1991), this fact has not been well marketed to the general public. Even
some engineers seem to feel that fracture mechanics is an unnecessary burden and an impractical complication. The net result is the retention of
the Galilean concept of fracture strength as gospel truth by those in other
sciences (see Chapter 5).
It is now known that energy, not stress, controls fracture. Simple examples illustrate this. The trouser-tearing test can be performed without special
equipment by tearing up a newspaper (Fig. 4.1). Hold one leg of a sheet of
the newspaper and attach a bulldog clip on the free-hanging lower leg. Add
a plastic bag that can be lled with weights such as ball bearings (Fig. 4.1a).
Once a sufcient amount of bearings has been added, the paper will tear.
The stress on the leg imposed by the mass of ball bearings, M, will equal
Mg/wt where g is the gravitational constant, w is the leg width and t, the
paper thickness. By experiment, it will be found that varying w, which
changes the leg stress, will make no difference to the force at which fracture
starts (Fig. 4.1b). On the other hand, increasing the thickness of the paper
Introduction
89
by tearing two to three sheets simultaneously instead of just one will require
double or triple the number of ball bearings respectively (Fig. 4.1c). The
onset of fracture is proportional to Mg/t, but independent of w and, therefore, of stress. The meaning of Mg/t is not force per unit thickness, as
might appear at rst glance, but energy per unit area and the understanding of it is in terms of conservation of energy. If the work done in extending
the crack in the paper a distance l is Mgl and all this work is converted into
a crack of area lt, then, with the toughness of the paper dened as R,
Rl t = Mg l .
This reduces to
Mg
.
t
In a universal testing machine, we do not usually measure l directly, but
use the displacement of the crosshead of the machine instead. This travels
twice as far as l because both legs are stretched. In addition, the force, F, is
given directly, so that
R=
2F
.
t
The same lack of dependence on stress is true for crack growth in any situation. For example, there is no dependence on the thickness of the peel
in the peeling test (Appendix A) (Kendall, 2001). Increasing peel thickness
decreases the stress, but does not alter conditions for its removal. The phenomenon of notch sensitivity provides another example (see Appendix A)
and there are also many indirect methods that demonstrate it. One is to
show that the fracture stress for a given material varies with particle dimensions. Kendall (1978a,b) showed this very clearly for compressive strength:
smaller particles are stronger than larger ones. This has been corroborated
in many other studies (e.g. Darvell, 1990) and is a well-known effect in
foods.1
So forces, always acting as stresses, are just the means to provide energy
to a particle: it is what happens to this energy that inuences whether this
particle fractures. Ordinarily, the energy is stored as elastic strain energy
within the particle and is available for release into a crack under the right
conditions. Alternatively, loading may produce permanent (plastic) deformation of the particle. From a fracture point of view, this is energy lost and
an efcient comminuting mechanism needs to avoid this if at all possible
because a large amount of energy can be involved. Energy may be lost in
R=
90
4 Tooth shape
other ways as well, e.g. as heat or kinetic energy (as when glass shatters and
small pieces y away), but both these effects can be an unavoidable consequence of needing to produce fracture very quickly. The last possibility
is that the work done on the particle can somehow be channelled directly
into fracture, so avoiding energy transfer to the bulk of the particle. This
is the only one to offer the (illusory) target of 100% efciency in loading.
loading geometry
While energy has no direction, the loading that imparts this most denitely has. There are three basic types of pure loading that could lead to
the fracture and fragmentation of solids: tension, compression and shear
(Fig. 4.2). A tensile loading involves a force acting perpendicular to the
surface of a solid and directly away from it (Fig. 4.2a). The opposite of a
tensile load is a compressive load. This also involves a force acting perpendicular to the surface of a solid, but straight towards it, rather than directly
away (Fig. 4.2b). The third possibility is that when the solid is loaded in a
direction parallel to its surface. This is loading in shear and is an angular
action (Fig. 4.2c). The movement can be linear, as in Fig. 4.2c, or involve
rotation, as in torsion (twisting). Pure torsion is a form of shear, but any
twisting in practice nearly always involves tension too.
These loading patterns are not the equivalent of internal stresses: all
these loadings can lead to combination of compressive, tensile and shear
strains inside particles as a reaction to them. Provided that Poissons ratio
is not zero, compression will lead to tension at right angles to the load and
shear at 45 to it. The Poissons ratio of solids is usually considerably less
than 0.5, meaning that particles will densify (reduce their volume) under a
Fig. 4.2 The three basic loading geometries of loading: (a), tension, (b), compression and
(c), shear.
Fracture geometry
91
compressive load. In such solids, the efciency of inducing tensile stresses

indirectly via such a load is generally low. Pure shear is dened as a state in
which the sum of tensile and compressive stresses in the specimen is zero, i.e.
the hydrostatic stress is zero. (Remember in such sums that tensile stresses
are given the opposite (positive) sign to (negative) compressive stresses; see
Appendix A.) However, all these pure loads are rare outside of mechanical
testing. Pure shear is extraordinarily difcult to produce anywhere, even in
a laboratory test. Pure tension and compression are extremely inefcient
methods of generating critical stresses. Why load all of a particle evenly if
the effect is aimed at a particular location? Bending, for example, results in
tension just on one side of a particle. The tensile stress is not even along the
bent length but peaks on the outer surface midway between its supports.
Bending is an efcient way to produce tension because it acts locally. In
constrast, indentation produces localized compression.
There are many ways to load a particle, such as by pulling, pushing,
bending or twisting, and all of these may be seen during ingestion. However,
once inside the mouth, the options reduce. Only pushing (compression)
is readily available unless there is some external restraint for the teeth to
act against. A tensile load inside the mouth would require moving parts in
both upper and lower jaws in order to get a grip. That is very complex to
arrange and has not evolved in mammals.
fracture geometry
Most, if not all, of the literature on tooth function suggests that tooth shape
can be understood by examining the details of loading, or the distribution
of stresses produced in the food particle. This is incorrect. The factors that
promote the production and growth of cracks within that particle are the
crux of the matter. In support, the next section shows how the geometry
of fracture is controlled by food particles, not by teeth.
Elastic fracture
Many solids maintain an elastic response to a loading right up until close to
fracture. If a crack forms and runs right through such a particle, dividing it
into two, then the fragments can be tted back together to form the size and
shape of the original particle. Compression, e.g. by the molar teeth, cannot
lead to crack growth directly because cracks or aws in particles are closed
by compression rather than opened. Compression tends to heal cracks and
its comminuting effect has to act indirectly via shear or tension. However,
92
4 Tooth shape
(a)
(b)
starter
notch
cracks
Fig. 4.3 A plate made of a homogeneous material has had a sharp-ended starter notch cut
into it at an angle that could favour crack extension in shear or tension in equal measure.
When loaded either in tension (left) or compression (right), crack growth from the ends of
the slot rapidly aligns along a direction in which tensile stresses are maximized: at 90 to
the tensile loading (left) or parallel to the compressive loading (right).
is it shear or tension that is responsible for crack growth? The experimental

and theoretical answer is unequivocal. Figure 4.3 shows the behaviour of a
homogeneous laboratory specimen in tension (Fig. 4.3a) and compression
(Fig. 4.3b), when a crack extends from the ends of a notch. The notch
has been inclined at 45 to the load to give the crack every opportunity to
grow in shear. It is found in experiments that, once a crack appears at the
notch tip on loading, the crack follows a short curved path and lines up
perpendicular to the direction of the maximum tensile stress. This path is
perpendicular to the load when that is tensile or parallel to the load when it
is compressive. Theoretical analysis supports these results by showing that
cracks grow essentially in the direction in which the elastic strain energy
stored within them is released most rapidly, where this rapidity is assessed
per increment of crack area growth, not per unit of time (Lawn, 1993). This
direction is always tensile in homogeneous elastic solids. If the applied load
is more complex, with a mixture of tension and shear for example, then it is
possible for the crack to travel for a short distance with a shear component
(as shown by the rst crack extension in Fig. 4.3). However, this is eeting
(Mai & Cotterell, 1989) and continued growth of the crack tends to follow a
direction perpendicular to the main tensile stress (Lawn, 1993). The theory
Fracture geometry
93
Fig. 4.4 The three modes of fracture. Only mode I fracture, where crack opening is associated
with tensile stresses ahead of the crack, is common in homogeneous materials. It is very
difcult to get fracture in either of the shear modes even in pure laboratory tests because
it is not easy to avoid an opening component at the crack tip.
is difcult (if not impossible) to solve for all except the rst crack extension,
but experiments seem conclusive: cracks tend to grow by opening up.
Generalizing from the above, there are three possible directions in which
a crack could grow. These are termed modes of fracture, each dened to
correspond with the three components of stress in front of the tip of a crack
that could result in its further growth (Fig. 4.4). Mode I fracture entails
tensile stresses in front of the crack tip that encourage the crack surfaces to
pull apart. This crack-opening mode is usually what is evaluated in
toughness tests. An example is when a large carnivore pulls meat off a
carcass. Mode II involves shear stresses in the plane of the section. An
example is provided, in theory anyway, by the action of a paper punch
or an incisal bite. Mode III involves out of plane shear stresses such as
those apparently produced by tearing a piece of paper (trouser-legging)
or by a pair of scissors (see Appendix A). In practice, fracture in neither
modes II or III is easily tested in isolation because it is extremely difcult to
prevent crack opening (mode I) involvement (because, as explained above,
this favours faster energy release). In practice, there is almost certain to be
a strong opening element. In fact, there is little reason to invoke modes II
or III as having any relevance to the action of the teeth. Even a guillotine
94
4 Tooth shape
or a pair of scissors produces mode I cracking in an elastic solid (Atkins &

Mai, 1985).
Thus, I cannot see any general grounds whatsoever for supposing that
the design of mammalian teeth is to maximize shear (Shaw, 1917). Sharpbladed teeth will not produce shear cracks in either mode II or III. Yet it
can hardly escape anyones attention in reading the dental literature that
shear lies at the magical heart of dental function for most researchers. This
is an error. The direction of maximum shear stress in a solid is not going
to be the direction in which cracks in food particles generally grow. Even
if it were, then such a theory could not explain dental diversity: all teeth
would look the same.
Some may feel that all this is restricted to homogeneous foods. Could
there be something special about the structural make-up of some food
particles that makes a heightening of shear stress within them advantageous
for cracking? Well, heterogeneous food particles will, by denition, contain
interfaces between their structural components. These interfaces might
have a very low toughness. If so, then cracks will tend to deect along them.
If this deection lies at an angle to the main direction of the bite force,
then this will involve shear and some mode II or mode III component (as
indicated, for example, in Fig. 4.18). However, the point is that the direction
of deection is the natural path of the crack. No aspect of dental design,
such as the presence of a sharp blade, is required to produce this, and the
original path of the crack will be recovered if only a slight weakness appears
in a favourable direction. Otherwise, a crack oriented at an angle to the
main direction of bite force will tend to arrest (stop).
It is generally unlikely that any tooth would possess features designed to
promote such natural crack directions. If it did, then the most logical dental
strategy would be to attempt to open up the interface rather than induce
slip. There are some instances where natural fracture paths in potential
foods are designed to facilitate fracture. Very often, seeds have these in
order to open up and germinate. However, even seeds have design features
to avoid being opened. To paraphrase: they need to open, but not be
opened (Lucas et al., 1991b). Fleshy fruits are also exceptional: they are
designed to be eaten, but only the esh and only by target consumers. In
fact, these fruits appear to be craftily designed in order to keep competitors
(i.e. fruit thieves) out (Cipollini & Levey, 1997). Otherwise, any potential
food particle is part, or all, of an organism that is designed to resist fracture.
The threat may not come from a predator (it could be wind, for example)
but any natural path of fracture in that particle is actually likely to form
part of its defences. If this logic is accepted, then teeth that let cracks follow
Fracture geometry
95
that path will fail to fragment the tissue. An example is to try to bring down
a tree by following the grain of its wood, rather than aiming across it with
a bladed axe or saw. The latter is a much tougher path, but it is the only
way to fragment it. This is the way that beavers (Castor ber a rodent) do
it. Leaves show this protective design too: cracks deect along their veins
as a way of stopping the loss of a whole leaf.
To sum up, we can assume that elastic foods fracture under the inuence
of tensile stresses. When these cracks are straight, then it makes sense
also to refer to this as mode I fracture. However, most commonly, a crack
zigzags through a food structure. There is no point at all considering modes
of fracture when this happens, which is to say: the mechanism of fracture
prevention in foods is what matters for understanding tooth form, not fracture
geometry.
Plastic fracture
When tissue volume is conserved during deformation (i.e. Poissons
ratio is 0.5), then the structural units in that tissue must be sliding over
each other: in other words, this is shear. Plastic deformation is exactly like
this. Tissues that show large amounts of plastic deformation do not really
crack even though there are changes in their external surface area as they
distort. However, very few plant or animal tissues display plastic deformation to this extent because it would be incompatible with their structural
role in the organism. Limb bone provides an example. It is known that
it could have a far higher crack resistance than it does (a comparison between the antlers of deer and their limb bones shows this Currey &
Brear, 1992), but achieving it would either reduce Youngs modulus considerably, so lowering the rigidity of the locomotory system, or cause yielding
that would interfere much more by producing uncontrolled change in the
shape of bones. (Shape control is less crucial for antlers and these do show
substantial plastic deformation J. D. Currey, pers. comm. in Purslow,
1991a.) Plastic deformation is possible though in any solid when loading
is conned to a sufciently conned space such as during indentation
(Atkins & Mai, 1985). Even so, any crack(s) that start at or near the plastic
elastic boundary of this indentation grow along lines of tensile stress (Lawn,
1993).
So I conclude this section, as the last, by predicting that the geometry
of fracture will not explain anything at all about mammalian teeth. The
word shear does not explain tooth shape any more than do crushing and
grinding. Semantically, words like these are simply shorthand descriptions
96
4 Tooth shape
of successful fracture events, often connoting a whole range of conditions

surrounding such events, but denoting only one thing: fracture. They do
not explain the factors required for successful fracture they cannot if
their meaning assumes it and so they have no place in an analysis at
all. There is a very large set of breakage synonyms, many more than 100
in the English language, which can be dragged into functional descriptions
of dental actions. None has any analytical value, being used as a kind of
descriptive shorthand. The fault in seeing otherwise, and the reason why
food properties are rarely invoked as causes of variation in tooth shape, is
very clear from the following quote:
The fact is that mammals use their teeth to acquire and prepare their food and
that their molars are in this sense adapted to certain kinds of activity and beyond
this to certain types of diet. (Simpson, 1936)
The activities to which Simpson refers are shearing and grinding. He

gives complex denitions for these in terms of tooth shape and movement,
but which do not include foods at all. The interpolation of activities such
as these between teeth and diet is the central reason why studies of dental
dietary adaptation have remained stagnant. Without foods, all that can be
done is to eulogize teeth for being so good at what they do.
All further sections of this chapter assume that the shape of teeth is an
evolved response for overcoming the toughening mechanisms inside foods
that frustrate their fracture and that these mechanisms lie at the heart of
the diversity of dental form.
fracture limits
Ashby (1989, 1999) showed how to analyse the possible limits to the success
or failure to fracture particles in a systematic manner. When a force produces a displacement, work is done. Forces always have to act as stresses,
but displacements are not always equivalent to strains because they could
easily be off the axis of loading. Immediately, these simple considerations
generate two possible limits to successful fracture: a food particle could fail
to be fractured either because there is insufcient stress or because there is
not enough displacement. An example, modied from Lucas et al. (2000),
illustrates the difference.
Suppose an elephant were attempting to remove a branch from a tree
by bending that branch downwards. The animal might be able to bend
it all the way down to the tree trunk, but not succeed in breaking it.
This is displacement limitation: there was sufcient stress, but the elephant
Basic roles of the dentition
97
ran out of displacement. The alternative is that the branch barely bends
because the elephant could not generate sufcient stress. There was a
lot of displacement still available but not enough stress. This is stress
limitation.
Mastication appears most often to be a displacement-limited activity
because that is what restricts the efcacy of loadings that push tooth surfaces
together. This may apply to success/failure to fragment a particle once, or to
a limitation on the number of times that a particle can break into separable
pieces, within any particular chew. Whichever, this is a displacement limit.
In contrast, ingestion is never going to be displacement limited. An elephant
is not going to stop attempting to break the branch just because bending
it vertically downwards fails to detach it. It could simply pull, or pull and
twist the branch simultaneously, persisting until successful. In the same
way, a primate that is ingesting a food particle by incision will only attempt
to load a particle between upper and lower incisors if there is sufcient
displacement to cause fracture. Otherwise, unless the particle is very small,
it would select another method. I will assume then for the theoretical
analysis, that mastication is displacement limited while ingestion is stress
limited.
basic roles of the dentition
Teeth at the front of the mouth have different shapes to those at the back.
The reason for this could be due partly to these different limits, but also due
to the differing roles that anterior and posterior teeth play in food intake.
Figure 4.5 suggests a hierarchy of activities necessary for the management of
solid particles. A primary requirement is grip for without this there can be
no management at all. The secret of grip is friction. Grip allows the tongue
to direct and transport food particles. The teeth of most non-mammalian
vertebrates are designed only for this function. Grip for transport may
require crack initiation within a food particle. Most mammalian teeth are
designed for fracture. If this need goes further to produce food particle
size reduction (i.e. fragmentation), then there are two alternative strategies.
Both a nal particle size and shape might be required, a strategy that
we could call sculpture. If no particular food particle shape is required
other than that some regularities may emerge from repeated size reduction,
then the process can be termed comminution. To emphasize the difference,
sculpture is fracture designed to produce a fragment(s) of a particular shape
while comminution is multiple fracture producing random fragmentation
(Lowrison, 1974).
98
4 Tooth shape
Fig. 4.5 A chart showing the various basic tasks that the dentition has to perform in the
mouth. See text for explanation.
Putting the last two sections together, the use of the incisors and canines
in a rst bite, with or without subsequent sculpture (fracture designed to
produce a fragment(s) of a particular shape), involves different demands
to comminution (multiple fracture producing random fragmentation) by
Basic tooth shapes for comminution
99
the postcanine teeth. Mastication will be discussed rst because conditions

there are better dened than during ingestion.
basic tooth shapes for comminution
The easiest way to set up the principles of construction of the basic tooth
features is to use a specic example. However, the analysis is quite general,
depending only on the relative ease of initiating and propagating cracks
within the food particles and the need to produce separate fragments rather
than just cracks. I need to assume that the loading is close to optimal, so
that some combination of force, displacement and/or work is minimized.
Loading an isotropic homogeneous food with a minimal force requires that
the contact area between tooth and food particle be minimized because this
keeps contact stress high. However, the shape of a food particle cannot be
precisely controlled during ingestion and it then changes continually as it
is fragmented. The exact loading conditions inside the mouth will change
from chew to chew. Thus, control of contact areas relies completely on
tooth shape.
Large particles
Imagine a large food particle one that requires reducing in size in
order to be swallowed lying on a at supporting surface. The cheapest
method of starting a crack in this particle, regardless of its properties or
structure, is to push a point into it (Fig. 4.6a), a point being a surface, both
dimensions of which are very small. A point contact produces a very large
stress in the food particle for minimal load. In dental studies, of course,
pointed features of tooth crowns are called cusps. What inuences the design of these cusps? If cracks start at cusp tips, then it may be thought
that cusps should be sharp. However, somewhat ironically as far as some
of the dental literature is concerned, a sharp cusp is likely to suppress
cracking. In even the least tough materials, if the stress can be conned
to a small enough volume, then the solid will deform plastically rather
than crack. This is exactly what happens in indentation. Imagine that the
cusp tip behaves just like a spherical indenter. The fracture load that produces cracking is proportional to indenter radius (Frank & Lawn, 1967).
Since the area of indentation is proportional to the square of the radius,
this means that sharper indenters elevate stresses, and thus tend to cause
plastic indentation rather than cracking. Assuming this to apply to teeth,
100
4 Tooth shape
Fig. 4.6 Considerations for understanding the design of dental features and basic features.
The main arrow gives the direction of the force. (a) The apparently logical starting point
a sharp narrow-angled cusp indenting a food particle with minimum stress. (b) A sharp
cusp may suppress cracking in the food particle by raising the local stress; in contrast,
a blunter cusp is more likely to promote cracking. (c) A narrow-angled cusp will encounter higher friction than one with a wide angle. (d) The basic elements of a dentition:
a cusp, wedge (symmetrical to the direction of the force) and blade (asymmetrical). None
of the features can be derived without considering the mechanisms by which cracks grow
in foods.
then blunt cusps are more efcient at producing fractures than sharp ones
(Fig. 4.6b).
It may also appear logical to suppose that the cusp be narrow-angled,
again so as to reduce the force. Though this may be so, it results in higher
101
friction. If the included angle of the cusp is 2 (Fig. 4.6c), then the increase
in the force due to friction as the cusp indents the particle is given by
Ffric = Fnofric (1 + cot )
(4.1)
where Fnofric is the force in the absence of friction, is the coefcient

of friction and the term cot refers to the cotangent of the half-angle,
i.e. 1/tan (Hankins, 1925). Widening the cusp angle reduces friction and,
while the force itself will increase due to the larger area indented by the
cusp, it is more likely to promote fracture rather than plastic deformation.
Also, both a wide angle and a blunt cusp tip are also features that help
protect the cusp from fracture.
The other tooth surface, the lower one in Fig. 4.6, is just a at surface.
However, there is no reason why it should not also act to promote fracture
and can, therefore, also be cusped. Clearly, the direct opposition of two
cusps, one upper and one lower, is an unstable arrangement because, after
fracturing the food, the cusp tips will contact and damage each other.
An array of cusps, alternating as in Fig. 4.9, seems logical. The spaces
between the cusps can be curved so as to retain food fragments generated
by propagating cracks. This allows repeated fracture within one compressive
movement.
Without knowing more about the food, prediction about cusp form is
stuck here because some foods could crack easily, while others not crack
at all. What food properties inuence this? To derive these requires some
assumptions. A crack in a food particle that is laid across these cusps may
start: (a) remote from cusps, e.g. by bending against a three- (or more)
point cuspal support or (b) adjacent to a cusp tip as the particle is indented
(Fig. 4.7).
I make two basic assumptions about fragmentation:
(1) Food particles are loaded relatively late in the closing phase of jaw movement.
It follows that the displacement during which a particle can break up is
small and will limit fragmentation. Displacement limitation will apply.
(I ignore the possibility of fragmentation during jaw opening when food
stuck to the upper and lower teeth could fracture as the teeth separate).
(2) Fragmentation of particles on loading follows the storage of elastic strain
energy in these particles, the release of which later pays for growing cracks.
This assumption might seem unnecessary, but it excludes the local loading of thin rod- and sheet-like food particles by tooth cusps (considered
later).
Figure 4.8a shows the geometry for the three-point bending of a simple
beam-like food particle. The example is drawn by Ashby & Jones (1996).
102
4 Tooth shape
(a)
(b)
cracking remote from a cusp
(c)
cracking at a cusp tip
(d)
wedging open an arrested crack
Fig. 4.7 A schematic diagram to illustrate cracking. Three cusps are shown in (a), and
cracking can occur remote from a cusp (b) or at a cusp tip (c); in the latter case, the crack
may arrest and can then be wedged open by the cusps further movement (d). (Reprinted
from Food Quality and Preference vol. 13, Lucas, P. W., Prinz, J. F., Agrawal, K. R. & Bruce,
I. M., pages 203213, Copyright (2002), with permission from Elsevier.)
If E is the elastic (Youngs) modulus of the food, then the deection shown
in Fig. 4.8b is
=
3Fl 3
4Ebt 3
(4.2)
while the maximum stress (sometimes referred to as the modulus of rupture), halfway between the supports on the tension side, is
3Fl
.
(4.3)
2Ebt 2
The force, F, can be removed by combining Eqns 4.2 and 4.3. Moving all
the food properties to the left-hand side, then the criterion for the initiation
F =
103
Fig. 4.8 Cracking remote from cusp tips. (a) Cusps modelled as producing a threepoint loading of a food particle. (b) The relevant particle dimensions as the particle
bends through a deformation . (c) The crack, of length a, develops remote from the
cusps. (Reprinted from Food Quality and Preference vol. 13, Lucas, P. W., Prinz, J. F.,
Agrawal, K. R. & Bruce, I. M., pages 203213, Copyright (2002), with permission from
Elsevier.)
of cracking at less than an arbitrary displacement (i.e. with displacement

limitation) is
2t
F
> 2 .
(4.4)
E
l
For failure to the point that an engineer would be interested in, which is
usually the point of crack initiation, a combination of food properties that
determines the response of the food is (F /E ).
However, the interest is not just in crack initiation, but in the resistance
to crack propagation because this inuences the formation of separated
fragments. The rst possibility is that the food particle cracks distant from
a cusp at the point of the maximum tensile stress. Once cracked, the stress
eld in Fig. 4.8b is given by
K IC = c F (a )0.5
(4.5)
where c is a dimensionless constant that depends on the geometry of the

loading. Then, dividing through by E and noting KIC (ER )0.5 (see
Appendix A), where R is the food toughness, gives
0.5
R
F
c (a )0.5 .
(4.6)
E
E
If the number and size of fragments is limited by the displacement available, then (R/E )0.5 is the combination of food properties that inuence
104
4 Tooth shape
crack propagation, and thus the possible fragmentation of the food. It

forms a displacement-limited fragmentation index that corresponds to
(F /E ) at the onset of cracking.
In the above scenario, the cusps are really acting only as supports. There
is no need for them to be sharp and it would be better if they were blunt
in fact because this would increase their longevity. Their height needs to
be sufcient so that the food particle does not bottom out before failing.
This could be achieved by making cusp height proportional to typical
values of (F /E ) in the diet. For example, materials with a high Youngs
modulus generally have a low value of (F /E ). This suggests that extensive
consumption of such foods would result in selection for low molar cusps
because the deection, , would be small at fracture.
Fractures may also start at, or very near, the cusp tip (Fig. 4.8c). Then
(ER )0.5 = c F (a )0.5
(4.7)
where c is probably close to 1.0 (Marshall & Lawn (1986) give c = 1.33
for indentation). If these cracks run straight through the material, then
fragmentation is limited only by stress, not displacement. However, these
indentation cracks may arrest before fragmenting the particle. An arrested
crack can only be continued by displacement of the cusp into the particle
so as to wedge the particle open (Fig. 4.7d). Unsurprisingly, this follows a
displacement criterion with
0.5
R
0.866uw 1.5
(4.8)
= 2
E
a (1 + 0.64w/a 2 )
where w = 0.5b and u is the width of contact of the cusp at the point at
which the crack of length a starts to grow (Hoaglund et al., 1972).
To summarize, if stress is limiting, then (ER )0.5 is the probable criterion
for the fragmentation response of a food solid while if displacement is
limiting, it is (R/E )0.5 . Note that both of these indexes have dimensions. For
reasons given above, mastication is likely to be dominated by displacement
limits, but evidence for this will be offered a little later.
Cusped postcanine teeth will only succeed in fragmenting food particles
if the value of (R/E )0.5 is low. If instead the index is high, then cracks in a
food particle will quickly arrest. Even if a cusp penetrates a particle of such
a food so as to reinitiate a stalled crack, that crack is then unlikely to spread
laterally. For such foods, tooth form should adapt such that the pointed
tip of a cusp is extended in one dimension so as to form an edge, a surface
with one dimension that is very small. A wedge is an example of an edged
105
design. It is oriented symmetrically with respect to the force (Fig. 4.6d),

so that stalled cracks can be reinitiated by the edge. Provided that the long
dimension of the wedge exceeds any of the particle dimensions, this ensures
particle subdivision.
If cracks arrest quickly, then the sharpness of the wedge becomes very
critical in re-establishing the crack. In this way, the edge of the wedge
will remain close to or against the crack tip all the way through the food
particle. However, while the food particle will be subdivided into two, the
end result will be that the wedge contacts the opposing tooth surface, so
damaging it. Wedges are not durable and need to be redesigned as blades.
The difference between a blade and a wedge is that, in the latter, the force
is aligned parallel to one surface (Fig. 4.6d). If the opposing tooth surface is
similarly designed, then the surfaces parallel to the force can pass each other.
These surfaces need to contact each other as they pass or the particle will
not separate into two (Fig. 4.9). Wedges appear unlikely as components in
a dental battery, with cusps and blades being more frequent, but a possible
role will be assigned to them a little later. So the two basic components of
the working surfaces of teeth then are likely to be cusps and blades. Blades
would be expected in matched sets, one on the upper dentition and one on
the lower, but cusps would not. Being adapted to foods that crack, cusps
do not have to contact after food fracture.2 They will produce fragments
ahead of their path of travel in a chew and rather than have a at resting
surface for particles that lie under cusps, it seems more logical to design this
lower surface as a retainer allowing fragments to be rebroken in the same
chew. A basin-like structure, i.e. a fossa, serves this purpose. The postcanine
dentition is thus expected to consist of patterned arrays of cusps and blades.
The analysis so far is summarized in Fig. 4.9.
The effect of particle shape and size
When a particle of any solid gets sufciently small, then it ceases to crack
and yields instead. The size at which it does this is called the brittle
ductile transition. The food property group that controls this deformation transition was worked out by Kendall (1978a) and is (ER )/ y 2 ,
where y is the yield stress. The transition particle size at which cracking behaviour in a solid starts to be frustrated by plastic deformation, is
given by
d =c
ER
y2
(4.9)
106
4 Tooth shape
Fig. 4.9 Chart showing the basic relationships between food properties and dental features.
Blunt cusps are associated with foods of low (R/E )0.5 that crack relatively easily. Cusps can
be aligned as upper and lower alternates (to avoid mutual damage) or into fossa as pestles
and mortars. Blades have to contact the supporting surface on the other side of particle
because they represent optimal designs to break foods with high (R/E )0.5 that obstruct
cracks. To prevent excessive damage, the supporting surface has to have a reciprocal
matching shape, forming a pestle and mortar type of arrangement. The upper and lower
dentition would, therefore, be expected to possess pairs of edges. If these were like wedges,
107
where d is a linear dimension and c is again a constant. For compression,

Kendall (1978b) showed that c = 32/3. However, its value depends greatly
on the general conditions of loading. Atkins & Mai (1985) have generalized
the arguments and show that for indentation, the constant is likely to
be >2000.
Below the transition size, energy stored in food particles cannot be transferred into cracks. Thus, a particle that cracks above this size threshold, and
which might be comminuted with a cusp, will not crack below this size
limit and might require a blade. This transition might set a limit to food
particle comminution (e.g. the low millimetre range identied by Lillford
(1991)), but it also suggests another simple way of looking at the range of
food properties. If any food particle were sufciently large, then it should
crack due to work done by the teeth on the particle being stored as elastic
strain energy, which then pays for any crack that might initiate. Below the
threshold size, energy storage is not sufcient to do this and so the work
done by the teeth must pass directly into crack growth. The food property
group that controls resistance to mastication is no longer (R/E )0.5 , as in
the analysis above, because that depended on elastic energy storage, but R
alone. In those circumstances, chewing would become like a kind of (always
relatively inefcient) toughness test.
Experimental validation
Corroboration of the simple theoretical approach advocated above is currently being sought by experiments on human subjects. The fragmentation
of particles that results from one chew is a direct measure of the breakage
function. If this can be expressed in a form dimensionally consistent with
the food property grouping (R/E )0.5 , then it can provide a test of the above
theory.
Fracture changes the surface area of particles. However, this is not itself
a measure of the breakage function because the surface area of a particle of
any given shape depends on its volume. The surface area per unit volume of
a particle is often called its specic surface in comminution studies. Taking
Caption for Fig. 4.9 (cont.) symmetrical about the direction of the force, then there would
be considerable toothtooth wear. Blades are asymmetrical allowing their faces to brush past
each other. The arrows on the pestle and mortar and double-bladed units indicate that
some lateral movement will be required in the former because the loose-tting arrangement
will reduce friction, the latter to make sure that blades contact as they pass.
108
4 Tooth shape
two food particles, a large one labelled x and a smaller one called y, with
the ratio of any of their linear dimensions being , then the ratio of their
specic surfaces is
2
= 1 ,
(4.10)
3
which means simply that the smaller a particle is, the greater is its surface
area to volume ratio. This is trivial, but suppose now that both particles x
and y are subjected to displacement-limited fracture. If the displacement
needed to start a crack in x is ux , while that to crack y is uy , then linear elastic
fracture scaling (Chapter 5) shows that the ratio of these displacements
uy
= 0.5 .
(4.11)
ux
There is, in fact, a long tradition in the mining and powder science industries of using an index with these units (i.e. the reciprocal of the square root
of particle size) called Bonds work index (Bond, 1952, 1962). A considerable body of careful work supports it as an indication of the work required
in industrial comminution processes (e.g. Rose & Sullivan, 1961).3 Thus,
the square root of the specic surface (or, alternatively, the reciprocal of the
square root of particle size) will be used to dene the relative success of fracture under these conditions. Is the fragmentation of particles in chewing,
described in this way, related to (R/E )0.5 ?
Agrawal et al. (1997, 1998, 2000) tested this on human subjects. These
studies have been summarized by Lucas et al. (2002) and are only described
briey here. The methodology was basic. The toughness and Youngs modulus of 38 relatively homogeneous foods (foods whose heterogeneity did
not appear to affect fracture growth) were measured using mechanical tests
described in Appendix A. The stressstrain curves of these foods were relatively linear. The measurement of toughness included energy expended
both in elastic (reversible) fracture and small irreversible (plastic) deformation even though the theory above was constructed only including the
former. Standard bite-sized pieces of these foods were then produced which
ve subjects each made one bite on, using the distal premolarrst molar
teeth. This biting procedure was repeated once for each food. Saliva dissolves some of these foods very rapidly and so each particle was shielded
from the oral environment during biting by placing it inside a sealed latex
bag (as described by Mowlana & Heath, 1993). The actual bite was thus
on the latex, which transferred the stress through to the food particle. It is
possible that this covering disturbed the results, but the latex appeared to
109
grip the food particles effectively. On opening the bags and tipping their
contents onto a Petri dish, this loading had obviously damaged all foods,
with most lying in fragments. These fragments were not rearranged other
than to make sure that fragments did not cover each other. Images of these
fragments were taken by a video camera held directly above the centre of
the dish and fed into a computer that analysed the apparent surface area
of these particles. The initial surface area was taken to be the area of the
face of an unbroken particle; volume was assumed not to change after
a chew.
The relationship of (R/E )0.5 to change in the square of the specic surface in these foods is shown in Fig. 4.10. The line represents inverse proportionality (because (R/E )0.5 is resisting crack growth) that appears to1
provide strong support for the theory. Foods with an index >25 mm 2
Fig. 4.10 Experiments on six human subjects with 38 foods shows that the breakage function, the estimate of fragmentation, and the food property index (R/E )0.5 are inversely
proportional to each other. This food index appears to describe resistance to fragmentation
1
during human mastication, although above an index of 25 mm 2 , particles were distorted
rather fractured. Data points 19 are nuts; 1027, cheeses; 2832, fruits and vegetables; 33
36, breads; 37 a type of soyabean curd; and 38, monocrystal sugar. (Redrawn from Food
Quality and Preference vol. 13, Lucas, P. W., Prinz, J. F., Agrawal, K. R. & Bruce, I. M.,
pages 203213, Copyright (2002), with permission from Elsevier.)
110
4 Tooth shape
were plastically distorted and not broken into discrete fragments. This may
indicate the limit of effectiveness of a cusp-based dentition like that of the
human.
Further papers by Agrawal et al. (1998, 2000) on a subset of these foods
relate (R/E )0.5 positively to the electrical activity of the anterior part of the
temporalis muscle, a portion of the jaw-closing musculature that lies close
to the skin, and also to the degree of lateral movement by the mandible
during the closing phase of chewing. Although there is, as yet, no published
information on the limiting particle size, i.e. on the deformation transition
controlled by (ER ) y 2 , preliminary evidence obtained by K. R. Agrawal
(pers. comm.) on food fracture behaviour during incisal bites in humans
suggests that such a transition in some foods is often in the low millimetre
range. This gives a value for the constant in equation 4.8 that suggests
loading with a strong compressive element (Kendall, 1978b; Atkins & Mai,
1985).
specic inuences of foods on dental form
Tooth sharpness
The sharpness of teeth has been one of the overriding obsessions of studies
of the mammalian dentition, being measured as the radius of curvature,
, of the tip of a structure (Lucas, 1982a). However, it is not the only relevant measure of the shape of a tip because this takes no account of how
rapidly this tapers from the bulk of its supporting structure (Freeman, 1992;
Sheikh-Ahmad & McKenzie, 1997; Evans & Sanson, 2003). A survey of the
sharpness of mammalian teeth found that it tended to be affected by tooth
size and that the bladed molar teeth of carnivores and insectivores were
usually noticeably sharper than those of herbivores (Popovics & Fortelius,
1997). Experiments by Freeman & Weins (1997) on models of bat canines
proved that sharpness was a strong inuence on the forces needed to penetrate apple esh. How can this be understood? The principal inuences
on tooth sharpness are probably the extensibility (i.e. the strain to fracture)
of foods and their Poissons ratios because both of these can inuence the
sharpness of crack tips.
Imagine a crack, oriented at right angles to the load, at the centre of
a large rectangular strip of a homogeneous elastic material subjected to
uniform tension (Fig. 4.11a). Modelling this crack as a very narrow ellipse,
Inglis (1913) showed the effect of this crack was to elevate the primary tensile
stress (the stress in the direction of the load, perpendicular to crack length)
Specic inuences of foods on dental form
111
Fig. 4.11 (a) Tissue strip with linear elastic behaviour containing a centrally located elliptical
crack (crack length being 7.7 times larger than crack width) being pulled uniformly. The
graph on the right shows that the effect of a sharp crack is to elevate the primary tensile
stress at the crack tip far above the average in the specimen (Inglis, 1913). (b) The elongation
is now 100% and the Poissons ratio is 0.3. The crack tip has rounded considerably due to a
secondary tensile stress that develops in the direction of the crack. A higher Poissons ratio
would magnify the blunting.
at the crack tip far above the average prevailing in the tissue as a whole. This
is shown in Fig. 4.11a, where if the crack length is 2a (the 2 implying that
the crack is double-ended) and its width is w, then the stress at the crack
tip is raised to [1+2a/w] times the average stress in the tissue. The radius
of curvature at the crack tip being , then the stress multiple can also be
expressed as [1 + 2(a/)0.5 ]. Thus, the blunter the crack tip, the lower the
stress and the less chance there is of crack propagation.4
Now if this tissue is very extensible (i.e. with a large strain at fracture),
the crack tip will blunt as the strain increases. The cause of the blunting
is another tensile stress I will call it the secondary tensile stress that
112
4 Tooth shape
Table 4.1 Poissons ratios reported for various plant and animal tissues
Tissue
Animal
Cow teat skin
Aquatic salamander skin
Wall of dog carotid artery
Intersegmental membrane of
female locust
Plant
Cork
Calophyllum inophyllum leaves
(parallel to veins)
C. inophyllum leaves
(perpendicular to veins)
Watermelon
Highest Poissons
ratio (usually
at low strain)
Source
1.41.6
2.5
0.30.4
1.0
Lees et al. (1991)

Frolich et al. (1994)
Dobrin & Doyle (1970)
Vincent (1981)
0.0
0.3
Gibson et al. (1981)

Lucas et al. (1991a)
0.0
0.5
Miyazato et al. (1994)
Standard cellular solid
0.33
Gibson & Ashby (1999)
Human tooth tissues

Enamel
0.3
0.23
Reich et al. (1967) cited

in Waters (1980)
Renson & Braden (1975)
0.3
0.5
Ashby & Jones (1996)

Vincent (1981)
Dentine
Engineering materials
Metals
Unlled rubber
develops parallel to the long axis of the crack. Right against the crack tip,
the secondary tensile stress in the material is zero (because there is empty
space within the crack side and there is nothing there to be pulled on).
However, a very short distance in front of the tip, this stress rises above
the average primary stress, subsequently declining rapidly (Fig. 4.11b).5 It
is this secondary stress that blunts the crack.
The propensity to blunt cracks depends not just on extensibility, but also
on Poissons ratio. The higher Poissons ratio is, the more a crack tip can
blunt for any given stress. Normally, Poissons ratios in engineering solids
are around 0.33, but they can differ very markedly from this in biological
tissues. In animal soft tissue, the ratio can be >1.0 (Table 4.1), the major
cause being the arrangement of collagen bres in the connective tissue.
These are often loosely connected in the form of networks and may realign
in relation to the stress (Purslow et al., 1984). Network geometries can
113
produce very high Poissons ratios (Fig. 4.12a) so that when the longitudinal
strain at fracture is high, there is considerable elastic blunting of natural
cracks. Figure 4.12c shows the effect of this on an unnotched and notched
tensile specimen.
All extensible materials can get some protection from fracture by this
elastic blunting effect, which can elevate their toughness considerably. Its
efcacy can be demonstrated experimentally by constraining the curvature
of the crack tip so as to prevent this blunting. Lake & Yeoh (1978) show, for
example, that the apparent toughness of rubber is reduced in proportion to
the sharpness of a razor blade pressed against the crack tip. A free-running
crack in rat skin has a toughness of 1420 kJ m2 (Purslow, 1983), while that
cut with scissors (blade sharpness 1.6 m) is only 0.59 kJ m2 (Pereira
et al., 1997). This is about a 20-fold reduction, converting a tough material
into nothing like one.6
Finally, Purslow (1991a) has presented a quantitative argument suggesting
that a J-shaped stressstrain curve (Fig. 4.13), such as exhibited by most
vertebrate soft tissues, adds to this crack-blunting effect. This point will be
returned to in Chapter 7.
Generalizing all this to dentaldietary adaptation, the postcanine teeth
of carnivorous mammals need to have sharp-bladed features to suppress
the toughness of vertebrate soft tissues. To practical limits, the sharper the
blades, the lower the cost, but whatever the sharpness, they need to travel
all through the tissue to stay in contact with the crack tip (to prevent it
blunting) and will end up contacting the opposing tooth in the other jaw.
Toothtooth wear is, therefore, inevitable.
In Fig. 4.14, soft tissue is going to be subdivided between two at blades.
However, as the blades close, the food quickly extends beyond the ends of
the blades, a function both of great extensibility and a high Poissons ratio,
producing a slit in the tissue but failing to subdivide it. One solution to
this is to make these bladed teeth much longer than the material that they
are going to comminute. However, the need for them to be outsize can be
reduced greatly by a concavity of the blades so that their ends contact well
before their centres. The carnassials of carnivores are like this (Fig. 6.2). The
same principle applies to mammals with an invertebrate diet. The cuticles
of insect larvae are sometimes very extensible and Vincents (1981) study
conrms that very high Poissons ratios are possible in some circumstances.
The sides of tribosphenic molars can look very similar to those of carnassials
with inexed blades (Fig. 7.3). However, the toughness for insect cuticles
appears to be much lower than the soft tissues of vertebrates (Appendix B).
Many have commented on the inexion of blades on the molars of
114
4 Tooth shape
under strain
(a)
(b)
under strain
UN-NOTCHED
NOTCHED
under strain
under strain
(c)
Stress
pre-test
pre-test
Strain
(d)
Callerya atropurpurea
seed coat
Intsia palembanica
pod
Fig. 4.12 Tissue geometry, Poissons ratio and non-linear elasticity. Open cellular structures
have a massive effect on all major mechanical properties (Cox, 1952; Gibson & Ashby, 1999),
including Poissons ratio. (a) An open diamond framework that produces a high Poissons
ratio, the value depending on the strain (Frolich et al., 1994); (b) an alternative arrangement
whereby the walls can tuck into the structure under load so increasing density with
zero Poissons ratio; (c) simulates the effect of both non-linear elasticity and Poissons ratio of
115
DISPLACEMENT-LIMITED
J-shape
r-shape
Stress
Stress
STRESS-LIMITED
r-shape
Stress
linear
J-shape
Strain
Strain
Strain
Fig. 4.13 The stressstrain curves of materials can be characterized as either r-shaped, Jshaped or linear. The cross indicates the fracture point. Simply due to curve shape, as higher
loads, rubber maintains a relatively constant stress at a wide range of strains and the opposite
for skin (Purslow, 1991a). It is suggested in Chapter 7 that r-shaped curves correspond to
stress-limited defences, while J-shaped curves (much more common in biological tissues)
represent displacement-limited defences.
carnivores (Fig. 6.2) (Savage, 1977) and insectivores (Fig. 7.3) (Crompton
& Kielan-Jaworowska, 1978) and their role in trapping food, but the true
diet-based explanation of this need seems to reside in the above.
Plant tissues do not have any ability to blunt cracks elastically in this
manner. Their Poissons ratios tend to be low because of their cellular
construction (Wilsea et al., 1975), with values averaging 0.3 (Table 4.1).
Also, they take much smaller strains to fracture and most tissues have almost
linear stressstrain curves. In consequence, the molars of herbivores do not
have to be as sharp as those of carnivores and insectivores. There may be no
general need to have blades at all on herbivore teeth and certainly no need
Caption for Fig. 4.12 (cont.) 1.0 on a strip of vertebrate tissue under tensile loading. The
strip extends greatly at low stress. This blunts any notch e.g. produced by an indenting tooth.
The effect is seen in rubber (Lake & Yeoh, 1978) and fresh spring roll skin (Sim et al., 1993),
but is greater in vertebrate soft tissues because Poissons ratios are higher. Sharp-bladed teeth
frustrate this mechanism. (d) The soft seed coat of Callerya atropurpurea (Leguminosae)
has a thin outer layer (at top) that is fully connected, but deeper in the coat, connections
formed during development have been lost, so producing an open framework. Pressure on
the outside of the coat with a blunt object like a cusp does not induce much tension in this
framework. Poissons ratio will be very low and a bladed tooth is required to open the coat by
indentation. However, if the coat could be subjected to e.g. trouser-legging, as in ingestion
using the forelimbs, then cracks would grow very cheaply more cheaply than with a blade
(Appendix B). (e) The pod of Intsia palembanica (Leguminosae) shows a similar pattern of
defence to the seed coat: groups of bres (arrowed) course internally but are only loosely
connected to other structures. Scale bars, 100 m.
116
4 Tooth shape
Fig. 4.14 (a) A piece of animal soft tissue lies between two at blades. Loaded as in (b), it
deforms beyond the ends of the blades due to high Poissons ratio and high strain. In (c), an
inection or concave curvature allows the ends of the blade to contact prior to much strain
on most of the particle, so trapping it while the strain is low as in (d).
for inexions or concavities. Experiments on plant tissues with scissors with

blades of varying sharpness suggest that this has only a relatively minor effect
on toughness values, e.g. an 800% reduction in blade sharpness appears
to raise apparent toughness values in across-grain tests on woods by only
37% (Darvell et al., 1996; Lucas et al., 1997). Occasionally, a free-running
crack in plant tissues can cost less than one driven by a blade, implying that
natural cracks can be sharper than those induced by scissors blades. This is
particularly true for cracks that propagate between cells (see next section).
Very low Poissons ratios in plant foods can also inuence tooth form.
In a food that behaves like cork (Gibson et al., 1981), with a zero ratio, it
will be impossible to generate tension indirectly via the compressive effect
of an indenting cusp or blade: cracks will not open up and there is likely
to be considerable friction against the tooth as the plant material densies
under load. The optimal solution is again a sharp bladed tooth. However,
the included angle of the blade might be much smaller (limited by the
117
possibility of its own fracture) because most of the energy loss is likely to be
sub-surface and not affected much by lubrication: in cork, the coefcient
of friction increases with load for just this reason (Gibson et al., 1981).
Structural connectivity
It might be thought logical that a food particle would evade fracture best
by having a tightly connected structure that provides maximum resistance
to indentation by teeth. In fact, slight disconnection can elevate toughness
provided that the scale of this detachment is kept within critical limits. The
limits are due to disconnected regions acting like aws that must be kept
small to avoid great reduction in the stress at failure. The strategy works
well in articial composites (Atkins, 1974). Gordon (1980) discusses it with
respect to tensile structures, showing how the separation of load-bearing
elements by a very low-modulus matrix can prevent crack propagation and
produce notch insensitivity (Fig. A.9). In the limit, fracture mechanics is
defeated like this and Galileo rules (see Chapter 5).
Structural disconnection is common in vertebrate soft tissues (Purslow,
1983, 1991a), but not in plant tissues.7 Some leaves show notch insensitivity
if they are gripped in their entirety (veins and all) and then pulled (Lucas
et al., 1991a; Vincent, 1992), but otherwise, continuum mechanics seems to
predict the fracture behaviour of fully connected plant tissues very effectively (Gibson & Ashby, 1999). There are exceptions though (Fig. 4.12d).
The soft seed coat of Callerya atropurpurea (Leguminosae) has a fully connected, but relatively thin, outer layer. However, most of the inner tissue
has lost connections that were formed during its development. Accordingly,
pressure on the outside of the coat with a blunt object does not produce
much tension in this inner tissue it is too disconnected for that and a
blade is required to open it. The pod of Intsia palembanica (Leguminosae)
shows a similar pattern of defence, with internal bres only loosely connected to other structures.
Patterns of fracture in relation to structure in plant tissues
Without the need for a musculoskeletal system, plants protect every cell
by developing a semi-crystalline wall inside the cell membrane. This cell
wall is a mechanical composite, consisting of cellulose, hemicelluloses and,
sometimes, lignin. Individual cells are glued together loosely with pectins
or rmly by lignin. The major component taking the load is the stiffest element and that is cellulose, consisting of nanometre-sized crystalline brils
118
4 Tooth shape
combined into microbrillar bundles. Hemicelluloses and lignin act as a

binding matrix to glue these together. While hemicelluloses are hydrophilic,
lignin is a heavily cross-linked hydrophobic polymer that waterproofs tissues and connects up all the wall elements. This interferes with transport
processes across the wall: if lignication is extensive, the cell dies. Lignication starts in the middle lamella and proceeds to the primary, then the
secondary, walls.
The purpose of the cell wall is to provide each cell with the potential to
support its shape. However, many cells have only a thin primary wall. Cell
stiffness in this case is achieved, not by the wall alone, but by maintaining
cell contents under high pressure (between 1 and 5 MPa; Tomos & Leigh,
1999). The critical factor that prevents bursting is the toughness of the
wall. To this end, the cellulose microbrils in it are laid down in lamellae,
with the orientation of one layer unrelated to that in the next (McCann &
Roberts, 1991; Carpita & Gibeaut, 1993). The primary cell wall is usually
thin, but can be locally thickened, as in collenchyma, a tissue that supports
growing tissues. More generally though, cells with thick walls also contain
a secondary cell wall. This wall usually consists of three distinct parts, S1,
S2 and S3, laid down in that order. The bulk of the secondary cell wall
(50% or more of its thickness) is usually the S2 layer (Fengel & Wegener,
1989). The cellulose microbrils throughout this layer are parallel to each
other and, in elongated cells specialized for mechanical support such as
sclerenchyma bres, they tend to wind at a small angle (say, 10 30 ) to
the cellular axis (Preston, 1974). When there is secondary cell wall present,
lignin is nearly always added to the hemicellulose glue.
Extensively lignied cells are just dead mechanical supports.8 The mechanical advantage of forming such complex structures does not lie in their
exemplary stiffness (i.e. Youngs modulus) because this is bound by what is
called a rule of mixtures: any composite cannot be stiffer than the stiffest
component or less stiff than the least stiff component. There is also no
overall gain in strength (however that is dened) from composite structure
either because it is bounded in a similar way. The single major property to
escape these limitations is toughness, which is set by structure much more
than by material content. The toughness of an articial composite, for
example, is typically 100010 000 times tougher than any of its separated
components (Atkins, 1974; Harris, 1980).
In general, the major nutrients of plants are inside the cells, so a mammal must push a crack through the cell wall to release these. The ease of
doing so depends greatly on wall toughness and the amount of cell wall
in the tissue. Nutritionists know the cell wall content of tissues as bre
119
(Van Soest, 1994, 1996) and measure it as a proportion of their dry weight.
The term bre content will be avoided here because of a potential confusion with a plant cell type that is called a bre. Instead, I will dene
the amount of cell wall in a plant tissue, Vc , as a proportion of its fresh
volume.9
Lucas et al. (1995, 1997, 2000) reported toughness data on 82 plant tissue
and plant-derived materials (Appendix B), varying from watermelon esh
(Vc = 0.0031; its thin-walled cells being >0.5 mm in diameter) to seed
shells, the walls of which are so thick that they can occupy 95% of the
tissue (i.e. Vc = 0.95). The study used scissors cutting tests (Appendix A)
because this technique can be used on a wide variety of tissues, regardless of
their shape and size, and can help to tease two major causes of toughening
in plant tissues apart.
When a crack runs through an object, it disrupts a small volume of material around the crack tip, permanently disturbing it. This plastic zone is
one of several zones of disruption that can be dened, depending on the
material (Lawn, 1993). Here I will distinguish small-scale fracture events
within the plant cell wall (which is what cracks), expressed whatever the
thickness of the tissue, from disruption in the surrounding cellular framework. The latter must be progressively diminished whenever the thickness
of a specimen is reduced below the normal size of this plastic zone. Thus,
when a specimen is thick enough, toughness is independent of thickness
(inset in Fig. 4.15a), but as it is progressively thinned, toughness will drop
in proportion to thickness because the volume of tissue being affected also
diminishes. There will be a threshold thickness, which is usually around
0.51.0 mm in plant tissue, where a plot of toughness versus specimen
thickness seems to plateau. The exact threshold is difcult to dene with
precision, but plots like the inset in Fig. 4.15a can yield two values fairly
accurately: an intercept toughness, derived by extrapolating a regression
line to zero thickness, and the slope of this line. The former is sometimes
called the essential work of fracture (Atkins & Mai, 1985) and probably
represents elastic fracture mechanisms that act within the cell wall. The
slope of the regression line on the other hand provides a total estimate of
the extent of the plastic behaviour that plant tissues can display.
The intercept toughness for all 82 tissues is plotted in Fig. 4.15a, where
it is seen to be proportional to Vc . By extrapolating from the densest seed
shells, which have Vc = 0.95, to Vc = 1.0, an indirect estimate of the
toughness of the cell wall itself can be obtained. This is 3.45 kJ m2 , a
value that can be interpreted as the intrinsic toughness of the cell wall.10
This gure is a surprise because it is exceptionally low for a composite.
120
4 Tooth shape
Fig. 4.15 Two separable components of toughness in plant tissues. (a) The toughness of
plant tissues in relation to the amount of cell wall they contain (dened as Vc , the fraction
of fresh tissue volume that the cell wall occupies) and also tissue geometry. The inset shows
schematically how cutting tests are conducted on individual tissues, so potentially dening
an intercept, slope and plateau. The intercept toughness describes elastic fracture in cell wall,
while the slope describes the ability of cells to deform (buckle) plastically. The main graphs
show results for a wide range of plant materials. At left, intercept toughness is proportional
to Vc , irrespective of tissue geometry. At right, plastic buckling in woody brous tissues
absorbs considerable amounts of energy. Tissues with isodiametric foamy cells cannot do
this. (b) A attempt to pull these two effects, cell wall and cellularity, together by assuming
that plasticity is conned to a spherical zone, of diameter 1 mm, around the crack tip. The
overall toughness of woody tissues, calculated as the sum of woody toughness and cell wall
toughness, is shown to be up to 10 times that of foamy tissues of the same Vc , but peaks at
Vc = 0.80.9, declining thereafter.
121
Copying natural principles, articial composites can be made that are

100 times as tough as this (Gordon & Jeronimidis, 1980). Even wood
tested by standard methods (Jeronimidis, 1980) is 10 times tougher than
this estimate for the cell wall. How can this be?
The answer, according to Jeronimidis (1980), is that a truly cellular mechanism provides most of the toughness of woody tissues. It is known that, at
critical stresses, structures like the helically wound S2 wall collapse irregularly and irreversibly into the cellular lumen (Page et al., 1971), so absorbing
large amounts of energy (Gordon & Jeronimidis, 1980; Jeronimidis, 1980).
It appears that this plastic buckling mechanism, operating probably just in
front of the crack tip, is the factor that elevates the toughness of woody
tissues well above cell wall toughness levels.
The effectiveness of the events in the cellular framework in setting toughness levels is clear in Fig. 4.15b where the toughnessthickness slopes for
82 tissues are plotted against Vc . The plastic toughening mechanisms
appear to behave very non-linearly. Tissues with very low Vc have no secondary cell wall and, therefore, no potential to buckle plastically at all.
From about Vc = 0.20.8, the mechanism absorbs energy roughly linearly
with Vc . At the peak toughness, at Vc 0.8, buckling provides about
10 times the toughness that cell wall alone can. Above the peak, toughness
starts to plateau and falls precipitously at Vc > 0.9 or so. This is almost
certainly because the cell wall is then so thick that the buckling mechanism
is frustrated by there being insufcient space (almost no cell lumen) for the
wall to buckle into. Thus, woody tissue such as many seed shells with
Vc = 0.900.95 do not really behave mechanically like woods. Despite these shells having an almost identical chemical structure to woods
(Preston & Sayer, 1992), the woody toughening mechanism hardly operates at all. They are exceptionally stiff and with a high yield strength, but
they are not tough at all. In fact, the graph in Fig. 4.15b, which attempts to
put all the toughening mechanisms back together, suggests the toughness
of the densest seed shells reverts to levels due to the cell wall.
All this has important repercussions for tooth shape. Foods with very
low Vc , like fruits, will not have much crack resistance, so teeth designed to
break them can be cusped. Woody tissues with Vc = 0.20.8 are progressively more resistant, so blades will be needed (Fig. 4.15b). When
Vc > 0.8, crack resistance is lost again because woody toughness does
not operate, and so cusps may sufce once more. This could explain very
well why the dentitions of both frugivores and seed-eaters very often have
rounded cusps, while those of leaf-eaters have sharp blades (Kay, 1975b,
1978; Kay & Hylander, 1978). If so, this mechanical analysis shows plainly
122
4 Tooth shape
why there can be no recourse to the chemical bre content of plant foods
to explain the dentitions of herbivores: this will not describe the difculty
of fracture. Lignin on its own is often touted as a tough material by nutritionists, but it cannot be so. Toughness is a consequence of the composite
cell wall and tissue structure, not of individual chemical components (Lucas
et al., 2000).11
The best evidence for the predominance of plant tissue structure over
cell wall content in setting toughness is from the seed shell of Mezzettia
parviora (Lucas et al., 1991b, 1995). The shell has three distinct zones, an
outer one composed entirely of parallel bres (zone I in Fig. 4.16), an inner
Zone II
Woody ruminations
invading endosperm
Zone I
Zone III
Zone III
woody plug
Toughness (kJ m -2)
escape
channel
embryo
15
Zone I
10
Zone II
Zone III
10 mm
0.1 0.2 0.3 0.4 0.5

Section thickness (mm)
Fig. 4.16 The mechanical behaviour of Mezzettia parvifolia seeds. The shell is mainly composed of bundles of bres that course randomly (zone II, scale bar, 100 m), but those
placed most supercially are normal to the surface (zone I, scale bar, 100 m). There is
also a narrow band of stubby cells (zone III appearance shown in Fig. 4.17c) that runs all
round the shell and connects to a woody plug. This band connects to a woody plug that
facilitates germination while being so narrow that a mammalian predator cannot utilize it to
enter the seed. Each zone has a similar intercept toughness and Vc , but very different capacity for plastic buckling. Invertebrates operate at a smaller scale than the cellular aspect of
toughness the circular holes seen were produced by scolytid beetles that enter randomly
anywhere around the seed. (Redrawn from Lucas et al. (1995) The toughness of plant cell
walls. Philosophical Transactions of the Royal Society London, B 348: 363372.)
123
one in which bundles of bres appear to course randomly (zone II) and
lastly, a band of short stubby cells (called brachysclereids) encircling the seed
(zone III). All these zones have roughly the same cell wall volume fraction
(Vc 0.940.95). Figure 4.16 shows tests on each zone with specimens of
different thickness. The intercept toughness for the three zones is about
the same, being very close to that of cell wall itself as would be anticipated.
However, the inuence of cellular organization is very clearly shown in the
toughnessthickness slopes. Fracture of the bres in zone I at right angles
to their long axis gives the highest slope; zone III the lowest slope (not
signicantly different from zero, in fact) with zone II intermediate. The
cause of these differences is structural: long bres in zone I can buckle
whereas short stubby cells in zone III cannot. The randomly directed bres
in zone II are often not oriented such that they can buckle and so lie
between these extremes. Nothing about cell wall chemistry explains this
it is purely a function of cellular organization, thus demonstrating how
important tissue structure is.
The relationship of plant tissue structure to defence against predation in
plants in discussed in Chapter 7, but it is perhaps pertinent here to note that
the anisotropy of woody tissues does not allow plants any easy compromise
between protection from attack from all directions (as often implied by
seed shell structure) and from one direction only (as in attack on wood in
trees). The former is the probable reason for bre arrangements such as
zone II in seed coverings. Yet random bre directions are not the midway
compromise that they might appear. Their Youngs moduli provide one
reason for this.
The Youngs moduli of woods, loaded along the grain, are proportional
to Vc . However, when the load is applied across the grain, it is proportional
instead to Vc 2 (Gibson & Ashby, 1999). Given that Vc < 1.0, this means
that a mammal that attacks wood by indenting or bending it will tend to
encounter a lot less stiff a structure than if it just pulls on it. The toughness
though is the same. Williamson & Lucas (1995) add the netting analysis of
Cox (1952) to this so as to explain why the Youngs moduli of the random
bre networks of seed shells are only 26 GPa, nothing like the 30 GPa if
the cells could all be loaded along their long axis.
The above discussion of plant tissues all assumes that cracks travel
through cells. Yet a free-running crack will select that path in a material in
which energy is released most rapidly (Lawn, 1993). If Youngs modulus is
constant, then this path can be predicted by comparing the relative toughness of intracellular and intercellular paths. However, if the modulus varies
too, as it does in woody tissues because of their considerable anisotropy,
124
4 Tooth shape
then the strain energy release rate in different directions becomes extremely
complicated to calculate (Atkins & Mai, 1985).
Experimental observations on woods show that cracks pass directly across
cells without deviating when Vc < 0.2 (Ashby et al., 1985). However, when
Vc > 0.2, cracks zigzag, alternating between being stopped by an interface
and nding a sufciently cheap path across cells (Ashby et al., 1985). The
toughness of the middle lamella, the intercellular glue, is not known in
thin-walled tissue like parenchyma because no one has succeeded in driving
cracks between cells, but in woods and seed shells, the cost of a direct path
is low, 0.10.2 kJ m2 , and independent of Vc (Jeronimidis, 1980; Ashby
et al., 1985; Lucas et al., 1991b).
If an intercellular path provides an avenue for particle fragmentation,
as it does in many seed coats, then it makes sense for a seed predator to
try to direct cracks that way. The point about tooth design here is that
such a path will be precluded by sharp bladed teeth because these will
drive cracks straight across cells at far greater cost. So blunt cusped teeth
do the job much better if this path leads to fragmentation, as shown by
the experiments shown in Fig. 4.17. However, in any tissue where bres are
aligned in one direction, blades are required to cross them. These blades
need to be sharp, not because that is critical for suppressing any particular
toughening mechanism, but for retaining the desired crack path, which
will deviate from it very easily (as part of the plants defences).
-2
0.1-1 kJ m
(a)
(b)
(c)
Fig. 4.17 Intercellular fractures in plant tissues from blunt starter notches. (a) Shows a
three-point bending set-up with a blunt notch. In (b), a sharp crack propagates from this
notch, running between cells. Schinziophyton rautanenii seed shell (courtesy of C. R. Peters).
Scale bar, 100 m. (c) Shows a typical builders rubble surface of between-cell fracture in
zone III of Mezzettia parviora seed shell (see Fig. 4.16 for explanation). Scale bar, 100 m.
The cost of these cracks (i.e. their toughness) depends on how circuitous the crack is, but
0.11.0 kJ m2 compares to 20 kJ m2 if sharp bladed teeth push cracks through these cells.
For woods, the difference in cost can be >100. The endosperm of Mezzettia seeds is traversed
by woody struts called ruminations. These are reminiscent of the plicidentine found in
some non-mammalian vertebrate teeth (Shellis, 1981).
125
Fig. 4.18 Interlocking bres in a seed shell. (a) An indentation in a cusp may propagate
an intercellular crack in a seed shell with low (R/E )0.5 . Scale bar, 1 mm. In (b), the crack
has deected (producing a strong mode II component from an original mode I crack) and
then arrested because of interlocking bres. Scale bar, 1 mm. The probable crack arrest
mechanism is shown in (c) (scale bar, 100 m) and (d) in which bre bundles interlock.
The suggested evolutionary response would be the development of cusps with pronounced
marginal ridges that rst penetrate the seed and then wedge it open against this resistance.
These ridges are seen in the cusp design of (a), where they send off cracks (arrowed) that
will grow further as the cusp penetrates the shell. In so doing, the cusp and ridges act rather
like a convex wedge.
The overall cost of intercellular fracture is affected by crack deections. A circuitous intercellular crack path (Fig. 4.18) may cost 10 times
that of a straight fracture between cells. At maximum, the cost is
1 kJ m2 , which is still a lot cheaper than an intracellular path (compare this to Mezzettia seed shell zone I toughness values for thick sections
in Fig. 4.16). More important than this though is what the interlocking nature of woody bres can do to obstruct crack growth in what is otherwise a
very low-toughness tissue. Figure 4.18a shows an indentation by a cusp that
has ruptured a seed shell, but crack growth has been obstructed sufciently
for the shell not to break cleanly (Fig. 4.18b). Figure 4.18c shows how this
is mostly because interlocking bres catch each other across the crack, so
126
4 Tooth shape
impeding its growth. The logical response to this in a seed predator is to

convert a circular cusp form in one that has ridges (as shown in Fig. 4.18a),
which act to wedge open the crack after the initial fracture. Such an explanation makes considerable sense of the presence of ridges on the sides
of cusps in many mammals. It is possible to go further and suggest that
such a cuspal design is actually like a wedge, where the wedge is somewhat
convex in outline (rather than straight), the opposite of a concave (inexed)
blade.
Living non-woody plant tissue has a very low toughness as simple calculations show. Take parenchyma, the tissue containing most of the food
value of plants. This thin-walled and homogeneous tissue always cracks
through cells. Often Vc 0.05, giving a tissue toughness of (0.05 3450)
172.5 J m2 , conforming to results from experiment (Vincent, 1990).
The lower toughness limit for parenchyma is probably about that of ripe
watermelon fruit esh, being 1030 J m2 (Appendix B). The Youngs modulus of parenchyma depends on its turgidity. Turgidity involves the active
transport of water into a plant cell to pressurize it and maintains the shape
of a cell as though it was a water-lled balloon with the help of only a very
thin primary cell wall. However, most turgid parenchyma has a Youngs
modulus of 110 MPa (Mohsenin, 1977) and, as shown in Fig. 4.10, tissue like this (e.g. many fruits and plant storage organs) can be broken down
rapidly by cusps.12
Fracture can be associated with extensive bursting of cells and uid release
(Peleg et al., 1976). The energies involved in fracture can be so low that teeth
may not even be needed because the at surfaces of soft tissues, like tongue
and palate, can fracture cells. Fruit bats can do this by compressing fruit esh
between the tongue and hard palate and two ape species, the chimpanzee
and the orang-utan, sometimes compress similar tissue between the labial
surface of the incisors and lips (Walker, 1979). If the soft tissues of the
mouth can fracture these foods, then it is also probable that the soft tissues
of the gut can do so too, so much fragmentation of such fruit in the
mouth is immaterial.13 However, many types of fruit esh do not release
their contents so easily. There can be a variety of reasons for this. If fruit
is unripe, it may crack only in one dened plane (Peleg & Gomez-Brito,
1977). At ripeness, the middle lamellae may break down, allowing cells to
slide past each other. The mode of failure then changes from intracellular
fracture to plastic shear (e.g. Heyes & Sealey, 1996).
If the major role of the posterior teeth is not to reduce particle size,
but burst as many cells as possible, then the optimal arrangement is to
have blunt tooth surfaces because they contact more cells than sharp ones.
127
Fig. 4.19 The left column (a) shows rounded cusps acting on thick turgid parenchyma,
while at right (column b), cusps are acting on thick shell. (a) Many cells fracture under
a symmetrical blunt cusp. Fracture lines need have no direction because the object is not
tissue fragmentation. A cusp tting into a loosely tting fossa is the design that expresses
most juice. (b) Cracks in seed shells are affected by their moisture content. In wet shell,
cracks pass neatly between cells. In drier tissue, however, they pass into the cell wall, even
though the major part of the path remains intercellular. Scale bars, 20 m.
However, if cells are not tightly bound, as in some ripe fruit esh, and
slide across each other easily, then the tissue needs to be enclosed or it will
simply ow out of the way. Reciprocal curvatures of tooth surfaces, such as
a loosely tting cusp-in-fossa or pestle-and-mortar arrangement, provide
the necessary compression but with channels for uid escape that would
otherwise impede tooth movement. The cusp would have no need for any
marginal ridges. Figure 4.19 contrasts features of tooth cusps designed for
low Vc and high Vc plant tissues. For the latter, moisture content is a minor
variable, although it does affect the cleanness of intercellular fractures in
woody seed shells: cracks run into the cell wall in dry tissue but pass cleanly
between cells when the tissue is saturated (Williamson & Lucas, 1995).
128
4 Tooth shape
Fig. 4.20 Results of Choong (1997) on the toughness of stacks of laminae of mature leaves
of Castanopsis ssa with a scissors test. Individual laminae (points surrounded by an ellipse)
have variable toughness, but this is always much lower than the toughness of two to four
leaf layers. Above 0.8 mm, toughness appears to plateau. The midrib and secondary veins
were excluded from these tests.
Thin sheets and rods

When particles become very thin, these tissues will no longer store enough
energy to make cracks propagate. The transition thickness for plant tissue
appears to be between 0.5 and 1.0 mm (Fig. 4.20). This is important. The
lamina thickness of most mature leaves ranges between 0.15 and 0.45 mm,
almost certainly below the critical deformation transition thickness, so
cracks do not spread.14 Just for this reason, leaves have to be fragmented
with blades rather than cusps. Structural heterogeneity adds another: the
mesophyll of mature leaves is where the nutrition is, but a thicker-walled
epidermis surrounds it and the leaf usually contains thick woody veins that
prevent effective pressure on thinner-walled tissues (Choong, 1996).
ingestion
There is much less to say about the shape of teeth designed for ingestion
than mastication because these shapes are fairly simple, linking back to
reptiles because of their usual single cusp form. Most mammals have small
peg-like incisor teeth that simply grip particles in order to ingest them.
Ingestion
129
Their working surfaces can sometimes be broad and rough, as in ungulates

and some primates, to grip leaves. Among living mammals though, broad
spatulate incisor teeth (Fig. 2.3) in both upper and lower jaws are unique
to the Anthropoidea, a taxonomic group containing the apes and the Old
World and New World monkeys.
Canines are extra large pegs, very prominent in carnivores, for controlling
prey items, even for killing them, prior to ingestion. Whether for killing or
controlling, carnivores need to avoid fragmenting food particles with their
canines because they would lose their prey in so doing. Blades are needed
to fragment animal soft tissues, so the incisors and canines of carnivores
are not like this otherwise their prey would escape (albeit with elongated
wounds).15
Canines are generally conical, but with a recurved (actually spiral) form.
The upper and lower canines of carnivores are generally long and projecting
teeth, acting at larger gapes than the other teeth (Chapter 5; Fig. 6.2). The
degree of projection and the relative sizes of upper and lower canines vary
greatly, but the uppers are always the longest. The essential load on the
tooth is not just due to indentation of their tips, but to an antero-posterior
force (Simpson, 1941; Smith & Savage, 1959). The cross-sectional shape of
the canines tends to reect this, typically having oval cross-sections with
the long dimension being antero-posterior. Carnivores that break bones
must deal with the vagaries of the direction of the bite force depending on
the orientation of the bone vis-à-vis the tooth. These tend to have circular
cross-sections (Van Valkenburgh & Ruff, 1987).
A classication of ingestion
The major problem with investigating ingestion is the variety of ways
in which it can be achieved. The only classication that I know is that
of Osborn et al. (1987) and Ungar (1992), which has been employed
successfully in eld studies of primates (Ungar, 1994; Yamashita, 2003).
Figure 4.21 is a modication of the OsbornUngar scheme showing
various ingestion methods, the actual details of which vary with food
type.
There are two major possibilities:
(1) A food particle is fractured between the anterior teeth without any other
applied force.
(2) The particle is gripped by the anterior teeth while the animal uses tension
(and a variable amount of torsion perhaps) against an external restraint
to fracture the particle. This external restraint could include pulling a
130
4 Tooth shape
Fig. 4.21 Categorization of incisal use by primates. (a) Ingestion without incision.
(b) Using the incisors for grip with any fracture being remote to the incisal surfaces.
(c) Fracture close to or at the incisal surfaces. From left to right: fracture of a small seed,
fracture of a stalk or shoot, peeling a fruit, removing fruit esh from a seed, peeling a fruit
and taking leaves off a branch as the latter is drawn through the teeth. This is not a comprehensive list of ingestion methods; primates, for example, often bite through leaves detaching
their distal parts while leaving the leaf bases on the plant (Dominy, 2001). (Adapted from
Osborn et al. (1987) and Ungar (1992) with permission.)
fruit, leaf or branch off an attached plant or pulling against an item

held by a forelimb.
As stated early in this chapter, the rst possibility is displacement
limited and would presumably not be employed unless the particle would
fracture under such a loading. Otherwise, the second possibility would be
chosen. Thus, it is probable overall that ingestion is subject only to stress
limitation, not displacement. Agrawal & Lucas (2003) have offered a critical
test of the stress-limited hypothesis, involving fracture of food particles between the incisors in humans. They compared in vivo conditions as a crack
Ingestion
131
propagates from an indentation produced by the anterior teeth (the stress

at fracture multiplied by the square root of the depth of incisal penetration) with the in vitro measurements of (ER )0.5 . There was a strong linear
relationship, supporting a recent view expressed in food science that this
is probably close to what humans sense as food resistance during ingestion
(Vincent et al., 2002).16
Factors affecting ingestion
Successful ingestion depends critically on grip and therefore on high friction. This is generally opposite to the postcanines where friction adds to
the load burden. To this end, the front teeth are less likely to be lubricated as well by saliva than posterior teeth and they may stick out of the
mouth to avoid being wetted altogether. Grip on plant leaves (by anterior and posterior teeth) tends to be complicated by features of the cuticle
(Chapter 3). Enamel decussation may improve grip because of the raised
ridges that it generates through wear (Fortelius, 1985; Boyde & Fortelius,
1986; Rensberger, 2000). The exposure of dentine, as in ungulates, is important because dentine alone has a much rougher surface than enamel,
the coefcient of friction being three times higher against a standard
(Table 6.2). A mixed dentineenamel surface provides much greater grip
than enamel alone. Some herbivores, both living and fossil, have replaced
their anterior teeth by surface-dry keratinized lips. J. F. Prinz (pers. comm.)
has found that the coefcient of friction between opposing dry mucosal
surfaces may be as high as 1.5. A saliva coating brings this down to very low
values, however, except perhaps when food tannins interact with it. Some
ungulates have lower teeth acting against an upper lip, while others have
lost their front teeth altogether, probably to enhance grip.
The enamel of some canine teeth, e.g. extinct sabre-toothed carnivores,
can have rough serrations that probably help grip (Abler, 1992), particularly
on skin. Mammalian hair is itself very tough, as evidence from horse hoof
(Bertram & Gosline, 1986) and human ngernails (Pereira et al., 1997)
shows, and it is possible that tooth serrations are related to consumption
of very hairy prey.17
The ease of ingestion of foods with high or low Poissons ratios might
be very different. Vertebrate soft tissues, for example, would exhibit maximum toughness under pulling or twisting because they would blunt cracks
(Fig. 4.12), while something like a cork would fracture much more easily
than under an indenting tooth. Animals that remove bark, like some rodents
and primates (Vinyard et al., 2003), have to indent it with their anterior
132
4 Tooth shape
teeth. The properties of barks should be important for understanding their

anterior tooth form.
concluding remarks
This chapter has only touched on a few of the possibilities that relate the
mechanical behaviour of foods to tooth shape, but the only major feature
of food to have been excluded is that of tissue age. Toughness depends on
structure, but structure takes time to develop. So young tissues, whether
plant or animal, always make for foods that are much easier to process in
the mouth than older ones. Dental features predicted for the consumption
of some foods, such as leaves, may not apply if those leaves are very young.
Tooth size
overview
The molar teeth of early mammals were less than a millimetre in length
while those of a living African elephant are several hundred times larger.
What is the reason for this? Well, it is probably true to say that, despite
being a prime focus of investigation for a century or more, the adaptive
signicance of such tooth size differences is not understood. The initial
approach here is to consider constraints not just on tooth size, but on the
size of all mouthparts, that are imposed on the digestive system by the
energy requirements of the mammalian body. The argument leads quickly
to what engineers call scaling arguments. Biologists call them this too, but
they also use the term allometric analyses. Both terms describe changes
in the shape of animals required by change in their size. The predictions
made here appear to be novel because they scale the mouth directly to
the food particles that it ingests and not to whole body size (which is the
norm in such studies). However, the arguments are couched in standard
allometric form because this is the nature of the available data. Postcanines
are treated separately from incisors and canines and both general arguments
and specic roles are covered. The overriding philosophy is that physical
properties of mammalian diets explain not only tooth size, but also the size
of most orofacial structures.
introduction
Tooth size was postulated in Chapter 2 to be an evolved response to variation in the external physical (or surface) attributes of foods. Paradoxically,
only the last sections of this chapter follow this logic, the early part examining instead whether tooth size adjusts to changes in the mechanical
properties of foods. This apparent illogicality is explained by the difculty
in separating size from shape in solid objects. Although the distinction may
133
134
5 Tooth size
seem very clear, with the former being dimensional measurements and the
latter angles and ratios, the distinction is actually blurred. For example,
a critical parameter of tooth shape is sharpness even though this is just a
radius of curvature, a linear measurement (Chapter 4). As soon as the size of
a tooth is related to the size of anything else, like a neighbour or the mouth
and body that contain it, we can begin to think of this as a description of
shape, not just size. Rather than get stuck on this, I will let arguments lead
the way.
theoretical relations between body mass,
food intake and metabolism
The limits to the general size of both the oral cavity and its critical contents, the tongue and teeth, stem from the energetic requirements of the
mammalian body and the arrangement of its digestive system so as to supply these. Previous arguments (given in Chapter 3) have impinged on this
issue briey, but they now need to be discussed in detail. I have remarked
that the mammalian digestive system is composed of a sequence of steps
and that the rate of the slowest step will determine the rate at which the
whole sequence runs. It follows from a reductio ad absurdum that the rate
at which mastication and swallowing must run is that which will satisfy the
bodys metabolic needs. This must constrain the size of all the anatomical
components taking part in these activities, including the oral cavity and
the teeth. Firstly though, how can the energy requirements of mammals be
calculated?
Large-scale studies of the metabolism of mammals were initiated during
the era following the First World War (Kleiber, 1932) and have continued
since. These studies have established that the basal metabolic rate of animals,
i.e. the rate of energy consumption at rest, is generally related to the body
mass (M) of a mammal raised to the power 0.75, i.e. to M 0.75 (Kleiber, 1961).
This relationship is not easy to understand and criticisms of it as a purported
fact have been frequent. For example, heat loss is the principal problem that
mammals have to contend with because they maintain a body temperature
that is usually far above their surroundings. Heat loss is via surfaces. If we
assume an invariant density to the mammalian body, such that its mass is
proportional to its volume, and also that the body shape of all mammals is
similar, then the surface areas of different sized bodies will be proportional
to M 0.67 . This prediction follows from simple geometry. If we measure
any dimensions of a large and a small geometrically similar mammal and
compare them, these will obviously be in a xed ratio. Suppose that the
Theoretical relations
135
larger mammal has linear dimensions times that of the small mammal.
The surface areas of the large mammal will be 2 bigger than that of the
smaller mammal and its overall volume, 3 bigger. Accordingly, surface areas
rise (2 /3 ) as fast as volumes. Given that their masses are proportional
to their volumes, this is equivalent to writing that surface areas scale as
M 2/3 = M 0.67 .
The exponent for metabolic rate appears to exceed this exponent, thus
appearing to destroy a trivial explanation. However, this did not satisfy
Benedict (1938), who felt that the data were sufciently inaccurate that a
0.67 exponent was just as reasonable as 0.75. Unfortunately, this particular
challenge has been left behind as the data have been extended and linetting techniques improved (LaBarbera, 1989). The higher exponent has
now been supported very strongly and a dependence of energy requirements
on geometry disproved.1
Most recently, some proponents of a method generally called phylogenetically independent contrasts have doubted that any xed body-mass
exponent exists. Essentially, their contention is that there is a signicant
degree of inertia to evolutionary change in organisms that results from
common ancestry. Thus, after any given time period, closely related animals
would resemble each other much more strongly than more distantly related
ones. An elaborate method of compensating for this presumed inertial bias
has been worked out and the adoption of this approach to energetics can
indeed destroy the simplicity of Kleibers result (e.g. Lovegrove, 2000).
There is no space to discuss this here, but I will assume that this nihilism
is wrong. Readers who are on the bandwagon are of course free to disagree
with me and can doubt many of the relationships printed in this chapter.
Of course, it would help in the defence of the 0.75 exponent if it had received some fundamental explanation. Many have tried, but nothing yet
seems sufcient.2 Nevertheless, I will base the early part of this chapter on
existing measurements rather than theory and assume that the rate at which
mammals need energy scales roughly as M 0.75 .
Early attempts to relate masticatory function to body metabolism focussed only on the area of the working surface of postcanine teeth in mammals, searching for evidence that postcanine tooth area also scaled as M 0.75 .
There is no good reason for this argument, but it was proposed (Pilbeam &
Gould, 1974; Gould, 1975), criticized immediately (Kay, 1975a), persisted
with (Pilbeam & Gould, 1975; Martin, 1979) and then dropped for good after an onslaught of data showed that it was impossible to sustain (Fortelius,
1985 and many others). Kay (1975a, 1978) and, later, Calder (1984) pointed
out that other aspects of the process have be included to make a coherent
136
5 Tooth size
theory. These additional factors include the rate of chewing as a primary

factor. Larger mammals chew more slowly than smaller ones, which may
put a premium on making every chew more effective in these mammals as
compensation. It is clear from this early work that pure empiricism, even if
it hits the front pages, is not an approach that will last. More appropriate are
fundamental arguments, such as that advanced by Fortelius (1985, 1990),
which are centred, as is this book, on the food intake as the stimulus for
adaptive change.
Relations between body mass, tooth size and energy requirements
Fortelius (1985) argued as follows. Assuming all foods to have a standardized
energy density, Kleibers work shows that the masticatory process needs to
provide the gut with volumes of food per unit of time proportional to
M 0.75 . Larger masticatory systems have jaws that traverse greater distances
than smaller ones and so they will tend to have slower chewing cycles. In
fact, many body rhythms, such as the rate of respiration and the heartbeat,
decline with body mass by an exponent of 0.25, i.e. M 0.25 (Peters,
1983). Fortelius assumed this exponent might also apply to chewing rates
and argued from it that mammals should ingest a mass (volume) of food
proportional to M 1.0 each time the masticatory process operates. In other
words, if the need is proportional to M 0.75 , but the rate of processing scales
as M 0.25 , then mammals should ingest an amount of food per unit of
time proportional to M 0.75 / M 0.25 = M 1.0 . (An implicit assumption here
is that all mammals feed for a similar proportion of every 24-hour period.)
As Fortelius remarks, if mammals need to ingest food volumes proportional to their own bulk, then every body structure directly involved in acquiring food should exhibit geometric similarity to accommodate this food
bulk. Thus, the volume of the oral cavity (strictly, the intra-oral volume)
has to accommodate this food batch and would, therefore, be expected to
scale as M 1.0 . The opening of the mouthslit has to be large enough to admit
this food into the mouth, but it is an area, not a volume, and would thus be
proportional to M 0.67 , not M 1.0 . The maximum gape angle of the jaw, a dimensionless measure, would be unaffected by size (scale as M 0.0 ) provided
that jaw lengths scale as M 0.33 , which they should. In order to achieve the
necessary gape, the lengths of jaw muscles, both those that actively open
the jaw and those permitting the jaw to open by stretching, should also
scale as M 0.33 . Those mammals that regulate the ingested mouthful of food
using their incisors, like primates, should have their relevant lengths (their
mesiodistal dimensions) proportional to M 0.33 . The height of a carnivores
Effort required to fracture food
137
canines should scale with the same exponent in order that all carnivores
can inict the same amount of damage on their prey. And so on.
Once food is in the mouth, it needs to be processed by mastication.
Fortelius continues by predicting that the size of the postcanine teeth of
mammals should also show geometrical similarity, i.e. that the areas of
the working surface of the postcanines should be proportional to M 0.67 ,
since this surface then matches that of the food. However, this is a premature
conclusion because it omits an examination of the central feature of the
process: the reaction of the food batch to mechanical loading.
how the effort required to fracture food
varies with animal size
Elastic fracture
To add mechanics to the analysis, I need to make four initial assumptions:
(1) Foods are ingested as single particles, sized in proportion to a mammal
that feeds on them.
(2) Food toughness is constant.
(3) Foods have linear stressstrain curves up to fracture, with negligible
yielding.
(4) Foods are homogeneous.
Thus, the mouthful that any mammal ingests, whether this mammal is
large or small, is considered to be composed of a single, geometrically
similar, linear elastic, homogeneous food particle with invariant material
properties. The adaptation of the size of the jaws and teeth is with respect
to this particle.
Figure 5.1 shows the loading of a large food particle termed x, and a
smaller one, labelled y. These particles are congured as cylindrical columns
loaded in compression, although any geometry or loading regime could
be considered. The results of loading the particles are shown on a force
(F )displacement (u) graph where the loading slope is called the stiffness
of the particle and given by F/u. The work done can be calculated from
the triangular area under the graph in Fig. 5.1, which is 21 Fu. This work is
transferred entirely to the particle and stored as elastic strain energy prior to
food fracture.
The compressive displacements that x and y endure act in proportion
to their linear dimensions while the forces produced act over their surface areas (proportional to the squares of their linear dimensions). The
work done on the particles is given by the product of the force and the
138
5 Tooth size
Fig. 5.1 The compressive forcedisplacement relationship for two intact solid cylindrical
specimens of the same shape and material, but differing in size. Each specimen is being
compressed within a linear elastic range.
compressive displacement and is proportional, therefore, to the cube of

linear dimensions. Thus, dening the ratio of any linear dimension of x and
y is , then the force, compression and the work done on x is (Atkins &
Mai, 1985):
2 (force on y ) (elastic compression of y )
= 3 (elastic strain energy absorbed by y ).
(5.1)
This relationship says that forces act over areas to produce stresses. The
production of identical stresses within the two particles requires that the
linear dimensions of x must be 2 times those of y. Since compressions
depend on linear dimensions, the work that is done (force multiplied by
compressive displacement) is 3 larger in x. This work, the product on the
left-hand side of Eqn 5.1, is absorbed by the particles as elastic strain energy
and carried by their volume, which is also 3 times bigger in x. This has
broad support from experimental evidence on industrial materials (Atkins,
1999).
Galileo (1638) spotted the importance of size to mechanical problems in
Renaissance times. DArcy Thompson (1961) discussed Galileos principle
of similitude extensively in a now-classic account of scaling. It is based
totally on similarity of stress levels. The most frequently given example in
the literature is that of weight-bearing columns: the larger the column, the
fatter it will have to be to do the same job. The explanation is that the
weight of the column, proportional to its volume, grows with size increase
as 3 , while the cross-sectional area of the column, which takes the load, is
only raised as 2 . Thus, in order to take the weight, the cross-section must
139
grow disproportionately large in bigger columns. This analysis has been

applied to the shapes of mammalian limb bones (DArcy Thompson, 1961)
and to the size of muscles and their ability to produce forces so as to move
these bones (Hill, 1950). It remains one of the central examples of shape
change with size in animals and the mechanical arguments stemming from
it pervade biology. According to Gould (1966), Galileo effectively founded
the study of allometry by making such observations.
Studies of how to keep deformation within limits are perfectly appropriate whenever similarity of stress or displacement, keeping strains within
elastic limits, is adopted for the design of a structure. For example, articial
structures such as skyscrapers can be engineered safely by keeping these
principles in mind. As long as the loads on any part of the structure stay
well below that which would produce buckling, or do not take the material
beyond its yield stress (a stress unaffected by size), the structure should
retain its shape indenitely. Engineers may, in some circumstances, be less
conservative and design against the onset of fracture (such as the leak before burst criterion for pressure vessels Atkins & Mai, 1985), but they do
not generally consider designs that prevent fast crack growth (propagation)
because this is far too risky. Buildings, bridges, aeroplanes or oil reneries
are considered to have failed long before this point is reached.3
This construction mechanics, then, is based largely on the control of
deformation. It contrasts greatly with the destruction mechanics of the
masticatory apparatus of mammals. Changing the shape of a food particle
by chewing is pointless unless the particle fractures and then fragments. It
is fracture/fragmentation, and not elastic response, that forms the design
criterion for mastication because the oral cavity is engineered for the total
destruction of foods, not something half-hearted like shape change. The
mechanical relationships that prevail during the deformational domain
do not apply after fracture and I will show this now by extending the
mechanical analysis to the scaling of slow-growing elastic fractures. Gurney
& Hunt (1967) solved this and their reasoning is followed here.
Figure 5.2 shows a cracked particle loaded in three-point bending. This
loading pattern is broadly relevant to how the alternating cusps of the upper
and lower postcanine teeth in humans could load a food particle during the
closing phase of a chewing cycle. I have used it as an example in Chapter 4
and will continue with it here. Again, it is arbitrary and the geometry does
not really matter because the argument is akin to a dimensional analysis.
Note that the particle contains a central crack caused by this loading.
Suppose a small displacement du on this particle, produced by a bite force
F, now extends this crack to produce a new increment of crack area, dA.
140
5 Tooth size
Fig. 5.2 A cracked particle, loaded in three-point bending, in which an existing crack will
propagate under a small displacement, du (not shown), so as to extend the area of the crack
from a sharp notch by an increment dA.
As before, the work done on the particle is 21 Fdu, but part of this work has
gone now into developing the crack. If the toughness (the work required to
grow the crack by a unit area) is R, then the work done in growing the crack
is RdA. As a result of crack growth, the energy stored in the particle will
drop. However, this contained energy is not entirely depleted and has not
fallen to zero. The dotted line in Fig. 5.2 shows that the particle is now more
compliant, i.e. its forcedisplacement slope being shallower. The drop in
strain energy within the specimen after cracking is 21 d(Fu).
The sum of the work done by all forces during such static conditions
should be zero. Thus,
Fu Rd A 21 d (Fu) = 0.
By putting 21 d (Fu) = 21 (Fdu + udF ), dividing both sides by F 2 and then
rearranging, we get
d u
F 2 = 2R/
.
(5.2)
dA F
We now consider two particles again, a large one (x) and a small one (y).
However, this time scaled cracks are introduced into them such that the
141
Fig. 5.3 Two geometrically similar cracked particles of the same material loaded by forces
Fx and Fy in three-point bending.
ratio (large/small) of any linear dimension of the outer surface of these

particles and of the dimensions of their cracked surfaces is also . The
crack area is Ax in the large particle and Ay in the small one (Fig. 5.3). From
the above elastic situation, the compliance of x and y, the inverse of their
stiffness, must be related by

uy
ux
.
(5.3)
= 1
Fx
Fy
These compliances change with respect to growth in crack area by

uy
d
d
d
ux
ux
= 2
= 3
.
dAx F x
dA y F x
dA y F y
(5.4)
Note that crack growth rates, i.e. growth with respect to crack area, in the
two particles are related by 3 . We can now use this result in combination
with Eqn 5.2 to nd the relationship between the forces on the two particles
producing crack growth, which is
F x2 =
2R y
2R x
=
.
uy
ux
d
1 d
3
dAx F x
dA y F y
(5.5)
The forces are, therefore, related to the toughness of the particles by

0.5
1.5 R x
Fx =
Fy .
(5.6)
Ry
If both large and small particles have the same toughness, i.e. Rx = Ry , then
Fx
= 1.5 .
Fy
(5.7)
142
5 Tooth size
Note straightaway that the ratio of these forces is now not proportional
to the ratio of the areas over which they act. Nevertheless, to obtain the
stresses, the force must be divided through by area. For y, this area is still
2 greater than that for x, so the ratio of the stresses is
x
= 0.5 ,
(5.8)
y
and, going back to the compliances, the scaling of the displacements is
ux
= 0.5 .
(5.9)
uy
The reason for the importance of fracture should now be clear. The ratio
of the forces on large and small particles is not 2 , as in Eqn 5.1, but 1.5
(Eqn 5.7) because larger particles fail at smaller stresses (Eqn 5.8).
The force, displacement and energy absorbed in crack formation in
particle x (Atkins & Mai, 1985) are
1.5 (force on cracked y ) 0.5 (displacement of cracked y )
(5.10)
= 2 (energy absorbed by cracked y ).
The right-hand side of Eqn 5.10 says simply that the energy absorbed by
growing cracks depends on crack area, which is to say a ratio of 2 . In
contrast, prior to fracture (Eqn 5.1), energy absorption depended on 3 .
It is because of this dimensional discrepancy that cracks are so dangerous:
they can feed on an abundance of stored energy (Atkins, 1999).
Now take a large mammal, denoted by X, and a small mammal, denoted
by Y, ingesting single food particles termed x and y respectively. Going back
to Fortelius (1985) analysis and the rst assumption, if these particles are
typical food items, then they should be in direct proportion to the sizes
of these mammals, i.e. x/X = y/Y. These mammals need to grow crack
areas in these particles at the same rates per chew. Both the body masses
of the mammals and those of the food particles that they are dealing with
differ by 3 , where is now the ratio of any of the linear dimensions of
either the mammals or their food particles. We can deduce from Eqn 5.7
that the larger mammal X need not produce bite forces that are 2 greater
than mammal Y, but instead only produce forces that are 1.5 as big. Its jaw
adductor muscles are generating this bite force and their ability to do this
is proportional to the physiological cross-sections of these muscles. Thus,
these muscle areas will also be related by 1.5 .
We can now convert any of these ratios to classic biological scaling by
remembering, from above, that the volume of mammal X is 3 that of
143
mammal Y. Provided that these mammals have the same density, then
their body masses will be related by the same ratio as their volumes. Thus,
masticatory adductor muscle cross-sectional areas are predicted to be proportional to M 0.5 (i.e. 1.5 /3 ). This contrasts with geometric similarity,
which would have muscular cross-sections proportional to M 0.67 . Fracture scaling (if I may start to call it that) does not require muscle areas
this large because larger mammals are eating weaker ( larger) food
particles. I predict, therefore, that any large mammal will have proportionately smaller muscle cross-sectional areas than its absolutely smaller
counterpart.
Jaw-closing muscles act to move the lower jaw such that the teeth press
on the food. The working surfaces of the teeth make an area of contact
with the food that dissipates the load. Nothing very denite needs to be
specied about the actual area of contact. The important point is that the
area of toothfood particle contacts, whatever they may be, cannot increase
at a faster rate than the scaling of bite forces produced by the jaw-closing
muscles or this would set off force ination. Thus, I predict that the
working surface areas of teeth that fracture and fragment food particles will
be proportional to M 0.5 .
I now need to introduce a bit of bio-jargon. When part of an organism
is predicted to scale by a factor less than that which preserves geometric
similarity, the result is called negative allometry. Such a body part would
be proportionately smaller in an absolutely larger animal. The converse,
where the larger animal is predicted to have a proportionately larger body
part, is termed positive allometry. Lastly, when there is no change of
shape with size, it is called isometry. I will adopt this jargon from now on
because it can ease description once these simple denitions are grasped.
It should be clear now that the fracture scaling of body parts results in
negative allometry. For example, while geometrically similar muscle areas
scale isometrically, i.e. M 0.67 (Hill, 1950), fracture theory predicts that
they should be M 0.5 .
Virtually all allometric analyses predict either positive allometry, isometry or negative allometry, but never a mix. The interesting feature of a
fracture analysis is that it does predict a mix: some aspects of the mouth are
geometrically scaled so as to accommodate food particles ( M 1.0 ), while
others are negatively allometric because they are determined by the fracture scaling of these same particles. The arguments thus become complex.
However, this complexity is also to the benet of this analysis because it
predicts a genuine change in the shape of mammalian body parts with size
that other theories really do not. Mixed rules are, as far as I can see, the
144
5 Tooth size
Table 5.1 Selected body-mass exponents of masticatory muscles and
oral structures predicted from an elastic fracture scaling analysis
Variable
Predicted exponent of body mass
Angle of jaw opening (gape)

Area of mouthslit
Area of the hard palate
Cross-sectional area of mandible
Diastema
Height of carnivore canines
Intra-oral space
Jaw length
Mesiodistal length of herbivore incisors
Muscle cross-sections
Muscle lengths
Muscle mass
Postcanine tooth area herbivores
Surface area of the dorsum of tongue
Tongue volume
0.0
0.67
0.67
0.5
0.08
0.33
1.0
0.33
0.33
0.5
0.33
0.83
0.5
0.67
1.0
only way to rescue allometric analyses from the substitutes bench in the
study of shape in biology.
What are the predictions? I give one example of how to calculate these
and then place the rest in a table, discussing them with the evidence.
The lengths of jaw-closing muscles need to be geometrically similar (
M 0.33 ) to stretch sufciently for the equally wide gapes that food particles
( M 0.33 ) enter the mouth. However, the (physiological) cross-sectional
areas of these muscles that generate the bite force need to scale only as
M 0.5 . Thus, adductor muscle mass is suggested to scale as M 0.5 M 0.33 =
M 0.83 . The result is that a muscle like the masseter should be short and
fat in a small mammal, but longer and skinnier in a larger one.4 Table 5.1
summarizes a number of the predictions like this from the theory.
A caveat
Alexander (1985) touched on a big problem with scaling theories that suggest
departures from geometric similarity. Over the range of body sizes seen
in mammals, body designs at the size extremes become absurd. Elastic
similarity, a theory associated with McMahon (1973, 1975), is somewhat like
this. Alexander (1985) shows that a size jump from a shrew to an elastically
similar elephant results in the latter having such fat legs that these would
Evidence for fracture scaling
145
not have space to t onto its body. This appears (appeared to me at any rate)
to kill the theory, but this may not be so. Allometric arguments are often
given an implicit upward direction, from small to large, because that is the
general direction of evolutionary change (Alroy, 1998). However, from a
non-biological perspective, it is also reasonable to argue this the other way
around from big to small. Starting with the elephant, what would the legs
of an elastically similar shrew look like? The answer is remarkably thin,
much thinner than they actually are, but its legs would be able to t onto
its body. The point is not that I wish to resurrect elastic similarity, but that
allometric analyses do not predict the absolute size of any organ in any
one organism, merely their size relative to that in other organisms. The
logic behind Alexanders caveat seems in fact to explain a lot about dental
evolution when there is reduction in body size.
The scaling theory based on food fracture given here is even more complex than envisaged in Alexanders objection to non-geometric scaling because I am suggesting a mixed isometric and negative allometric scaling
pattern and just for one small part of the body, the oral apparatus. It could
not be applied to the locomotor apparatus or musculoskeletal elements of
the trunk because they are not designed according to fracture similarity.
Food fracture is a uniquely complex and essential aspect of mammalian
life.
evidence for fracture scaling
The data in the literature use varying estimates of body size. The usual
measure is body weight, but headbody length is often employed as well.
Damuth (1990: his Fig. 12.1) studied a large sample of ungulates and showed
that headbody length is isometric to body weight and could be used to
represent whole body size for dental studies. Less satisfactory is the use
of whole skull length though, simply because this is dominated by the
mesiodistal dimensions of the teeth. Skull length is negatively allometric
to whole body size in mammals (Martin, 1990). Basicranial head length
(the length of the brain case) is too: larger mammals have relatively smaller
brain volumes, so their brain cases are also relatively smaller (Fortelius,
1985). There is also a considerable headache introduced by variation in the
line-tting techniques. These are beyond the scope of this book, but have
been reviewed by LaBarbera (1989). He advocates the reduced major axis
(Kermack & Haldane, 1950) because this method incorporates variation in
both the body part under consideration and body mass when constructing
trend lines. A least-squares regression, the conventional line tted by a
146
5 Tooth size
statistical package and reported by many researchers, can be converted to a

reduced major axis value very easily (the slope itself, but not its condence
limits). However, another technique, the major axis, is also found in the
literature and this is difcult to relate to the other techniques. Of course,
all matters little if the data points cling tightly to the line and the larger the
range of body sizes incorporated in the data, the more obvious the trend is
going to be.
There are more problems than this. Valid tests of allometric predictions
about body functions that are directly involved in feeding are complicated
by the fact that diets are not independent of animal size. There have been
several debates about how general or particular the dataset ought to be (e.g.
Kay, 1975a versus Pilbeam & Gould, 1975). Animal and plant feeders do not
predate similar mechanical targets, so it is reasonable to separate them. The
need to distinguish between types of plant-eaters is more complex. Very few
small vertebrates, including primates (Kay & Covert, 1984), are folivorous
(leaf-eating) because it is difcult to t a ruminant digestive system into a
small animal. Despite this, a very small bird called the hoatzin has it (Grajal
et al., 1989) and microtine rodents (lemmings and voles), which all weigh
<0.3 kg, possess something very similar (Hume, 1994). Nevertheless, very
small mammals may well be excluded from certain diets. Are any diets
excluded for large mammals? The answer is yes. Few large mammals are
insectivorous, unless they feed on social insects, because it is difcult to
catch them fast enough to full energy requirements (Kay & Covert,
1984). The largest terrestrial mammals generally consume structural parts
of plants, although even these (e.g. lowland gorillas: Rogers et al., 1990)
devour eshy fruits whenever they nd them. Bearing in mind that it is
difcult to know exactly how to homologise diets, I will simply keep to
a traditional herbivorecarnivore split in making a general review of the
evidence.
assumptions about the food intake
As with the rest of this book, I will consider the food intake rst, and follow
with the required changes in tooth size.
The rate of ingestion and the size of the oral cavity
There is ample evidence that the rate of ingestion of food volume is
proportional to M 0.75 (Peters, 1983; Shipley et al., 1994). However, the
Assumptions about the food intake
147
exponent for the scaling of chewing rates (the time needed to complete an average chewing cycle) is not well established. Large mammals
chew more slowly than smaller species, but taken together, the data of
Hendrichs (1965) and Fortelius (1985) suggest an exponent lying between
M 0.19 and M 0.23 (Fortelius, 1985). Kay (1978) makes a similar statement, but does not present the raw data. In contrast, Druzinsky (1993)
suggests a much lower exponent of M 0.13 . The stride frequencies of
limbs scale very similarly to this in mammals, which led Druzinsky to
suggest that jaw movements are less like body rhythms such as heart
rates and more like locomotor patterns. However, the reduced major axis
slope for all the data that he tabulates, his own and that of Fortelius
combined, provides 35 species and a body-mass exponent of M 0.17
(Table 5.2). This does not support either Fortelius nor Druzinskys argument very well and there is considerable variation not explained by
the t.
I do not know either if the internal volume of the oral cavity is proportional to body size (i.e. scaling as M 1.0 ). However, the weight of the
gut in mammals, which would be proportional to gut volume, has a bodymass exponent of 0.94 (Peters, 1983) and the volume of the rumen and the
capacity of the whole gut in ruminant animals is also isometric to body
size, i.e. M 1.0 (Demment & Van Soest, 1985; Owen-Smith, 1988), which
provides massive support for Fortelius viewpoint. The (nominal) surface
area of many other compartments of the gut in mammals appears to scale
as M 0.75 (Peters, 1983; Martin et al., 1985), suggesting that the food passage
time down the gut scales very similarly to other body rates (i.e. M 0.25 ).
While some evidence pertains to this (Robbins et al., 1995; Van Soest,
1996), there is an urgent need for more. Fermentation can add enormously
to the time of food retention, but models of mammalian gut function
otherwise seem to assume a constant food passage velocity (Alexander,
1991).
The analysis of feeding can become far more complex than this though,
needing to take into account differences between mammals in the time
devoted to feeding every day (Mysterud, 1998). However, the assumptions
about gut sizes and food passage times are not critical to the arguments
advanced here. There seems no point to stop and scratch my forehead over
this, so I go back a step to the mouth.
Indirect evidence about the volume of the oral cavity comes from gape.
Many mammals can open their mouths to very similar gapes, about 6070
according to Herring & Herring (1974), which is in accord with predictions.
148
5 Tooth size
Table 5.2 Chewing frequencies of mammals in relation to their body size

Species
Myotis lucifugus (little brown bat)
Suncus murinus (Asian house musk shrew)
Tupaia glis (common tree shrew)
Rattus norvegicus (Norway rat)
Pteropus giganteus (Indian ying fox)
Saimiri sciureus (squirrel monkey)
Aplodontia rufa (mountain beaver)
Tenrec ecaudatus (common tenree)
Cavia porcellus (guinea pig)
Mustela putorius (domesticated ferret)
Otolemur crassicaudatus (thick-tailed galago)
Pedetes capensis (springhare)
Oryctolagus cuniculus (domestic rabbit)
Didelphis marsupialis (common opossum)
Felis domesticus (domestic cat)
Macaca mulatta (rhesus monkey)
Marmota monax (woodchuck)
Ateles sp. (spider monkey)
Capra hircus (domestic goat)
Sus scrofa (miniature pig)
Canis familiaris (domestic dog)
Capra ibex (ibex)
Hemitragus jemlahicus (thar)
Homo sapiens
Oreamnos americanus (mountain goat)
Capra falconeri (markhor)
Ammotragus lervia (Barbary sheep)
Equus hemionus (Asiatic wild ass)
Bos grunniens (domesticated yak)
Tapirus terrestris (Amazonian tapir)
Bison bonasus (European bison)
Bos taurus (domestic cattle)
Camelus bactrianus (Bactrian camel)
Equus caballus (domestic horse)
Loxodonta africana (African elephant)
Body mass (g)
Chewing frequency (s1 )
7
41
150
200
480
550
583
700
900
1 350
1 500
2 300
2 500
2 500
2 500
3 500
3 790
6 000
20 000
22 300
36 287
40 000
50 000
60 000
60 000
64 000
66 000
210 000
250 000
272 155
300 000
476 272
500 000
650 905
2 812 273
4
5.46
4.2
5.21
1.72
2.8
4.83
2.62
6.13
4.59
3.18
3.15
3.61
2.56
3.25
2.99
1.72
3.07
2.4
3.03
3.16
1.68
2.12
1.3
1.28
1.72
1.71
1.25
1.15
1.45
1.26
1.45
1.1
1.27
0.97
Source: Data from Hendrichs (1965), Fortelius (1985) and Druzinsky (1993).
The hard palate traverses the length of the oral cavity. Ravosa (1991) measured its anteroposterior dimension in 108 species of primates, nding it
proportional to M 0.316 , close to isometry. Druzinsky (1993) also conducted
a mini-survey of palatal length in the mammals in his chewing frequency
study with an identical result.
Assumptions about the food intake
149
Ingested particle size

Is it correct to assume that larger mammals feed on larger single food
objects? Among carnivores, larger predators seem to feed on equally large
prey (Peters, 1983). Frugivores may show a similar pattern. Fleshy fruits are
designed for feeding on by vertebrates because these animals can disperse
their seeds effectively. Janson (1983) shows that fruit size is an important
component of dispersal syndromes and Gautier-Hion et al. (1985) show
that this characteristic affects fruit choice by mammals. Thus, for example,
small bats eat smaller gs than larger bats (Wendeln et al., 2000). However,
the net needs to be cast wide geographically to establish this because it is
not always true at a local level (e.g. for primates: Ungar, 1996). On a broad
evolutionary scale, the dispersion of tree species in tropical forests can best
be explained by larger frugivores eating larger fruits, very often containing
larger seeds, which they transport shorter distances than small frugivores
like bats and birds achieve with small, smaller-seeded, fruits (Howe, 1989).
Eriksson et al. (2000) document a substantial evolutionary increase in fruit
size during parts of the PaleoceneEocene, which could be related to the
evolution of larger mammals capable of dispersing the seeds of these fruits.5
A major stumbling block though is leaf-eating, where the relationship
between food item size and mammalian body size breaks down completely.
Plants do not want their leaves to be eaten and it would be absurd to
expect them to scale these to the size of the likely herbivorous threat because this strategy would probably encourage rather than deter the predator. There is no general tendency for bigger plants to have bigger leaves,
owers/inorescences or fruits/infructescences containing bigger seeds than
smaller ones (Cornelissen, 1999). This is not to argue the impossible. A tiny
plant (size in plants often being assessed, somewhat confusingly, by height)
obviously cannot have gigantic leaves. However, the biggest ower
(Rafesia) sits on the oor of the rainforest, while the biggest leaves in
that environment are on (often young) plants that are of about human
height (Turner, 2001).
The term browser means consumer of leaves other than grasses. The
major browsing pressure for the last 65 million years is thought to have come
from tiny invertebrates, not mighty mammals (Coley, 1983). Even so, the
ecological factors that determine tree leaf size are probably more to do with
light reception and climate than with defence against predation (Turner,
2001). In contrast, the major grazing (grass-eating) pressure has probably
been mammalian and the average size of grass leaves is probably inuenced
by this, becoming tiny (in volume) compared to their consumers.
150
5 Tooth size
It would appear certain that the rst assumption of the theory, that foods
are ingested as single particles, sized in proportion to mammal that feed on
them, is well and truly breached at this point. Nevertheless, plant-eating
mammals still need to nd food quantities proportionate to their size and
probably forage on the basis of food patches instead of individual leaf size.
If the tongue stacks these items during chewing such that stresses could be
transferred across them when loaded by the teeth, a batch of leaves could
perhaps be treated as a unit. Even so, comparisons would have to be limited
to other leaf-eaters.
Food toughness
It seems long to have been assumed that larger mammals take in food of
lower quality than smaller ones, especially herbivores (Bell, 1971; Jarman,
1974). Gould (1975) gave this as one possible reason for predicting a positive
allometry of tooth size in mammals. Quality is an imprecise term, but one
sense of the word connotes energy content. Foods of lower energy density
would denitely have to be eaten in larger quantities than richer food.
However, another use of the term low quality is to connote something
about the mechanical properties of foods, as in tough meat or brous
plant food. These uses of the word are relevant here and provide an alert
about the validity of assuming that toughness is constant. A tougher food
intake means higher bite forces (Eqn 5.6).
What about the converse: a reduction in food toughness? A reduction
in bite forces, i.e. of the jaw musculature, is a logical evolutionary response, but this would amount to little if tooth size did not reduce faster.
Why do frugivores (meaning fruit esh feeders) often have such small
cheek teeth (Freeman, 1988)? The answer could be that their foods are not
tough (Appendix B). This reasoning could apply equally well to piscivores
(sh-feeders). The toughness of muscle tissue in mammals and sh can be
compared very crudely by considering their collagen contents. Sato et al.
(1986) show in 22 species of sh that collagen constitutes 0.342.19% of
fresh muscle weight, while in Atlantic salmon, it is 2.9% (Aidos et al.,
1999). In mammals, gures of 1.99.6% are given for the muscles of cows
(Nguyen & Zarkadas, 1989) and 2.845.89% for those of domesticated pigs
(Zarkadas et al., 1988).6 These gures do not determine relative toughness,
but they do indicate that the sh muscle will be much less tough than
mammalian muscle, even if their collagen is chemically identical (which
it is not: sh collagen is generally much more soluble). Accordingly, tooth
size might be reduced in piscivores for this reason. The argument might be
Herbivory
151
extended much further to nectar feeders that probably hardly need cheek
teeth for most of their diet. Freeman (2000) shows that nectarivorous bats
have greatly reduced cheek teeth.
So, variation in food toughness affects predictions about tooth size. In
particular, a reduction in food toughness probably means a reduction in
tooth size. The application of this to cooking and human evolution in spelt
out in Chapter 7.
Food homogeneity
This is an obviously false assumption all foods are heterogeneous. The
issue though is anisotropy, i.e. food structure at a scale relevant to fracture
processes. The relevant scale for fracture-resistance mechanisms in plant
tissues is very small, at the cellular level. Full toughness in a plant tissue
is already expressed in tissues that are >1 mm in thickness (Chapter 4).
This is the approximate size range of swallowed particles and so does not
affect the validity of the assumption. However, there are fracture protection
mechanisms in animal soft tissues that act at a much larger scale and which
do reect directly on this homogeneity assumption. Accordingly, I will
consider predictions of fracture scaling given in Table 5.1 for herbivores
separate to carnivores.
herbivory
Jaw muscle size
Unlike for limb muscles (Alexander, 1985), there are no systematic data
on the sizes and shapes of jaw muscles. An ideal group for study would be
the bovids (the cattle family), but I do not know of enough data to test
the theory, either on the lengths of opening and closing muscles or their
physiological cross-sectional areas. However, there are some data on muscle masses. Cachel (1984) presented body-mass exponents for masticatory
muscle masses in 31 adult primates. These exponents vary from 0.755 (standard error 0.12) for the anterior digastric, an opening muscle for which
my whole mass predictions are weakest, to 0.959 (0.102) for the anterior
part of the temporalis. The rest of the temporalis, the posterior part, has a
body-mass exponent of 0.903 (0.115) while that for the masseter is 0.869
(0.122). The exponent predicted by fracture theory is 0.83 (Table 5.1),
which is clearly lower than most of these values, although not really
precluded by them. In contrast, limb muscles are at least isometric or show
positive allometry (Alexander, 1985), which helps perhaps to indicate that
152
5 Tooth size
different constraints must be operating on the jaw muscles than on those

in the limbs.
However, fracture scaling theory is much more specic than this. It
predicts that the masseter, for example, should change from a short fat
block in a small mammal to a longer narrower band in a very large one.
Fitting the latter onto the skull of a large mammal might involve raising
the jaw joint higher above the plane of the working surface of the teeth to
accommodate muscle length. This is a curious prediction because much
effort has been spent in trying to explain why the height of the jaw joint
varies such a lot in mammals. Fracture scaling suggests that it might be
size-dependent. Some data pertain to this in primates. Although there is
no clear evidence that larger primates have higher jaw joints (Lucas, 1981;
Smith, 1984), the highest joints do tend to be in the largest species, e.g.
all the living great apes including humans, the extinct apes Gigantopithecus
and Sivapithecus and the large extinct gelada baboon, Theropithecus oswaldi.
Grazers (grass-eaters) have higher jaw joints than browsers too, which may
reect body-size differences (Mendoza et al., 2002).
Bite force
There are very few mammals for which maximal bite forces are known
accurately. Much of the literature seems more concerned with trying
to nd record forces than with systematic investigation. The data in
Table 5.3, obtained by such diverse methods as contraction induced by
electric shock, voluntary bites on transducers and the results of bite mark
analysis (described in Lucas et al., 1994), give no great condence in any
exponent.
Demes & Creel (1988) took another approach, estimating the physiological cross-sectional areas of the masseter and temporalis muscles and
the lengths of their lever arms from landmarks on skulls. From this, they
calculated bite force equivalents for species of hominoid primates. These
values were then checked against real bite force data on the long-tailed
macaque and the human (Table 5.3) with good agreement. Hill et al. (1995)
conrmed the validity of this for Japanese macaques (Macaca fuscata) using a bite mark analysis. Individual monkeys were seen trying to open a
woody gall with their teeth, sometimes succeeding, otherwise failing. In
doing so, they left many bite marks on gall fragments that could be used
to calculate the bite forces that they were developing with their anterior
teeth. These calculations agree well with the bite force equivalents as estimated by Demes & Creels method. Anyway, the important point here is
Herbivory
153
Table 5.3 The maximum bite force reported for a variety of mammals
Species
Norway rat (Rattus
norvegicus)
Cat (Felis domesticus)
Long-tailed macaque
(Macaca fascicularis)
Japanese macaque
(M. fuscata)
Rhesus macaque
(M. mulatta)
Human (Homo sapiens)
Orang-utan
(Pongo pygmaeus)
a
Location
Incisors
Canines
Carnassials
Molars
Incisors
Premolars
Molars
Molars
Maximum bite Body weight

(kg)
force (N)a
Source
50.2
0.56
Robbins (1977)
227.9
274.4
333
(females)
487
3.8
3.6
Buckland-Wright
(1975)
Hylander (1979a)
11.08.0
Hill et al. (1995)
242 (males)
185 (females)
400
2000
(females)
7.7
5.4
65
36
Dechow & Carlson

(1990)
Carlsson (1974)
Lucas et al. (1994)
The data include muscle-twitching experiments, but are not limited to them.
that Demes & Creel (1988) found that tooth areas are isometric to these
bite force equivalents, which supports a critical assumption made in the
fracture scaling theory that bite forces match tooth sizes in mammals
with similar diets.
Incisor tooth size
Those anterior teeth involved in sculpting the bite size, the incisors of herbivores, appear to be isometric to body mass e.g. incisal widths of bovids
(M 0.33 : Illius & Gordon, 1987) and of cercopithecoid primates (M 0.312 :
Hylander, 1975; although this slope, from least-squares regression, is certainly an underestimate). These results are in accord with expectation
(Table 5.1).
Postcanine tooth size
The largest dataset is on small mammals, mostly rodents, with the area
of the rst lower molars of 288 species plotted against headbody length
(Creighton, 1980). The slope is equivalent to M 0.61 , considerably below
isometry but well above an exponent of 0.5. There are many other studies,
but Fortelius (1985) summarizes them well and adds his own superlative
data. Generally, body-mass exponents range from negative allometry to
isometry in herbivores and primates (Table 5.4). While the data do not
154
5 Tooth size
Table 5.4 Abridged data assembled by or originating from Fortelius (1985,

1990) on the scaling of postcanine tooth size a in herbivores with body mass
Mammalian group (sample size)
Exponent of body mass (standard error)
Hystricomorph rodents (14)

Primates (77)
Uppers 0.72
Uppers 0.62
Lowers 0.69
Uppers 0.65 (0.06)
Uppers 0.65 (0.07)
Lowers 0.62 (0.07)
Uppers 0.58 (0.07)
Lowers 0.57 (0.07)
Uppers 0.62 (0.07)
Lowers 0.62 (0.09)
Uppers 0.65 (0.04)
Lowers 0.62 (0.04)
Uppers 0.56 (0.14)
Lowers 0.43 (0.13)
Ungulates (43)
Non-selenodont artiodactyls (20)
Selenodont artiodactyls (22)
Bovids (13)
Rhinoceroses and hyraxes (8)
Bilophobunodonts (5)
a
Postcanine tooth size is estimated as the sum of the area of all premolar and molar
teeth. All slopes are major axes through the data points.
really support the theory very strongly, they do suggest that exponents tend
often to run signicantly below isometry, a sentiment with which Martin
(1990) seems to agree.
The diastema
Fracture scaling predicts that, in larger mammals, the cheek teeth will be
housed in oversize jaws. A simple consequence of this is that empty spaces
should appear, e.g. between the anterior and posterior teeth or behind the
posterior teeth. Several herbivorous mammalian orders, such as the Lagomorpha (rabbit family), Rodentia, Artiodactyla (even-toed ungulates like
cattle) and Perissodactyla (odd-toed ungulates like horses), have a gap (a
diastema) between the premolars and anterior teeth (Fig. 3.10). Fracture
scaling predicts that this diastema should be size-dependent, with a exponent of M 0.08 (Table 5.1). No studies have measured the size of the diastema
directly. Indirect evidence comes from the study of Williams (1956) on the
relationship between the length of the molar tooth row and the length of
the lower jaw, but his data do not seem to show size-dependence. However, the diastema is larger in grazers than browsers (Mendoza et al., 2002),
which might reect the larger body sizes in the former.
Herbivory
155
The diastema gured prominently in the theoretical work of Greaves

(1978), who was concerned with the optimal location of the postcanine
teeth in ungulates. The mesiodistal length of these teeth is, he says, constrained by the effectiveness of the bite force. The tooth row cannot extend
too far posteriorly or else the jaw joint would dislocate because its articular
surfaces would be distracted (Chapter 3). The anterior limit is not xed,
but is constrained by a diminishing bite force the further forward a bite
is focussed. This theory could account for the lack of inuence of diet on
the mesiodistal length of the postcanine row in many mammals (Fortelius,
1990). However, even Greaves doubted that it could explain the presence
of the diastema. Instead, he also considered the idea that if a long jaw is
required for gape, but the length of the postcanine row is constrained by
characteristics of the bite force, then a diastema might arise as a consequence. This anticipates the argument offered here, but for very different
reasons.
The only other explanation for a diastema, as Greaves relates, is that it
provides a mechanism for separating the function of anterior and posterior
teeth, e.g. in rabbits (Ardran et al., 1958) and rodents (Landry, 1970). Rodents possess soft-tissue anges that help to seal off the posterior part of
the mouth from the front of the mouth where they bite some items, like
seed coverings, that they do not mean to ingest. However, there must be
simpler ways of isolating the mouth from the actions of the anterior teeth
the modiolus (Chapter 2) does this, so the explanation given by fracture
scaling seems more plausible.
The jaw bones

The bones that support the dentition obviously need to be built to take
masticatory stresses. Fracture scaling predicts that the cross-sectional area
of the jaw is related to M 0.5 while the jaw height, if this took an equal
share of load-bearing responsibility, would scale as M 0.25 . Smith (1983),
Hylander (1985) and Bouvier (1986) discuss this in anthropoid primates,
but Ravosa (1991) gives the largest dataset, nding that the height of the
mandible under the second molar has an exponent of about 0.36. An
obvious reason for this is the need to prevent excessive jaw deection under
load: the more that the jaw deects, the more this impairs the alignment
of upper and lower teeth during mastication. If jaw length scaled as M 0.33 ,
while the depth of the mandible grew only as M 0.25 , then the jaws of larger
mammals might become unacceptably exible.7
156
5 Tooth size
the best test
Most size changes in evolution are enlargements. It is difcult to generalize

but these changes are probably slow and thus involve subtle changes in
diet in a mammalian lineage, both in the external and internal physical
properties of foods. Any such dietary shift concomitant with size change
reduces the value of scaling predictions. What are needed are rapid changes
in which there is little time for dietary adjustment. Very much faster size
changes are known in some populations of mammals trapped on islands,
where larger herbivores seem to respond not by enlarging, but by reducing in
size.8 There are quite well-documented examples in living primates like this
(Fooden & Albrecht, 1993), but the best examples of dwarng are mostly
from fossil evidence. The best documented are probably insular elephants,
with many instances of an evolutionary reduction in body size down to
just a few per cent of that of their ancestors (Roth, 1990). It is thought that
such rapid change is an adaptation to the diminishing size of food patches,
i.e. not to a change in diet but to a change in food distribution. If so, this
provides a perfect test for the predicted negative allometry.
Maglio (1972) claimed that dwarf elephants have proportionately larger
teeth, but the problem with a test on this is that elephants bring their
molars into the mouth one by one. As each tooth wears out, it is shed,
so making postcanine tooth size a difcult concept to quantify. Luckily,
other mammals provide examples too. Gould (1975) examined data from
a short lineage of four hippopotami. He showed negative allometry, i.e.
that the smaller hippos had proportionately larger teeth. The slope was
apparently equivalent to M 0.58 , higher than predicted in Table 5.1. However,
basicranial skull length was used as a measure of body size. This is likely to
have inated the slope (Fortelius, 1985) and the body-mass exponent may
actually be close to 0.5. Fortelius (1985) also discusses dwarng in several
lineages of domesticated mammals where evidence for negative allometry of
the teeth with dwarng appears strong. The only objectors, in fact, appear
to be Prothero & Sereno (1982), who studied dwarf rhinoceroses. They
state that these mammals did not have relatively large teeth, but relatively
smaller skulls instead. From the standpoint of dental function, there isnt
much difference between the two interpretations, but their case needs more
evidence from other groups.9
the appearance of jaws of different sizes with linear
elastic fracture scaling
The evolutionary implications of fracture scaling are a lot easier to grasp
with illustrations than with words. Figure 5.4 shows the effects of size
Jaws of different sizes
157
Fig. 5.4 A 15-fold body mass increase and its effect on jaw and tooth form in a male Old
World monkey, Miopithecus talapoin. The ancestral form at left is expanded (at right)
either by linear elastic fracture (left hemi-jaw) or geometric scaling (right hemi-jaw). This
fracture scaling produces lots of free spaces around the postcanine teeth, which has been
lled arbitrarily by expanding the lower anterior premolar. The talapoin was chosen at
random, but is the smallest known Old World monkey, including the fossil record (Delson
et al., 2000).
increase on the lower dentition of a male Old World monkey, the talapoin (Miopithecus talapoin). In Fig. 5.4, the real jaw (at left) is called the
ancestor, while the scaled enlargement, a 15-fold enlargement in body
mass, is called the descendant and lies to the right. The descendant diagram shows two hemi-jaws, one geometrically scaled and the other scaled
according to linear elastic fracture. The anterior dentition and the length
of the jaw adapt to gape, so are always geometrically scaled and t the jaw
perfectly. However, linear elastic fracture scaling results in a proportional
decrease in postcanine size. The excess jaw space that this creates could
either accommodate a diastema, e.g. between the anterior premolar and
the canine, or else the whole postcanine segment could be shifted forwards
to leave an enlarged space behind the last molar. Nothing about fracture
158
Theropithecus gelada lower jaw
5 Tooth size
Mandrillus sphinx lower jaw
Miopithecus talapoin lower jaw
Fig. 5.5 A comparison between the lower jaw of the male talapoin monkey and those of
two baboon species shows that the lower anterior premolar has a large extension (arrowed)
in the baboons. The space for this can be created by linear elastic fracture scaling of the
postcanines shown in Fig. 5.4.
scaling suggests how this arrangement should be achieved, but one possibility (shown in the gure) is that the front part of the lower anterior
premolar could be expanded. This expansion is seen in larger Old World
monkeys as depicted in Fig. 5.5, where the hemi-jaws of two large Old
World monkey species, the gelada (Theropithecus gelada) and the mandrill
(Mandrillus sphinx), are shown to the left of the talapoin jaw, all brought
to the same size by simple geometric enlargement. The extended ridge of
the lower anterior premolar rubs against the posterior ange of the upper
canine, so sharpening it (Zingeser, 1969). It is nothing to do with diet and
is much more pronounced in the baboons than in the talapoin, perhaps
simply because fracture scaling creates space for this feature.
A modest size increase reveals the truth of Alexanders caveat quite
quickly: jaws and teeth do not t properly when geometric similarity is
departed from. However, the consequences of size increase do not seem
critical. In contrast, downsizing seems to produce far more devastating
changes.
Figure 5.6 takes the identical procedure shown in Fig. 5.4, but instead
reduces the body size of the talapoin by 15-fold. Linear elastic fracture scaling
now produces obvious postcanine crowding and the last molar does not
have enough bone to seat itself on. If this consequence is accepted, then
dwarsm involves serious problems: the need for proportionately larger
cheek teeth, but without the space to house them. If so, then it could only
evolve under great duress. Alexanders prediction is borne out but, rather
Carnivory
159
Fig. 5.6 The effect of a 15-fold body-mass decrease on jaw and tooth form in Miopithecus
talapoin. Linear elastic fracture scaling produces dental crowding with insufcient
space in the jaw for the last (third) molar. See text and compare to the size increase shown in
Fig. 5.4.
than indicating that fracture scaling is impossible, suggests that it could

lead to tooth loss in dwarng mammals. As stated in Chapter 2, mammals
do have variable numbers of teeth and this theory could explain it. This is
returned to in Chapter 7.
carnivory
Non-linear elastic behaviour
The third assumption of the fracture scaling theory was that food particles
had linear elastic behaviour. This is not so for the vertebrate soft tissues
that carnivores eat, which display very non-linear behaviour (Chapter 4).
However, a remarkable feature of fracture scaling is that it is capable of
dealing with this behaviour. The following analysis of non-linear elastic
behaviour was worked out by Mai & Atkins (1975) and is only summarized
here.
Linear elastic behaviour assumes that stress is in proportion to strain:
=
160
5 Tooth size
DISPLACEMENT-LIMITED
J-shape
r-shape
Stress
Stress
STRESS-LIMITED
r-shape
Stress
linear
J-shape
Strain
Strain
Strain
Fig. 5.7 The various types of stressstrain curves exhibited by biological tissues have been
pressed into a format whereby the stress ( ) and strain () are related by = k n , with k
being a constant (Purslow, 1991a). In the text, k is assumed to be 1.0. When n < 1, the curves
are r-shaped, and when n > 1, J-shaped.
where = E, the Youngs modulus. Non-linear behaviour in materials

such as animal soft tissues can be modelled by using a power law:
= kn
where k is a constant and the exponent n > 1.0 for the concave J-shaped
curves of most animal soft tissues, unity for most plant tissues and animal
hard tissues, but <1.0 for materials like rubbers (Fig. 5.7).
Mai & Atkins (1975) show that the equivalents of Eqns 5.75.9 for such
tissues are
Fx
= (n+2)(n+1)
(5.11)
Fy
x
= n/(n+1)
y
(5.12)
ux
= n/(n+1)
uy
(5.13)
When n is set to 1.0, this trio of equations reduces to Eqns 5.75.9. If n

gets large, then the ratio of forces approaches rather than 1.5 , stresses
move toward 1.0 rather than 0.5 and displacements approach rather
than 0.5 . Thus, the predicted exponents in fracture scaling vary with the
deformational behaviour of the tissue.
Yamada (1970) shows many forceelongation curves for mammalian soft
tissues for which n = 3 is a very approximate t.10 Assuming this to be typical, then the length of the blade of a carnassial tooth of a carnivore (Fig. 6.2),
Carnivory
161
as an example, would be predicted to scale as M 0.21 , considerably lower

than for a linear elastic response (with M 0.25 ). However, it is important to
understand exactly what Eqns 5.115.13 do: they predict the scaling of forces
for propagating already-formed cracks in geometrically scaled particles, not
the difculty in initiating these cracks. Yet, it is crack initiation that is especially difcult in animal soft tissues (Chapter 4). The analysis may then be
inappropriate. Further, it may be invalidated completely if the structure of
foods is sufciently disconnected as to be notch-insensitive (Fig. 4.12 and
Appendix A). The next section discusses this aspect of food behaviour in
relation to tooth size.
Structural connectivity
Disconnecting parts of a tissue results in a structure that does not respond
to fractures in the way that true solids do (Chapter 4). The test of this,
described in Appendix A, is called notch sensitivity. Notch-insensitive
structures do not transfer strain energy between structural elements and
so defeat fracture mechanics by making the fracture stress independent of
size. Galileo was right, but just in this special case. However, foods as a
whole could be a special case are any foods actually notch-sensitive?
I claim that plant foods basically are, mostly being cellular solids of a
completely connected closed-cell variety (Gibson & Ashby, 1999). Examples like the seed coat and pod in Fig. 4.12 are probably exceptional.
Large-scale heterogeneities such as the veins in a grass leaf can produce
something close to notch insensitivity when the whole leaf is loaded
in tension (Vincent, 1982), e.g. during ingestion, but this will not be
the case during mastication. So fracture scaling seems to apply to plant
tissues.
It is the connective tissue of vertebrates that are largely notch insensitive. In consequence, the complexities of non-linear elasticity are unlikely
to apply to carnivores at all. Instead, we should expect that their teeth,
such as the carnassials of cats, are not fracture scaled, but geometrically
scaled. Van Valkenburgh (1990) shows that the length of the lower carnassial tooth, the rst molar, of felids (members of the cat family) and mustelids
(weasel family) is approximately isometric to body mass. The reduced major axis for felids gives M 0.35 (with a sample size of 16 species) while that
for mustelids is M 0.31 (19 species). However, there is considerable variation
between taxonomic groups, undoubtedly due to the dietary diversity of the
Carnivora.11
Evolutionary effect on postcanine tooth size

(a) Linear elastic fracture (b)
Log-linear behaviour
(c)
Linear elasticity with

elastoplastic fracture
Plastic fracture or
completely
notch insensitive foods
(d)
Strain
0.5
always very
negative allometry
Strain
[(n+2)/(n+1)]/3
negative allometry but

as n
0, isometry
Plastic fracture
yield point
Strain
0.5
M +M
[(n+2)/(n+1)]/3
negative allometry
between (a) and (b)
Stress
Stress
Stress
Stress
decreasing n
yield point
Strain
0.67
always isometry
Fig. 5.8 The possible types of stressstrain curve of major foods in mammalian diets and their inuence on postcanine tooth size. Mammals
that eat foods with some denite elastic response and which have structural connections capable of transferring loads will show some negative
allometry of the dentition. A diet of food that is completely plastic or with complete notch insensitivity (i.e. complete disconnection of
cload-bearing elements) will result in dental isometry. Positive allometry is not possible.
Scaling analysis and tooth size
163
Other mechanical behaviours exhibited by foods

Any true solid will normally have sufcient elasticity to crack. However,
fruit esh does not have to support a load or even maintain its own shape
and so is relieved of the need for elastic response. In many species of fruit,
the esh can become almost completely plastic at ripeness. This tissue
would not really be fractured by teeth, but rather ow away from them.
The dental task may involve trapping this tissue in order to get individual
cells to burst. In this way, it is possible to argue that the work done on a
plastic tissue is actually proportional, not to surface areas, but to the volume
of tissue being processed (Atkins, 1999). This volumetric action leads to a
prediction of isometry of the postcanines of frugivores (if the fruit tissue
behaves this way), as going back to Eqn 5.1 shows.
As a nal possibility, foods could exhibit a combination of linear elastic
behaviour, followed by yielding and a small amount of plastic deformation.
This is elastoplastic fracture, for which Atkins (1999) gives effectively
tooth size M 0.50 + M [(n+2)/(n+1)]/3 .
(5.14)
overview of scaling analysis and tooth size

Figure 5.8 sums up the various possibilities for the scaling of tooth size.
They vary from strongly negative allometry to isometry.12 Positive allometry
appears to be impossible and rather than represent the norm for the scaling
of tooth size in mammals, it probably indicates a dietary shift in larger
mammals towards tougher food.
It might be felt by this stage though that the number of alternative scaling
possibilities acts against the usefulness of such theories, making them seem
innitely exible and capable of being twisted to provide an exponent for
every circumstance. This is false: a mammals dentition reects what it
eats. The diet dictates the dental optimum and the better that this can be
characterized, the more precise will be the theoretical prediction.
Accordingly, dividing up mammals into herbivores and carnivores is
obviously inadequate: tooth size predictions can only be made with condence on species groups with homologous diets. Without knowing what
diets are homologous, it is impossible to predict tooth size. Will folivores
have larger teeth than carnivores or frugivores or vice versa? Right now,
it is impossible to predict this. Did this analysis predict anything at all? I
answer yes, but such predictions lie only in large-scale trends. The disparity between the scaling of postcanine tooth size and jaw size that follows
164
5 Tooth size
negative allometry is a good example of this and will be returned to in

Chapter 7.
the importance of sexual dimorphism
All the preceding arguments were derived from considerations of
metabolism, in terms of the dimensions of the mouth and its component parts for processing ingested food particles at an adequate rate. Body
weight entered the arguments only as a surrogate for basal metabolic rate.
Its use though presents considerable statistical difculties when analysing
data for mammals with pronounced sexual dimorphism (sexual differences
other than the genitalia). In some species, the body weight of males can
be double that of females on the average. If the body weights of enough
individuals in those species could be plotted as histograms, then there
would be a bimodal distribution attributable to sex: the males on one side,
females on the other. Sexual dimorphism can go either way, but in mammals the biggest differences are always associated with larger males (Ralls,
1977). The overall mean of such a bimodal distribution is meaningless because it represents the weight of very few animals males would be larger
than the mean while females would be smaller. In such mammals then,
the measurements of males and females have to be considered separately.
This does not just apply to body mass. Bimodal distributions are a
feature of the sizes of other body parts in these mammals provided that
their distributions are normal. This includes the teeth (e.g. Zwell &
Pilbeam, 1972; Lauer, 1975). Lauers study includes the extreme case of sexual dimorphism in Papio hamadryas cynocephalus, a primate baboon species
with a wide distribution in sub-Saharan Africa (and including a population
of very small-bodied forms, the kindae baboons). However, rather than
being a nuisance, sexual dimorphism can actually help establish whether
there is really any critical relationship between variables like tooth size and
body weight.
Inconsistencies arise almost immediately when such an analysis is performed. Females in primate species where the males are largest appear to
have relatively larger postcanine teeth (Kay, 1978; Harvey et al., 1978). Lucas
et al. (1986b) and Cochard (1987) both investigated this in detail and found
that the larger the male is relative to the female, the smaller is the relative
size of the males teeth. This is actually an extraordinarily regular feature
that can be predicted by the degree of dimorphism in body weight (Fig. 5.9).
In the extreme, males have only 60% or so of the tooth size predicted for
their weights.
The importance of sexual dimorphism
165
Fig. 5.9 Relative male upper postcanine tooth size (the sum of the areas of the last premolar
and all the molars) plotted against the ratio of male to female body weight for 49 species
of primates. The larger the male is relative to the female, the relatively smaller is the size
of its postcanine teeth. Postcanine tooth size was estimated as the sum of the average area
(buccolingual width mesiodistal length) of the last maxillary premolar and the upper three
molars. Data for males and females were kept separate. Postcanine tooth size in females was
then regressed against mean female body weight. Relative male posatcanine tooth size was
then estimated as observed tooth size for a given body size divided by that expected from
the regression line for females. The relationship is highly signicant (r = 0.86; p < 0.001).
(From Lucas et al. (1986b).)
This is all very difcult to explain from an energetic perspective. In sexually dimorphic species where energy requirements have been well documented, males tend to have higher basal metabolic rates than females even
after correcting for body weight differences (Benedict, 1938; Dale et al.,
1970). This is the opposite of the trend in relative postcanine tooth size.
Pregnancy and (much more so) lactation increase the energy requirements
of females. These factors could perhaps explain the trend in Fig. 5.9, but
if so, then why in primates where males and females are exactly the same
body size are their postcanine teeth also exactly the same size? Females of
these species still get pregnant and lactate. Additionally, there are some
mammals in which females are larger than males. These are not many, and
the difference is not great, but it is enough to try to ascertain if in these
166
5 Tooth size
species relative tooth size then favours males. Lucas (1980) and Fortelius
(1985) assert that this seems to be so. This result suggests very clearly that
there is no critical relationship between postcanine tooth size and body size
and I am forced to conclude that, unless some highly systematic variation
in energy requirements or diet between the sexes is found, tooth size cannot
actually be dealt with accurately in the allometric format above. This does
not mean that any of the above arguments are wrong, because they related
mouth size to food properties, merely that the method of testing it is not
correct.
To go back to the beginning, the mammalian body requires a food
input delivered at a certain rate. The teeth adapt to the exact form of that
input in order to deliver on that rate. What is needed now then is further
documentation of that input and this is the direction to which the chapter
now turns.
evidence that postcanine tooth size affects

the rate of food breakdown
Before wandering there though, it seems important to hit the ground with
some basic facts. Since this book is claiming that foods break down in the
mouth at rates that can be inuenced favourably by the right combination
of tooth shape and size, this seems the right point to lay down evidence
showing that tooth size really does matter. Otherwise, it might be queried
whether a little cusp on a molar, like a hypocone for example, matters one
way or another to the survival of a mammal.13
There is some physiological evidence for the importance of postcanine
size in ruminants (Perez-Barberia & Gordon, 1998b) and there is a recent
report claiming that molar tooth size has an effect on tness in a wild
population of mantled howler monkeys, Alouatta palliata (DeGusta et al.,
2003). However, it is a century of work on this topic in dentistry that comes
to my aid and which is fairly conclusive in showing that small variations in
tooth size exert a substantial inuence on the rate of food breakdown. The
evidence can at times be tricky to interpret because variation in the methods
of measurement can introduce distortions. In particular, rather arbitrary
measures of the rate of food breakdown, like masticatory efciency and
masticatory performance (Manly & Braley, 1950), exaggerate differences in
food breakdown rates produced by different individuals. The following is
a precis of dental ndings.
Postcanine tooth size
167
The rst paper to report a correlation between the rate of food breakdown
and the area of the postcanine teeth (in humans) was authored in the late
1940s (Yurkstas & Manly, 1949). Their work culminated in a review by
Yurkstas (1965). It is worthwhile documenting the techniques used. The
state of the working surface of the postcanines was estimated by placing
a sheet of opaque dental wax between upper and lower cheek teeth and
asking subjects to bite on it. The wax was thinned by the bite to a degree
that depended on the proximity of parts of the upper and lower teeth.
This plastically distorted wax sheet was then placed under a light of xed
intensity and the proportion of light transmitted through it recorded. Then,
each subject was asked to make 30 chews on a given volume of roasted
peanuts and spit out the fragments. The expectorate was washed through
a 2.0 mm aperture sieve and the percentage of peanut fragments passing
the sieve employed as an index of food breakdown called the masticatory
performance. Results for a large number of subjects showed a very strong
correlation between masticatory performance and dental state (effectively,
the size of the postcanine tooth surface). The relationship was non-linear,
but this might simply be due to the food sizing procedure. Helkimo et al.
(1978) and Kayser (1980) have conrmed this result with slightly different
techniques, while Wilding (1993) has found that even differences in tooth
size in one individual on the left- and right-hand sides of their mouths will
inuence breakdown rates in unilateral chewing.
All this work gives little information though on which parts of the dentition are most critical for food breakdown. Experiments by Lambrecht
(1965) on the size of the postcanine segment of articial dentures were
more specic, proving that a reduction of just 1.0 mm in buccolingual
width of the working surface was sufcient to reduce chewing efciency
by as much as 20%. Luke & Lucas (1985) examined a range of characteristics of the oral cavity of 32 subjects in relation to the rate of breakdown of the median particle size of raw carrot. They studied the shape
of the dental arcade, the height of the palate and intra-oral volume in
addition to postcanine tooth size. The most important factor by far was
tooth size. However, they found that females have higher rates of breakdown for their postcanine tooth areas than males (Fig. 5.10) making it
important to place the teeth in their oral context (e.g as a proportion of
the oral surface) and not consider them purely in isolation. In summary,
all papers on this subject agree that small modications to the working
surface of the molars have a strong effect on the physiological rate of
breakdown.
168
5 Tooth size
Fig. 5.10 The rate of breakdown of raw carrot particles in 32 human subjects as the reciprocal
of median particle size plotted against their postcanine tooth sizes shows that females have
generally higher rates of breakdown than males on this basis. (After Luke & Lucas (1985),
with permission from the British Dental Journal.)
the external physical attributes of foods

and tooth size
The external physical attributes of foods, as may be recalled from Chapter 1,
describe the food surface. This includes:
(1) The extent of the food surface, obtainable from knowledge of food particle
size and shape and the total volume of food particles in the mouth, which
affects the instantaneous rate of mastication.
(2) The properties of the food surface. This category includes stickiness (both
between food particles and between a particle and the oral mucosa) and
abrasiveness. These properties will both affect the rate of mastication
averaged over a time period. However, stickiness and abrasiveness operate on different timescales. The stickiness of food particles changes
during the time-span in which a mouthful is being processed, as saliva
coats food particles and may be absorbed by them. Stickiness is important in forming a food bolus because it helps to clear the mouth of
particles during swallowing (Chapter 3). In contrast, the abrasiveness of
foods acts over a much longer duration the lifespan of the individual
leading to wear that may cause the rate of mastication to decline. This
is dealt with in Chapter 6.
The concept of breakage sites
169
All these food surface attributes are united by their effect on the probability of particles being broken by the teeth. Tooth size is the most important
oral adaptation to these characteristics because of its dominant inuence on
the probability of particle fracture (the selection function). Thus, a study
of the adaptation of tooth size to diet requires a systematic examination
of each of these attributes. The immediate problem though is a working
denition of relative tooth size. Tooth size relative to body weight is clearly
not adequate for anything this specic. More logically, in terms of dental
function, postcanine tooth size should be considered relative to the size of
the oral cavity as a whole since the probability of fracture should depend
on the proportion of the oral surface given over to the teeth. However,
the fracture scaling arguments above suggested that the size of the oral
cavity could scale differently to tooth size, such that the postcanine teeth
would not occupy the same proportion of the oral surface in all similarly
adapted mammals. This is a problem to which I will offer a specic solution
later on.
the concept of breakage sites
The folding of tooth surfaces into cusps, fossae, crests and ridges makes
some parts of these surfaces inevitably more likely to contact food particles
than others. This has led to the concept of breakage sites (Chapter 3),
hypothetical regions of the tooth where food particles get contacted and
broken (Lucas et al., 1985, 1986c; van der Glas et al., 1992). Here, I describe
several implications of this concept for understanding an adaptive response
in terms of tooth size. In any of the following theoretical examples, only
the external attribute being discussed is presumed to be varying.
Food volume (mouthful)
Whenever the number of food particles exceeds the number of breakage
sites, then it is obvious that not all particles can be broken in that chew. It
follows immediately that the greater the excess of food particles, the slower
is the rate of food size breakdown per chew simply because fewer particles in
a mouthful get broken in any chew. Accordingly, larger mouthfuls must be
reduced in size at slower rates. There is good evidence in humans that this
is correct (Lucas & Luke, 1984a). The probability of fracture of all particles
appears to be reduced by an increase in food volume being processed. However, volume for volume, it takes a smaller number of chews (and a smaller
number of swallows) to ingest food in larger mouthfuls than in smaller ones
170
5 Tooth size
(Lucas & Luke, 1984a). There is an upper limit, of course, when the jaws
get jammed but no mammal appears to try to chew food in such quantity.
If a mammal can nd food in large enough patches to ingest large mouthfuls, then it will not experience any difculty to getting the necessary food
throughput to the gut. However, if a mammal can only nd food in small
patches, it faces a problem. The evolutionary answer for a mammal adapting
to the ingestion of small mouthfuls is to increase the number of breakage
sites on its postcanine teeth, i.e. to evolve larger teeth. Even if only a limited amount of tooth surface were then utilized per chew, the benet in
increasing its food throughput would be important.
Food particle size
In general, larger particles get fractured much more commonly than small
ones. We know this in humans from data on the selection function (Fig. 3.4)
(Lucas & Luke, 1983; van der Glas et al., 1987). This leads to the rate of
food breakdown being far greater with large particles than with smaller
ones. Thus, for any given food mouthful, the most important inuence on
the rate of food particle size reduction is food particle size. Generalizing,
we can argue that the smaller the particle size in the food intake, the slower
the rate of breakdown per chew.
This is not necessarily any problem though. The smaller particles are,
the less mastication is needed to swallow them because ingested sizes are
closer to the hypothesized particle size threshold that can be swallowed
(discussed in Chapter 3). The evolutionary importance of this only arises if
ingested food particles are chemically sealed. It is then necessary to ensure
that all particles are broken at least once, or else they would simply become
dead weight in the gut, passing the entire length of the gastrointestinal tract
undigested. For two equal volumes of chemically sealed particles ingested
into the mouth, one with large particles and the other small, then the larger
the difference between these particle sizes, the slower would be the relative
rate of processing the small particles, i.e. making sure that all are broken
prior to swallowing. Thus, if a mammal habitually ingests small chemically
sealed particles, then it would benet from increase in postcanine tooth
size to increase the probability of fracture.
Food particle shape
Food particles are ingested in a wide range of shapes, from relatively isodiametric particles (i.e. with all three dimensions being more or less equal)
The concept of breakage sites
171
like many seeds or eshy fruits, to sheet forms (with one dimension being
small), such as such wind-dispersed fruits or leaves, to rod-like particles
(with one dimension being large), such as stalks, stems or branches. The
shape of a particle must be very important initially because teeth act on
food surfaces and the presentation of that food surface is very dependent
on shape. The rate of size reduction also depends very greatly on shape.
If a sphere, rod and sheet are compared, then the rate of production of
new surface will depend really on their smallest dimension. The initial
advantage that a sheet-like particle like a leaf has, with its very high specic
surface, over a sphere, for example, is rapidly lost because fracture of the
sphere exposes surface more quickly than fracture of the sheet. The most
important effect of food particle shape on postcanine tooth size point is
likely to be that sheet-like and rod-like foods are in reality small objects
because their volumes are small. If they are sealed particles, then their
efcient processing might require enlarged postcanines.
Stickiness
The stickiness of a food is produced by a combination of its own surface
properties, any juice liberated from it during food comminution, and saliva.
Particles of most fruit esh and animal soft tissues stick together very
easily and make a swallowable bolus rapidly. However, the cuticle of leaves,
a hydrophobic waxy layer (Barthlott & Neinhuis, 1997), will resist the
leaf surface being wetted in the mouth and inhibit bolus formation. This
probably has considerable implications for tooth size, but not with equal
effect along the tooth row.
During chewing, the tongue throws food particles laterally (Fig. 3.7). A
sticky bolus, e.g. of fruit esh, will not spread along the tooth row. This
puts a premium on postcanine surface in the central part of the tooth row
(e.g. in the rst molar region) where most of the food particles would be
expected to be broken. Foods like leaves or carrot, however, would disperse
much more evenly along the tooth row, because they do not form a bolus
easily, suggesting that a more even distribution of tooth area is then optimal. Thus, a long narrow tooth row is suited for non-bolus-forming foods
while a wide short dental arch would be best for a mammal eating sticky
food.
Table 5.5 tabulates these predictions while Fig. 5.11 shows them graphically. For a test, it remains just to try to put real food names to these
attributes. Just a brief discussion sufces here. Leaves are small chemically
sealed sheet-like particles that do not form boluses. So leaf-eaters (folivores)
172
5 Tooth size
Table 5.5 The relationships between variation in the external physical
(surface) attributes of food particles and tooth size
Food attribute
Particle volume
Particle size (when chemically
sealed)
Particle shape
Bolus formation
Factor favouring large teeth Factor favouring small teeth

Small
Small
Large
Large
Sheets/rod
Non-bolus-forming
Isodiametric
Bolus-forming
Fig. 5.11 The expected relationship between the surface attributes of food particles and
postcanine tooth size. Large teeth means a tooth row with equally large molars and a
tendency for the premolars to be molarized, while small teeth means a tooth row with
most of its surface area located in the centre of the row around the rst molar.
Canine size
173
should have large postcanine teeth. All else being equal, the smaller the
leaves that are consumed, the larger should be their teeth. Folivores that
eat shoots and branches, i.e. rod-like particles, should also have the largest
teeth. In contrast, fruit-eaters and carnivores should have relatively small
teeth. Frugivores tend not to destroy seeds (see p. 217), but leaf-eaters
commonly do. Seed-eaters should also have large teeth. Quite a lot of
evidence appears to support the above. Freeman (1988) shows that many
frugivorous bats have dentitions with wide arches and short tooth rows. This
is also true in many primates and Kay (1978) shows that folivorous primates
tend to have larger postcanines than frugivores. Carnivorous mammals tend
to have reduced postcanine dentitions centred on one pair of carnassial
teeth, which also appears to t predictions.
In summary, very simple theories about the evolutionary effect of the
surface attributes of foods produce predictions about tooth size in different
dietary groups much more readily than the allometric theories described
earlier in this chapter.
canine size
The canines of most mammals are used to fracture the integument of a target
so as to produce a deep and penetrating wound. Fracture, not fragmentation, is their major design characteristic. Early insectivorous mammals
had enlarged canines in both deciduous and permanent dentitions for this
purpose (Romer, 1966) and carnivores have them too. In some New World
monkeys, they are used to break into the hard outer layer of fruits (Chiropotes satanas and Pithecia pithecia: Kinzey & Norconk, 1990) but this is
unusual in plant feeders.
In anthropoid primates, several herbivores and some marine mammals,
the canines tend to be much larger in males than in females. Even when
there is no sexual difference in the permanent canines, as in gibbons (the
lesser apes: genus Hylobates), the disparity in size between non-projecting
deciduous canines and their permanent replacements betrays a role outside
of feeding, in social behaviour in fact (Plavcan et al., 1995). The major context for canine function in these mammals appears to be during conicts
between males (and sometimes between females) where they form potent
weapons. They are rarely used to kill though because there is usually strong
regulation of disputes, something that has been cogently explained by game
theory (Maynard Smith & Price, 1973; Maynard Smith, 1979). There appears to be active selection for canine size enlargement in the females of
some living primates (Hayes et al., 1996).
174
5 Tooth size
Whatever their role, canines require a large gape in which to operate, a

fact that relates the size of these anterior teeth back to the early part of this
chapter and that provides a good test of this. Herring (1972) was the rst
researcher to emphasize this in a study of Old World suids (pigs) and New
World tayussuids (peccaries). Later, she established that most mammals
generally manage gapes of 6070 of rotation of the mandible (Herring &
Herring, 1974), which will limit canine projection to some fraction of this
angle (otherwise they could not hit anything).
There are exceptions to this gape limit. The hippopotamus can produce
immense gapes (Herring, 1975) and many fossil mammals, such as sabretoothed carnivores (Emerson & Radinsky, 1980), must have been able to
do so as well. This is no easy matter to arrange though. One corollary in the
hippopotamus (Herring, 1975), as also in many molossid bats (Freeman,
1981a), is slack cheeks in order to allow them to stretch sufciently during
yawns. When the mouth is closed, the cheeks and lips are accid or wrinkled, possibly impairing their use in sealing the front of the mouth during
food processing. The jaw joint can also be set very low, e.g. in marmosets
(Vinyard et al., 2003), so as to reduce stretch in the masseter and medial
pterygoid muscles (Fig. 2.16).
Canine size and gape
The upper and lower canines must clear their tips well before the maximum
gape is reached. Figure 5.12 illustrates this using the skull of an Old World
monkey (the gelada, Theropithecus gelada), where two important angles
of jaw opening are distinguished, the angle at which the tips of upper and
lower canines just clear each other, termed c , and that at which the canines
are normally used, u .14 The latter angle is likely to be at or around the
maximum gape. There are no obvious adaptations to enlarged gapes in
primates like this. None appears to have wrinkled lips features that might
interfere with food manipulation and certainly preclude food wadging in
orang-utans and chimpanzees (Walker, 1979).
The gape angles required to just clear the tips of upper and lower canines
( c ) can be measured on skulls. Results on a large sample of primates and
a small sample of carnivores are shown in Fig. 5.13 (Lucas et al., 1986b).
Gapes are identical in male and female carnivores of the same species, but
range from around 10 in ursid (bear) species, through intermediate values
in canids (dogs) to 20 or more in felids (cats). Female primates have small
canines, but male primates have a very similar range to those of carnivores,
with the same maximum of 23 . Note that these gapes only clear the canine
Canine size
175
Fig. 5.12 (a) The living male gelada baboon has large upper and lower canine teeth.
(b) These clear each other (in this specimen) at a jaw joint rotation of approximately 17 , a
gape called c in the text. In this situation, the teeth are useless. (c) The gape, u , at which
the teeth might be used. This has been achieved in the diagram by extending the skull on
the neck by 30 and by depressing the lower jaw 35 . The combination of neck and jaw
opening muscles to this gape makes no difference to the actual rotation of the jaw joint (i.e.
c = 65 ), but the way that a mammal achieves it may play a substantial role in affecting
the relative sizes of the upper and lower canines.
tips. They do not offer any possibility of hitting anything: that possibility
only begins at this gape.
The major point to stress about these c values is their variation is limited,
so supporting a key contention of the earlier part of this chapter that the
anterior dentition is isometric to jaw length (Lucas, 1982b). However, what
explains the variability seen? One answer is the inuence of variation in the
height of the jaw joint, which will now be modelled.
Suppose that the jaw joint lies on the plane at which the teeth bite
together and the length of the jaw between the joint and the canine is l
(Fig. 5.14a). If the jaw is opened through an angle , then relative to its
position when the jaw is closed, its vertical displacement is
y = l sin
while the horizontal retraction is given by
x = l (1 cos ).
Carnivores all essentially have their jaw joints on this plane, which is optimal for the wide gapes that they require for food acquisition. However,
many herbivores have high jaw joints, which effectively restrict their gapes.
Unsurprisingly, low jaw joints tend to be associated with diets requiring
large gapes and with large canines and vice versa (Chapters 2 and 3). As
176
5 Tooth size
Fig. 5.13 The average gapes at which the tips of the upper and lower canines of living
anthropoid primate and carnivore species are just clear of each other. Sexual dimorphism
in the canines of the anthropoids is obvious: the canines of females in most of these species
do not project very far. The canines of the male primates and carnivores have a similar gape
distribution.
shown in Fig. 5.14b, when the jaw joint is raised above the plane of the bite,
then the ratio of y/x is increased, so reducing the potency of the canines as
weapons (Herring, 1972). Setting the jaw joint a distance h above the bite
plane, the vertical canine displacement becomes
y = l sin h(1 cos ),
(5.15)
while the horizontal retraction is given by

x = l (l cos h sin ).
(5.16)
Even without plotting anything out, it is obvious from Fig. 5.14 that the
higher the jaw joint (i.e. larger h/l ), the greater the y/x ratio at any value
Canine size
177
Fig. 5.14 (a) A mammal with a jaw length, l, and a jaw joint level with the bite plane of
the molar teeth rotates its jaw through 60 , moving its lower canine a vertical distance y
and a horizontal distance x. (b) The jaw joint is now raised by a distance h. This reduces
the y/x ratio at any given gape, resulting in the lower canine moving further backwards
compared to (a).
of . Despite large canines being apparently incompatible with high jaw

joints, h/l ratios of 0.5 have been measured in male anthropoid primates.
It seems likely that this would limit their canine size. Accordingly, Fig. 5.15
plots h/l ratios for male and female primates versus c values. There is
no relationship in females (Smith, 1984) because they usually have small
canines, but there is a signicant and quite strong negative relationship in
males. This relationship could just be an accident. To show that it is not,
the analysis must go further and predict the gape angle at which canines
are used ( u ).
I make the assumption that canine size adapts such that u corresponds
to a particular value of y/x, in common for all species in Fig. 5.15. As shown
in this gure, the trend line through the c data does not correspond with
slopes of equal y/x.15 However, correspondence can be obtained if all the c
values are multiplied by a constant m, such that
u = mc .
There is only one value of m that will do this. If m turned out to be
<1.0, then the gape theory simply cannot hold. On the other hand, if m
>>1.0, then this would provide a lot of support because obviously, u >>
c (see Fig. 5.12). As shown on Fig. 5.15, the t is m 2.8. Taking the
maximum c for any living primate or carnivore, then u = 2.8 23 =
64 , which further suggests that the theory is plausible.
178
5 Tooth size
Fig. 5.15 A plot of the gape angle at which the tips of the upper and lower permanent canines
of male primates just clear ( c ) versus the height of the jaw joint, h/l. The relationship is
negative (n = 51, r = 0.50; p < 0.001). The solid trend line was produced by multiplying
values of c by a constant m until it matched that of the dotted line, which is one of a
family of curves of constant y/x generated from Eqns 5.15 and 5.16. The value of m turns
out to be 2.8.
Are there any relevant physiological data on maximum gapes in

mammals? Carlson (1977) states that female rhesus macaque monkeys
(Macaca mulatta) can only open their jaws to about 30 . It is possible,
but unlikely, that males can open their jaws further. c = 30 for male
rhesus macaques should be compared to a u of 42 predicted from the
theory (i.e. 2.8 c ). This t is improved when an extra feature, not so far
included, is considered. The head of the condylar process of the mandible
of some primates, including rhesus macaques and humans, translates anteriorly during sufciently wide jaw openings. This does not affect measured
values of c , but it does increase y/x ratios. The average maximum values for
anterior translation in humans and rhesus macaques (Carlson, 1977) seem
about 1015% of the jaw length, l. Though the correction is not entirely
satisfactory, all h/l ratios in Fig. 5.15 could be assumed overestimated by
Conclusions
179
15%. This reduces m to 2.5 and changes u for male macaques from 42
to 37.5 , which again seems reasonable. Going further, could the theory
explain the size of the canines in sabre-toothed carnivores? Drawings suggest that the large upper canine cleared the small lower one at about 50 .
Its jaw joint was xed and thus c = 50 2.8 = 140 , which also seems
satisfactory.
There are a considerable number of extinct primate species with high jaw
joints and reduced canines, including the fossil dryopithecine ape Sivapithecus indicus (for which h/l 0.5, unpublished data), Theropithecus oswaldi
(Jablonski, 1993), particularly later members of the species and Oreopithecus bambolii, for which a reconstruction in Moyà Solà & Kohler (1997)
gives h/l 0.60, a record in primates. This primate probably had a very
small mouthslit. It will not have escaped readers attention that humans
do not have projecting canines in either sex either. Our species has a relatively high jaw joint (h/l 0.33), although the data shown in Fig. 5.15
suggest that this would not be sufcient to explain why our canines are
very small. Humans also have very small mouthslits and consequent small
maximum gapes (23 : Baragar & Osborn, 1984). This is discussed further in
Chapter 7.
I have only analysed one possible inuence on canine size. While it
seems to predict the maximal (optimal) size of these teeth, they are of
course free to vary in size below this maximum. Nothing has been said
about why there is often a big disparity between the sizes of the upper
and lower canines. In some mammals, such as mouse deer (chevrotains),
muntjacs, Chinese water deer, sea cows and walruses, there is only an upper
canine. It is possible that some of these mammals use their canines with
their mouths closed. Walruses do this, for example, when they use their
canines to move across ice. However, why in most sabre-toothed carnivores
is the upper much longer than the lower? The answer probably depends on
whether a canine bite is made more by the neck muscles than by those of
the jaw.
conclusions
Mammals are what they eat and there is no point in trying to understand
tooth size without having rst characterized the properties of the diet. Jaw
and anterior tooth size are linked to the size of ingested food particles, but
the sizes of the postcanine teeth are not. They depend on the deformation
behaviour of food particles and their external physical characteristics. Kay
(1978) stated that a coherent theory which accounts for molar size must
180
5 Tooth size
explain why tooth dimensions are negatively allometric with respect to

metabolic body size (mass3/4 ) and why all [tooth] dimensions do not
follow the same regressions. That is a pretty tall order, but I suggest that
the theories contained here are indeed capable for tackling such a problem.
There is a vast literature on tooth size, but little quantication of the diet.
That has to change if headway is to be made.
Tooth wear
overview
It will be recalled from the last chapter just how dependent the rate of food
breakdown is on tooth size. However, the exposed surfaces of the teeth are
subject to wear, either from food, grit or opposing tooth surfaces. Wear
threatens to destroy tooth shape, decreasing the rate of food breakdown
and so jeopardizing a mammals health. Insofar as the food input is responsible for this wear, the rate must depend on the characteristics of the food
surface, i.e. on the external physical attributes of foods, because wear involves small-scale events involving the interaction of surfaces. Maintaining
the argument from Chapter 1, I expect the evolved response to this threat
to lie in adaptation of tooth size. This chapter follows through the logic of
this argument and also tries to identify the major causes of tooth wear.
introduction
Wear is the loss of volume of an object and results from a number of
processes rather than just being one process in itself. It may often involve
fracture, but the chemical dissolution of stony objects like teeth also comes
under the same general heading. Teeth wear mostly on their working surfaces, but any other tooth surface exposed to the oral cavity can wear too.
Movement of food by the tongue and cheeks against the teeth is sufcient
to cause a small amount of wear (Fox et al., 1996; Ungar & Teaford, 1996),
enough anyway to result in the gradual loss of perikymata (a pattern of
ridges and troughs present on a newly erupted enamel surface). The most
important of these other surfaces though are those where adjacent teeth in
the dental arch make contact.
There have been three distinct tools for investigating wear: the naked eye,
the light microscope and the scanning electron microscope. Gross wear, as
observable with the naked eye, has lacked much quantication, but this
181
182
6 Tooth wear
Table 6.1 The pH of fruit esh eaten/avoided by long-tailed
macaques ( Macaca fascicularis) in Bukit Timah, Singapore
Species
pHa
Status in the diet
Calophyllum pulcherrimum (Guttiferae)

Garcinia forbesii (Guttiferae)
G. parvifolia (Guttiferae)
Nephelium lappaceum (Sapindaceae)
Tinomiscium petiolare (Menispermaceae)
Fibrauria tinctoria (Menispermaceae)
Mezzettia parvifolia (Annonaceae)
Tetrastigma lawsoni (Vitaceae)
Pellacalyx saccardianus (Rhizophoraceae)
Gnetum microcarpum (Gnetaceae)
3.03.5
2.53.0
3.03.5
4.6b
5.05.5
5.56.0
4.04.5
3.0
5.0
5.0
Eaten
Eaten
Eaten
Eaten
Eaten sporadically
Eaten
Eaten sporadically
Eaten
Avoided
Eaten
a
b
Measured with half-unit pH paper.

Data from Poon (1974).
is now changing rapidly using geographical techniques designed originally

for contour mapping (Zuccotti et al., 1998; Jernvall & Selanne, 1999; Ungar
& Williamson, 2000). However, it is the scanning electron microscope that
has yielded most insight. Initiated 25 years ago (Walker et al., 1978), this
dental microwear analysis can resolve actual wear features (Teaford, 1994).
Microwear studies have proved that mechanical processes are almost certainly the predominant cause of wear, leading to the pitting and scratching
of tooth surfaces.1 The other major possibility is erosion, the term given to
the chemical dissolution of tooth tissue by acids, particularly from fruits
(Table 6.1). This tends to produce a matte surface, representing the uneven
loss of mineral from enamel rods, depending on their orientation. It is
probable that saliva buffers acid so as to prevent much tissue loss, but I do
not know.
As a topic in mechanics in general, wear remains very incompletely
understood. Although understanding grew rapidly in the early to mid parts
of the twentieth century due to interest in friction (Bowden & Tabor, 1950),
there remains much that is not understood. Accordingly, it is difcult to
predict wear rates. All this becomes obvious when you start to read around
the subject. Books or articles on wear tend to be long, to subdivide the
subject into categories that are not clearly connected, and to be specic
rather than general. This is all very sad for dental studies because the general
state of the dentition present in an individual mammal of any species,
picked out at random, is worn. Unless the factors that control its wear
Wear terminology
183
can be understood, the mammalian dentition cannot really be understood.

Mammals have tossed away the reptilian option of many tooth generations
and gambled on just a couple, so the longevity of teeth is vital.
I will limit my observations here to some that could inuence thinking about the causes of wear in biological terms. I cannot do any better
than others in fundamental terms, but this is my brief shot at a difcult
subject.
wear terminology
Dental researchers often use the term attrition to describe the wear of
the working surfaces against each other, while wear caused by intervening food is called abrasion. With either the naked eye or a magnifying
glass, abraded molar surfaces have an irregular matte surface, while attrition seems to give rise to at glossy surfaces. The latter shiny surfaces are
called facets and can be seen clearly either with the naked eye or at low
magnication under oblique lighting. Facets tend to develop in similar positions on the teeth of a given species because malocclusions are rare and
so opposing teeth always contact in the same areas (Mills, 1955). Butler
(1952) initiated the study of facets and started a system of nomenclature
for them in the molar region. This work has since been extended to a wide
range of mammals and their evolution in early mammals has been plotted
(Crompton, 1971). The shape of wear facets provides a strong indication of
whether features of the working surfaces of teeth are operating as cusps or as
blades.
Under the higher magnication available with a scanning electron microscope, the difference between abraded surfaces and attrition facets begins to
blur. Many facets that look featureless at lower power are actually scratched
and pitted when magnied sufciently (Teaford & Walker, 1983). Clearly,
food must be entering the space between upper and lower tooth surfaces.
This is only right and proper of course because this is what teeth are there
for, but many questions arise from it. Why do facets and abraded surfaces
look different? Are facets caused more by the contact of opposing teeth
than by food, or are some of the assumptions wrong and food capable of
producing glassy surfaces? Indeed, what is the predominant cause of tooth
wear?
Since I want to get at causation, I will redene the terminology slightly
here. What dentists call attrition, I will call toothtooth wear, and what
they label abrasion, I will refer to as foodtooth wear. This allows the
term abrasion to be specialized here to mean mechanical wear involving
184
6 Tooth wear
fracture. It will not be limited in this chapter to contacts in which food

intrudes. Additionally, a pit will be dened as a roughly isodiametric feature
on a worn tooth, while a scratch is elongate (Teaford, 1994).
mechanical wear
Any mechanical process whereby small fragments are lost from a large
block is likely to be called abrasion. This type of wear was long thought
determined by the relative hardness of contacting surfaces. Hardness itself
is now known not to be the cause, but the root of much wear is still
believed to lie in indentation, which is of course how hardness is measured
(Appendix A). The ability of one object to indent another is determined
by their relative hardness, but loss of volume depends on other properties
too, such as toughness, Youngs modulus and surface features that inuence
friction (Kendall, 2001). This is entirely logical: if wear involves loss of tissue
by fracture, then toughness must play a part in this. A general theory of
mechanical wear that incorporates toughness is surely necessary to explain
how, for example, mammalian skin and hair can wear enamel (Aitchison,
1946; Rose et al., 1981), but it has not been discovered yet.
Toothtooth wear
Teeth are very often the hardest substances in the mouth and are, thus, potentially serious wear agents unless their motion can be controlled precisely.
This potential has been proven by examination of the wear of the molars of
the guinea pig (Cavia porcellus). This mammal erupts its molars in utero and
makes empty jaw movements with sufcient force before birth to produce
toothtooth wear. Teaford & Walker (1983) show that worn molars are basically featureless at surfaces (presumably well washed by amniotic uid),
conrming that toothtooth wear produces facets. Teaford & Walker (1983)
and Walker (1984) have also run simple in vitro experiments on toothtooth
wear, showing that the wear of enamel depends on the orientation of its
component rods to the tooth surface: rods normal to the contacting surface
wear faster than those that are parallel to it.
Humans also commonly wear their teeth without any food being present,
but they do this after birth, not before it. Such events are called bruxism.
They usually take place at night, in a small proportion of the population,
and are the presumptive result of psychological stress (Lavigne et al., 2003).
In a number of papers, Every (e.g. Every, 1970) hypothesized that all animals
need to brux in order to preserve the sharpness of their teeth, reasoning that
Mechanical wear
185
if foods blunt teeth, toothtooth wear is there to resharpen them.2 He called

bruxism in mammals thegosis, a term implying that this is normal rather
than pathological behaviour. His theory still attracts worldwide interest and
has been critically reviewed by Murray & Sanson (1998).
I do not wish to go into the wilder aspects of Everys thinking about
the direction of jaw movement in thegosis (he thought that these were
opposite to those in mastication there is no evidence for this), or his
view that the dentition is basically designed as non-alimentary weaponry
(a theory that could only emerge from dentistry). He was correct about the
wear of some of the anterior teeth, although all that did there was to name a
behaviour that had already been observed. The wear of the canines of male
Old World monkeys (shown in a fantastic series of photographs in Every,
1970) is almost certainly due to specic sharpening activity. Zingeser (1969)
described the facets that develop from this, but there are other undoubted
instances of thegosis, such as the process of wearing of the lower incisors of
rodents, which can only result from a protrusion of the jaw beyond a point
where it could conceivably be used in ingestion (Osborn, 1969). However,
there is no evidence though that sharpening activities are more widespread
than this in mammals and they do not include the postcanines. Why then
did it have a vogue? The underlying cause of its eeting popularity seems
to be that everyone believed that teeth needed to be sharp. The reasoning
would be something like this: if foods break at xed stresses, then sharp
teeth will possess the smallest contact areas and so generate these stresses at
the lowest forces. Accordingly, efcient teeth should be as sharp as possible.
Yet, tooth sharpness cannot be maintained because contacts with food will
blunt them. A search for separate sharpening activities thus seems logical.
This reasoning is awed, as Chapter 4 should make quite clear, but
there has never been a clear refutation. Instead, researchers have looked
for alternatives. Osborn & Lumsden (1978) published the most famous of
these, proposing that teeth are maintained sufciently sharp by dietary use.
Foodtooth wear
The biggest threat is thought to come from quartz, which is the extraneous
grit of microwear studies. There is very strong evidence for its inuence
as a dental wear agent not just in terrestrial mammals, but also in arboreal
mammals in tropical forests (Ungar et al., 1995). For herbivores, a close
second is the opaline amorphous silica deposited in many plant tissues.
This has been found on the teeth of fossils (e.g. Gigantopithecus blacki, a
Pleistocene ape species from China: Ciochon et al., 1990) and prehistoric
186
6 Tooth wear
humans (Fox et al., 1996) and passes through the gut of living mammals
into the faeces, e.g. of sheep (Baker et al., 1959) and hyraxes (Walker et al.,
1978). In any plant, the shape and size of these phytoliths varies with the
cells that contain them (Ball et al., 1993). Some plants also contain crystals
of calcium salt, such as those of calcium carbonate and calcium oxalate
(Metcalfe & Chalk, 1950). The latter are more common and there is strong
circumstantial evidence for them as a wear agent in prehistoric human
populations (Danielson & Reinhard, 1998). A factor whose importance
is completely unknown is the practice of geophagy the consumption
of soils that is widespread in mammals (e.g. in primates: Oates, 1978;
Krishnamani & Mahaney, 2000; Dominy, 2001).
Additionally, woody tissues themselves may wear the teeth, certainly
doing so in rodents and being the cause of the need for ever-growing
incisors. Some fruit esh contains isolated stone cells, which produce
a gritty feel (Lucas, 1991). For carnivores, there is bone (Van Valkenburgh
et al., 1990) and for insectivores, the cuticle (Schoeld et al., 2002). Freeman
(1979, 1981a) rst pointed out clear differences between types of bats that
eat insects with unsclerotized (e.g. moth) or sclerotized (e.g. beetle) cuticles.
The latter produce many pits on teeth (Strait, 1993c), more than on primate
leaf- or fruit-eaters. Similar wear is associated with fracture of bone in the
mouth (Van Valkenburgh et al., 1990).
indentation and wear
At the heart of every pit and scratch is probably an indentation process.
This section discusses how indentation leads to wear and why scratches are
far more probable wear features than pits. When two particles of dissimilar
materials are pressed together, then the harder one will indent the softer
one. Permanent indentation (due to plastic deformation) starts when the
stress in the softer solid exceeds 1.1 times its yield strength, i.e. 1.1 y , and
full plasticity takes place at 2.8 y (Atkins, 1982). However, if the differences
in hardness between the two materials are not very great, then both will
be affected by the contact. To prove this, I need to distinguish between
the yield stresses of the harder and softer solids. For convenience, the yield
stress of the softer solid will be termed Ys and that of the harder one, Yh .
If full plasticity in the softer solid is not achieved before the harder solid
reaches 1.1Yh , then the harder solid will also yield. The boundary between
a one-sided effect and mutual indentation is, therefore,
2.8Ys = 1.1Yh .
(6.1)
Indentation and wear
187
Provided that the harder solid has a yield stress (2.8/1.1) = 2.5 times that of
the softer solid, then it will not be affected by the indentation. However,
it is not very convenient to deal only in yield stresses because these can
be difcult to measure. Hardness, the mean pressure that produces an
indentation of unit area (Appendix A), is much simpler. The types of
materials that could indent tooth tissues are all going to be true solids (i.e.
with negligible airspace), so I will assume for simplicity that the hardness
of potential wearing agents is always going to be a common multiple of the
yield stress. I will take this multiple to be 2.8 so as to comply with Atkins
(1982).3
Enamel microwear
I will assume initially that microwear features are plastic indentations. The
force, F, required to make a pit comes directly from the denition of
hardness and is
F = Hr 2
(6.2)
where H is the hardness of enamel and r the pit radius. Taking the hardness
of enamel as H = 3700 MPa (just an average value it is highly variable
as shown in Appendix B) and a typical pit radius as r = 2.5 m (Teaford,
1988), then F = 0.073 N, which must be small compared to masticatory
forces and very small compared to the bite forces known for any mammal
(Chapter 5). Thus, many such pits could form simultaneously were suitable
particles to be present in the food intake.
However, the pit might also be formed from a contact without any real
indentation, the rough pit being the product of something like a cone crack
(Lawn, 1993). An estimate for this force is

ER 0.5
2
(6.3)
F = 6r
a
where r is the radius of contact, a is the length of the crack and E and R both
refer to the properties of the enamel (Sharp et al., 1993). Inserting r = 2.5 m
and an arbitrary, but constant, value for a = 5 m, and taking typical
values for enamel from Appendix B, E = 90 GPa and R = 13 J m2
(remembering that enamel nearly always fractures along rods), then
F = 0.033 N, which is even smaller than for plastic indentation.
The main effect of a sliding contact against the abrading particle, so as
to form a scratch, is to reduce the forces involved. These forces depend,
188
6 Tooth wear
Table 6.2 Frictional characteristics of tooth tissues
Friction
Coefcient of friction ()
Reference
Enamelenamel
Diamondenamel
Diamonddentine
0.100.42
0.14 (0.02)
0.31 (0.05)
Douglas et al. (1985)

Habelitz et al. (2001)
Habelitz et al. (2001)
however, very strongly on the coefcient of friction, , between the surfaces.

Table 6.2 gives some values from the literature.
Sharp et al. (1993) give the force for plastic indentation as
F =
w2 H
.
(1 + 92 )0.5
(6.4)
For a scratch of width 2w = 2 m (a typical value: Teaford, 1988), taking

= 0.14 from Table 6.1, then F 0.011 N, which is even smaller still.
Alternatively, if there is fracture without indentation, then the appropriate relation is

ER 0.5
w2

.
(6.5)
F =
13
a
1.12 61 + 16
This relation assumes Poissons ratio to be 0.33 (Sharp et al., 1993), which
is probably a little high for enamel and the types of materials that can wear
it (Lawn, 1993). However, for the same 2 m wide scratch, creating a 5 m
crack, then F 0.001 N, which is the lowest of the four force estimates.
Appendix B gives hardness values for dental tissues and various possible
causative agents of pits. Few materials look capable of pitting enamel (many
more could scratch it), but there are a lot of candidates lined up to damage
dentine.
The biggest threat overall is from quartz with about twice the hardness
of enamel, while opaline amorphous silica is about 1.6 times as hard.
Both quartz and opaline silica are in a range whereby they could indent
enamel, but they have <2.5 times its hardness and so would themselves
permanently deformed. This mutual indentation has been subjected to
considerable research by Atkins (1982) and results in mean pressures that
are considerably below those when only one particle deforms. However,
the process also depends on the geometry of loading. Most tooth-abrasive
particle contacts are likely to be of a ball-on-at type whereby a piece
of quartz dust or a roughly spherical phytolith presses against a larger,
189
relatively at, tooth surface. Assuming this geometry, then enamel that is
being indented by a roughly spherical phytolith of hardness 1.6 times that
of enamel, becomes fully plastic at just 2.2Ys (Atkins & Felbeck, 1974).
Equation 6.1 now becomes
(2.2/2.8)Ys = 1.1Yh ,
reducing the pit-forming force to 0.057 N. Nevertheless, I can sum all
this up and, while accepting that these are very rough calculations, suggest that scratches will form much more readily than pits on enamel. The
difference in predicted forces is so great that this conclusion is not really
affected by substituting other values for the coefcient of friction (e.g. =
0.41; Table 6.2). The predicted scratch predominance on worn enamel is
supported by the literature.4
Teaford (1994) reported data on the enamel microwear of 20 species
of primate, with widely varying diets, nding that scratches are always
more frequent than pits. However, the formation of scratches requires a
sliding contact, so the only occasions on which pits would be predicted to
predominate would be when there is just compression. There is support for
this as well. Gordon (1984) showed in the chimpanzee that pits are more
frequent on the enamel of third molars than the others probably because
they are closer to the jaw joint and so to the centre of rotation. There are
more pits on the molars of frugivorous primates than on those of folivores,
suggesting that there is less tooth sliding during mastication in the former.
The extent of wear areas on the molars of Old World monkeys, a group with
great dietary diversity, seems to support this. Crushing areas are relatively
larger in frugivorous primates, while shearing facets are correspondingly
enlarged in folivorous primates (Kay, 1978), and presumably covered in
scratches.
Dentine microwear
Dentinal surfaces are very rarely scrutinized in microwear studies because
the surface is so intrinsically heterogeneous that quantication of microwear
features is made very difcult (P. S. Ungar, pers. comm.). It also collects
debris, which is hard to remove (M. F. Teaford, pers. comm.). Nevertheless, the same calculations can be made as for enamel. Ignoring mutual
indentation for the moment and, taking typical values from Appendix B
for dentine properties (H = 650 MPa, E = 20 GPa, R = 270 J m2 ) with
= 0.31 from Table 6.2, provides a fascinating comparison with enamel
(Table 6.3).
190
6 Tooth wear
Table 6.3 Approximate forces (in millinewtons) for the creation of
microwear features, either by plastic indentation or surface fracture
Pit (5 m diameter)
Enamel
Dentine
Scratch (2 m width)
Plastic indentation
(Eqn 6.2)
Fracture
(Eqn 6.3)
Plastic indentation
(Eqn 6.4)
Fracture
(Eqn 6.5)
73
13
33
41
11
1.5
1
1
It seems that when their surfaces are compressed against an abrasive,

enamel is much more likely to suffer microfractures than dentine. This
seems again to be borne out by the evidence. Most published micrographs
of enamel pits seem to have ragged edges. Dentine is easily indented, but
this may often be just plastic modication. However, wear could result from
other fracture possibilities, which depend very much on the sharpness of
the abrasive particle (Lawn, 1993).
Similar to enamel, scratches on dentine are much more likely than pits.
However, in a major surprise, Table 6.3 suggests that although scratches
form at lower forces than those in enamel, fracture damage under sliding
forces is probably equally likely in both tissues. No data in the literature
appears to relate directly to this, but this result can be interpreted as showing how the material properties of enamel and dentine support the stable
construction of teeth after the dentine is exposed.
Relevance to dental wear in herbivores
The dentition of most herbivores is so designed as to expose dentine at the
earliest opportunity. Guinea pigs do this in the womb, while the molars of
mice erupt with enamel-free areas already present on their cusp tips. Other
herbivores, lacking both these characteristics, have very thin enamel at the
cusp tip (e.g. in the hippopotamus: Luke & Lucas, 1983), quite unlike in
humans, for example, where cuspal enamel is thickest. This also ensures
that dentine will be open to wear processes as soon as possible. The need
for dentine (and, in some species, cement) on the crown surface is that it
allows the edges of enamel ridges, which wear more slowly, to stand proud
of the rest of the crown and act as short blades on thin vegetation (Fig. 6.1).
Dentine is always thought to wear away more quickly than the enamel
because it is softer. However, why do enamel ridges not get progressively
191
Fig. 6.1 The form of an upper molar tooth in a bovid artiodactyl, such as a cow. (a) The
working surface sits on top of a tall crown. Each of the four elongate cusps can be identied
by light grey areas representing half-moon shaped exposures of (grey) dentine surrounded
by enamel ridges. The whole crown is originally covered with cement. The black areas
represent holes. (b) This shows a schematic section through part of the tooth showing how
the dentine (labelled d), and cement (c), wear down so that enamel ridges (e), stand proud
of these tissues. Note that the surfaces of the dentine and cement lie just below that of the
enamel. (c) A schematic diagram to show the path of movement of a lower molar across an
upper, which has a strong mediolateral (or transverse) component.
192
6 Tooth wear
higher? Surely this must follow if dentine always wears away more rapidly.
One explanation is that the enamel takes the brunt of the load, so relieving
dentine depressed beyond a certain distance. However, the ridges do not
seem high enough for this to be plausible. How can this be explained?
When rst exposed, dentine is level with enamel and subject to both
compressive and sliding forces. The data on pitting suggest that dentine will
wear faster than enamel in these circumstances. However, once depressed
below the enamel ridges, dentine is largely immune from this compression
and more subject to sliding forces. The results in Table 6.3 imply that
dentine could then wear no faster than enamel. There are no data on this,
although experiments and observations reported by Greaves (1973), Costa
& Greaves (1983) and Walker (1984) all agree that dentine on the edge of
an enamel exposure can get heavily scratched.
All this ignores mutual indentation, which is more important for dentine
than enamel. Insect cuticles, bone and the woody tissues of plants have similar hardness to dentine (Appendix B) and could wear it substantially, but
leave enamel unmarked. This is of some considerable interest in herbivores
because, under some circumstances, the control of the enamel/dentine wear
ratio goes wrong. Some diets fed to domesticated horses leave very high
enamel ridges, obviously wearing the dentine but not the enamel. A veterinary surgeon has to come in and le the ridges down, a procedure often
called oating the teeth.
The general effects of friction on all this can be noted. For usual values
of , where 1.0, the effect is greatest on fractures during sliding
(Eqn 6.5) where the force reduces. This could have important implications for browsers that ingest leaves with substantial quantities of tannins.
Whether eaten because they speed up digestion (Aerts et al., 1999) or because browsers cannot avoid tannins altogether, the interaction of tannins and salivary proteins, particularly those proteins that adsorb onto
the enamel surface (Jensen et al., 1992), is liable to increase the rate of tooth
wear. True grazers encounter few such compounds (Owen-Smith et al.,
1993).
Unfortunately, too little is known about the properties of cement on the
tooth crown in herbivores to add this into the analysis. It could act as an
important test. While there is no reason to suppose that it would not behave
somewhat like dentine in this regard, none of the above is anything like the
last word on the mechanics of wear. Equation 6.5 has an exact look, with
constants expressed as fractions, because it derives ultimately from a very
complex, apparently unsurpassed, piece of applied mathematics (Hamilton
& Goodman, 1966). Yet, the problem is so complex that experiments are
Inuence of dental microstructure
193
denitely needed. Some of these have been reported on homogeneous

ceramics. Lee et al. (2001) show on ceramic coatings that the effect of
sliding contacts on indentation fractures are real they can reduce the
load, for example, in forming cone cracks but the differences may not
be anything like as great as my simplistic view suggests. Interestingly, Lee
et al. make reference to in vitro wear studies on human teeth, effectively
questioning whether it is necessary to model actual jaw movement patterns
with quite the rigour that has marked some research on the subject (e.g.
DeLong & Douglas, 1983).
The size of plant abrasives
The silica particles in plants are very small, generally being <50 m in size
(Ball et al., 1993). The marks that they produce on teeth are much smaller
because only a tiny part of their surface contacts the dental tissues. A lot
of these particles pass through into the faeces undamaged. It is entirely possible then that their size protects them against being fractured because they
lie below the brittleductile boundary referred to Chapter 4. If material
property values for quartz from Appendix B (E = 70 GPa, R = 2 J m2
and y 2500 MPa) are inserted into Eqn 4.9, with the constant c = 2300
for indentation (Atkins & Mai, 1985), then the brittleductile boundary
lies at 50 m. Rough as this calculation is, it suggests that phytolith size
might be a deliberate anti-herbivore strategy by plants because it obviously
enhances the wear potential of these particles if they can avoid being fractured by rst contact with the dentition. Dust particles are generally an
order of magnitude smaller than plant silica (Lawn, 1993) and would be
expected to cause smaller wear features.5
the inuence of the microstructure of dental
tissues on wear
It has long been known that enamel structure is very variable in mammals
(Boyde, 1964; Osborn, 1974) and there is increasing interest in explaining how adaptive this variation is (Janis & Fortelius, 1988; Rensberger,
2000). The decussation of enamel rods in many herbivorous mammals
(e.g. in rhinoceroses: Rensberger & von Koenigswald, 1980; Fortelius, 1985;
Boyde & Fortelius, 1986) produces irregular surfaces that help to grip food,
not just by increased friction, but also via microindentation. In voles, the
maintenance of the height of the enamel crests above the exposed dentine
looks to depend on an abrupt change in decussation on a minute scale
194
6 Tooth wear
(von Koenigswald, 1982). In another rodent, the beaver (Castor ber), the
worn lower incisors develop a step that follows the hardness contours of
the dentine very closely (Osborn, 1969). However, some microstructural
features appear to make little sense (e.g. the pattern of rod decussation of canine enamel in baboons: Walker, 1984) and a coherent argument is needed
that relates these features to diet and that appears to me to be lacking at
the moment.
wear on other tooth surfaces
Abrasive wear appears typical of the working surfaces of teeth, but is not
limited to them. Collagen bres, called transseptal bres, hold adjacent
teeth in a dental arch together. Should one tooth, supporting most of the
load, be depressed into its socket relative to its neighbour, wear is likely
to result at their point of contact. This approximal wear is not caused by
transseptal bres, which are weak and need to be remodelled by cellular
activity so as to adjust their length. The attractive force between neighbours
is mostly a result of a bite force that is directed mesiodistally. The classic
investigation of the angulation of the bite force was by Osborn (1961) on
human subjects. With the upper and lower teeth apart, Osborn found that
it was possible to push thin metal strips between the approximal surfaces
of teeth without effort. During a bite, these could only be removed by
applying a tensile force (which Osborn measured). By going around the
dental arch and performing the same test, Osborn deduced that the teeth
in the arch must be forced mesially by the bite. Approximal wear is then
explained by the bite being supported by particular teeth that move into
their sockets, so rubbing against their neighbours.6
It has been calculated that Plio-Pleistocene hominins lost only 6% of their
molar tooth surface at maximum to approximal wear (Wood & Abbott,
1983), but much heavier losses are found in some modern and prehistoric
human populations (Murphy, 1959). Van Reenen (1982) showed that 25% of
the length of a mandibular rst molar could be lost to approximal wear in the
lifetime of a San Bushman individual in southern Africa. His calculations
seem to show that a mandibular third molar would erupt approximately in
the same position in the jaw that the second molar had occupied when it
erupted several years earlier. There is no logical reason why approximal wear
should mirror that on the working surface of the teeth (meaning that the
former could exist without the latter), but van Reenen showed an effective
association between the two, with a total lifetime loss of about 60% of
second molar crown volume.
Wear and dental efciency
195
Curiously, as approximal wear surfaces develop, they tend to form reciprocal curvatures on neighbouring teeth, rather than being at. The concave facet appears always to be found in the tooth that erupts rst. It is
probable that the tooth erupting second brushes past the rst as it erupts,
so spreads its wear over a larger surface area than the former, eventually
invaginating it.
Loading of just one part of the postcanine dentition with food is particularly likely in mammals that consume small food particles in small
mouthfuls. Also, teeth within one functional group may also not come into
contact with food simultaneously during a chew. Osborn (1982) showed
that the helicoidal plane of occlusion in several anthropoid primates, particularly some hominins, brings the molar teeth into food contact in a distal
to mesial sequence (i.e. third molars, then second molars, then the rst).
Counter-intuitive though this may seem, the suggestion works perfectly
on paper and could be explained by a dietary adaptation requiring close
to the limit of bite force, a factor that C. R. Peters (e.g. Peters, 1981) has
consistently suggested to explain the evolution of hominins.7
Other surfaces of the teeth wear, but much less so. The actions of the
tongue and cheeks can push food particles against the lingual and buccal
surfaces of foods with enough force to wear them (Ungar & Teaford, 1996).
Buccal wear of the molars may be accentuated in mammals with cheek
pouches and labial wear of the incisors by the wadging of food particles by
chimpanzees and orang-utans (Walker, 1979).
wear and dental efciency
In its early stages, wear imperils tooth shape. I would assume, a priori, that
this would be detrimental to efciency, but there is no need to speculate
because there is evidence on this matter. Gipps & Sanson (1984) showed
unequivocally in the ringtail possum (Pseudocheirus peregrinus) that dental wear decreased chewing efciency and that this impaired digestion.
Lanyon & Sanson (1986) conrmed this in the marsupial koala (Phascolarctos cinereus) and Logan & Sanson (2002) have now gone further and
associated dental wear and the reduction in digestive efciency that it
produces with impaired chance of a male nding a mate. This link to
reproductive success represents one of the most important ndings in the
dental literature because it provides a direct demonstration of selective
advantage.
At later stages of wear, the functioning of the entire dentition is threatened. There is a presumption among many dental researchers that its loss
196
6 Tooth wear
spells the end of a mammals life. This point should probably not be pressed
too hard all mammalian physiologists are likely to feel that the particular
body system that they study is the one crucial for an organisms survival. All
systems cardiovascular, neural, urinary, respiratory or whatever show
clear signs of ageing, so which of these is generally responsible for limiting
a mammals lifespan? Surely, each system is optimized to last just a lifetime
and no more? The evidence that dental wear can limit lifespan in the wild
though is really quite strong, involving many anecdotal reports from shrews
to elephants.
The teeth of wild shrews are often worn down to the gums at an age
when captives appear perfectly healthy. These wild shrews starve to death.
Elephants bring their cheek teeth into action in a paired upperlower sequence, from anterior to posterior. Once the last such tooth pair comes into
occlusion, the remaining lifespan of the animal depends on these teeth or,
again, they starve. Hypsodont mammals are just as vulnerable to a mismatch
of wear rate. An impressive time series of zebras that used to be exhibited
in the Natural History Museum in London showed that the oldest zebras
generally have teeth that are very heavily worn.
Another line of evidence is provided by observations on carnivores. These
mammals use one pair of teeth, called the carnassials, to fracture soft tissues. Carnivores that break bones, like dogs (Fig. 6.2) or hyenas, use molars
that are posterior, or premolars anterior, to the carnassials respectively, the
difference in development of these teeth being thought to be related to
gape. Van Valkenburgh (1996) observed large carnivores in the wild and
showed that most mammals did use the teeth that anatomists had always
predicted that they would. Observations by Berkovitz & Poole (1977) on
ferrets, a feral mammal closely related to polecats, suggest that the wrong
choice could be fatal. Caged animals fed on rat hard pellets blunted their
carnassial teeth. When the animals were subsequently fed mice, which
are part of their normal diet, they were unable to bite through them
easily.
adapting to wear
The mammalian dentition may be built to withstand wear, but it is nevertheless under undeniable threat. The principal physiological adaptations
against it are extremely ne interdental sensitivity to detect abrasive particles (as discussed in Chapter 3) and the salivatory response found by
Prinz (in press). The former may trigger the latter and, although the copious salivation of ruminants is supposed basically to provide buffered
Adapting to wear
197
Fig. 6.2 The jaws of a canid (dog family) in three staged poses: (a) Gripping a prey item
with its incisors and canines; (b) fracturing soft tissues with its carnassials and (c) breaking a
bone with the molars lying behind the carnassials. Note the different gapes involved in each
of these activities. If this mammal attempted to fracture bone with its carnassials, it would
blunt the blades.
uid for the rumen, another potential benet may be to wash abrasives off
the teeth.
The commonest evolutionary response to a wear threat is hypsodonty,
particularly of the postcanines (Stirton, 1947; van Valen, 1960; Janis &
Fortelius, 1988). Many herbivorous mammals show this to a greater or
lesser degree (sometimes only in the last molar, on which hypsodont indices
are often based: Mendoza et al., 2002) and it is usually associated with
grass consumption although there can be errors in blithely assuming this
(Macfadden et al., 1999). The family of horses, the Equidae, is a classic
example of a lineage that tracks the development of hypsodonty (Simpson,
1953). The family rst appeared in the fossil record of the Eocene about
198
6 Tooth wear
55 million years ago. Eohippus, the rst genus, had short-crowned teeth.
During the Miocene, the family diversied into a large number of forms,
all with high postcanine crowns. This development has consistently been
associated with the development of savanna and consumption of plants,
notably grasses, which contain large quantities of opaline silica. Logical
though all this may be, the key experiments look still to be done.
Increase in the crown heights of incisors and canines may also represent
adaptations to wear. The spatulate incisors of primates wear extensively in
some species, including hominins where the cause has been the subject of
considerable debate (Hylander, 1977; van Reenen, 1982; Ungar & Grine,
1991; Ungar et al., 1997). The canines of mammals can be heavily worn
(e.g. in great apes: St Hoyme & Koritzer, 1971; Dean et al., 1992), but more
usually they are subject to gross fracture, as indeed any tooth can be (e.g.
Frisch, 1963; Van Valkenburgh, 1988; Van Valkenburgh & Hertel, 1993).
The other major wear adaptation that is generally recognized is the
thickening of the enamel. The enamel of hominin teeth is relatively thick
compared to that of the herbivores or carnivores sketched above. The thickest part is over the cusps, obviously to protect them. Other mammals with
this low-cusped tooth form (bunodonty, it is called) show a similar arrangement, such as robust australopithecines (Gantt & Rafter 1998). There is
some disagreement though about whether this might be for resisting wear
or very high stresses (Kay, 1981).
Less well recognized is the possibility that the surface area of the postcanines could also be a response to wear (Lucas et al., 1985), which DeGusta
et al. (2003) provide some indirect evidence for. The logic is simple: abrasiveness acts at the foodtooth interface and it can, therefore, be grouped
with other external physical characteristics of food particles like particle size,
shape and stickiness. The only difference between them is that abrasiveness acts over a long periods of time, not within one masticatory sequence
(Fig. 6.3). The argument goes like this: suppose that food particles are broken during chewing at one of a nite number of breakage sites located along
the postcanine tooth row. Whenever food particles outnumber these sites,
they must, in any one chew, compete to be broken, i.e. not all particles
can be broken in that chew. As the sites are utilized again and again over
the lifetime of the animal, they gradually wear and eventually some breakage sites may be lost because of this. It follows then that loss of breakage
sites reduces the probability of particle fracture and slows the rate at which
the food is broken down. This will delay bolus formation, set swallowing
back and so reduce the rate of energy acquisition by the body.8 A logical
evolutionary response would be to increase the number of breakage sites by
Adapting to wear
199
Fig. 6.3 A map of the surface attributes of foods in mammalian diets and their anticipated
effect on the adaptation of tooth size. The diagram is identical to that of Fig. 5.11, except
that adaptation has been expanded to include abrasion, i.e. a factor that acts over a long
time-span.
increasing tooth size. For any given typical mouthful, this increase would
decrease the rate at which any one of these sites is used and would, therefore,
decrease their individual wear rate. The result is: larger teeth last longer.
The identication of breakage sites is not difcult: areas of wear on
the tooth row show where they are. There could be many sites at any
individual wear facet, for example and, in general, the larger the facet, the
more breakage sites will be found there. There is no need to go further
than that with dening these sites: the important point is that foodtooth
contacts do have xed locations because food is moved around the mouth.
In contrast, toothtooth contacts are completely xed: the same teeth will
always contact each other.
200
6 Tooth wear
There is now some evidence, from the study of DeGusta et al. (2003)
on howler monkeys, that molar size is positively correlated with longevity
in a mammal and there are also numerous examples of mammals with
relatively large molars. Examples include robust australopithecines, the recently extinct ape Gigantopithecus blacki (Pettifor, 2000) and gomphotheres
(Maglio, 1973), and the warthog, which has a last molar that probably equals
the area of the rest of the postcanine tooth row. Could molar sizes like these
be wear-resistant features? Why then do these mammals also have rather
thick enamel? Are increases in tooth size and enamel thickness just parallel
wear adaptations? I suggest not.
In the previous chapter, I argued that the type of food input that places
the biggest strain on the rate of energy acquisition by the body was the ingestion of chemically sealed small particles (such as seeds) in small mouthfuls.
I also suggested that the rate of breakdown in the mouth could be increased
by an evolutionary enlargement of the surface area of the teeth. These other
characteristics are also surface attributes and it is entirely consistent to add
abrasiveness to these. What I did not mention in Chapter 5 though was the
potential disadvantage of doing this. If there is little food in the mouth,
but large teeth, then there is a greater chance of toothtooth contacts. The
evidence from humans is that there would be about 50 ms in which dangerously high forces could build up between these teeth before inhibition
in the jaw-closing musculature protects them. In these circumstances, it
is possible to argue that an increase in enamel thickness is actually not an
adaptation to foodtooth wear, but to toothtooth wear in such mammals,
being the only way in which the effects of this auto-wear could be countered. Teeth always contact each other in the same locations: there is no
parallel to the circulation of food particles; so changing tooth size makes
no difference to the rate of toothtooth wear. Once started, the argument
must be pressed to its logical conclusion: toothtooth wear might be more
damaging to the dentitions of some mammals than food or grit. After all,
teeth are generally the hardest structures present in the mouth. They are also
composed of relatively tough ceramics and are capable of inicting heavy
wear on each other. This is exactly why the dentitions of non-mammalian
vertebrates are so arranged as to avoid this.
summary
While I did not say much in this chapter, the literature on tooth wear is
vast. There are big problems explaining the enormous range of phenomena
that has been reported because of limited theory. That given here is very
Summary
201
basic, but to step out further does not seem justied. It is worth mentioning
though that wear can be pictured very differently to the account given here.
Fracture, the separation of objects, is opposed to the general trend in the
universe, which is to attract objects so that they stick together (Kendall,
2001). Foodtooth or toothtooth wear could start by adhesion with wear
only resulting when one breaks away by a fracture line that lies off the
interface. Such adhesive wear is known (Ashby & Jones, 1996) and could
explain the plucking of rods (Walker, 1984) and why tough materials like
woody tissues (e.g. in leaves) and keratin can apparently polish teeth.
The evolution of the mammalian dentition
overview
This chapter is intended to be a chapter of ideas, mixing fact with suggestions that, although seemingly logical and based on the previous chapters,
might require a lifetimes work to substantiate in any detail. More precisely
stated then, it is speculative. When biologists do this (speculate) on the
adaptive signicance of changes in the form and function of plants and
animals by setting these adaptations in an ecological context, they tend to
refer to it as a scenario. In keeping with this theatrical jargon, this chapter goes distinctly off-lineage at times, mentioning the dinosaurs even,
but once the scenery is painted and the script sufciently advanced, the
mammals are swung in for their top billing.
By analogy to the sharpest way of controlling the content of television
programmes, those who really do not like it can simply shut the book: there
are generalizations beyond redemption here. However, all this stems from
the best of intentions, which is to offer new possibilities for explaining some
of the major trends in mammalian evolution, trends that are documented
very largely via the dentition. I try to portray the evolution of the dentition
of mammals in terms of the principles expounded in the previous chapters.
However, it would be impossible to cover every lineage and foolish to claim
that I can understand the range of forms that exist or have existed on the
basis of current knowledge. Instead, I have divided mammals up into a
small number of dietary adaptations and attempt to understand these. I
nish with an overview of the dentition of primates and what this can say
about the evolution of humans.
introduction
I assume here that the dentitions of mammals adapt to the fracture properties of their diets. For this to be so, mammals must either make feeding
202
Types of mechanical defence
203
decisions on the basis of perceived food texture or else select food on sensory grounds that correspond to a distinct range of textures. It is implicitly
assumed here that one or the other is correct.
Animals and plants do not want to be eaten.1 To prevent this, they defend
themselves in various ways. Some structures are obviously defences against
large predators (e.g. plant spines and thorns), while others are probably
barriers against a wider range of environmental threats, such as the skin of
mammals or the bark of trees. Whatever the primary role of these coverings
though, they also tend to show patterns related to predator avoidance,
such as spines, thorns, horns and thick dorsal (back) skin. Even the teeth
themselves are weapons, at least in part (Every, 1970). Regardless of whether
any feature is organized specically as a mechanical support (like bone
or wood) or primarily for defence, similar design principles are likely to
apply to avoid fracture (Ashby, 1989, 1999). The construction of the basic
mechanical indices was shown in Chapter 4. All the start of this chapter
does is to extend their use to strategies that organisms develop to avoid
being eaten.
types of mechanical defence

There are two basic types of strategy to avoid mechanical damage (Ashby,
1989, 1999):
(1) A structural component could be organized such that it breaks at a
stress higher than an animal could achieve without its own structure failing. These stress-limited defences depend on preventing cracks
from starting and are achieved either by a high yield stress, y , i.e.
high hardness, or by a high value of the square root of toughness
multiplied by Youngs modulus, (ER )0.5 . Structures defended in this
way are generally referred to by biologists as hard (Lucas et al.,
2000).
(2) A structural component might be constructed so that it fails at such
a large displacement that it limits the degree of detachment that an
animal could achieve in a given time. These displacement-limited defences
depend on preventing cracks from growing. The most important features
of such displacement-limited defences are either the toughness of the
animal or plant part alone if it is very thin, or the square root of
the ratio of its toughness to Youngs modulus, (R/E )0.5 , if it is fat.
Biologists tend to refer to structures like this as tough (Lucas et al.,
2000).
204
At rst sight, stress-limited defences seem to be the best alternative because

they hold out the prospect of the organism evading any damage at all.
However, they are an all-or-nothing strategy. Placing all resources into the
prevention of crack initiation risks catastrophe once a crack is started. A
high yield stress means that a large amount of elastic strain energy can be
absorbed, which is then released very easily into even the smallest crack,
resulting in almost instant fragmentation. Small cracks accumulate through
fatigue as well as from single bites (this is why engineers need toughness
in their construction materials: Gordon, 1991), so this defence has denite
risks. In contrast, a displacement-limited defence allows feeding to begin,
by allowing cracks to start, but avoids detachment by frustrating crack
growth. Without detachment, there can be no ingestion.
In Chapter 4, I suggested that stress-limited designs were likely to be important in deterring the rst bite of an animal while displacement-limited
designs would act better against repetitive fracture and fragmentation during chewing. The following (deliberately non-mammalian) example illustrates the ecological pressures that might have led to one or the other of
these defences.
The big ght: dinosaur versus invertebrate
Herbivores come in all sizes, but when we think really big, we think of
herbivorous dinosaurs (Sereno, 1997). Throughout most of the Mesozoic
era, these reptiles were the major consumers of plants. Being very large
compared to the foliage that they ate (often large compared to the size
of the entire plant that they were consuming), just one feeding bout (even
one bite) could be life-threatening to that plant. Additionally, dinosaurs did
not chew.2 If all this is put together, then defence against the rst bite and
subsequent crack initiation must have been everything to a plants survival.
Faced with such mega-herbivore pressure, it is likely that Mesozoic plants
began to develop stress-limited defences. What did these defences look
like?
A large dinosaur probably ingested leaves by pulling on them, but with
torsion and bending as potential complications to simple tension. To prevent crack initiation under these circumstances, plant leaves need to place
dense tissue as externally as possible because this sandwich arrangement
confers high hardness and stiffness (Gibson et al., 1988). Yet this mechanical defence cannot be spread evenly in the leaf because thick-walled cells
are opaque and impede the passage of light to underlying photosynthetic
organelles. Since the latter provide energy keeping the plant alive, woody
Types of mechanical defence
205
tissue has to be restricted in distribution, mostly being disposed as strands

lining veins (hydraulic conduits) because this keeps them out of the way of
photosynthesis.3 To resist tension or bending most effectively, the optimal
orientation of these woody tissues would be parallel to the direction of the
bite because this would maximize (ER)0.5 . A relatively narrow leaf design
reduces any chance of a bite by a big animal from all but one direction.
Thus, any plant that is attacked predominantly by large herbivores would
be likely to evolve long thin leaves with parallel longitudinal veins that
stand proud as ribs. Thorns and spines may be mounted on veins as extra
deterrents. However, these sharp structures need to be made from hard
materials, and hardness requires density. Cell wall is not really very hard, so
plants often employ silica, particularly in the tip of these structures, because
it is very hard, usually abundant and quick and cheap to embed in tissues
(Raven, 1983).
The ora of the Mesozoic is now glimpsed only in remnants, but towards the end of the Mesozoic, non-owering plants like ferns dominated
the land. The owering plants (angiosperms) that now dominate the worlds
ora had begun to evolve in the Mesozoic, particularly some of the monocotyledonous forms (monocots), such as palms. I envisage much of this
vegetation being populated by trees with a hard surface to both their trunk
and branches with frequent spines. Many of these plants have little lateral
expansion to their foliage and the central trunk tends to be pithy with protection conned to the outer layers. The great advantage of stress-limited
defences is their self-sufciency. Placing all investment in countering crack
initiation, there is no logical point in dealing with the consequences of
growing cracks, those that open cells, because these incidents trail the critical event.
The dinosaurs died out 65 million years ago and very different defoliators
took their place: invertebrates. In modern oras, these animals are responsible for most of the damage to the foliage of angiosperms (Coley, 1983;
Leigh, 1999). They are generally smaller than the leaves that they attack,
so one bite would not be serious.4 The most logical defence to the biting
variety would be to erect displacement-limited defences in order to stop
the detachment of a bitten part or simply slow the animals progress if not.
Dicotyledonous angiosperms have relatively wide leaves with a lattice-like
arrangement of their veins. Woody bres accompany these veins to defend
the leaf from potential attack from all directions.5
Displacement-limited defences depend on toughness. As stated in
Chapter 4 though, this depends on structure and structure takes time to develop. So young leaves need some other protection prior to the construction
Key:
plant storage organs
human foods
fruit
(a) INGESTION
seed contents
(b) MASTICATION
grass leaves
tree leaves
seed coats
increasing
difficulty
Toughness (J m -2 )
10
increasing
difficulty
10
BLADES
WEDGES
blades
cusps
10
10
CUSPS
10
1 -6
10
NO
TEETH
NEEDED
10
-4
10
-2
10
Youngs modulus (GPa)

5
10
horse hoof
10
antler
insect cuticle
3
10
10
nacre
cartilage
bone
dentine
2
hydroxyapatite
woods (along grain)
enamel
10
gel
1 -6
10
soybean
curd
10
-4
woods (indented)
skin
C
E
R
A
M
I
C
S
Toughness (J m -2 )
skin
Toughness (J m -2 )
10
woods (across grain)
10
insect cuticle
3
10
10
nacre
cartilage
bone
dentine
enamel
10
quartz
10
-2
10
quartz
1 -6
10
10
-4
10
-2
10
Fig. 7.1 Dentaldietary adaptation expressed in terms of plots of toughness versus Youngs modulus of potential food items to express their
relative resistance to (a) ingestion and (b) mastication. Data for human dental tissues and an abrasive (quartz) are included for comparison.
See text for explanations.
The basis for dentaldietary adaptations
207
of mechanical defences. Extensive work on leaf development (from Coley

(1983) to Coley & Kursar (1996)) shows that this protection is chemical.
There is an entirely logical association between the deployment of chemical
defence and displacement-limited mechanical defence because the former
is almost entirely intracellular while the latter permits such cells to be
opened. It is equally logical to expect noxious chemicals to be absent in
plants with stress-limited defences because, not only are these quick to
erect, but cells only open when cracks grow and this type of mechanical defence is intended to prevent that from ever happening.6 Plants like
cycads and monocots appear to contain few toxic chemicals, not I suggest because the latter are modern, but because they do not t into their
defensive pattern. Silica is often viewed as an archaic defence, organized
before the chemical warfare of angiosperms was organized, but I do not
agree.
I will now consider mammalian dentitions and diets using the same kind
of arguments about defences to predation.
the basis for dentaldietary adaptations
The toughness of many potential mammalian foods is plotted against their
Youngs moduli on logarithmic axes in Fig. 7.1. Most of the data are from
Appendix B, but a substantial amount of unpublished data has been added,
particularly for fruits and seeds. More plant data would make a total muddle. There are few good data on the toughness of animal soft tissues, but
the latter are difcult to incorporate anyway because of the frequency with
which they exhibit non-linear stressstrain relationships. For example an
ellipse is shown for the rat skin data of Purslow (1983) simply to encompass
the initial and nal modulus, which is to say that the slope of the stress
strain curve changes by a factor of about 103 from the onset of loading to
fracture.7
The use of the graphs in Fig. 7.1 is very different to that made of most
bivariate graphs in biology. The points do not fall on any line, nor are
they intended to. They sketch out a starry sky, the fundamental reason
for this being that while Youngs modulus is framed by material content
and the rules of mixtures that these impose, toughness is totally free from
such constraints. Unless study is restricted to a small group of tissues with
similar structures, when these properties are correlated with each other and
predictable (Ashby, 1998), then the sky at night is the result. This is a
central fact for the thesis of this book: it is the underlying cause of dental
diversity.
208
Lines showing equal values of (ER )0.5 and (R/E )0.5 have been drawn on
the graphs in Fig. 7.1, denoting resistance either to ingestion (Fig. 7.1a) or
mastication (Fig. 7.1b). These graphs include the same foods, but are not
identical due to considerations of anisotropy. For example, woody tissue
is likely to be pulled during ingestion. Thus, in Fig. 7.1a, woody tissue
is represented by across-grain and along-grain ellipses with the modulus
and toughness values matching that loading, i.e. the across-grain ellipse
consists of the modulus along the cellular axis and the toughness across
the cell, vice versa for the along-grain ellipse. Wood is probably mostly
indented in mastication though, so the single ellipse refers to the major
problem in fragmenting it loading perpendicular to the cellular axis with
toughness still being across the cell.
The smallest graphs (boxed with a dotted outline), positioned above the
vertical axes of the main graphs, indicate the degree of difculty that foods
present to the dentition. This is upward and to the right for ingestion, but
upward and to the left for mastication. Thus, a food that is difcult to
chew may not have been so difcult to ingest and vice versa. Note, if the
general theory is correct, that it would be impossible for a prey organism
to be equally defended against both.
The intermediate-sized graph in Fig. 7.1b has greyed-out food symbols,
superimposed on which are sectors bounded by thick solid lines that indicate dental adaptations to diets of which these foods would be typical. The
most critical division is the (R/E )0.5 line separating blades from cusps and
wedges. The line is equivalent to that shown in Fig. 4.10 from experiments
on humans and reects a point where particle fragmentation seems to become impossible without blades. This must actually be a graded change
rather than a threshold, but the theoretical point has to be made by imposing the latter. On the cusped side of the fence, foods with higher toughness
probably require marginal ridges on the cusps so as to form wedges. All the
above refers only to large food particles. For the special case of thin particles
(rods and sheets), the horizontal line shows that the criterion for fracture
is not (R/E )0.5 , but toughness alone (Fig. 4.9). If these foods have high
toughness, then blades are required to comminute them whatever their
modulus, while cusps are probably satisfactory if toughness is sufciently
small (the horizontal boundary is set at R 200 J m2 , a suggestion of
C. R. Peters). Finally, some foods are sufciently low in (R/E )0.5 or R that
postcanine teeth are not necessary at all. The position of this sector on the
graph is a pure guess.
I have less to say about ingestion because so little is known about it, either
experimentally or observationally, making it difcult to generalize. The
Insectivory and carnivory
209
underlying principle is that when (ER )0.5 is sufciently high, ingestion gets
progressively more difcult. However, the multitude of options available
at the front of the mouth means that many other factors come into the
picture. Experiments are currently under way to describe thresholds, but
critical factors in whether food is fractured between the teeth or simply
gripped and fractured against an external restraint must include friction.
When a food particle is sufciently thin, it will no longer fracture under
indenting incisors, but can still be pulled to fracture by gripping it and
acting against an external restraint such as (directly or indirectly) against
the rooting system of a plant or against the feeders upper limb. The reason
for this is the constant in Eqn 4.9 is very much lower in tension than
in indentation thin particles may fracture when pulled, but not when
indented (Atkins & Mai, 1985).
I will now consider the main dentaldietary adaptations of mammals,
referring to Fig. 7.1 whenever tooth shape is being considered. It must be
borne in mind though that nothing will replace the need for eldwork that
measures the mechanical properties of the actual diet. Just the age stage at
which foods are eaten makes an enormous difference to these properties.
insectivory and carnivory

The evolution of early mammals and changes in tooth number
Mammals evolved from synapsid reptiles in gradual evolutionary stages.
These fossil stages appear to have been perfectly viable, with their lineages surviving for millions of years (Hopson, 2001). The sequence started
with the synapsid reptiliomorphs (often called mammal-like reptiles) about
300 million years ago in the Late Carboniferous and culminated in the emergence of basal mammals, the Mammaliaformes, about 210 million years
ago. This was not the end of the trail though. The earliest mammal that
can be labelled eutherian (the kin of modern mammals) emerged about
125 million years ago (Ji et al., 2002). Eutherian mammals only started to
diversify and become dominant land animals at the end of the Cretaceous
about 65 million years ago.
Major changes in the head and neck region distinguish the extensive
sequence of fossils that mark this transition. In the mouth of an early
synapsid, just as in modern-day lizards, there sat a large number of teeth of
simple conical form. These unicusped teeth had no roots and were anchored
into position by being fused (ankylosed) to bone. The teeth were shed
210
periodically, being replaced rapidly by another tooth in the same position.

Osborn (1971b) described all the teeth occupying the same position in the
mouth of a reptile as members of one tooth family. Early synapsids had many
tooth generations per family, with each succeeding tooth being larger than
its shed predecessor. This size increase is usually explained by the relatively
constant growth rate of non-mammalian vertebrates, the size of each new
tooth being proportional to the size of the growing jaw (Berkovitz, 2000).
This obviously increases the chance of catching food.
Synapsid teeth sat in arches, more or less as in mammals, arranged around
the margins of the tongue. However, early on, the oral cavity had no roof,
allowing the tongue to gaze upwards at the upper respiratory tract. The
reptilian tongue, and probably that of synapsids, contains muscles that can
only act anteroposteriorly, so as to propel a food particle backwards towards
the pharynx. Their pharynx has no specic musculature for swallowing, appearing to rely on supercial subcutaneous muscles to help move a single
ingested food particle towards the stomach (Smith, 1992). The upper and
lower teeth of a synapsid did not contact and the lower jaw moved vertically. The teeth were partly covered by cheeks on their lateral sides that
hung loosely without muscle (there being little need for ne control of the
buccolingual direction of food movement because synapsids did not chew)
in order to accommodate a large gape.
In contrast, the mouth of the earliest mammals consisted of fewer teeth,
in only two generations, housed in bony sockets. The teeth were clearly
differentiated into three tooth classes: (1) peg-like incisors, (2) large canines
(in both deciduous and permanent tooth generations) and (3) postcanines
that possessed several cusps. The cheeks appear to have been muscular and
there was a hard palate above a tongue that was capable of side-to-side food
movement. Food was usually shifted to the left or right sides of the mouth
by these tongue movements. The lower jaw was narrower than the upper,
such that it had to move a fraction sideways in order that teeth on the
non-active side would stay out of contact. The pharynx itself had a novel
musculature designed to take a bolus of fragmented food particles quickly
down to the oesophagus.
It is generally thought that the major selective pressure for all the above
anatomical changes resulted from a dietary focus on insects. Mammals
evolved alongside the angiosperms, a group of plants relying on insect
pollination rather than on wind (the agent used by most other plants). In
order to attract adult insects, angiosperms developed owers containing
nectar as an attractive reward for adult insects. However, for an active
pollinating agent like an insect to have been superior to passive means
211
(i.e. the fortuitous accident of wind), then all growth stages of insects
would have to be supported by a food source. The nectar of angiosperm
owers could not supply nutrition to insect larvae and it seems that the
latter have often favoured their young leaves as a food source.8
Angiosperm-linked insects offered mammals several food choices. They
could, for example, have searched tree bark for adult insects running to and
from owers or instead they could have concentrated on larvae located on
leaves. Tree shrews are living insectivores with dentitions somewhat similar
to those of early mammals and which illustrate some of the variety of niches
available to insect feeders. Pen-tailed tree shrews (Ptilocercus lowii) seem to
catch their prey on tree bark as they rush along the trunk or large branches.
Other tree shrews, e.g. in the genus Tupaia, are more likely to investigate
foliage during day and night (Emmons, 2000).
Insectivory in early mammals would have entailed many changes to
body form but a major one, apparently neglected in the literature, was undoubtedly body size reduction. Insects probably were (as they still are) very
small, so mammals would have had to reduce in size from their synapsid
ancestors in order to consume them.9 There was a range of sizes in early
mammals, presumably reecting a size range of insect prey. The smallest, Hadroconium, from about 195 million years ago, weighed in at only
2 g, but even a large Sinoconodon specimen weighed <500 g (Luo et al.,
2001). At sizes like that, heat loss via body surfaces is critical, thus putting
a premium on rapid digestion. Also, insects are (and probably were) active
mainly at night. To maintain nocturnal activity levels, mammals would have
needed to control their body temperatures because locomotor stamina in
the coolest part of the day requires an elevated body temperature that reptiliomorphs lacked (Hopson, 2001). In order to supply these energy requirements, the front end of the gut had to start processing insects into smaller
pieces.
Once mammals began to masticate food, the scaling arguments of
Chapter 5 suggest that the sizes of the jaw and anterior teeth were decoupled from those of the postcanines because fracture mechanics, not
gape, determines the size of the cheek teeth. The size disparity would have
depended on the mechanical properties of the insects that they ate, but,
almost certainly, this would have led to negative allometry of the postcanine dentition and thus to enlargement of the posterior teeth in a dwarng
mammal. The result would have been congestion of the jaw, something
that could have resulted in tooth loss. Small vertebrates do not live as long
as larger ones and so a reduced lifespan in early mammals could also have
been responsible for a reduction in the number of tooth generations. Thus,
212
Fig. 7.2 A schematic view of tooth loss from reptile to mammal due to reduction in body
size. In life, the jaw grows in size, but the diagram is scaled so that tooth families can be
visualized. Lost teeth are shown as being greyed-out. These include tooth families (vertical
in the diagram) and whole generations (horizontal). In reptiles, tooth size is proportional to
jaw size, which is an adaptation to prey size. In a smaller mammalian descendant, the size of
the anterior teeth and the jaw is still determined by the same selective pressures, but the size
of the posterior teeth is relatively enlarged because of fracture scaling (Chapter 5). Posterior
tooth loss is suggested, but this is somewhat arbitrary. The reduction in the number of tooth
generations follows from a reduced lifespan.
the loss of tooth number and tooth generations in early mammals could be
explained by dwarsm, as Fig. 7.2 suggests.10
Molar shape in early insectivorous mammals
In this book, I have consistently tied the adaptation of tooth shape to the
fracture properties of foods, the requirement for particle fragmentation in
mammals being a function of relatively high basal metabolic rate. The fossil
evidence for the evolution of early mammals from synapsids suggests that
basal metabolic rates increased only gradually from reptilian levels, with
most intermediate forms only having partial control of body temperature
(Hopson, 2001). It is probable that changes in tooth form during this lineage
represent change, not in the mechanical properties of the diet, but in the
213
relative amount of fracture/fragmentation required to service a particular

metabolic rate.
The dental features required for fracturing insects appear to depend
mostly on the mechanical properties of the cuticle.11 Insect cuticles have
complex architectures (Vincent, 1981). Though based on chitin, their properties depend very much on the degree of a process called tanning or sclerotization (Vincent & Hillerton, 1979). Tanning is an age change in most
insects (Schoeld et al., 2002), converting a low modulus material (e.g.
Youngs modulus of 0.015 GPa: Vincent, 1980), probably with a J-shaped
stressstrain curve, to one that is effectively linear and with a much higher
modulus (Vincent, 1980, 1981). Yield strength rises dramatically with the
degree of tanning (Hillerton et al., 1982; Schoeld et al., 2002). Yet, these age
changes are not inevitable. The adults of some groups, like the Lepidoptera,
have very lightly tanned cuticles while parts of the cuticle of beetles can be
very heavily tanned. Which type of prey did early mammals eat?
There was a consistent trend in early mammalian evolution towards the
tribosphenic form where molars have sharp cusps linked by sharper concave blades (Crompton & Sita-Lumsden, 1970; Crompton, 1971). Evidence
from modern insectivorous groups (Freeman, 1979, 1981b; Strait, 1993a,b)
suggests that blade development is much more prominent in forms that
eat softer insects. Inexed blades were associated in Chapter 4 with foods
that break at high strain and that have high Poissons ratios. It is very likely
that synapsids and early mammals tended to eat Mesozoic insects with
lightly tanned cuticles probably larvae. These cuticles are tough enough
(1.5 kJ m2 ; there are few published data) not to propagate cracks easily:
they are a displacement-limited defence. Making a fracture in the cuticle
of any size would require a sharp structure, while continuing this to obtain
fragmentation would require the blades to move through the food and nally to contact. If they did not, then the crack would arrest. It is likely to
have been the need to fragment insects with these properties that led to the
need for tooth contact ( occlusion).12
Prinz et al. (2003) show that one subdivision of an insect (Tenebrio molitor
larva a mealworm) with a lightly tanned cuticle is sufcient for extensive
digestion by gut enzymes.13 The cusped teeth of early synapsids (Fig. 7.3a)
may only have needed to make a small opening to obtain a sufcient rate
of digestion from it. Such fracture does not require the teeth to contact. As
metabolic requirements increased over time and body size decreased, molar
shape seems to have been adapted to increase that exposure by turning
fracture into fragmentation. That needed blades, which were developed
between cusps. Initially, in an early mammal like Morganucodon, there
214
(a)
(b)
(c )
stylar
Prd
area
Me
Pa
Pad
Med
Pr
dD
Fig. 7.3 (a, b, c) Schematic view of the evolution of molars in early mammals (see ickart
for a side view). Upper teeth have unlled cusps. (a) The spaced single cusps of a basal
synapsids. (b) The occlusion of an early mammal such as Morganucodon, in which upper
and lower molars had three cusps arranged in a mesiodistal line that had to wear against
each in order to align properly and make a blade. (c) The triangulated molars of a mammal
like Kuhneotherium t together without wear, forming blades on mesial and distal surfaces
of the teeth. (df ) The upper and lower molars of Didelphis marsupialis, a polyprotodont
marsupial. The stylar region of upper molars evolved very early and its relative extent on the
working surface is a clear marker for insectivory in many mammals, e.g. microchiropteran
bats (Freeman, 2000). Scale bars, 1 mm.
was no preformed occlusion and the upper and lower teeth, consisting essentially of a mesiodistally aligned row of three cusps (Fig. 7.3b), worked
against each other to produce an edge. At best, this might have produced
one subdivision of the insect. Preformed occlusion was developed in later
forms and there then followed a gradual tendency to increase the number of blades that could act simultaneously so as to increase the amount
215
of fragmentation. Presumably, higher rates of energy acquisition and diminishing size drove these trends. Triangulation of the cusps (Fig. 7.3c),
such as in Kuhneotherium, positioned blades at angles so increasing the
chance (selection function) of a blade hitting the insect and speeding up
the process. The subsequent evolution of the tribosphenic form (Fig. 7.3d)
has been described by Crompton & Sita-Lumsden (1970) who show how
it is designed so that several blades can act almost simultaneously on either
side of any given tooth with lateral jaw movement and waggling of their
mandibular symphyses.
The tribosphenic form has cusp-in-fossa alignments. This suggests the
additional consumption of insects with more heavily tanned cuticles, probably adult insects. A tanned cuticle is more towards a stress-limited design,
likely to fail by rapid crack propagation, so rendering the need for blades
unnecessary for this aspect of the diet.
The need for both blades and cusps on the molars of tribosphenic mammals is explained by the position of the ellipse for insect cuticles in Fig. 7.1b,
which straddles the dental divide between blades and cusps: because tanned
cuticles will crack easily while untanned ones will not, so needing blades.
The great problem with all of this is that very little is still known about
the diversity of insects in the diet of living mammals, let alone those in the
Mesozoic, or their mechanical properties (Strait & Vincent, 1998). What
information there is does not suggest that insect cuticle is anything like as
tough as the veins in leaves, for example. Exactly how much fragmentation
of an insect is or was necessary is difcult to estimate. Modern insectivores
that have molars that approximate a tribosphenic form chew insects only
briey (e.g. the Western tarsier (Tarsius bancanus), a nocturnal primate
observed by Jablonski & Crompton, 1994).
Just a glance at the dentitions of the living Insectivora, the ragbag old
order of small mammals (now disbanded), covers up dietary diversity belied by its nametag (Strait, 1993a). Little detail is yet known about the
consumption of invertebrates other than insects, but this certainly seems to
hold many keys to the understanding of the diversity of mammaliaforms.
Some hints come from the tinted enamel referred to in Chapter 2. The
Soricidae (the shrew family, which includes the smallest living mammals),
have two dental types: the red-toothed Soricinae and the white-toothed
Croidurinae. Red or brown pigmented enamel is known to contain iron
deposits, probably always from breakdown products of haemoglobin. A
group of authors including Akersten have consistently alluded to this red
pigment being tied to the consumption of earth-dwelling prey like earthworms and some insect larvae (Akersten et al., 2002). Siliceous grit is usually
216
harder than dental enamel, but the red-tinted enamel of soricines contains
goethite, a very hard mineral, that apparently gives the radulae of some molluscs a hardness of about 6000 MPa (Runham et al., 1969). This should
resist abrasion by grit better than plain white hydroxyapatite-based enamel
(Akersten et al., 1984). Plausible as this is, there is little support as yet from
microindentation. Nanoindentation studies (Appendix A) are called for
because this red layer is thin the teeth as a whole are very small and
the hypothesis is eminently sensible. Akersten et al. (2002) point out that a
Palaeocene multituberculate (Lambdopsalis bulla), in a mammalian lineage
separate to therians, also had red-brown enamel, indicating at least that this
terrestrial niche was available at that time.
Tooth form in carnivores
Of all the major dietary streams in mammals, carnivores (vertebrate feeders)
and piscivores (sh-feeders) do the least chewing. Generally, their anterior
teeth resemble those of early mammals and, in some sh-feeders, the whole
dentition is similar to that of a reptile. The critical teeth for ingestion are
the canines, which provide the simplest link back to reptiles because of their
simple single-cusped form. They show the same kind of conical tooth form,
with a recurved (actually spiralled) shape. The upper and lower canines of
carnivores are generally long and projecting teeth, although their degree
of projection varies greatly. The uppers are longer than the lowers. The
canines act at larger gapes than the other teeth (Fig. 6.2). The essential load
on them in a carnivore is not just due to indentation of the tips, but to
an anteroposterior pull (Simpson, 1941; Smith & Savage, 1959). The crosssectional shape of the canines tends to reect this, in that they typically
have oval cross-sections with the long dimension being anteroposterior
(Van Valkenburgh & Ruff, 1987). Carnivores that use their canines to
break bones must deal with the vagaries of the direction of the bite force
depending on the orientation of the bone vis-à-vis the tooth (Rensberger,
2000). These tend to have circular cross-sections (Van Valkenburgh & Ruff,
1987).
Vertebrate soft tissues almost certainly will all lie to the bladed side of
Fig. 7.1b. Most carnivores have just one blade on either side of the mouth
the carnassial. They do little processing of these soft tissues because they
are readily digestible even in large pieces. Generally, they only seem to manipulate one particle of food in the mouth at a time and all that the tongue
has to do is organize this so that it lies between the carnassials. Very often, it
appears that fracture between the carnassials is an ingestive process because
Plant feeders
217
the bite is made on the carcass with most of the food outside the mouth.
The carnassials always have the inexed blades referred to in Chapter 4
(Fig. 6.2). Their size indicates the probable linear dimensions of swallowed
pieces and their position is presumably linked to the gape necessary for their
use (Fig. 6.2). Some carnivores, such as hyaenas and canids, also fracture
bones. Mineralized tissues have very low (R/E )0.5 ratios and will fracture
the carnassials if these are used to break them (Chapter 6). Hyaenas have
developed blunt conical premolars in front of their carnassials, while canids
use molars distal to them.
plant feeders
Frugivory (fruit-eating)
As mentioned in Chapter 4, fruit esh is undefended because it is intended
to be eaten, being the reward offered to a frugivore for dispersing its seed(s)
away from the parent plant.14 The arrangement is the equivalent of employing animal pollinators. In the Cretaceous, before dinosaur extinction,
angiosperms probably populated areas with relatively still air close to water.
Some early angiosperms were actually aquatic, but by the late Cretaceous, it
is clear that there were a considerable diversity of forms on land (Wing et al.,
1993). Being in relatively still air not only needed a pollination agent other
than wind, but an alternative dispersal agent for seeds. Even today in tropical rainforests, animals disperse most of the seeds, this independence from
wind having facilitated the development of such stratied forests (Eriksson
et al., 2000). Mammals diversied very rapidly after the dinosaurs died out
65 million years ago. Seed and fruit size records suggest that diversity peaked
very early in mammalian diversication during the late Palaeoceneearly
Eocene epochs (Collinson & Hooker, 1991; Eriksson et al., 2000), suggesting that mammals became signicant consumers of fruit (frugivores) at that
time.
The seeds of a plant contain its embryos the potential members of its
next generation so these have to be heavily insured against death or else the
plant species may not survive. Attack by many kinds of animals including
mammals, both large and small, can devastate a seed crop at various stages
of development. The attraction for animals is basically the energy reserve
intended for germination, this being either carbohydrate (starch) or lipid
(in the form of oil) in nature (Leighton & Leighton, 1983). Angiosperms
have developed varied defences to counteract this threat. Many produce
seeds in vast numbers, rather as some animals do with eggs, sacricing
218
Table 7.1 The characteristics of fruit skins and peels consumed by primates
Species
Thickness (mm) of outer

covering (type)
Fruit width
(mm)
Alangium nobile (Alangiaceae)

Cyathocalyx ramuliorus (Annonaceae)
Mezzettia parvifolia (Annonaceae)
Willughbeia coriacea (Apocynaceae)
Gnetum microcarpum (Gnetaceae)
Garcinia forbesii (Guttiferae)
Fibrauria tinctoria (Menispermaceae)
Calamus luridus (Arecaceae)
Xerospermum sp. (Sapindaceae)
1.8 (peel)
1.7 (peel)
0.6 (skin)
5.0 (peel)
1.5 (peel)
2.9 (peel)
1.9 (peel)
0.6 (peel)
2.2 (peel)
25
21
52
30
14
27
20
14
20
Source: Lucas & Corlett (1991).
the majority to ensure the survival of a few (Curran & Leighton, 2000).
There is plenty of evidence that others produce seeds that contain very toxic
chemicals (Bell, 1984; Waterman, 1984). Although mechanical protection
is also common, it seems rare to nd extensive chemical and mechanical
defences together.15 A minimal amount of mechanical protection to seeds
always seems to be present, presumably in part so that mammals can distinguish seeds from esh in the mouth by their hardness (Corlett & Lucas,
1990).
I will follow a simple classication of fruits on mechanical grounds.
Fleshy fruits can be distinguished from dry (eshless) ones, while the former can be subdivided into those with a peel versus those with a skin
(Janson, 1983). Dry fruits are generally either wind- or rodent-dispersed
(Turner, 2001). A peel is a generally thick outer covering that separates
cleanly from the underlying esh at ripeness (Table 7.1). Primates are associated with these protected fruits (Janson, 1983; Gautier-Hion et al., 1985;
Leighton, 1993). Familiar examples of these mechanically protected fruits
are citrus fruits with their thick peels. These can be compared to unprotected fruits that have only a skin like an apple, for example, and which
can be eaten by a wide variety of consumers including birds, bats and
primates.
There have been many attempts since Ridley (1930) to match up the
varied form of eshy fruits with features of their dispersal agents (including
mammals), but there appear to be a lot of mismatches. Some of these are
probably due to inertia on the plant side of the fence (Janzen & Martin,
1982; Jordano, 1995; Herrera, 2002), but another issue is the behaviour
Plant feeders
219
of mammals whenever there is a fruit glut. Terborgh (1983) studied the

behaviour of primates at a tropical rainforest site in Peru. Whenever there
was a fruit glut, many mammals and birds would feed on these, with no
apparent relationship between the anatomical form of these animals and
the fruit that they ate. However, in periods when fruit was very scarce,
i.e. when resources diminished and survival became difcult, each primate
under study moved to a resource that distinguished it from the others and
which did make considerable anatomical sense. For example, individuals of
Cebus apella, a capuchin monkey with a large jaw and low-crowned, thick
enamelled cheek teeth ate palm nuts. More lightly built squirrel monkeys
(Saimiri sciureus) foraged for gs, while tamarins (Saguinus spp.) searched
for nectar (Terborgh, 1983). This fallback food hypothesis, also called a
keystone food resource hypothesis, seems to apply particularly to frugivory
and could explain anatomical mismatches (Terborgh, 1986). The idea that
the dentition itself was designed for such critical resources appears to be
due to Rosenberger & Kinzey (1976).
Figure 7.1 shows that fruit esh is very variable in its properties, due as
much to cell size as to cell wall thinness. Its stiffness is a product of its
turgidity (Chapter 4), but once cells burst, this is quickly lost. Teeth may
not be needed to process many fruits because even the at surfaces of soft
tissues around the mouth can fracture cells (as suggested in Fig. 7.1b). Fruit
bats do this by compressing some types of fruit esh between the tongue
and hard palate.16 If the soft tissues of the mouth can fracture these tissues,
then it is also probable that the soft tissues of the gut can too. There is little
point then in the fragmentation of a lot of types of fruit esh by chewing.17
However, even if moisture-laden, some esh does not give up its moisture
very easily (a feature that is well known in domesticated tropical fruits too:
Peleg et al., 1976).
The spatulate incisors of anthropoid primates are relatively large in fruiteating species (Hylander, 1975), making it likely that the trait evolved in
association with this type of diet. One of the more plausible ecological links
in fact is that between anthropoid primates and the consumption of fruits
with peels (Janson, 1983; Gautier-Hion et al., 1985; Leighton & Leighton,
1983; Leighton, 1993). However, the work of peeling itself (Appendix A) is
generally very low and peels themselves are neither very thick (Table 7.1)
nor very tough. How does a plant ward off fruit thieves with fruit like this?
The answer appears subtle. Successful peeling is only possible at full ripeness
and requires the control of cracking such that the fracture plane remains at
the junction between peel and esh. This can only be achieved by getting
a bladed tooth into the interface between the tissues to guide the crack.
220
Additionally, peel thickness seems generally to lie below the deformation

transition described by Eqn 4.9 (Chapter 4) in other words, cracks will
not propagate within the peel when it is penetrated and so the blade of a
spatulate incisor is needed to remove it.
The molar teeth of most frugivores are generally rather small (Kay, 1978;
Strait, 1993b; Freeman, 2000), being low-crowned with rounded cusps (Kay,
1975), presumably designed for cellular fracture rather than fragmentation.
The diets of primate fossils are commonly categorized on this basis and the
introduction of fruit into the diets of insectivorous early mammals seems
marked by a reduction in blade length and a rounding of cusps (e.g. Kirk &
Simons, 2001). However, there can be a lot of variation in the dentitions
of frugivorous primates as Fig. 7.4 indicates. A male gibbon is larger than
a male long-tailed macaque, yet the latter has a larger dentition all round.
How can this be understood? I suggest by looking harder at how these
primates treat seeds.
Patterns of seed treatment
Seed treatment in general, whether consumption or avoidance is practiced,
is of major evolutionary importance and it should be anticipated that there
are considerable dental adaptations in frugivores for dealing with them. I
will categorize mammals on the basis of how they treat seeds and see if
this helps in understanding tooth size variations. The classication derives
from the study of the behaviour of wild long-tailed macaques (Macaca
fascicularis), a Southeast Asian Old World monkey commonly used in
biomedical research (Corlett & Lucas, 1990; Lucas & Corlett, 1998). Seed(s)
can be:
(1) Ingested, then swallowed almost immediately. The intact seeds pass
straight through the gut and are defaecated. (Seed swallowing)
(2) Held by the hand while the esh is cleaned off the seed by the incisor
teeth. (Seed cleaning)
(3) Ingested into the mouth where the esh is removed entirely or in part
by the postcanine teeth. The seeds are then spat. (Seed spitting)
(4) Ingested, then chewed with the postcanine teeth until swallowed. The
seeds would be destroyed by this action, rendering their content digestible. (Seed destruction)
These four behaviours, the swallowing, cleaning, spitting and destroying of
seeds, were observed in just one social group of long-tailed macaques, but
they did not occur with equal frequency (Corlett & Lucas, 1990). Clearly,
individual animals have some sensory criteria for deciding how to tackle
Plant feeders
221
Fig. 7.4 The contrast in dental morphology produced by the consumption of very similar
fruits by sympatric primates, but which treat these foods in very different ways. The longtailed macaque, an Old World monkey, has cheek pouches that it uses to store esh-covered
seeds, retrieving these one by one, chewing off the esh and then spitting the seeds out.
The gibbon tends just to swallow seeds with the esh, later defaecating the former.
the seeds in a fruit. Long-tailed macaques destroyed the seeds of all the
dry fruits that they ate, but avoided breaking down the seeds of almost all
eshy fruits.18
The treatment of eshy fruits appeared to depend on fruit and seed size.
Lucas (1989) postulated four size thresholds to explain the behaviour of
these monkeys, which can be generalized. In Fig. 7.5, fruit size thresholds are
denoted as F1 and F2 , with F1 F2 , while seed size thresholds are symbolized
222
Fig. 7.5 The processing of single-seeded fruits by long-tailed macaques. The mouth of these
monkeys is viewed from above, with the lips to the right and the passage leading to the
pharynx shown at left. Soft oral tissues are shown in grey. These monkeys have highly
sensitive cheek pouches that communicate with the vestibule through narrow openings.
The esh of a fruit is shown as a dark band around the single seed. The presence or absence
of a fruit peel is ignored. (a) Shows what happens to fruits that can t into the mouth (i.e.
of a size F1 ), but which have seeds that are too large to t into a cheek pouch (>F2 ) and
too big to swallow (>S2 ). (b) Shows the oral passage of a fruit that can t into a cheek
pouch and also be swallowed (although the two thresholds are independent). (c) Shows the
thresholds that apply to a dry fruit (shown just as though it was a seed, this containing
all the nutritive value).
Plant feeders
223
as S1 and S2 , with S1 S2 . F1 and S1 are both ingestion thresholds. Feeding

on fruit could be understood in terms of these four thresholds.
Macaques would pluck large fruits with their hands and bring them to
their mouths. The size of the opening to the mouth (mouthslit) controls
the entry of large fruits. If the fruit minus any peel was too large to t
into the mouth, i.e. >F1 , then it would be continued to be held by the
hand(s) and cleaned in front of the mouth. There is a relationship between
seed and fruit size (Lucas & Corlett, 1991; Dominy, 2003), but it is not very
strong. A eshy fruit can contain one, several or many seeds, although of
course, only large fruits can contain large seeds. If seeds within a fruit of
size >F1 had a size >S1 , then they would be cleaned outside the mouth and
dropped. If however, the seeds <S1 , then they might enter the mouth so
that more esh could be stripped from them.
Any fruit of size <F1 would be ingested. All cercopithecine Old World
monkeys, including long-tailed macaques, possess muscular cheek pouches
located opposite the molar teeth for storing food (Figs. 7.4 and 7.5). These
pouches get distended when lled with fruits, which were the only item
of natural food long-tailed macaques were seen to place in them. The
pouches are probably smaller than the oral cavity proper. So, dening the
maximum size of each cheek pouch as F2 , we can say that fruits <F2 could
be stored in them. Long-tailed macaques appear to retrieve fruits from
these pouches one at a time, pushing them back into the centre of the oral
cavity and removing the esh with the teeth (Fig. 7.4) while they move to
another food source. Once seeds had been cleaned, long-tailed macaques
either spat the seeds or swallowed them. Only the smallest seeds were swallowed and found in the faeces and an intra-oral seed detection size, S2 ,
can be postulated, above which seeds were spat and below which they were
swallowed.
The size of the fruit and seed size thresholds in macaques appear to
be F1 3040 mm, F2 1015 mm, S1 30 mm and S2 34 mm,
all dimensions being maximum seed width (i.e. the second largest object
dimension).19 Thus, the seed size thresholds, S1 and S2 , are a decade apart in
macaques, which explains why these monkeys spit >70% of the seeds that
they process (Corlett & Lucas, 1990). Despite the variety of seed behaviours
that they display then, long-tailed macaques are seed-spitters, probably to
try to spit the seeds of all eshy fruits that they can.20
The evolutionary point of this is that mammalian frugivores can be
classied into spitters, cleaners, swallowers or destroyers, the basis being the
fate of either the majority of seeds or those of critical (fallback) fruits if
this is thought to be more important.
224
What dental or oral changes are necessary for the efcient processing of
seeds in these various ways? The coincidence of S2 with the oral mucosal
thresholds discussed in Chapter 3 makes it likely that all mammals possess
a similar threshold. So seed spitters and seed predators are very likely to
swallow very small seeds simply because they do not detect them. However,
few mammals other than cercopithecines appear to use this lower detection
threshold to actively eject seeds. The singular ability of cercopithecines to
spit seeds appears to be connected to their possession of highly sensitive
cheek pouches. Most other mammals are probably limited to either cleaning
or swallowing seeds. The process of cleaning seeds requires a level of manual
dexterity present only in anthropoid primates with thumbs, such as longtailed macaques, and apes (e.g. the chimpanzee: N. J. Dominy et al., unpubl.
data). Anthropoids with reduced thumbs, such as spider monkeys, gibbons
and colobines, have very limited ability to clean seeds. So, most mammals
probably have S1 = S2 , meaning that they will swallow any seed that they can
ingest. This is likely to be controlled by the size of the mouthslit, which is in
turn a reection of the maximum gape of the animal. For African elephants
(Loxodonta africana), this means that they can ingest and swallow Balanites
wilsoniana seeds that are 88 mm in seed length and 47 mm in seed width
(Chapman et al., 1992).
The tooth sizes of anterior (incisor) and posterior (postcanine) teeth
of frugivores can be deduced by making some simple assumptions. The
time available for processing fruits is assumed to be limited, so that any
behavioural changes towards one category of seed processing or another
would result in changes that reduce processing time. Support for these
considerations comes from Ungars (1994) study of four Sumatran primates
in which he suggests that the time spent in processing foods with the incisors
may be important in determining their size.
Further, from Chapter 4, I assume that higher primates have spatulate
incisors to cope with protected fruits (those with peels), such that the
common ancestor of all these primates peeled these fruits efciently. The
most important dimension of a spatulate incisor is its width (i.e. its
mesiodistal dimension). For the postcanines, changes in either the buccolingual or mesiodistal dimensions of their working surfaces could change
their efciency.
Seed swallowers peel fruit with their incisors, but do little with their posterior teeth, since esh plus seeds are quickly swallowed. Seed swallowing
is probably the ancestral behaviour pattern, requiring less in the form of
manual or oral adaptations than the others. Most lineages of eshy fruits
Plant feeders
225
probably evolved in the Eocene before anthropoid primates, but protected

fruits (a small percentage of those available in every site yet studied) probably co-evolved with anthropoids. Seed cleaners do most of their oral processing at the front of their mouths with their anterior teeth, whereas seed
spitters, perform extensive processing of fruits with both anterior and posterior teeth. Designating the size of anterior teeth as A and the overall size
of the posterior teeth as P, and employing for small and + to mean
enlarged, then the following dental congurations for the three categories
of primate can be anticipated:
SEED SWALLOWERS
AP
SEED CLEANERS
A+P
SEED SPITTERS
A+P+
These represent three of the four possibilities for combining these symbols. What about AP+, a primate with small anterior, but large posterior, teeth? It has been established for a long time (Hylander, 1975; Kay,
1975) that primates with this dental conguration are likely to be predominantly folivorous (leaf-eaters) rather than frugivorous. It would, of course,
make a far cleaner classication if there were a way to incorporate these
apparent folivores into the classication. The way to do this is to recognize that when most of these folivores eat fruits, they destroy their seeds.
Many colobines are seed destroyers, eating leaves for the most part, but
also consuming large quantities of seeds when these are available (McKey
et al., 1981; Bennett, 1983; Harrison, 1986; Davies & Baillie, 1988). The
leaf-eating sifaka, Propithecus verreauxi, is also a seed predator (Overdorff
& Strait, 1998; Yamashita, 2000).
It is possible, therefore, to nish off the classication like this:
SEED SWALLOWERS SEED CLEANERS SEED SPITTERS SEED DESTROYERS
AP
A+P
A+P+
AP+
This addition makes logical sense if the processing of fruit by a seed destroyer is considered. Seed destroyers usually either concentrate on dry
fruits, where the anterior dentition has little role in processing or on unripe fruit. There is no peeling involved because this does not detach from
the fruit and, as pointed out by Leighton (1993), when a primate targets an
unripe fruit, it does so for the nutrients contained in the seed. So, fruits
are immediately thrown to the posterior teeth, which take an extra processing load, as compared to a seed swallower or seed cleaner, while the
anterior teeth do not. The evidence is very strong that colobines have small
anterior, but large posterior, teeth (Hylander, 1975; Kay, 1975). The gelada
226
baboon (Theropithecus gelada), a cercopithecine, is another example of

a folivore/seed eater (Dunbar, 1977; Iwamoto, 1979) and has an AP+
dentition too (Fig. 5.5) (Jolly, 1970).
Most cercopithecines other than the gelada baboon are seed spitters.
Extensive study of the redtail monkey (Cercopithecus ascanius) conrms
this in Kibale, Uganda (Lambert, 1999). In contrast, gibbons appear to be
seed swallowers, avoiding fruits that will be t into their mouths (Whitten,
1982). These thumbless primates have an AP dentition. New World
spider monkeys (Ateles spp.) have small thumbs, but this is sufcient to
allow them to clean some fruits (van Roosmalen, 1980). They have small
posterior teeth their last molar is occasionally absent (Zingeser, 1973)
and could be classied as A+P.21
In contrast, the great apes seem to be seed cleaners. Certainly, chimpanzees show a much greater propensity to swallow seeds than do redtail
monkeys (Lambert, 1999), but they are also equipped with hands that facilitate the processing of fruits in front of the mouth. Large fruits are processed
in this way (N. J. Dominy et al., unpubl. data). However, outside primates,
frugivorous bats are the best example of seed cleaners and Freeman (1988)
shows that the dentitions of some microchiropteran bats are very denitely
A+P in form.
A note on seed treatment by humans is applicable here. Some readers
may feel, from personal experience, that it is impossible to categorize seed
treatment by modern-day humans for example, some people spit grape
seeds, others swallow them while still others destroy them. However, as
stated in Chapter 1, populations that have control over their food supplies
may not provide a good guide to the behaviour of humans with seeds in the
past (any more than with diet as a whole) because the leisurely oral activities
that we exhibit could be maladaptive under natural conditions. One clue
to selective pressures being eased is the very fact that seed treatment is so
variable. Ridley (1930) gives a very clear account of the behaviour of hunter
and gatherer groups on the Malay Peninsula and states that they generally
swallowed seeds, even one as large as those of the durian (Durio zibethinus,
Bombacaceae).
Molar tooth shape in frugivores
Frugivores, whether seed destroyers or not, tend to have blunt-cusped molar
teeth because most fruit esh lies in the cusp sector in Fig. 7.1b. The
principal reason for this was given in Chapter 4, where it was shown that
Plant feeders
227
both fruit esh and seed shells often have a very low toughness (Figs. 4.15
and 4.17). Furthermore, if the objective is to break open as many cells are
possible, cusps do this much faster than blades.22 Thus, the molars have
few prominent crests ( shearing crests analysed by Kay (1975, 1978; Kirk
& Simons, 2001)) because cracks spread easily. An exception may be the
noticeable marginal ridges of mammals that feed on seeds whose shells need
to be wedged open (Fig. 4.18).23 Several seed shells lie in the wedge sector
in Fig. 7.1b. The cusps may be more bulbous in seed destroyers because the
bite force may not be aligned with the direction of jaw movement when
very stiff foods are being chewed (Rensberger, 2000). The enamel of the
cusp tips may also be thicker (Kay, 1981; Walker, 1981), both these features
being adaptations preventing tooth fracture. Consumption of a lot of highmoisture fruit is likely to result in very loose-tting features on upper and
lower molars to allow the expressed juice to escape. As was pointed out
in Chapter 2 concerning saliva ow in the mouth, this juice is probably
subject to (biaxial) extensional ow, which means that its effective viscosity
can be very high, so impeding tooth movement if no release valves are
available.24
Seed contents, on the other hand, do not express juice and tend to
fragment rapidly. The match of cusps and fossae is also important here,
but for different reasons. Cusps are only effective for the fracture of food
fragments when these lie inside a certain fracture zone around those cusps
(Lucas & Luke, 1984b). The size of these zones depends on the ratio of
particle size in relation to cuspal dimensions. Particles lying outside these
zones will be missed by cusps because the low coefcient of friction between them will cause slipping even with contact. Bulbous cusps will have
larger zones of action than small ones. In fact, it is probably optimal for
the cusps to be bulbous for this reason in order to avoid particle slippage.
Rose & Sullivan (1961) worked this out in detail when analysing the action
of certain types of industrial comminution machinery and, if their analysis can be transferred to the dentition, then the maximum size limit for
comminuting seed fragments with cusps may be only 0.20.4 times the
cuspal diameter. However, when seed contents are tough, then wedges are
probably needed, as seen in colobine molars (Lucas & Teaford, 1994). In
support of all this, seed contents tend to fall into the cusp or wedge sectors
of Fig. 7.1b.
Seed predation is of immense ecological and evolutionary importance. The list of seed-predating mammals is long and includes primates such as the otherwise leaf-eating sifakas (Overdorff & Strait, 1998),
228
the colobine Old World monkeys (McKey et al., 1981; Bennett, 1983;
Davies et al., 1988) and pitheciin New World monkey genera, Pithecia
and Chiropotes (Kinzey & Norconk, 1990, 1993). If the list is further extended to primates that are occasionally seed destroyers (but none the less
important for that), then certain capuchin monkey species (Cebus apella:
Terborgh, 1983), break large palm seeds while gelada baboons concentrate
on grass seeds (Dunbar, 1977; Iwamoto, 1979). Seed-eating (Jolly, 1970;
Peters, 1979, 1981, 1987) or hard fruit-eating (Walker, 1981) has consistently been suggested as a diet for members of the lineage of primates
leading to us the hominins so this behaviour has deserved a lot of
attention here.
Despite the above list, rodents are the most important seed predators
because they often move seeds (sometimes in their cheek pouches, but this
depends on seed size, a crucial factor for understanding rodent behaviour:
Theimer, 2003), prior to consuming them. Though seed destroyers, they
may forget to consume a signicant portion, thus being seed dispersers as
well (Turner, 2001). Rodents tend to feed on mechanically defended seeds.25
However, they are extremely small animals and so cannot develop the bite
forces that pigs (Curran & Leighton, 2000) and peccaries (Kiltie, 1982),
important seed destroyers in Asia and Central/South America respectively,
can generate. Their ever-growing incisors make sense in terms of abrading
woody tissues, albeit at the cost of considerable loss of tooth tissue (Lucas &
Peters, 2000).
Herbivores
Any mammal that feeds extensively on the structural, rather than reproductive, parts of plants can be called a herbivore. The category includes
bark and woody tissue feeders like (some) rodents and elephants, root
specialists such as pigs, and leaf-eaters such as the ruminant artiodactyls
(e.g. bovids the cow family) and the hindgut fermenting perissodactyls
(e.g. zebras). I will deal with plant storage organs briey in the sections on
human evolution, but even without considering these, the range of foods
in herbivore diets is very broad. The dentitions of the mammals that eat
them are correspondingly diverse, being at a maximum between 45 and
30 million years ago (Jernvall et al., 2000). To get clues that lie at the root
of this diversity, which is all that can be done here, herbivore diets need
to be classied in mechanical terms. Unfortunately, descriptions of plant
mechanical defences have often been narrowly focussed on features like
Plant feeders
229
Table 7.2 List of jaw and dental features that differentiate grazers
from browsers
Jaw and dental features
Grazers
Browsers
Height of jaw joint

Shape of incisal row
Size of incisal row
Muzzle morphology
Width of premaxilla/width of palate in molar region
Robusticity of mandible
Diastema
Hypsodonty
High
Straight
Large
Broad
High
High
Long
High
Low
Curved
Small
Narrow
Low
Low
Short
Low
Source: Mendoza et al. (2002).
spines and thorns, i.e. on analogues of the weapons that animals have, so
the literature does not always help much. The conventional way to examine
herbivorous diets has been in relation to nutritional factors like bre content
(Milton, 1979; Shipley & Spalinger, 1992; Van Soest, 1994), plant chemical
defences (Janzen, 1978; Waterman, 1984) or their combination (Waterman
& Kool, 1994). Body size has also gured prominently (Owen-Smith, 1988;
Van Soest, 1996), although interest in this factor appears to be on the wane
(Gordon & Illius, 1994). Recently, factors that inuence the rate of oral
processing of foods have been considered in relation to feeding preferences
(e.g. Janis & Ehrhardt, 1988; Perez-Barberia & Gordon, 1998a,b, 2001),
providing a link to the present discussion.
Folivores are usually separated into grazers (grass-eaters) versus browsers
(consumers of other foliage). Although this distinction ignores mixedfeeders, it is very useful for dening dietary extremes. The major predators
of grasses seem to be large mammals. Being small compared to these herbivores, grasses seem to show stress-related defences with relatively narrow
blades, parallel venation, an often-extensive pattern of silica deposition in
the tissues and little defensive chemistry (Owen-Smith et al., 1993). Browse
mostly consists of dicotyledonous angiosperms, which is predominantly
attacked by relatively small invertebrates and so exhibit displacementlimited defences: relatively wide laminae, reticulate venous network, probably rare silica deposits and an often-extensive defence chemistry. Accordingly, the prediction is that the oral morphology of grazers should
be ingestion-dominated while that of browsers should be masticationdominated. Table 7.2 is adapted from a table in Mendoza et al. (2002) and
230
indicates consistent differences between the jaws and dentitions of grazers

and browsers. It would appear that grazers generally have larger squarer
muzzles, paralleled with larger squarer front ends to their jaws containing
larger incisors than browsers. This arrangement increases the volume of food
per bite when collecting from a basically two-dimensional mat (Illius &
Gordon, 1987; Janis & Ehrhardt, 1988; Solounias et al., 1988; Solounias &
Moelleken, 1993) and accords well with ingestion dominance.26 Prehensile
lips seem more prominent in browsers, very possibly because of greater
problems in manipulating tree leaves, which often have very smooth surfaces. The coefcient of friction between such leaves and unwetted mucosa
is probably much higher than that between the teeth, which may explain
why many herbivores have developed horny pads for ingesting leaves to
replace teeth. It is unsurprising that grazers have higher-crowned (more
hypsodont) postcanines than browsers because they probably ingest more
silica (Janis, 1995).
Tooth size in herbivores
Most herbivores consume foods like leaves, twigs, barks, etc. that are shaped
like sheets and rods. Ingestion is still controlled basically by (ER )0.5 , but
these foods do not store much strain energy when fractured across their
thinnest dimension by the postcanines, so their toughness, R, is obviously crucial for understanding resistance to fragmentation. Such foods
are among the most difcult items both to ingest (Fig. 7.1a) and to chew
(Fig. 7.1b), (considering R, not (R/E )0.5 ), but the degree of woodiness is
critical to this judgement. Undoubtedly, it is because mature leaves, twigs,
etc. are so tough, that herbivores deliberately select immature stages when
available. Figure 7.6 indicates the problems that browsers face in nding
enough of this tissue. Generally, immature tissue is located only on the
ends of a branch, twig or stalk. If only a very small amount of tissue is
selected, this can have a very low toughness. However, the more proximal
the rst bite on these structures, the tougher the food. This rst bite with
the anterior teeth will tend to face a tougher obstacle than the postcanines
will face on average.
The most obvious repercussion of this patchy distribution of preferred
tissue is on scale. Many authors have commented on the need for smaller
ruminants to take immature plant parts because their digestive systems
cannot handle mature parts (e.g. Van Soest, 1996). However, the toughness
cline shown in Fig. 7.6 means that a larger herbivore, generally ingesting a
larger mouthful, will tend to eat food that is of higher average toughness.
Plant feeders
231
Fig. 7.6 A food fragment (Enterolobium cyclocarpum, Leguminosae), consumed by a New

World howler monkey (Alouatta palliata) in Costa Rica, to illustrate a general point about
the problems of scale in herbivory. The larger the mammal, the more proximal on a plant
part its bite point is likely to be, and the tougher the tissue will be.
This is exactly what both Bell (1971) and Jarman (1974) supposed: larger
herbivores are more likely to eat food of lower quality, where quality in
this sense is the inverse of toughness. The scaling arguments in Chapter 5
can be reintroduced here. Recalling Eqn. 5.6 and rearranging so that the
ratio of forces of a larger mammal eating food particle x versus a smaller
mammal eating food particle y gives
0.5
Fx
1.5 R x
=
.
Fy
Ry
(7.1)
If there were no difference in body size, then = 1, and the ratio of bite
forces would be proportional to the square root of the ratio of the toughness of their foods, but applied to differently sized herbivores, positive
allometry of postcanines results if differences in average food toughness
are great enough. Much would hinge on the selectivity of the herbivore
in question. Perhaps the least selective feeders are elephants and this argument might explain why elephant molars are so large that their mouth
can only accommodate one of these teeth at a time. They are the only
232
living mammals to adopt this strategy and it might be forced on them by

the above arguments. However, the food data are simply not available yet,
which stops this line of discussion. If grasses were tougher than browse, and
grass can be very tough (Appendix B; Fig. 7.1) (Vincent, 1991; Wright &
Illius, 1995), then this might generate larger molars in grazers. On the other
hand, grazers are often larger than browsers (although there are exceptions), which might explain their higher jaw joints and longer diastemata
in Table 7.2 (these predictions are listed in Table 5.1).
Postcanine tooth shape in herbivores
If any plant food were thin (<0.5 mm) and sufciently tough (i.e. tougher
than the dividing horizontal line in Fig. 7.1b), blades would be required
on the postcanines to comminute them. These blades though are very different from those on the molars of insectivores or carnivores, being very
low and produced by the exposure of at least two dental tissues. They
evolved from cusps with expanded marginal ridges. These cusps generally
have thin enamel (equally thin or thinner on the cusp tips than on the
anks, although some species have no enamel at all there) and are very low
in height. Thus, just a small amount of wear exposes the links between
cusps so to form what is called a loph (Jernvall et al., 2000), an enamel
crest that surrounds, often very irregularly, an enclosed pool of exposed
dentine. The surface is often further complicated by the spread of cement
onto the tooth crown and this together with variations in the hardness
of dentine combine to form a very rough surface (Fig. 6.1) that grips the
food particles. The enamel crest stands slightly proud of the other tissues
(Chapter 6) and acts as a blade against similar crests on opposing teeth,
being driven across them by an almost completely transverse jaw motion.27
The key food properties for understanding this type tooth form in herbivores appear to be their sheet/rod format, their frictional properties and
their toughness. Unfortunately, there are very few data yet available on any
of this.
A major complicating factor for friction is the presence of tannins, which
often have high concentrations in young tree leaves and unripe fruits,
and which can combine with some salivary proteins to precipitate them
(Chapter 3). The relative tannin consumption of mammals can often be
surmised from the black stain that tannins leave on their teeth, the stain
being produced presumably by the combination of tannins with salivary
proteins adsorbed onto the enamel surface. The result must be a roughened
enamel surface and increased friction.
Human evolution and diet
233
human evolution and diet

The theory developed in this book should apply to the evolution of the
dentition in any group of mammals. What limits its application mainly is
the lack of quantication of diets. I have space only for one case study to try
to show its power and that concerns human evolution. The major dietary
difference between humans and other mammals lies in food preparation
techniques and cooking which modify food properties prior to ingestion.
The classic form of explanation for human specializations is an analogy to other mammals. However, it is getting a little difcult to provide
new dietary suggestions because previous efforts have spanned the whole
herbivorecarnivore gamut. According to different authors, we might have
been specialized herbivores (Jolly, 1970), generalized omnivores (Hatley &
Kappelman, 1980), scavengers (Szalay, 1975), hunters (Ardrey, 1961) or
sh/shellsh eaters (Morgan, 1972). While some may feel that these are
old references that have been attened under the weight of new nutritional ndings, recent literature shows no sign of consensus (e.g. the edited
volume of Ungar & Teaford, 2002). It is not my intention here to nd some
strange new alternative diet: I have the same gut feeling about this as about
fads in Western societies. Instead, I am more interested in the direction that
the magnetic logic of this book turns the human dietary compass. At the
outset though, it must be recognized that there are intractable problems
with some issues, not least the antiquity of key practices, such as cooking,
which is very difcult to establish.
The face and dentition of modern humans
Both the face and teeth of modern humans (Homo sapiens sapiens) are
much smaller than those possessed by primates of similar body size, so
these structures must have been reduced in size at some point in time.
The only general model for this reduction is that of Brace (1963, 1964)
who argued that the most likely effect of any mutations on structures no
longer being maintained by selective advantage would be to reduce their
size. The problem with this general explanation is that there is no test of its
applicability: it is last ditch. I attempt an alternative here, but it is specic
rather than general and does not apply to all aspects of structural change.
The fossil record shows very clearly that some size reduction has been very
recent, while part of it dates back much further. For example, while the
length of the cheek tooth row is very short in modern humans, it is also
short in anthropoid primates compared to others (Martin, 1990). Getting
234
a general perspective on this should improve the chances of a viable overall

explanation.
Facial and dental characteristics of hominins
Five genera of hominins28 seem to have existed at various points in time:
Sahelanthropus, Ardipithecus, Australopithecus, Paranthropus and Homo. A
monospecic lineage runs from Sahelanthropus, the earliest known hominin, dated close to 76 million years ago (Brunet et al., 2002) through
Ardipithecus, living around 4.4 million years ago, putting these genera at
the ends of the same line (Wong, 2003). Species of the gracile Australopithecus ourished between 4.2 and 2.3 million years ago while those of the
robust Paranthropus lived between 2.3 and 1.4 million years ago. Both these
australopithecine genera were not direct ancestors of modern humans. The
oldest member of the genus Homo is H. rudolfensis from Koobi Fora, Kenya,
known from 2.41.8 million years ago. Other early members of the genus
were H. habilis, between 1.9 and 1.6 million years ago at Olduvai Gorge,
Tanzania and H. ergaster, from 1.7 to 1.5 million years ago at Koobi Fora,
Kenya. Homo erectus, ancestral to H. sapiens, is rst known from about
1.8 million years ago.
I will be brief here, concentrating only on major features of these forms.
The incisors of hominins are broad and spatulate, but they vary greatly
in size and in orientation in the jaws. They are large and procumbent (i.e.
stick out) in Sahelanthropus (Brunet et al., 2002), but are relatively small and
more vertically oriented in robust australopithecines. However, orientation
is not well understood and varies with wear (Ungar et al., 1997), making
incisal size a better target for explanation. The canines in male hominins are
always small, projecting only 45 degrees of gape beyond the other teeth. In
this, they are unique among the apes. The postcanine teeth of all hominins
are low-crowned. Their cusps are also low, blunt (e.g. in modern humans:
Lucas, 1982a) and thick-enamelled, with the enamel being thickest over the
cusps (Martin, 1985; Gantt & Rafter, 1998). The enamel of the molars is also
quite wrinkled in appearance, somewhat resembling that of orang-utans
and their fossil relatives, which are noted for this characteristic (Chaimanee
et al., 2003). The cusps tend to have rounded cross-sections and are very
bulbous (i.e. they increase their cross-section rapidly away from the tip).
The lingual cusps of the upper molars and the buccal cusps of the lowers
are particularly broad. The jaw joint of hominins is situated high above the
occlusal plane and jaw movements in chewing involve considerable lateral
excursions.
235
Some hominins show these traits more prominently than others. The premolars vary in their degree of molarization, by which I mean that they vary
in the proportion of the working surface of the postcanines that they occupy.
The relative sizes of the molars also vary substantially and have variable size
gradients. If the surface areas (mesiodistal length multiplied by buccolingual width) of each of the three molars, from mesial to distal, are symbolized
as M1, M2 and M3, then in some populations of modern humans, the sizes
of the lower molars can be expressed as M1>M2>M3, whereas in robust
australopithecines, M1<M2<M3 (Robinson, 1956; Keyser, 2000). The last
lower premolars of the latter have an expanded root support as compared
to contemporaneous Homo, so adapting this tooth to take greater stresses
(Wood et al., 1988). Also, the posterior elements (talonids) of the mandibular molars in robust australopithecines are noticeably enlarged in these
animals (Wood et al., 1983), something not explained by their size (Hills
et al., 1983). They also seem to have made greater side-to-side excursions
with their mandibles than some other hominins (Grine, 1981).
Potential explanations for some of these trends have already been presented in the theoretical sections. The ubiquitous presence of rounded cusps
on the postcanine teeth implies that the hominin diet basically consisted
of foods with low (R/E )0.5 . Further, even though fracture strength is a food
particle size-specic trait, low cusps also indicate low F /E. Bone ts these
descriptions (Fig. 7.1; Appendix B) and has been suggested as a hominin
specialization (Szalay, 1975), but is scarcely likely without consumption of
vertebrate soft tissues, the properties of which are the antithesis of a match
to hominin molar form (Fig. 7.1b). Seed shells could have been important,
supporting the views of Jolly (1970) and Peters (1979, 1981, 1987). The shells
shown in Fig. 7.1b are closer to those that Peters envisages for a robust australopithecine diet rather than the grass seeds that Jolly favoured. Not so
far from seed shells on Fig. 7.1b are seed contents (the part of the seed that
would be chewed in order to be swallowed) and plant storage organs: all
these foods lie well within the (R/E )0.5 range that Agrawal showed would
fracture and break down rapidly in the mouths of modern humans (Lucas
et al., 2002). Some fruit esh also resembles that of plant storage organs in
texture and would also t hominin diets.
Experiments on human chewing indicate that both jaw-closing muscle activity (Agrawal et al., 1998) and the degree of lateral movement
in chewing (Agrawal et al., 2000) increase with food (R/E )0.5 , which
is again supportive of foods with high (R/E )0.5 being important in the
diets of hominins like robust australopithecines. While bone is abrasive to
teeth, seed shells are apparently not particularly so (Peters, 1982). Storage
236
organs, particularly if these were underground (Hatley & Kappelman, 1980;

Wrangham et al., 1999) would have abrasives attached to them. Some
of this might support a case for thick enamel on the basis of abrasion
resistance.
So far, this just sketches out explanations for one food characteristic
internal mechanical factors that affect the breakage function. However,
what about the effect of external physical characteristics, such as the size
and shape of particles and ingest food volume (the mouthful)?
The number of teeth in the human mouth
It may be recalled from Chapter 5 that the size of the working surface of
the cheek teeth has a big effect on the selection function (the probability of
particle fracture). Food breakdown efciency can be predicted in humans
just by counting the number of functional cheek teeth, counting each molar
as two units and a premolar as one (Helkimo et al., 1978). Unfortunately,
compared to early mammals or even to some primates, hominins are short
on teeth, having two incisors, one canine, two premolars and three molars
on each side of the upper and lower jaws at most. This is a considerable
reduction from the number in early mammals, though this did not happen
recently: all anthropoid primates have far fewer teeth than early mammals.
Why?
In Chapter 5, I suggested that the reduction in tooth number in early
mammals was due to dwarng. It seems reasonable to expect any reduction in teeth in other mammalian lineages to be explained on similar grounds. An example of how this may operate is illustrated by
the last instance of tooth loss prior to the emergence of the hominins.
Hominins are catarrhines, a taxonomic group also containing Old World
monkeys and apes. The catarrhines have a common ancestry with
platyrrhines (New World monkeys), diverging something like 4035 million
years ago. It is hard to dene the date it might be much earlier (Tavare et al.,
2002).
New World monkeys have an extra premolar in both upper and lower
jaws compared to catarrhines. Osborn (1978) gave a highly plausible developmental explanation for this premolar loss. His view is sketched in
Fig. 7.7. He suggested that each tooth class (incisors, canines or postcanines) is developmentally distinct. As it grows, it competes for jaw space
with other classes. In Osborns view, whenever there is some (undened)
pressure on jaw space, the last teeth to form in each series will be the rst
to be lost. The upper diagram in Fig. 7.7 shows two generations of teeth
237
Fig. 7.7 The process of reduction in tooth number in the hypothetical ancestral catarrhine
primate according to Osborn (1978). In the upper diagram, the postcanine teeth of a catarrhine ancestor are pictured as a growth series with two generations. The rst generation
includes both deciduous and permanent molars, the premolars being second-generation
replacements of deciduous molars. The stem tooth is the rst tooth to develop that is
the last deciduous molar (coloured grey). The lower diagram shows the problems that a
dwarfed catarrhine would face due to fracture scaling (and, therefore, negative allometry)
of the postcanine teeth. The last teeth in the series, the most posterior premolar and molar, fail to form, thus promoting the most posterior deciduous molar, the stem tooth, to
permanence.
in a hypothetical catarrhine ancestor. The rst tooth to form is the stem

tooth of that class. The last postcanines to form are the most distal molars, premolars and permanent incisors. Under pressure for space, some of
these teeth may not form. Osborn noted that there are known statistical
associations like this. If the third molars in humans do not form, the next
238
likely teeth to fail to form are the lower posterior premolars and, following
this, the upper lateral incisors (Garn & Lewis, 1962). Osborn suggested
that the third molars and most distal premolars were the teeth that were
actually lost in the earliest catarrhine. Thus, the rst molar of any member of this group is actually a retained deciduous molar that has been promoted to permanence because its second-generation replacement no longer
forms.29
This is not the traditional view, which has the most anterior deciduous
molar and premolar being lost simultaneously. I have always supported
Osborns view as being more likely, but what trigger precipitates tooth
loss? The only one that I can suggest is dwarng. Sadly, the fossil evidence
for early anthropoid evolution is still too limited to examine this. Some
early anthropoids were very small, like Eosimias (Beard et al., 1996), which
probably weighed only about 100 g, but other early Eocene nds with
anthropoid afnities appear to be somewhat larger, as are the North African
parapithecids (Fleagle, 1999). Yet this remains a viable possibility in my view,
because it should be remembered that dwarng involves relative changes
and is not tied to any absolute size band.
A lost premolar here or here does not take tooth numbers back to those
in an ancestral mammal. The overall suggestion then is that the evolution
of mammals in general, and primates in particular, has been punctuated
by periods when, surviving in shrinking isolated locations, populations of
some species were subject to intense pressure to reduce their body size so
as to cope with smaller dietary patches. This resulted in dental crowding
that led to tooth loss. When conditions ameliorated, perhaps as barriers to
movement were removed and the climate changed, these species could
then have re-expanded slowly, both in a geographic and a somatic sense.
These species were the ancestors of the major lineages that followed them
(Fig. 7.8). On the direct line from the earliest mammaliaform to modern
humans, some 200 million years, I hazard a guess that there might have
been three or four dwarng events. Although there has been no dwarng
within hominins, we show the signs of these past events.
Molar size gradient in primates
The important point about reduction or loss of a single tooth in a class is that
it will affect the size of its neighbours because their growth is co-ordinated
(Sofaer, 1973, 1977). Tooth size gradients are usually smooth (certainly so if
Osborns denitional criteria are followed), so the penultimate tooth would
also be affected, and so on, leading to a change in the gradient of tooth size in
239
Fig. 7.8 The suggested pattern of body size evolution in some mammalian lineages. Although the direction of body size changes has certainly been slow enlargement for most of
the last 65 million years (this is called Copes law: Alroy, 1998), there may have been short
periods when body size had to decrease very rapidly. Dental crowding would then have
resulted with the loss of some of the teeth. While some of these dwarfed mammals probably died out, others could have been ancestral to later forms. Reduction in tooth number,
which the theory suggests should be abrupt and not drawn out, is the suggested evidence
for this.
a series. In extreme reduction, presumably prior to loss, the nal tooth may
not develop to its full shape potential (Osborn, 1978), as when the upper
lateral incisors in humans fail to form their typical spatulate shape (Sofaer
et al., 1971a,b). The size relations of the molars are remarkably variable in
primates, so the size gradient seen within a dental class may reect features
of the food intake too.
The second part of Chapter 5 dealt with predictions of tooth size in
response to changes in the external physical characteristics of the diet in a
manner independent of any change in the size of the rest of a mammal.
Specic dietary inuences are unlikely to affect all tooth classes in the same
way and are unlikely to result in general tooth loss just reduction or
enlargement within a tooth class. However, to keep a smooth gradient,
then just as reduction proceeds in the permanent dentition from distal to
mesial, enlargement should follow that trend but in reverse.
Table 5.4 suggested that an enlargement of the cheek teeth would follow a diet in which small, chemically sealed, non-sticky food particles
were ingested in small mouthfuls. To this, Chapter 6 added abrasiveness.
240
(b)
45
40
35
30
25
20
15
50
M3 area as % of molar area
50
%age contribution to total molar area
(a)
45
40
Key:
New World monkeys
Old World monkeys
Hominoids (apes)
35
30
25
20
15
20 25 30 35 40 45 50
M1 area as % of molar area
Fig. 7.9 The molar size gradients of 69 species of anthropoid primates (taken from Lucas
et al., 1986). The size of each tooth was calculated by summing the nominal surface areas of
each molar (maximal mesiodistal length multiplied by maximum buccolingual width), in
upper and lower tooth rows separately, and then expressing their percentage contribution to
the total molar area. Primates can be very sexually dimorphic, so data for males and females
were plotted separately (though there is no evidence of any difference between the sexes in
molar gradients). (a) The contribution of the area of the upper middle molar (M2 ) varies
very little in these primates, while that of the rst (M1 ) and third (M3 ) molars are inversely
related, each varying three times more in contribution than the second molar. (b) Data
for lower molars, plotting the percentage contributions of the rst (M1 ) and last (M3 )
permanent molars to the total molar area. These are inversely related, strongly supporting
that the molars have co-ordinated growth.
Dening food particle size in terms of particle volume, then the most
common small items in primate diets are seeds and leaves. These foods are
not only small in volume; they are indigestible unless they are opened.
Figure 7.9 shows that gradients in the size of molar teeth in anthropoid primates are obviously co-ordinated and that a large proportion of
the variation can be expressed by the ratio of the areas of the rst and third
molars (Lucas et al., 1986d). A high value of the M1/M3 ratio would mean
relatively small molars and vice versa. This ratio is strongly inversely correlated with the percentage of leaves plus seeds that have been reported in the
diets of anthropoids, calculated on a cumulative annual basis (Fig. 7.10a).
Sticky foods, like most fruit esh, quickly form food boluses, leading to
food being chewed on a very limited part of the cheek tooth row. Lucas
et al. (1985, 1986a) suggested that the optimum would be a buccolingually
wide tooth row with most of the working surface located in the middle
of the row (i.e. with a large M1, small M3). Such an association is found

Key:
New World monkeys
loge (M1 M 3 areas)
0.9
(b)
Old World monkeys

Hominoids (apes)
0.5
0.1
-0.3
-0.7
20
40
60
80 100
% leaves in diet
loge (breadth/length of M 3 )
(a)
241
-0.15
-0.20
-0.25
-0.25
-0.30
-0.35
-0.45
-0.50
20
40
60 80 100
% leaves in diet
Fig. 7.10 (a) The relationship between the percentage of leaves reported in the diet of
anthropoid primates and the logarithm of the ratio of areas of the upper rst (M1) and
third (M3) molars. There is a negative relationship that can be interpreted as suggesting
large molars (i.e. low M1/M3 ratio) are required for consuming leaves. (b) The relationship
between the shape (buccolingual breadth/mesiodistal length) of the lower third molar and
the percentage of leaves reported in the diets of Old World monkeys. This indicates selective
pressure on folivores for longer thinner molars.
in primates: the buccolingual width/mesiodistal length is positively correlated with the M1/M3 ratio. Furthermore, the area of the last premolar
contributes more to postcanine area (premolar area plus that of the summed
molar areas) when the M1/M3 ratio is high. Thus, most of the working surface is placed in the centre of the tooth row.
Leaves do not form into boluses when they are chewed because of the hydrophobic (and roughened) nature of their cuticular surface that dominates
their exposed surface throughout mastication (Chapter 3). So leaf-eating
primates will have long thin cheek tooth rows. In some Old World monkeys, the shape of the third molar alone appears to give a good prediction
of the proportion of leaves in the diet (Fig. 7.10b).30
This discussion suggests that there could have been great differences
in the relative proportions of seeds and leaves in hominin diets. Robust
australopithecines could have ingested large quantities of these items. This
supports Jolly (1970) who argued for an analogy between hominins and
geladas (where gelada really refers also to a large number of extinct species
in the genus Theropithecus: Jablonski, 1993). Some geladas have strong
resemblances to the robust australopithecines, as also does Gigantopithecus
blacki, an Asian ape (Pettifor, 2000). Virtually no mammal could live on
a continual diet of seeds, just as they could not for fruits, because they are
242
simply not around for long enough. Dunbar (1977) found that geladas ate
young grass leaves for a considerable portion of the year. Seeds and leaves
go together in the diets of leaf-eating colobine monkeys and even in some
ruminants (Bodmer, 1989).
Interaction between anterior and posterior tooth sizes
The last section emphasized the cohesiveness of response of a tooth class.
What about relationships between the sizes of these classes? These have long
been the focus of attention in hominins, quantitative study dating back to
Groves & Napier (1968). The relative sizes of the incisors and postcanines
of hominins are very variable. Sahelanthropus has large incisors, while those
of Paranthropus are very small. The molars of modern Homo sapiens look
like miniatures in relation to our body size, while those of Paranthropus are
relatively huge. One possibility for explaining this variation is by differing
seed treatments among hominins. In fact, differing seed treatments are
how sympatric mammalian species (i.e. those living at the same time in the
same location) can share fruit resources in tropical forests. This has some
particular interest for the study of hominins.
About 1.6 million years ago, in East Africa, a robust australopithecine
species, Paranthropus boisei, was sympatric with early Homo erectus, a species
that was (in most researchers eyes) ancestral to H. sapiens. Probably also
around in the same habitat was H. habilis, a species that probably died
out soon after this date. Homo erectus was signicantly larger in body mass
than the others, but its maximum bite force was low. In contrast, the
smallest hominin of the three, P. boisei, had the highest bite forces (Wood &
Collard, 1999a). That comment represents an extrapolation from skeletal
measurement, but such a reading of these measurements makes considerable
sense. How could three such similar organisms have co-existed without
dietary interference?
Doubtless, all hominins ate fruits. Early members of the genus Homo
had relatively large incisors, but small postanines. Earlier in this chapter, I
denoted this as A+P and as the dentition of a seed cleaner. In contrast,
P. boisei had massive postcanines and tiny incisors. This AP+ dentition
is typical of a seed destroyer. Just like this, it is possible for at least two
hominin species to partition the same food between themselves on the
basis of seed treatment. Early Homo species could have eaten fruit esh,
cleaned the seeds and dropped them, while the robust australopithecine
could have moved through later and consumed these cleaned seeds. Why
did robust australopithecines die out?
243
Well, later members of the genus Homo reduced the size of their incisors
as well, possessing a AP dentition, i.e. that of a seed swallower. If
H. erectus started to swallow whole seeds, then that would leave robust
australopithecines without a critical resource, unless, that is, they then
searched through . . . 31
Whatever, this section has painted robust australopithecines as seed destroyers in almost exactly the way that Peters (1987) has suggested.
Canines and premolars of hominins
All hominins have small permanent canines in both sexes, an important characteristic separating Sahelanthropus from other hominoids (Brunet
et al., 2002). The only other anthropoid primate species like this are some
New World monkeys. The titis (genus Callicebus) are very small monkeys
with very small canines (Kinzey, 1972), while the largest living New World
monkey, the woolly spider monkey, Brachyteles arachnoides (Zingeser, 1973;
Lucas et al., 1986b) is similar. Catarrhines (Old World monkeys, apes and
hominins) other than hominins have large projecting canines in males.
What precipitated their reduction in our ancestors?32 One of the commonest methods of estimating the size of the canines in studies of the dentition
of fossil hominins is to compare the size of their bases (mesiodistal length
multiplied by buccolingual width) with those of the molars (Wood, 1984).
As neighbours in a tightly packed jaw, with constraints over the position
and length of the postcanine tooth row (Greaves, 1978), such comparisons
might be inuenced by competition over available jaw space (Jungers, 1978).
I believe that this competition comes from molarized premolars.
Any tooth used in chewing needs to be placed between the tongue and
cheeks because this balance is essential in ensuring that food particles lie
on the working surface of the postcanine teeth as the jaw closes. The
regular and extended use of the premolars for mastication is indicated by
their molarization changes to their form to resemble the molars. This
increases their proportion of the postcanine working surface. These changes
are seen most commonly in the most distal premolar of primates. If the
premolars are used intensively for chewing, then it is probably necessary
for the anterior limit of the cheek, the modiolus, to be positioned further
forwards. This movement, however, jeopardizes the ability to gape widely
because the mouthslit must now be narrower.
The reason for molarizing premolars in primates is probably to do with
ingested food particle size. The premolars in the human dentition have
cusps but no fossae. These teeth normally provide preliminary breakdown
244
from large to medium particle sizes, while molars take size reduction further once food particles have been reduced to a size where the fossae can act
properly. Thus, most of the masticatory sequence is centred on the molars
(Wictorin et al., 1971). However, if there is something wrong with those
teeth, food is placed more anteriorly on the premolars (Lundberg et al.,
1974). This will result in lower chewing efciency for particles of around
5 mm or smaller and is due both to lower selection and breakage functions
(van den Braber et al., 2001). Molarizing the premolars will increase efciency with small particles, but probably to the detriment of larger ones.
Thus, I suggest that molarizing the premolars is an adaptation to the intake
of small particles.
Modern humans are distinguished by rather limited gapes compared to
other mammals, possibly managing only about 23 of condylar rotation
(Baragar & Osborn, 1984). This is very small compared to the average 60
70 gape that most mammals are capable of (Herring & Herring, 1974).
A smaller gape would not preclude werewolf-sized canines, but they could
only project about 89 before the tips cleared each other or else they would
not have been able to bite anything. The critical feature in humans is the
soft tissue at the corner of the mouth (which can tear when the jaws are
over-opened in some dental treatments: Smith, 1984).
I suggest then that the sizes of the canines are restricted by molarizing
the premolars. This shifts the modiolus forwards to allow the cheeks to
cover these teeth. That restricts the gape via a smaller mouthslit, and that
puts pressure on reducing canine size. To address this, an index of the
molarization of premolars is required. This is relatively easy: the postcanine tooth areas of the posterior premolar (conventionally termed P4)
and the three molars can be summed and the proportion of the surface represented by P4 then calculated. Lucas et al. (1986b) correlated c , the angle
at which the canines just clear, with this molarized premolar index in 41
42 species of primates. Negative correlations were found in males, but not
in females. Given that the restriction on canine size only applies to males,
and that all basic size inuences have been removed in the calculation of
these measurements, this is strong evidence for interaction between these
teeth.
Extinct hominins are among many mammalian species with molarized
premolars and, judging from modern humans, had small gapes. Accordingly, the canines had to be small. This reasoning then would have small
canines in hominins being a response to the enlargement of the premolars. This section supports Jollys (1970) general hypothesis of hominins as
small-object eaters.
245
Food preparation and cooking

There has been such substantial reduction of the face and the dentition in
the recent evolution of the genus Homo that it is tempting to call it facial
dwarng. Rapidly dwarng mammals seem generally to reduce their intake
as a reponse to shrinking food patches. This is not what I think happened
in the genus Homo though: like virtually all authors, I attribute this to the
effect of modifying food properties prior to ingestion, by food preparation,
preservation and cooking.
These practices characterize all modern human populations and change
the physical properties of foods very radically. In fact, they may have almost
taken the job of the dentition and masticatory apparatus away. Farrell
(1956) showed that many cooked foods could be swallowed by humans and
digested almost completely without any chewing.33
Raw vegetables in human diets are, rather unsurprisingly, an exception.
Raw carrot comes through the human (Farrell, 1956) and primate (Sheine,
1979) gut into the faeces in still-recognizable pieces. In fact, there is a lot
of evidence that food preparation and cooking techniques do not necessarily expose nutrients completely. For example, when peanuts are offered for consumption as whole roasted nuts, comminuted roasted meal or
peanut oil, the last-named preparation is digested more completely than
the other two (Levine & Silva, 1980). Levine and Silva interpreted this result as supporting the continued need for chewing in human populations.
However, other factors are probably in play other than food particle sizes
here. Oral stimulation inuences the efciency of fat uptake later in the gut
(Mattes, 2001; Tittelbach & Mattes, 2001) and oil and comminuted meal
will undoubtedly send a clearer fat signal to the body than intact nuts
(Chapter 3).
There is now general agreement that relative tooth size has declined for
dietary reasons during at least two time periods within the lineage leading
to H. sapiens. Early H. erectus is distinguished both from earlier members
of the lineage Homo and contemporary australopithecines by virtue of a
considerable reduction in the size of the postcanine teeth (Wood, 1992).
However, the anterior teeth do not appear to have reduced in size at this time
(Calcagno & Gibson, 1991), so giving early H. erectus an A+P dentition
(see above). Dental dimensions remained relatively stable in succeeding
H. erectus populations, until about 300 000 years ago, when the size of the
whole dentition seems to decline gradually up to the present and to an
AP dentition. Brace and co-authors have documented this decline very
precisely (Brace et al., 1987).
246
What could explain these patterns of dental reduction? The general

answer has always been declining use for two reasons: (1) the advent of
tool use and (2) cooking. There is evidence for stone tools from about
2.5 million years ago in Ethiopia (Semaw et al., 1997) and from East Africa
only slightly later (Wood & Collard, 1999b). Other types of tool use may
have developed much earlier (Panger et al., 2002), in part because some
populations of both orang-utans (van Schaik et al., 1996) and chimpanzees
manufacture tools, with demonstrated evidence for this dating back several
hundred years (Mercader et al., 2002). The rst use of re to cook is also
difcult to establish. There is clear-cut evidence from about 300 000 years
ago for the control of re and so for cooking these latter developments
seem to have been a prerequisite for population movements into temperate
regions but Wrangham et al. (1999) have recently suggested that H. erectus
was cooking in Africa from about 1.9 million years ago. While the details
may be contentious, there is an onus on this book to see if dental reduction,
indeed the reduction of the face as a whole, could be explained in terms of
food preparation and cooking.
Food preparation and pre-ingestive particle size reduction

An important aspect of virtually any food preparation technique is the
reduction in food particle size that it produces. A pestle and mortar is one
example of this and is one of the easiest of tool sets to fashion: a crude
equivalent of this is what chimpanzees use to break oil palm nuts in West
Africa (Boesch & Boesch, 1990). Even now in developed human societies,
fracture and fragmentation is an important and heavily overlooked aspect
of our behaviour. Comminution on an industrial scale is a massive drain
on energy in developed countries, accounting for perhaps 5% of the energy
generated by power stations (Lowrison, 1974). I will contend here that the
major selective pressure for tool development was not foraging or hunting,
but the reduction of food particle size.
It will be recalled from the start of this chapter that tooth size and jaw
size were proportional to each other in reptiles as the jaw grew, teeth were
replaced. In mammals, this is not so. Jaw size and the size of the anterior
teeth are proportional to food particle size because of the need to open wide
enough to ingest such particles. In contrast, the size of the posterior teeth
is set by criteria related to the fracture and fragmentation of these particles.
This does not make these structures independent of each other, but it does
give them different scaling exponents. In recent human evolution, it would
247
seem that these functions have been teased apart and a mismatch between
tooth and jaw size has resulted.
A diminution in ingested food particle size due to food preparation
decreases the need for gape. Better said, ingested food particle size should
be proportional to linear dimensions of the face.
Raw under cooked and the problem with cube roots
Cooking is more complicated and involves the scaling arguments of
Chapter 5. Suppose that it changes food toughness. Equation 5.6 can be
modied to suggest the effect that this change has on tooth size. The force
produced by a hominin chewing raw food, of toughness Rraw , will be termed
Fraw , while the force required for cooked food, of toughness Rcooked , will
be Fcooked . There is no need to invoke any difference in size of any sort,
so the size ratio becomes 1.0 and disappears as a variable. Thus, Eqn 5.6
simplies to

R cooked 0.5
Fcooked
=
.
(7.2)
Fraw
R raw
As explained in Chapter 5, the effect of change in food toughness would
be rst a demand on the bite force and the cross-sectional area of the jawclosing muscles that produce it. Tooth size is responsible for the area of
contact with food and should be proportional to that force. Thus, tooth
size (the product of mesiodistal and buccolingual dimensions) would be
proportional to the square root of the ratio of the toughness of cooked : raw
food. However, instead of tooth areas, it is probably better in a comparison
with the effect of food preparation to use linear dimensions: any tooth dimension should be proportional to the cube root of the ratio of cooked : raw
food, i.e.

R cooked 0.33
any postcanine tooth dimension
.
R raw
Now a scenario: suppose that hominins took up both food preparation
and cooking, the former reducing food particle size, the latter reducing
food toughness. Over time, the face and anterior teeth would reduce in
proportion to the degree of comminution with tools while postcanine tooth
size would reduce only as the cube root of change in toughness. I conclude
that changes in toughness with cooking would have to be colossal to match
the effect of tool use. So the face and anterior teeth would reduce in size
248
faster than the postcanines. Over time, such a mouth would be incapable
of accommodating all the postcanine teeth that tried to grow into it. This
is precisely what happens in modern human populations: there is often no
room for the third molar.
The effect of cooking on food
What, in a scientic sense, does cooking do to food? There are two general
types of cooking: in air (e.g. roasting) or in water (boiling). It is likely that
roasting preceded boiling because the latter demands a water source and
receptacles. Most primates do not actually seem to use water sources very
much, obtaining most of their water from food items, so it seems likely that
roasting is more ancient. However, I conducted some simple experiments
on both cooking methods to judge their relative effect.
Wrangham et al.s (1999) target food for H. erectus is starchy plant storage
organs. Peters et al. (1992) have made a compendious list of plants that
almost certainly were available about 2 to 1 million years ago, while Peters
& Maguire (1981) have made a specic investigation of this at a hominin
fossil site. I tested several supermarket vegetables not because these were
around at the time but because they might stand as examples of these food
types.
Potatoes were both boiled and roasted, while turnips were just boiled.
Figure 7.11 shows the results.34 After about 4 min of boiling (shown as
Fig. 7.11 The effects of boiling and roasting potato tubers on their toughness. Toughness
reduces with boiling beyond about 4 min to levels lower than with roasting. For method,
see note 34 to Chapter 7 (p. 304).
249
Fig. 7.12 Reduction in toughness on boiling of white onion (left), but not white turnip
(right). Both were tested with a 15 wedge.
Fig. 7.13 The development of a curvilinear J-shaped stressstrain curve in potato esh with
boiling. For method, read note 34 to Chapter 7 (p. 304).
circles), the toughness of the potato parenchyma started to decay exponentially. The state after about 10 min cooking represents the usual cooked
texture its toughness was 16% of the raw value. Roasted potatoes were
tougher than boiled specimens and had a casing that still had 50% of
the raw toughness. The toughness of the internal tissue of roasted pieces
was close to double that resulting from boiling. This may not be generalizable. Results of similar boiling experiments on white turnip, with white
onion and Chinese leaf vegetables (choy sim) show that turnip and leaves
remain tough after boiling, while others, like onion, follow the potato example (Fig. 7.12). However, there is a second element to this: the shape
of the stressstrain curve with cooking. On this, cooking has profound
and general effects. Figure 7.13 show that raw potato has an essentially
linear curve, which increased cooking times converts to a J-shape. The explanation is that cells lose their turgidity when they are cooked and their
framework gradually collapses during compression until cell walls are being
250
pressed against each other, when the tissue rapidly becomes much stiffer.
Cracks then pass around, rather than through, cells (Lillford, 2000). This
appears true of all the vegetable foods that I have tested though I hesitate
in claiming that it will be universal.
Cooking also has profound effects on animal soft tissues, but they cannot
be quantied as easily because the tissues are much more complex and the
effects of cooking more difcult to quantify. A tough piece of meat is one
that has a lot of collagenous connective tissue in it. Most of the fracture is at
the level of bundles of bres in connective tissue called perimysium, which
has been the focus of a lot of studies on meat toughness (Purslow, 1985,
1991b). This can be surprisingly tough, even in cooked meat (0.41.8 kJ m2 :
Purslow, 1985), but probably less so than coagulated muscle protein. One
way to estimate the toughness of meat would be to work out how much
perimysium there is in it. It is much easier though just to calculate total
collagen content. Chapter 5 gave data indicating that tetrapod muscle has
more collagen than sh, irrespective of real differences in collagen make-up
between the two vertebrate groups.
What cooking generally seems to achieve is a thwarting of the elastic
crack blunting mechanism (discussed in Chapter 4) by stiffening tissue up.
Uncooked skeletal muscle has a J-shaped stressstrain curve when pulled
along its bres (Fig. 7.14ad). Yamada (1970) and colleagues tested an
enormous range of raw mammalian tissues and found this type of deformational response to be by far the most common. However, after cooking,
muscle stiffens and stress is approximately proportional to strain (Fig. 7.14e)
(Purslow, 1991b). This will tend to connect up an otherwise purposefully
disconnected tissue (if this makes sense) and prevent much crack blunting. Cooking temperature is important because at temperatures where collagen starts to break down into a gel, myobrillar proteins (those responsible for muscle contraction) start to coagulate. Overall, the toughness of
any piece of meat cooked at high temperature seems to depend more on
the state of these muscle proteins than on collagen (Christensen et al.,
2000).
The effect of cooking on evolution of teeth
The basic shape of the human dentition has remained surprisingly stable
over the last few million years despite fundamental alteration in the food
supply. In contrast, tooth size has varied and varied dramatically. Simple
cooking experiments suggest that cooking has a substantial inuence on
food properties and probably, therefore, on tooth dimensions. However,
251
Fig. 7.14 The reduction in elastic crack blunting (see Chapter 4) in muscle tissue after
roasting. Contrary to the trend in plant foods, cooking turns a J-shaped stressstrain curve
into a linear one. Notch insensitivity is partially retained (Purslow, 1985).
it also looks clear that food comminution capacities of tools would far
outstrip cooking in morphological effect. Tooth crowding resulting from
the difference between these two effects does look like a reasonable way
of explaining dental crowding in modern humans. What changes in tooth
size could be anticipated though just from cooking?
With regard to the early cooking theory of Wrangham et al. (1999) and
the consumption of underground storage organs, the experiments support specic and general effects on dental dimensions. Table 7.3 indicates
252
Table 7.3 The effect of cooking on potato tissue and the suggested reduction in
tooth size of a hominin living on this food type
Cooked sample
Roast casing of potato
Inside of roast potato
Boiled (10 min)
a
Mean (s.d.) toughness and

range of cooked specimens
(J m2 )
Expected reduction in tooth

dimensionsa
(percentage of ancestral tooth size)
122.7 (21.7)
58.8 (13.3)
37.9 (4.4)
82%
64%
56%
The average toughness of the raw potato was 225 J m2 , so expected tooth size is
calculated as (mean cooked toughness/225)0.33 100.
predictions for postcanine tooth size for a cooked potato diet. The suggested
tooth size reduction is signicant. It is tempting to suggest that roasting
could explain the pattern of postcanine dental reduction in early H. erectus.
The anterior teeth would probably process the tougher roasted casing while
the posterior teeth would deal with both, with the greater volume of the
inner tissue predominating during mastication. Just on this basis it could
be predicted from Table 7.3 that anterior tooth size would not reduce as fast
as that of the posterior teeth. Boiling the potato would have been dentally
devastating and, if this were continued into soup, then the need for the
teeth would be removed completely. The nding that cooking does not
always reduce the toughness of vegetables makes investigations potentially
more valuable: it may be possible to pinpoint those foods on which cooking
has the greatest effect.
The changes in the stressstrain curve are also relevant. These can be
modelled, very approximately, by the non-linear elasticity scaling deriving
from Mai & Atkins (1975). Although the response is not actually purely
elastic, provided that the load is maintained to fracture (and not cycled until
eventual failure), it can be modelled in this way. Under such a monotonic
loading (Atkins, 1999), sensory receptors could not judge whether part of
the response is viscoelastic, plastic or elastic. Applying non-linear theory
from Chapter 5 to the potato curve for 10-min boiling (Fig. 7.13) yields a
stressstrain exponent of n 1.7. This would drop the body-mass exponent
of tooth size, say, to M 0.46 from M 0.5 for a diet of raw tissue when n = 1.
Modest tooth size reduction would result.
What are the consequences of cooking meat for hominins? It will be
recalled from Chapter 4 that there is a large discrepancy between the toughness of skin (or muscle) when there is a free-running crack versus the narrow
Language and the senses
253
radius of curvature of a crack tip kept conned by the presence of a sharp

blade. A free-running crack blunts, causing toughness to soar. Cooking
abolishes this and that probably allows humans with round-cusped molars
to consume meat. There is still the issue of getting the skin off, both before
cooking and after, because it can still be very tough due to its collagen content. I concur with many previous authors that the development of sharp
stone tools was very likely to have been associated with the skinning of prey
animals.
Summary of hominin diets
I do not intend to use a book account to express rm convictions on hominin diets because this was simply to indicate the depth that must be
involved in substantiating claims about the dentaldietary adaptations of
a mammal. A vexing issue is undoubtedly a date for the advent of cooking.
Wrangham et al. (1999) contend that this was early and I believe too that
this would be the only way to understand the decline of posterior tooth dimensions in early Homo. However, are starchy underground storage organs
unavailable for the human gut without cooking? Modern humans have
salivary amylase, but so too do most non-carnivorous mammals (Junqueira
et al., 1973). This suggests that some starch is available to mammals without
starch granules having to be modied by heat. The most likely source of
available amylose is in leaves because photosynthates are only stored there
briey before being converted to sucrose and translocated. Young leaves
that are just on the verge of being productive for a plant at the point in
the tropics where yellow or reddish young foliage starts to turn green may
have been very important in the diet of early hominins, as one of the above
sections indeed suggests. The roasting of storage organs must surely have
preceded boiling. Remembering that primates usually do not drink, boiling
may have followed a newly developed thirst on hot savannas. However, it
is also possible that it was a much later development associated with life
close to freshwater sources. The addition of sh to the diet, cooked or not,
would probably have had a dramatic effect on tooth size.
a word or two about language and the senses
The senses communicate to the body about the states of its external and
internal environment. Language, both non-verbal and verbal, provides humans with the chance to convey some of this information to other members of their social group. Sometimes, there may have been an important
254
need to share information about food, but other sensations have probably
always been personal. The latter may stay below the level of consciousness and terms may never have been coined for them. The following section is intended as a short nal discussion of language in relation to food
texture.
The evolution of texture perception
Identication of insect prey, coupled with careful evaluation and manipulation in the mouth, must have been critical to the success of early mammals,
much more so than to a reptile swallowing food particles whole. While
the sense of smell and vision are most important to diurnal reptiles, the
senses of smell, taste and texture, the latter two most likely rst developed
strongly in early mammals, were probably vital to these nocturnal insectivores. Taste release, e.g. of salts, free amino and fatty acids, would make
certain that the cuticle of prey had been penetrated (Lumsden & Osborn,
1977), while texture perception would be needed to enhance food manipulation in the mouth and to assess cuticular properties. The oral processing
of an adult insect might be a risky process because the cuticle of the jaws
(Hillerton et al., 1982) and legs can be considerably harder than the thorax
and abdomen, calling for precise control of jaw movements. The hardness
values reported by Schoeld et al. (2002) for the jaws of leaf-cutter ants
are probably high enough to cause the microwear seen by Strait (1993c) on
modern insectivores. Thus, there should have been considerable selective
pressure for an expanded sense of food texture over 200 million years ago.
Virtually all mammals seem likely to get important cues about food
from texture inside the mouth. For example, herbivorous mammals are
known to avoid foods with high bre content ( high Vc ). Yet, plant cell
wall is essentially colourless, tasteless and odourless, so how can herbivores
learn from their senses about it? The answer appears obviously to lie in
its texture. They probably avoid toughness, not bre (Choong et al., 1992;
Hill & Lucas, 1996). Seeds reveal themselves inside fruit esh in the mouth
by virtue of their extra hardness (Corlett & Lucas, 1990). This tactiledependent distinction might be necessary from the plant side of things
because the seeds of eshy fruits will survive gut passage, sometimes with
enhanced germination potential, if they are undamaged by teeth. From the
mammals side of the coin, the seeds might be toxic and important not to
damage. So many frugivores may avoid hardness. They may prefer or be
averse to the astringency (friction) that tannins produce and to the lipids
present in some foods, such as insects or certain fruit.
Aftermath
255
The origin of language in humans

The development of spoken language seems to me to be bound up with
texture, but not with texture in the mouth or even directly with food: I
mean textures associated with tool-making. I envisage that language originally developed in order to transfer skills that were acquired in both the
making and the using of tools to the next generation. Language was the
method by which the handfeel needed in making or using a tool could be
communicated. The result was the development of a texture terminology.35
Jeannorod (1997) has pointed out that the ability to make tools depends
just as much on visual as tactile information, so that an extensive language for both senses developed, generally linked to tool-making. Some
auditory terms may have been useful, but this sense, plus olfaction and
taste, have little value for this activity. Accordingly, the vocabulary of these
senses is limited. In the case of olfaction, it is almost entirely absent, with
virtually all communication requiring that partners have experienced the
same smells, connotation rather than denotation predominating. Hearing,
smelling and tasting were, and still are, probably of paramount importance
in mammals that forage nocturnally, while vision and touch are the senses
of greatest value to diurnal primates. Forelimb touch is heightened in most
higher primates, probably due to use of their thumb and ngers in sensing
the ripeness of fruit by palpation: fruit colour alone is an insufcient cue
(N. J. Dominy et al., unpubl. data). Yet the point here is that the size of
the vocabulary of a sense is not related to its acuity, but to the need to
communicate it. It is doubtful if there was ever a need to communicate
mouthfeel, for example. Thus, Lucas et al. (2002) have argued that the
texture vocabulary that is often assumed to refer to mouthfeel is actually
a hand feel terminology that developed with tool use. This could explain
why the training of human subjects in taste panels in the food industry to
make ratings of various aspects of food texture has proved so difcult: it
has been borrowed from a manual context.
aftermath
Freed from references, my writing style can attempt a comeback.
Whirr, gurgle, gaseous snort.
Well this is what happens soon after food has been swallowed. Thank
goodness that the mouth does not sound like this or it would have put me
off studying it ages ago. It is difcult to see why coffee machines have to
sound like it either, unless they wish to portray the noises that the ow
256
of our digestive system is going to produce minutes after imbibing their

products. The bowels just go on and on, very unlike a short snappy chew.
However, ultimately, an account of the action of the mouth, at sensory,
nutritive and general evolutionary levels, will only make sense in relation
to those food-sinking feelings that these noises reect.
My account of how teeth work was intended to be brief, but perhaps
chewing research is rather like mammalian feeding the getting ready for it
(foraging for ideas) and the aftermath (digesting the results) take far longer
than the doing of it. Even on the human side, the start (food preparation,
cooking, etc.) and the end (washing up and/or culinary organization) generally take far longer than the meal. These considerations alone suggest that
oral processing is extremely efcient. So why could I not match this brevity?
I cannot answer this except to say that the books length is perhaps justied
by the apparent fact that there are really no other books that treat this
subject in a comprehensive manner despite its importance. Some may feel
offended, but I contend that this is probably so. Contrast this with locomotion, which has been much more thoroughly studied, and on which there
are many good accounts. Locomotion and mastication are the two major
motor activities of mammals. Both are distinctive, but . . . mastication is
unique.
appendix a
Mechanical properties and their measurement:

material properties made easy
INTRODUCTION
Many biologists dont want to know much about the property testing of
materials because it sounds extraneous to many problems and something
that could be handled in a routine manner by qualied technicians. Why be
bothered? While it is certainly true that some biomechanical investigations
do not require any detailed knowledge of this, the material properties of
foods lie at the heart of the analysis of dental function. Accordingly, I suggest
that the reader at least skim this appendix before tackling Chapters 4 to 6.
The aim is to make material properties simple to grasp and fun to contemplate. I include basic concepts, some examples and a reference list with
more for those that need it.
THE BASICS
When a force loads a solid object, it distorts. In mechanics, this distortion

is called deformation. If the object deforms sufciently, then it cracks (fractures). To document these behaviours, mechanical tests need to record the
force applied to the object and the deformation that this produces. There
are two options: controlling the force to see how the deformation varies
or vice versa. Prior to fracture, there is little difference what is chosen, but
to understand what happens afterwards, it is best to control deformation
because this controls the rate of crack growth much more effectively.
Information about the behaviour of the deformed object can be recorded
conveniently on a forcedisplacement graph, such as that shown in Fig. A.1.
The displacement is generally a measurement of deformation in the direction of the force. The forcedisplacement curve of Fig. A.1 can be divided
into two major parts: a deformational domain and a fracture domain. A
force peak at the junction between the two indicates the instant of crack
initiation. I now try to show by trial and error why mechanical properties
257
258
Appendix A
Fig. A.1 Schematic representation of forcedisplacement curves for geometrically scaled

specimens of different sizes, all made from one material. The specimens are being pulled.
The shape of the specimens is typical of those made for tensile tests, being designed such
that they will not fail around the grips. It can be very difcult to shape biological tissues
like this, which is one big reason for the development of several of the tests described in this
appendix.
have the form that they do. Remember though that all these properties are
extrapolated from simple graphs such as Fig. A.1.
the deformational domain
Imagine pulling tests on a group of regular specimens labelled A, B and C,

all made of the same material and scaled replicas of each other (Fig. A.1).
Pulling produces a uniform uniaxial (unidirectional) tension in each specimen. In fact, a pulling test is usually called a tensile test. The test results
are the curves shown in Fig. A.1. The slopes of the initial portions of each
curve are linear and dene what is called the stiffness (force/displacement)
of these specimens. Clearly though, these slopes do not describe the stiffness of the material that all these specimens are made from because each
specimen has its own characteristic slope. The obvious reason for this is
their size difference. If a way could be found to account for specimen size,
then it might be possible to make the curves conform.
Mechanical properties and their measurement
259
Fig. A.2 Force has now been divided by the cross-sectional area over which it acts for
the three specimens in Fig. A.1. This size compensation does not make the curves on the
graph coalesce in a way that could be used to characterize the material from which they
are made.
If, instead of mechanics, we were dealing with chemistry, we would

understand that the characteristics of a chemical reaction depend not just on
the quantity of molecules that are available, but also on their concentration.
Analogous to a chemical reaction, it could be hypothesized that the extent of
deformation (i.e. the mechanical reaction) produced by the applied force on
A, B and C was caused by differing concentrations of the force in specimens
of different sizes. The force is applied uniformly to the specimens and is thus
supported by their cross-sectional areas (Fig. A.2). The force divided by the
cross-sectional area over which it acts is actually a basic unit in mechanics
called stress ( ). Provided that the specimen has not been stretched too far
(say less than 25% of its length), then using the original cross-sectional area
for this estimate is accurate enough to obtain the stress.
However, plotting stress against deformation, as in Fig. A.2, still does
not bring the loading curves for specimens A, B and C into congruence. In fact, it appears to overcompensate with the stiffest specimen
(stress/displacement) now having the lowest slope. So, the displacement
also needs to be scaled to specimen size. The simplest method of making
this correction is to divide the displacement by specimen length. It turns
out that this is again correct if the specimen has not been stretched too
far: the deformation divided by the original specimen length is the strain,
a dimensionless ratio. The symbol for strain is .
260
Appendix A
Fig. A.3 (a) The stress (force/cross-sectional area) is now plotted against displacement divided by original specimen length, which is the strain. The result is a common gradient at
low stresses, which can be measured to give an estimate of Youngs modulus. The crosses
represent the point at which each of the specimens breaks. Note that these stresses are different. (b) If a wide enough range of specimen sizes of a material could be tested, then this
graph would be the likely result, whereby the smaller the specimen, the higher the stress at
which it fractures.
Stressstrain curves for the specimens are shown in Fig. A.3a. The initial
slopes for all three specimens have converged, this common slope providing
a measure of the deformation resistance of the material from which these
specimens are made. If a test is stopped while deformation remains in
the region of this common slope, then the original dimensions of the
specimen are generally regained, indicating that this is an elastic property.
The ratio of stress to strain in this region is called either Youngs modulus
(after its discoverer) or the modulus of elasticity. The symbol for Youngs
modulus is E.
The crosses in Fig. A.3a mark the point of fracture (crack initiation) for
each of the specimens. These stresses are the fracture strengths ( F ). The
graph shows that these stresses are not the same for all specimens, being
slightly larger for the smallest specimen, which is thus a little stronger.
Choosing a much wider range of specimen sizes reveals the truth of this
fracture strength is not a true material property. Figure A.3b shows the fracture strengths of specimens A, B and C marked on a curve that also gives
the loci for specimens of other sizes if these were tested in the same way:
the smaller the specimen, the higher is F . Below some limiting specimen
size, specimens no longer crack at failure but, instead, show more extensive deformation that is not recovered on unloading, i.e. the deformation
is permanent. This permanent change is called plastic deformation. (This
phenomenon can also be called ductility a term not used in this book.)
261
Plastic deformation can be induced in any material, not just by changing

specimen size, but also by conning stresses to a very small part of specimen volume such as is done when making indentations (Fig. A.9). The
point of onset of plastic deformation, usually dened as a small percentage
decline in the slope of the stressstrain slope, is termed the yield strength,
y , and is independent of specimen size (as indicated in Fig. A.3a).
In a pulling test, fracture results in the specimen breaking into two. If the
two fragments can be retted so as to resemble the original specimen, then
this is elastic fracture. If the pieces are substantially longer and narrower
than the original due to plastic deformation, then this is plastic fracture.
Figure A.3a shows that the extent of plastic deformation is far greater in
specimen A than specimen C. If the amount of plastic ow is relatively
minor, as it usually is in biological materials, then this is called elastoplastic
fracture.
Thus far, we have described the deformational domain as characterized by
the Youngs modulus, the transition from elastic (recoverable) deformation
to permanent set by the yield strength and the point of crack initiation by
the fracture stress.
the fracture domain
A pulling test is not a good way of observing fracture because the crack
tends to rip through the specimen very rapidly. A three-point bending
test is better because it produces tension only on the lower side of the
specimen (Fig. A.4). The upper part of the specimen is compressed, so
controlling the cracks growth. (Chapter 4 points out the stresses that can
promote cracking. A compressive stress can never do this.) The specimen
has been deeply notched (pre-cracked), with the notch tip made as sharp as
possible.
Loading the specimen beyond the point where the crack extends from
the notch gives the forcedisplacement curve marked ABC in Fig. A.4.
The peak at B represents the point of crack growth. The force drops after
this, but does not reach zero i.e. crack growth is controlled. In fact,
the test can be stopped now and the displacement put into reverse, so
unloading the specimen and sending the forcedisplacement curve back
to zero in Fig. A.4. The sequence of events is marked out by ABCA. The
growth in the crack after this rst experiment can be measured and the
specimens reloaded. Ideally, the loadingunloading curve will then follow
the curve ACDA. After unloading, we can again measure the crack growth
that took place during this second loadingunloading episode. If we express
262
Appendix A
Fig. A.4 A notched specimen is being bent such that a crack extends from the sharp notch.
The specimen can be loaded and unloaded repeatedly and the new length of crack measured.
The graph shows two work areas, 1 and 2, that correspond to the increments of crack growth
shown in the specimen inset.
crack growth as crack area, i.e. as an increase in crack length multiplied

by specimen width, it will be found that the graphical areas ABCA and
ACDA are in proportion to the crack areas produced in each episode. Now
these graphical areas are the product of force and displacement, which
is work. These areas ABCA and ACDA are therefore work areas and the
work done in extending the crack is proportional to the new crack surface
area produced. The material property that controls the fracture domain is
called toughness, dened as the work done in producing unit area of crack.
Toughness is symbolized in this text as R. Cracking is therefore an energetic
process and the energy available in a cracked object controls the growth of
a crack or cracks within it.
The basis of the toughness concept is easily explained. Imagine the material being made up of a lattice of atoms joined by bonds (Fig. A.5). The
position of atom x lying within the material is balanced by the attraction of
the atoms lying around it. Now suppose a crack (produced in the gure by
a ctive blade of atomic width) ran through this material, breaking bonds
on one side of x. This disturbs the equilibrium so that x moves towards
the atoms to which it remains connected. Such movement does work and
thus, the production of new surfaces by cracking requires work. Physical
chemists refer to the cost of this surface production as the surface energy
of a solid. It is identical to the surface tension of a liquid, both these quantities having the same units. Toughness has this chemical surface energy as
its basis, but while surface production is its root cost, it also includes the
263
(b)
(a)
Crack
Surface
Fig. A.5 A solid consisting of a lattice of atoms connected by chemical bonds. The position
of atom x is being disrupted by a ctive knife blade of atomic dimensions that is severing
bonds on one side of this atom, so resulting in a new equilibrium position.
cost of any physical disruption around the crack in its estimation. While
Youngs modulus and strength were properties fully appreciated by the midnineteenth century, the concept of toughness was only formulated about
80 years ago. It remained unappreciated for some time after that and remains so in many areas of biology. This is ironic in that its energetic basis
makes it easy to integrate into issues central to biological theory, e.g. in
ecology, which deal with energetic concepts much more easily than with
stresses and strains. The next sections include formal denitions and more
detail.
YOUNGS MODULUS
Youngs (elastic) modulus is a measurement of the stiffness or rigidity of

materials. It is measured as the force producing unit deformation of a
specimen, normalized to the relevant dimensions of that specimen. These
dimensional corrections convert (1) force to stress by dividing by the crosssectional area of a specimen over which the force acts and (2) deformation
to strain by dividing by the original dimension of the specimen in the
direction of the force. More simply, the Youngs modulus is the initial slope
of a stressstrain graph. Strain has no dimensions and so the units of a
modulus are force per unit area. The SI units are meganewtons per metre
squared (MN m2 ) or giganewtons per metre squared (GN m2 ) where
264
Appendix A
M stands for mega (106 ) and G for giga (109 ). To improve presentation,
the pascal (Pa), which equals 1 N m2 , has been introduced. This is a very
small unit and MPa and GPa are used here to describe the modulus.
highly extensible solids
When a solid is very extensible, it is no longer sufcient to use these denitions of stress and strain. Instead, account has to be taken of the progressive
change of deformation as the material is distorted. It is simple to do this for
strain where it can be easily shown (e.g. Ashby & Jones, 1996) that it should
be calculated, not as , but as loge (1 + ), which is called the true strain.
It is only necessary to calculate it when the change in dimensions is very
great. The stress should be based on the instantaneous cross-sectional area
of the specimen, not the original area. It is difcult to do this except when
the volume of the material is conserved during loading when the stress can
be calculated as Fl/Al0 where l0 is the original specimen length and l is its
instantaneous length.
viscoelasticity
Elastic behaviour is, strictly speaking, an instantaneous response. However,

few biological solids are like this, usually taking time to recover their dimensions on unloading. The time lag is due mainly to their water content.
This type of time-dependent elasticity is called viscoelasticity, the analysis
of which is complex. If a viscoelastic solid is repeatedly loaded and unloaded without fracture, the result is considerable energy loss, which helps
to prevent fracture.
There is little about viscoelasticity in this book or about the inuence of
deformation rate on the properties of materials. It is a highly complex issue
on its own, without even considering fracture. There are undoubtedly areas
in biology where it is extremely important, such as when there are repetitive loadingunloading cycles (e.g. the regime that the periodontal ligament
endures), but under a fairly rapid masticatory movement, resulting in the
fracture of food particles during that same chew, then viscoelastic theory,
which specializes in the understanding of equilibria in such solids, adds inappropriate complications. Nevertheless, due to time-dependent behaviour,
the values of Youngs modulus obtained in tests can vary considerably with
change in deformation rate. It is important to measure stressstrain behaviour at an appropriate rate. For accounts of very time-dependent foods,
read Shama & Sherman (1973) and van Vliet (2002).
265
non-linear elasticity
Many biological tissues do not have linear stressstrain curves. Most raw
animal soft tissues are like this, as are many cooked plant tissues and biological gels. One way to analyse such tissues is to perform a linear regression
on plots of log stress on log strain such that = n (as in Chapters 4
and 5), allowing the exponent n to characterize the relationship. However,
many biological tissues with non-linear behaviour cannot be linearized in
this way. Otherwise, depending on the purpose of the investigation, either
the slope of the initial part of the stressstrain curve or that close to fracture
could be used to dene a modulus.
Most raw plant materials have relatively linear stressstrain curves. However, accidity and cooking makes them non-linear. A J-shaped concave
curve is seen in most vertebrate soft tissue (e.g. Fig. 4.13), boiled vegetables and gels, while a r-shaped curve tends to be found in cooked muscle
(Purslow, 1991b) and cheeses (Charalambides et al., 1995).
anisotropy
Many biological solids have complex structures whose mechanical response depends on the direction in which they are loaded. This directiondependent behaviour is called anisotropy, the opposite of isotropy. The more
complex the food behaviour, the more tests are needed to characterize it.
STRENGTH
The strength of a solid is the characteristic stress at which there is a dened

change in mechanical behaviour. It either has units of kPa, where k stands
for kilo (103 ), or MPa. There are two basic types of strength. The fracture
stress is the stress at which a crack initiates while the yield strength is the
transition from elastic to plastic behaviour in some part of the specimen.
the direction of stress
Stress is a vector quantity. Since any loading pattern can produce tension
and compression within a solid, signs have to be given to indicate their
differing directions. By convention, tensile stresses are positive, while those
of compression are negative. When a solid is suspended in a uid, it is
subject to compressive stress from all sides. This is a state of hydrostatic
pressure. The opposite, overall expansion, is not relevant to tooth function,
but the overriding concept is hydrostatic stress, negative if it is compressive,
positive if expansive.
266
Appendix A
TOUGHNESS
Toughness is a measure of the resistance to crack growth in a material. The

toughness of foods is critical for analyses of tooth shape. It is measured as
the work done in making new area of crack (i.e. the surface of one of the
two crack faces) and its units are joules per metre squared, i.e. J m2 , or
kJ m2 . It is an analogous quantity to the surface tension of liquids or the
surface energy of solids. However, toughness differs in its calculation in
that:
(1) The energy expended in the formation of new surface by cracking is
divided by only one of the new surfaces formed. In contrast, in surface
energy measurements in physical chemistry, the divisor is the area of
both surfaces.
(2) It includes the cost of disruption of material in the vicinity of the
crack, away from the cracks surfaces. Due to this, toughness values are
nearly always much higher than surface energy unless the crack surface
is glassy smooth. Fracture is only fully understood in such relatively
homogeneous solids and even these are complicated enough. The understanding of many fracture phenomena, such as when growing cracks
change direction or when they branch, is still limited.
Toughness is, at minimum, surface energy, but can be ten, one hundred, or
even thousands of times greater than this because of the spread of damage
around the crack. The single biggest contributor to toughening is plastic
deformation.
symbols for toughness
As soon as the fracture literature is entered, letters tumble out at you like
some kind of kindergarten test. Generally, there are Gs, Js, Ks and Rs to
juggle with, but there are also books that could hit you with a T or even a W
(Lawn, 1993). How to manoeuvre through this alphabetic nightmare? Well,
the preface of Atkins & Mai (1985) indicates how a lot of these terms can be
related to each other and their book describes at length the circumstances
by which their separable denitions evolved. The symbol G, for example,
has usually been restricted to solids with a linear elastic response, while
the quantity embodied in J refers strictly to non-linear elastic situations.
The term R is more loosely dened as the energy involved in crack resistance. In order not to lose readers, I will stick to R here. However, most of
the other letters refer to quantities that have the same units as R, differing
267
usually in the way that they are describing energy dissipation within a
awed, notched or cracked material under load. The mechanical energy
stored in a material that can help pay for crack growth is called the elastic
strain energy. However, there are many kinds of plastic or non-recoverable
processes, even in apparently simple brittle ceramics, into which energy
can be sunk. Some of these processes obstruct crack growth and thus raise
toughness. By and large, research on fracture represents the attempt to factor out that structural disruption which is intrinsic to crack growth from
that which is non-essential. Much of the problem arises from the fact that
the energy expended in crack growth is usually measured indirectly. While
an overriding denition of the essential work of fracture may eventually
emerge, there appears to be no across-materials consensus on this yet.
Sometimes, fracture symbols have modifying subscripts, such as Kic or
Gic . The c stands for critical value, the point at which a crack starts or starts
moving. The i, more rarely ii or iii, refers to the modes of fracture. (Modes
of fracture, types of failure and mechanisms of toughness are discussed in
Chapter 4.) The association of energy with direction can seem strange
because energy is not a vector quantity. However, Kic , which nearly always
has these subscripts, is very different as the following illustrates.
the meaning of kic
The effect of a thin aw on the average tensile stress, , in a large rectangular plate of a short sharp crack, oriented at right angles to uniform
tension, is shown in Fig. 4.11 where the aw is modelled as a thin ellipse.
This aw modies stress levels in its vicinity very greatly. At the crack
tip, the tensile stress at right angles to the long axis of the ellipse is much
higher than , but it declines to well in front of the sharp edge of
the aw. Nevertheless, this aw appears to weaken the plate so badly that
if fracture stress were the criterion for fracture, the plate would break into
two pieces immediately. However, experiments show that the aw will not
grow unless there is sufcient energy stored within the plate for this to
happen. This is given approximately by
2 a
,
E
which relates the energy per unit area used up by a crack that grows from
a aw of length a to the loss in stored elastic strain energy that funded this
growth. The equation can be arranged to give
R=
(ER)0.5 = (a )0.5 .
268
Appendix A
Note that the weakening effect of the aw depends on the square root
of its length. Kic , sometimes called the critical stress intensity factor and
sometimes fracture toughness, is simply symbolic shorthand for ()0.5 .
Its validity is limited to linear elastic situations where Kic 2 ER. Note that
K can be measured at any stress because crack propagation does not gure
in its denition. However, the critical value of K, i.e. Kc , is that when a
crack propagates from the aw in the plate. Kc will always be referred to
here as Kic because crack growth is usually in mode I (Lawn, 1993).
Many biological papers measure Kic , but it is becoming less common
to use this nomenclature because its denition excludes any plastic deformation and is difcult to employ on composite materials such as most
biological tissues. However, Purslow (1991a) has suggested a form suitable
for use on non-linear tissues whereby Kic can be replaced by ic with
IC F a n/(n+1) .
The term n is the exponent in power law non-linear elastic equations.
OTHER QUANTITIES
poissons ratio
Stress
When most materials are pulled, they get narrower; when compressed, they
get wider. Figure A.6 shows a specimen of original specimen length, l 0 , and
l0
l
w0
w
Fig. A.6 A particle is compressed, so reducing its vertical (longitudinal) dimension. However, as it does so, it spreads horizontally (laterally). The ratio of the lateral to the longitudinal
strain of a material is referred to as Poissons ratio.
269
width, w 0 being compressed to a new length, l, and expanded width, w,

such that Poissons ratio
v=
log e (w/w 0 )
.
log e (l /l 0 )
The value of log e (w/w 0 ) is called the lateral strain, lateral , and is positive
because it is an expansion, while log e (l /l 0 ) is called the longitudinal strain,
longitudinal , and is negative, being a contraction. However, normal Poissons
ratios are invariably converted to positive numbers by a preceding negative
sign, i.e.
lateral
.
v=
longitudinal
to make normal deformational responses positive. For many engineering
materials, Poissons ratio is about 0.3, reecting the fact that, when true
solids are loaded, they reduce in volume. Saturated biological tissues sometimes preserve their volume when they are deformed. If they are isotropic,
then v = 0.5. The tongue is supposed to be an example (Kier & Smith,
1985). A very high Poissons ratio of 1.0 or more can be measured in some
animal tissues while in plants, zero Poissons ratios are possible. In fact, they
might sometimes be lower than zero. When reported, a negative Poissons
ratio is not necessarily an error: it results from a type of cellular structure
whereby cell walls collapse inwards on compressive loading (see Fig. 4.12).
Some materials (auxetic materials) have been designed specically for such
negative ratios (Lakes, 1987).
hardness
Hardness is not a property in itself but a concept derived, like many other
scientic terms, from specializing a word used in everyday language. Used
loosely, it means resistance to deforming under indentation (Fig. A.9).
Indentation tests are the most ancient and simple of mechanical tests,
wherein a blunt or sharp indenter is pressed into the at surface of a solid. If
the force on the indenter is F, while the projected area of the indentation (the
area measured in the plane of the surface) is A, then hardness is dened as
H=
F
.
A
This is the correct expression for hardness values, sometimes called Meyer
hardness. However, it is not the value given in manufacturers look-up tables
270
Appendix A
because these tend to divide the force by the actual area of indentation. A
correction factor needs to be applied to make sense out of such data.
The units of hardness are those of stress. By itself however, hardness
is an arbitrary meaningless measurement (rather like if I had a hammer
mechanics). Meaning began with the work of Tabor (1951) who established
that hardness is an indirect measure of the yield stress within a material.
For materials where there is little change of volume upon the application of
load, then the hardness is three times the yield stress (Kendall, 2001). Where
a solid completely collapses in itself, however, then the hardness is the same
as the yield stress (Wilsea et al., 1975). Low-density plant tissue obeys the
latter relation as cells burst and collapse down to a pile of cell walls. In
contrast, seed shells, being very high-density cellular tissue, can produce
pile-up around the edge of an indentation, indicating that material is being
displaced. Whatever, there is always a certain amount of densication under
the centre of the indenter because hydrostatic pressures are so high.
Early tests involved large indentations but these macroindentations have
largely been replaced by micro-, and now nano-, indentations. Also, indenters with sharp tips, like the Vickers, Knoop and Berkovich geometries,
have increasingly replaced blunt spherical types. Hardness depends to some
degree on indenter geometry but understanding of what an indenter does
to a surface has grown dramatically (Lawn, 1993), extending the value of the
test. If resistance to indentation is totally elastic (i.e. the material springs
back after indentation), then the apparent hardness is controlled by Youngs
modulus. If the indentation is permanent, then the amount of plastic deformation depends on the yield strength of the material. Most uses of the
term hardness, both here and elsewhere, refer to permanent indentations.
However, even if the surface appearance of an indentation remains xed
due to plastic deformation, the deepest part of a sharp-tipped indentation
tends not to be so, resulting, for example, in the deepest part of an indentation formed by a sharp-tipped indenter rising up somewhat after the
load is released. Measurement of this recoil can be used to estimate Youngs
modulus.
Indentation tests are now routinely used to measure most of the material properties of ceramics (Lawn, 1993), being attractive both for their
simplicity and for their non-destructive characteristics. Many tests can be
made on the same specimen simply by spacing indentations sufciently
far (more than four indentation diameters) apart. The test has always been
the method of choice for investigating tooth tissues, but a rather sterile
literature has been rejuvenated recently by nanoindentation, a technique
developed in the last decade as an offshoot of the atomic force microscope.
271
This microscope uses a ceramic stylus, rather like that on the now-extinct
gramophone record player, located on the end of a cantilever beam, in order
to map a surface. To do this, very low forces (e.g. 108 N) are applied to
the stylus, which is mounted on a relatively springy (compliant) mounting. Measurement of mechanical properties requires a much stiffer setting
and a diamond stylus. Loads can range from 106 N to nearly 103 N. By
swapping indenting and mapping functions, Balooch et al. (2001) were able
to image the surface before and after indentation. The eld is advancing
rapidly.
It is perfectly feasible to measure indentation in mammalian eld studies,
preferably with a indentation of millimetre dimensions (Lucas et al., 2003).
Current research aims to estimate both the hardness and Youngs modulus
of plant materials.
notch sensitivity
If a notch is cut into a specimen of virtually any true solid, then its strength
will decline with an increase in notch length. However, this decline is
more rapid than can be predicted from the loss of cross-sectional area
(Fig. A.7). The non-linear form of the curve can be predicted from arguments in fracture mechanics. Such solids are said to be notch sensitive.
However, some biological tissues do not behave in this manner and decline
in strength in simple proportion to the loss of area (Fig. A.7). Such tissues
are said to be notch insensitive. Some food tissues are like this, a function of
their structural heterogeneity and the lack of rm mechanical connection
between their structural components (see Chapter 5).
ACTUAL TEST ARRANGEMENT
Before considering these, it is wise to know that there are a lot of subtleties
involved in this type of work. There may be faculty in your institution that
can provide practical help and there are also many standard techniques that
can be looked up in materials science books. Sometimes these standards
help biologists, but they can also be a hindrance. (Be aware that many of
the commonest industrial standards pre-date fracture mechanics and are
phrased in terms of strengths. Sometimes, these strengths are actually just
forces.) It can be difcult to turn biological tissues into the specimen sizes
and shapes demanded by some material standards. Additionally, certain
conditions which these standards aim to satisfy, such as plane strain (which
is not discussed in this book), are not likely to be related to the way that
272
Appendix A
Fig. A.7 When a notch, of length a, is cut into a tensile specimen, of width w and thickness
t, then the fracture stress (calculated as F/wt, i.e. force divided by cross-sectional area without
taking the notch into account) is reduced disproportionate to the loss of supporting area
represented by that notch. This phenomenon is called notch sensitivity and is central to
the tenets of fracture mechanics. However, a substantial number of biological materials are
not disproportionately affected by notches and show Galilean strength. These are termed
notch-insensitive tissues.
foods fail in the mouth. However, even if no standard test is possible, all
is not lost. The aim is not accuracy for the sake of it and the precision
that is necessary in biology always depends in the end on the tightness of
theoretical predictions. Many (but not all) estimates in biomechanics end
up with an accuracy of a factor of two: the real value could be half or double
that predicted. Generally, such a result is not something to be sniffed at.
Most eld tests require a miniaturized version of the universal testing machines found in engineering and food science laboratories. Darvell
et al. (1996) describe a tester that has since been expanded in range (Lucas
et al., 2001, 2003). The current version is shown in Fig. A.8. Together with
all its accessories, it ts easily into a suitcase. By comparison, laboratory
machines are often massive, the point being to make them so rigid that
only the specimen under test deforms. A portable eld tester obviously
has a much lower stiffness. Accordingly, specimens have to be very small,
273
Fig. A.8 The Darvell HKU tester and accessories for measuring the mechanical properties
of mammalian foods in the eld.
particularly if they are made from high-modulus materials. The other major issue with eld tests is the maximum load that it can take. So far, the
Darvell tester has only been tted with 10 N and 100 N load cells.
Data can be uploaded to a notebook computer giving immediate results. I have generally interfaced a 12-bit A-to-D PCMCIA card (DAQCard
6062E, National Instruments, Austin, TX, USA), displaying and analysing
the data using programs written in the Labview environment (National
Instruments). A suite of programs has been written for this purpose and is
available free from the author.
Really accurate work on materials demands that deformation be measured very accurately. However, deformation is constrained at the ends of
specimens. Tensile grips are an obvious example because they compress the
specimen locally making it very likely that it will fail around the margins of
the grips. To avoid this, it is common to make specimens wider at the grips
(Fig. A.1). There are a variety of grip surfaces, even involving pneumatic
action, so as to provide just sufcient compression to prevent slip.
Plates for compression should allow the specimen to slide freely. This is
encouraged by low friction. Some polytetrauoroethylene (PTFE) coated
tape on the plate surfaces can help this a lot. Bytac (Chemplast Inc., Wayne,
NJ, USA) is such a product: a 25-m thick layer of aluminium foil backed
with adhesive and coated with a 50-m thick layer of PTFE. Additionally,
some researchers lubricate plates with oil.
bend
notched
bend
wedge
cut
trouser
tear
pull
Force (F )
push
TOUGHNESS TESTS
MODULUS TESTS
gradient
u
Displacement (u)
Force
INDENTATION TESTS
loading
unloading
Displacement
Fig. A.9 Common testing geometries for estimations of Youngs modulus, toughness and indentation. The choice of
test depends on the ease with which specimens can be shaped and the control that they offer over crack growth. See text
for explanation.
275
Even with these safeguards, the only sure way to measure the strain
accurately is to measure the change in specimen height or length remote
from the ends of the specimen. This requires an extensometer, the most
accurate of these being a laser that detects the movement of reective
markers placed on the specimen. Such techniques are impossible in the
eld.
The basic form of most eld tests is shown in Fig. A.9. An excellent
cook book of methods has been produced by Vincent (1992), so this part
of the appendix is kept brief. Not all tests are listed here anyway. The instrumented microtome (Atkins & Vincent, 1984) and C-ring tests (Jennings &
Macmillan, 1986), to mention just two, are not included because they are
not yet feasible in the eld.
YOUNGS MODULUS
compression tests
All that is needed to get a value for Youngs modulus for a specimen loaded
in compression is to make a short cylindrical sample, making sure that
the ends are trimmed squarely. The ratio of cylinder height to diameter should be low (2 : 1 is ideal) in order to eliminate any possibility of
buckling.
tensile tests
Tensile specimens should be long and thin because the disturbing effect of
gripping the ends of the specimen is then reduced. Ideally, test specimens,
such as shown in Fig. A.1, should have a narrow waist, the strain in which
can be monitored with an extensometer. This renement is not possible in
the eld. If tested to failure, then a specimen should fail in the centre of
the specimen, certainly not at the grips.
bending tests
There are two types of bending tests three-point and four-point bending.
The latter has many advantages, but it can be difcult to set up in the
eld. The length of a bending specimen is called its span. The ratio of the
span to the depth (thickness) of the specimen in three-point bending should
be greater than 10 : 1 for relatively homogeneous materials and higher than
this if there is any obvious heterogeneity.
276
Appendix A
TOUGHNESS
As related above, toughness tests usually require a pre-existing crack or

notch. Rather than try to measure the aws in specimens, which is hopeless
(they are likely to be microscopic), an articial aw called a notch is made,
from which a crack will grow during the test. The sharpness of the notch
may matter: if so, it should be made sufciently sharp with a razor blade. In
practice, however, this often doesnt matter because the test crack tends to
extend from a small part of the microscopically irregular notch. However,
test pieces generally have to conform to a certain shape to get reliable
results and sometimes be above a certain size limit. These restrictions are
burdensome to a biologist. Foods come in all shapes and sizes and trying
to match denitions given in materials standards is often hopeless. Luckily,
there is a large class of fracture tests that have been developed that do not
require a notch: cutting tests. Given a choice, a materials scientist would
be very unlikely to recommend them because, for one major reason, they
often involve a lot of friction and for another, it can be difcult to decide
how much of the work done in such tests was actually required for fracture.
However, the action of teeth involves friction and, in any case, there are
some ways in which much of the friction can be factored out. Most of
these cutting tests derive from Atkins & Mai (1979) and involve blades and
wedges (which resemble features of dentitions). Often, the shape and size
of the specimen is not critical in these tests.
wedge
The wedge is one of the oldest fracture tests, dating from Obreimoff in the
1930s (Lawn, 1993). It is still recommended for fracture studies (Kendall,
2001) and has been introduced in a modied form into biomechanics by
Vincent et al. (1991) and Vincent & Khan (1993).
A wedge is a single blade pressed onto a block specimen that starts a
crack by bending two halves of the block apart (Fig. A.9). Crack growth
is usually stable. The choice of included angle for the wedge is arbitrary,
but a narrow angle (e.g. 15 ) seems to work best. The width of the block is
measured (that dimension that will be cut) and then the wedge is pressed
into it until a crack is started. The test can begin from this point, i.e. after
fracture. The crack will normally be stable in all but the stiffest tissues
and every unit movement of the edge of the wedge will result in the same
unit growth of the crack just in front of it. At the end of the run, the
277
depth needs to be measured to estimate crack area (this is generally given

by the depth of wedge penetration after the start of the test). However, the
work done involves not just fracture, but also friction and adhesion. To
estimate the latter two factors together, the wedge can be reversed and run
against the new crack surface. The work done in this second pass needs to
be deducted from that in the initial run, then divided by the product of
the block width and depth of wedge penetration to give an estimate of the
toughness.
trouser-tear
This test is only effective for tissues that will not deect cracks. The basic
arrangement is shown in Fig. 4.1. If there is little or no stretching of the
trouser legs, then with t being the specimen thickness, the toughness is
calculated as
2F
R=
t
where the force, F, should be constant during the test.
scissors
Scissors provide a practical eld test for fracturing thin sheets or rods, e.g.
leaves and shoots (Fig. A.10). The test is not well described in textbooks
and has been criticized on several occasions, so more is said about it here
than other techniques. A sheet or rod, usually <1 mm in thickness, is
placed between the blades of a pair of scissors. The work done during a
cut can then be recorded over a given displacement. This displacement
does not have to be the displacement of the scissors blade itself it can
be measured at any convenient location, such as the crosshead of a tester
that is driving the blade handle provided that the load is monitored at the
same place. The scissors can then returned to their original position and
the specimen removed. The scissors are then driven down again over the
same displacement to record the frictional work done between the blades
themselves. The work done in this empty pass is then deducted from the
work done in the rst pass to nd the work needed just to fracture the
specimen. All that remains is to measure the cut length in the specimen
(it is easier to cut right through it than to measure partial cuts) and its
thickness. The toughness estimate is the work done in fracture divided by
the area of cut. It is possible to run tests the other way around, with the
278
Appendix A
Fig. A.10 Measuring the toughness of leaves using a pair of scissors or a penetrometer. Both
devices require careful mounting and attention to their condition because the tests incur a
considerable amount of friction. The recording of a forcedisplacement curve is vital for
establishing the work done in the tests. With scissors, friction is almost entirely due to
metalmetal contact as the blades brush past each other in point contacts, so excluding the
compression of leaf tissue between the blades. In contrast, the penetrometer resembles a
punch and die separated by a small clearance in which leaf tissue gets compressed. (a) A
half-leaf specimen is cut by scissors along the dotted line, with the cut running from left
to right. Cutting the midrib rst allows both it and any secondary vein to be identied as
discrete items on a forcedisplacement plot. Their toughness can be obtained separately to
the rest of the lamina without dissection due to characteristic force peaks. (b) A penetrometer
generally involves a circular punch moving through a slightly larger hole cut in a circular
plate. The edges are usually very rough, particularly around veins, and this gets dragged
through the hole producing variable friction.
279
empty pass done rst (so ignoring the friction of the specimen against the
blades, which is often low). Regardless of the sequence of cuts though, the
scissors blades must be cleaned before the test (never between the cut and
empty pass) so as to get accurate results. A good pair of scissors is absolutely
essential. I have used hairdressing scissors made from cobalt steel because
plant tissues do not stain them.
When cutting young leaves the forces can be very low. Accordingly,
the force between empty blades needs to be low or errors will accumulate. Whatever the scissors used, it is important to remember that these
are not high-tech devices. The joint of the scissors is not stable when
the blades are very wide apart and the point contact along the blades
can be lost near full closure. It is sensible to work away from those
limits. Small variations in blade sharpness are not critical for work on
plant tissue, but the blades can be blunted quite quickly if tissues contain silica or silicates (present in some leaves and barks). The blades are
otherwise harder than most tissues on which they will be used. However, the blades will still wear against each other, this bladeblade contact being essential to avoid plant material being trapped between the
metal surfaces, a problem with penetrometer tests, (as discussed below).
Thin specimens are likely to be oppy. However, a very stiff specimen
will actually drive the blades together, resulting in more work being done
(and more metal being lost) than in an empty pass. A specimen this stiff
should be tested another way. All in all, it is a practical test that has stood
up well.
The scissors test was designed as a replacement for the popular penetrometer tests used by ecologists. It is only relatively recently that actual results of penetrometer tests have been tabulated in publications,
nearly always having been reported previously only as correlations. There
have been many designs, but they nearly all t the diagram in Fig. A.10
whereby a piece of biological tissue such as a leaf is punched through
a hole in a plate. The results are reported as a force, a stress (force divided by the area under the punch) or a force divided by punch perimeter. However, none of these analyses can be correct and there is also
friction to contend with, variously between specimen, punch and the
hole (Atkins, 1980). Without a clear analysis of how to understand tissue
trapped between the punch and the die at different phases of movement,
the technique seems wanting (Lucas et al., 1991a; Vincent, 1992; Wright &
Vincent, 1996). Despite this, the technique has its proponents (Aranwela
et al., 1999; Edwards et al., 2000), who have in turn suggested that scissoring
is also awed. I do not enter this argument here.
280
Appendix A
notched tension and bending
These tests are very well treated in Vincent (1992) and I make no attempt
to add a potted account of these tests here, particularly when examples of
these tests have been shown already in this appendix and in Chapter 5.
peeling
The bond between objects of dissimilar tissues or materials is referred to as

adhesion and the work to break it as peeling. The theory and practice of
peeling experiments are identical to those of fracture mechanics (Kendall,
2001). The peeling of a fruit, possible only when the outer layer strips
cleanly from the underlying esh, can be estimated by the simple geometry
given in Fig. A.9 with the work of peeling given by
Rp =
2F
w
where w is the width of the peeled strip.

friction
Friction is the resistance to motion between two surfaces at their interface produced by some (undened) interaction between them. It is far
from being properly understood, but it can be important to get some measurement of it for various aspects of tooth function. Figure A.11 shows
the classic view of friction whereby two bodies compressive force between
Fig. A.11 A block of mass M, so weighing Mg, is dragged across a surface by a force Fh .
With the vertical force Fv = Mg, the coefcient of friction = (Fh /Fv ).
281
them, Fv , require a frictional force, Fh , in order to slide. The coefcient of

friction is
Fh
=
.
Fv
Resistance to the start of movement tends to be greater than that needed to
continue it, leading to two distinct concepts: static and dynamic friction.
The coefcient of friction may not be constant, depending instead on the
forces involved (e.g. cork: Gibson et al., 1981). There are many approaches
to the measurement of , but simple devices will usually work. The device
shown in Fig. A.11 is a slider mounted on a low-friction track that can be
pulled along by winding the mechanical stage.
RELATING DENTAL FORM TO DIET
Investigations of the dentaldietary adaptations of mammals can be aided

on charts such as Fig. 7.1 where toughness (R) is plotted against Youngs
modulus (E ) on logarithmic axes. This chart, based on Ashby (1999), is
one of a large family of charts that could be constructed to represent the
obstacles that foods present to oral processing. Measurements like this
from eld studies on mammals are only just beginning, examples being
Yamashita (1998) on ve lemur species, Dominy (2001) on four African
anthropoid primates and the abstract of Strait & Overdorff (1996), who
were the rst researchers to take a full mechanical testing machine to the
eld.
HELPFUL REFERENCES
One of the biggest hurdles of entering the materials science literature is that
articles often lack citations, making it difcult to know who agrees with
whom. This is usually obvious in biology, but in the physical sciences, it
appears often to reect the lack of trips to libraries. Some books stress a
theoretical overview and while they may emphasize the need to do things
properly (i.e. in accordance with theory), they also indicate that accuracy
for its own sake is nothing to aim for. Other books may just make you (have
already made me) scared. I have found the following books optimistic and,
in places, exciting: Cottrell (1964), Ashby & Jones (1996, 1998), Gibson
& Ashby (1999), Atkins & Mai (1985), Kendall (2001) and Lawn (1993).
A good introduction to viscoelasticity is by Dorrington (1980). Brennan
(1980) and Bourne (2002) both describe the history of the mechanical
282
Appendix A
testing of foods. Vincent (1992) explains many modern techniques in detail

and is the only book that really concentrates on the mechanical properties of
biological tissues and how to measure them. A general text on biomechanics
that describes biomechanical techniques like electromyography and strain
gauge usage is Biewener (1992).
appendix b
Properties of teeth and potential foods

This appendix consists of tables listing properties mentioned in the text. As far as mechanical property measurements go, they are but a selection of those available in the
literature. While some of the property values represent many years of work by investigators, others are nothing more than preliminary estimates. Where there has been a choice,
I have opted for the most recent summary because some older work is probably more subject to variability due to state of preservation. Teeth provide an example. It is now
established that storing teeth in deionized water loses mineral, producing a 2030% decline in hardness inside a day (Habelitz et al., 2002). This means that it is certain that
some hardness values quoted in the literature are too low. Estimates for other properties might inadvertently have been elevated: e.g. antler, which is lower in modulus and
strength than limb bone, is higher in toughness (Currey & Brear, 1992), due to a slight reduction in its mineral content. Teeth extruding from the mouth may generally have
lower mineralization, e.g. elephant ivory versus human dentine (Rajaram, 1986), and are undoubtedly drier. Though the toughness of ivory is unknown, a lot of brillar
pull-out at fracture seems to imply higher toughness than that of human dentine (Rajaram, 1986).
MECHANICAL PROPERTIES OF BIOLOGICAL TISSUES
Tissue
Tooth tissues
Human
(Homo
sapiens)
Youngs
(elastic)
modulus
(GPa)
Enamel 50120
Dentine 2327
Peritubular
29.8
Intertubular
17.721.1
Yield
strength
(MPa)
Tensile
strength
(MPa)
Toughness
(J m2 )
Hardness
(MPa)
Enamel
Across rods: 200
Between rods: 13
Dentine
Across tubules: 270
Between tubules: 550
Enamel 25006000
Enamel 835
Dentine 250800
Peritubular
22302540
Intertubular 120520
Dentine
31104
Energy
absorption
at fracture
( J m3 105 ) Reference
Mostly compiled by
Marshall et al. (2001);
enamel from Cuy
et al. (2002);
toughness from
Rasmussen et al.
(1976); peri- and
intertubular dentine
from Kinney et al.
(1996); mantle
dentine much
Remarks
Enamel of
deciduous teeth
averages 85% of that
of permanent teeth
(Nose, 1961; N.B.
absolute hardness
values reported there
must be wrong).
Major reviews by
Braden (1976) and
Waters (1980)
appendix b (cont.)
Tissue
Youngs
(elastic)
modulus
(GPa)
Toughness
(J m2 )
Hardness
(MPa)
Yield
strength
(MPa)
Tensile
strength
(MPa)
Energy
absorption
at fracture
(J m3 105 ) Reference
Beaver
(Castor
ber)
Dentine 560
softer than rest of

primary dentine
(Renson & Braden,
1971)
Osborn (1969)
Sheep
(Ovisaries)
Enamel 26503750
Baker et al. (1959)
Remarks
Wear patterns match

hardness contours;
mantle dentine is
much softer,
400 MPa
Dentine 324725
Pigtail
macaque
(Macaca
nemestrina)
Enamel 3800
P. W. Lucas
(unpubl. data)
Average of
sub-surface enamel
and primary dentine
P. W. Lucas
(unpubl. data)
Average of
sub-surface enamel
and primary dentine
P. W. Lucas
(unpubl. data)
Average of
sub-surface enamel
and primary dentine
Dentine 590
Gibbon
(Hylobates
muelleri)
Enamel 2765
Dentine 580
Orang-utan
(Pongo
pygmaeus)
Enamel 3550
Dentine 480
Indian
elephant
(Elephas
maximus)
Ivory:
Dry 12.5
Wet 3.5
Other tissues
Bone (human
femur
unless
stated)
Antler bone
17.7
937 (bovine tibia)
Dry 17.1
5000
340400
Wet 7.5
Articular
cartilage
0.31
0.08 103
1401200
Rat skin
Initial 106
Final 6 103
Horse hoof
0.180.56
Trouser-tearing:
14 00020 000
Scissors: 590
5.510.7
Fruit peel
Gnetum
microcarpum
(Gnetaceae)
Scissors:
across sclereids 1485
along sclereids 642
6.59.5
110
8.7
36
4.9
99.2
4.0
188
13.5
108
14.6
Rajaram (1986)
Softer than human

dentine (Rajaram
1986); probably dry
in tusk; dry density
1.7 g cm3
Rajaram (1986);
Blackburn et al.
(1992); Norman
et al. (1992)
Rajaram (1986);
Currey & Brear
(1992)
Dry density
1.94 g cm3
Chin-Purcell &
Lewis. (1996);
Korhonen
et al. (2002)
Purslow (1983);
Pereira
et al. (1997)
Bertram & Gosline
(1986);
Kasapi & Gosline
(1997)
Korhonen et al.
(2002) give Poissons
ratios of 0.150.21
P. W. Lucas
(unpubl. data)
Dry density
1.86 g cm3
appendix b (cont.)
Tissue
Youngs
(elastic)
modulus
(GPa)
Pods
Derris
thyrsifolia
(Leguminosae)
Intsia
palembanica
(Leguminosae)
Seed coverings
Mezzettia
parvifolia
(Annonaceae)
zone II
Schinziophyton
rautanenii
(Anacardiaceae)
Cocos nucifera
(Arecaceae)
Elaeis guineensis
(Arecaceae)
Macadamia
ternifolia
(Proteaceae)
Callerya
atropurpurea
(Leguminosae)
Toughness
(J m2 )
Hardness
(MPa)
Yield
strength
(MPa)
Tensile
strength
(MPa)
1900
Air-dry 4.96
Wet 5.23
1437
851
2.94.9
17001900
205
P. W. Lucas
(unpubl. data)
Pod 0.8 mm thick
67
Lucas et al.
(1991b)
Woody shell: dry

density 1450 kg m3
51.65
45.66
Williamson &
Lucas (1995)
Woody shell;
moisture content:
air-dry 9.2% on
initial weight,
wet 20.5%
Woody shell
Woody shell
205235
26
1001000
Single edge notched
tension: 330
Trouser-tearing: 355
Scissors: 3620 (445)
Remarks
P. W. Lucas
(unpubl. data)
Scissors:
ripe brown 20004485
unripe green 255469
Scissors:
along pod width 4906
along pod length 6950
7
Energy
absorption
at fracture
180
2580
5.04 (1.12)
Jennings &
Macmillan
(1986)
P. W. Lucas
(unpubl. data)
Woody shell: dry

density 1300 kg m3
Flexible covering;
modulus in tension
Flexible covering
Cut made
perpendicular to
massive (100 m
diameter) sclereids
Moisture content of
whole seed 7.5%
Moisture content of
whole seed 8.6%
Moisture content of
whole seed 7.8%
Scissors: 3835
267
261
327
167
Woody plant gall

C-ring
On Distylium
(compression):
racemosum
0.565
(Hamamelidaceae)
Seed contents
Callerya
atropurpurea
(Leguminosae)
Gnetum
microcarpum
(Gnetaceae)
Brazil nut
kernels
Macadamia
ternifolia
(Proteaceae)
kernel
P. W. Lucas
(unpubl. data)
Scissors: 1500
Intsia
palembanica
(Leguminosae)
Gnetum
microcarpum
(Gnetaceae)
Albizia splendens
(Leguminosae)
Samanea saman
(Leguminosae)
Adenanthera
pavonina
(Leguminosae)
Alangium ridleyi
(Alangiaceae)
0.01
C-ring: 177
Wedge: 1285 (269)

Scissors: 947 (327)
Wedge: 1143
33.84 103
160.8
32.18 103
214.3
Hill et al.
(1995)
80
2.5
P. W. Lucas
(unpubl. data)
Cotyledon
Endosperm
appendix b (cont.)
Tissue
Leaves
Lolium perenne
(Graminaceae)
Youngs
(elastic)
modulus
(GPa)
Mollusc shell
Nacre
Insect cuticle
Locust mandible
Leaf-cutter ants
(Atta sexdens)
Hardness
(MPa)
Energy
absorption
at fracture
215402
Vincent (1982);
Greenberg et al.
(1989)
Choong (1996)
Longitudinal
0.1860.240
Transverse
0.0500.068
Whole lamina 410

Without veins 120
Veins 20006000
Midrib 20009000
Whole lamina 724
Without major veins
220300
Veins 6000
60
3501240
Longitudinal
0.55
Transverse
0.014
Castsanopsis ssa
(Fagaceae)
Calophyllum
inophyllum
(Guttiferae)
Toughness
(J m2 )
Yield
Tensile
strength strength
(MPa)
(MPa)
Without
veins 2
Across veins
720
140
175350
Light unsclerotized:
200
Dark sclerotized:
300350
Mandibular cusp
(containing zinc):
Light unsclerotized:
350
Dark sclerotized:
800
Jackson et al. (1988)

Hillerton et al.
(1982)
Schoeld et al.
(2002)
Remarks
Isolated tissues and pure components

Hydroxyapatite
101
30
5170
Zhang et al. (1997)
Elastin
0.0011
1.6
Resilin
0.002
Collagen
1.2
120
Periodontal
ligament
Disc of
temporomandibular
joint
Cell wall
(of woody tissue)
0.05
0.044
Gibson & Ashby

(1999); Lucas et al.
(2000)
3450
Parallel to
cellular axis:
25
Perpendicular to
this: 15
Opal phytoliths
58006000
Baker et al. (1959)

Baker et al. (1959);
Lawn (1993)
Soil particles
Quartz
70
70007750
Engineering materials
Diamond
1000
150 000
50 000
Alumina
371
Polymethylmethacrylate 3.4
100
640
560
65
110
Mild steel
14 000
220
430
210
Human foods
Initial 106
Gel (mung bean starch)
Spring roll pastry
(briey cooked)
0.522
Fresh 0.2 103 Trouser-tear: 456.9

Scissors: 139.7
Gosline et al.
(2002)
Gosline et al.
(2002)
Gosline et al.
(2002)
Rees & Jacobsen
(1997)
Tanne et al. (1991)
0.043
Ashby & Jones (1996);

Atkins & Mai (1985)
Atkins & Mai (1985)
Ashby & Jones (1996)
Atkins & Mai (1985)
Ashby & Jones (1996);
Atkins & Mai (1985)
Lucas et al. (1993);
C. G. Oates et al.
(unpubl. data)
Sim et al. (1993)
Extensibility 1.5
Extensibility 1.9
Extensibility 0.13
appendix b (cont.)
Tissue
Raw carrot
Cheese
Reduced-fat cheddar
Mozzarella
Parmesan
Youngs
(elastic)
modulus
(GPa)
Toughness
(J m2 )
1 day old
1.3 103
4.57 103
0.94 103
Trouser-tear: 235.5
Scissors: 207.8
440
172.7
0.15 103
2.26 103
70.7
233
Hardness
(MPa)
Yield
Tensile
strength strength
(MPa)
(MPa)
Energy
absorption
at fracture
0.0635
Agrawal (1999)
Agrawal (1999)
Agrawal (1999)
Agrawal (1999)
Remarks
PROPERTIES OF NON-WOOD PLANT TISSUES USED IN FIG. 4.15
Plant tissues
Citrullus vulgaris (Cucurbitaceae) red
watermelon fruit esh
Brassica rapa (Brassicaceae) green turnip
Solanum tuberosum (Solanaceae) Russet
Burbank potato
Gossypium sp. (Malvaceae) cotton hairs
Callerya atropurpurea (Leguminosae) cotyledon
C. atropurpurea (Leguminosae) seed coat
Leucaena leucophala (Leguminosae) pod:
inner brown layer only
Albizia splendens (Leguminosae) pod
Aleurites moluccana (Euphorbiaceae) seed shell
Schinziophyton rautenenii (Anacardiaceae)
seed shell
Mezzettia parvifolia (Annonaceae) seed shell
Zone I
Zone II
Zone III
Scheelea sp. (Arecaceae) palm nut endocarp
Cellulosic materials
Filter paper (Whatmans, UK) No. 1
Bank paper
Newsprint
Volume fraction
occupied by
Number
cell wall ( Vc )
of tests
Slope plastic work

Correlation (J m2 m1 ) mean
coefcient r (standard error)
Intercept cell wall

toughness (J m2 ) mean
(standard error)
0.0031
22
0.815
21.9 (3.5)
9.5 (6.8)
0.0192
0.0253
25
35
0.731
0.467
341.0 (66.4)
343.5 (113.4)
84.5 (43.2)
89.6 (50.7)
0.089
0.127
0.349
0.242
20
31
27
25
0.909
0.720
0.633
0.822
455.7 (49.3)
1 039.5 (186.2)
894.6 (217.4)
5 775.3 (799.4)
300.8 (101.2)
267.8 (123.5)
1 681.3 (210.7)
761.3 (194.1)
0.349
0.90
0.94
41
40
21
0.821
0.801
0.491
3 143.2 (349.8)
14 630 (1,778)
11 349 (4,617)
970.0 (259.2)
2 638 (515)
3 767 (1 005)
0.95
0.92
16
29
12
21
0.576
0.577
0.184
0.614
24 840 (9 421)
12 011 (2 822)
2 310 (3 905)
15 636 (4 734)
4 558 (1 935)
3 739 (680)
4 192 (1 136)
2 423 (1 532)
0.343
0.347
0.411
0.413
0.449
15
17
30
45
39
0.963
0.931
0.924
0.855
0.921
7 650 (597.6)
6 793.3 (689.5)
11 840 (923)
10 853 (1 005)
9 225 (640)
1 648.9 (217.3)
1 980.2 (313.8)
1 368.3 (218.9)
913.8 (108.6)
1 232 (147)
Notes
1 HOW TO GET EXCITED ABOUT TEETH

1. The most recent research, some of it still unpublished at the time of writing,
suggests that teeth may be older than envisaged above, evolving as part of an
adaptive explosion of body forms in the Cambrian.
2. Some ability to fracture food particles using the mouth or pharynx is not unique
to mammals. Some sh (Vincent & Sibbing, 1992), insects (e.g. locusts) and
birds (e.g. parrots) do it, as did an early reptile (Rybczynski & Reisz, 2001).
3. Some mammals do regurgitate food. Many of the ruminants, for example, do
this, but with the intention of chewing food further rather than expelling it.
Throwing up as a routine behaviour is exceptional in mammals.
4. This view has recently been challenged (Rich et al., 2002).
5. Sensory feedback from the masticatory muscles is sent back to the brain via a
single sensory neuron, i.e. by the fastest possible route. I do not know of any
instance of this outside of the trigeminal nerve (the cranial nerve conveying
these signals) in mammals. However, this will not offset massive differences in
body size an elephant could not respond at the speed that a shrew can.
6. Some taste buds are located in the inlet to the larynx in humans, which cannot
have anything to do with food quality. Instead, it seems to suggest that taste
sensations also help identify where food particles are, additive to the sense of
touch.
2 THE BASIC STRUCTURE OF THE MAMMALIAN MOUTH

1. Some authors refer to the deciduous molars as deciduous premolars (e.g.
Schwartz, 1974). I disagree, but this 1974 paper was signicant as marking the
rst substantial break with a long-standing tradition over dental homologies.
2. The mammals with the thickest enamel appear to have been the gomphotheres
(probable ancestors of elephants: Maglio, 1973) which had enamel on their
molars up to 5 mm thick in places. Both gomphotheres and their teeth were
very large, but it has to be remembered that the ameloblasts (cells that form
enamel) of a gomphothere were probably no different in size from those of a
shrew.
292
Notes to pp. 2660
293
3. Cracks in mature enamel then probably run through a rather gel-like matrix.
The toughness of a thick gel is of the same order as that of enamel (Lucas et al.,
1993). Averaged over the tissue though, enamel is about three times tougher
than pure hydroxyapatite (White et al., 2001), but about 40% less hard (Zhang
et al., 1997).
4. However, Foxs experiments only provide limited support of this hypothesis
because the predicted effect is not large. The main problem though is that Fox
tested whole teeth and, as shown later, dentine very denitely displaces uid
when it is loaded.
5. The term symphysis is the name given to a type of brocartilaginous joint
found in the midline of the dentaries of the lower jaw and of the pubic bones
of the pelvis (where it allows for substantial movement during childbirth).
6. One long-term denture-wearing subject that I measured by the same method
had only 5 mm of mandibular bone height left in the molar tooth region.
7. This bone is called the lamina dura in dental radiographs, being more radiopaque (dense) than the rest of the alveolar process.
8. An innocuous result that is equation 86 of Synge (1933). Albeit a classic analysis
by a famous applied mathematician, the paper is very difcult to read and the
change in meaning of symbols as the paper progresses does not help. I am
indebted to Waters (1975) for unravelling critical aspects of this article.
9. The periodontal ligament must provide the motive force for this, but quite
how it does it is unclear.
10. Large anteriorposterior movements are also seen in elephants and, judging
from tooth wear, in some extinct mammalian groups. The text sticks to rodents
because these have been studied physiologically.
11. The modiolus can be palpated in humans by placing a thumb inside the corner
of the mouth with the index nger over the skin. The modiolus is felt as a
thickening of the soft tissue.
3 HOW THE MOUTH OPERATES
1. Some dinosaurs have been suggested to be capable of mastication and may also
have had some control, albeit limited, over their body temperatures (Norman
& Weishampel, 1991). This latter point has been debated for about 30 years.
2. Image analysis is now taking over from sieving.
3. Despite the apparent ease of the selection concept, there are difculties in
practical measurement. Take the largest particle sizes, the squares, shown
in Fig. 3.4. Now suppose that, during any given chew in which it is being
calculated, these squares break to produce the smaller-sized circles, hexagons or
triangles. The selection function for the squares is easy to calculate because we
know unequivocally the fraction that is broken. However, when we come to do
the same with smaller sizes, then we are faced with the proportion of particles
that is broken down being counterbalanced by a proportion broken into this
size range from larger sizes such as the squares. If we take no notice of this, it
could result in a negative selection value. This would be absurd: particles are not
294
Notes to pp. 6071
getting bigger. The way around this problem has been to label particles, which
we can do by dyeing them (as indicated by the different tones of the particles
in Fig. 3.4) or by shaping them (as indicated by the fanciful shapes shown in
the gure). By doing such experiments in humans, it has been shown that the
selection function is quite stable over a huge time-span, well beyond the usual
physiological timescale (van der Glas et al., 1987).
4. These authors report that a solution to comminution equations evaded even a
mathematician of the standing of Kolmogorov.
5. Such distributions probably do typify foods sticking to the teeth, such as biscuits
and other products made from ours. Obviously, some very ne fragments are
produced, but I have never seen a study of this.
6. I am not advocating here that the locomotory substrate makes no difference
at all if it did, there would be no need for special running tracks for human
athletes (McMahon & Greene, 1979).
7. The space outside the dental arches is termed the vestibule, which is like a gutter.
The tongue rarely needs to sh food out from here though because the control
of this space by the buccinator seems to be so good that appreciable quantities
of food rarely fall into it. In humans, only wearers of articial teeth called
dentures or those with paralysis of facial muscles collect food in the vestibule.
However, in some mammals, the vestibule houses cheek pouches as in some
rodents and in cercopithecine Old World monkeys. These act as designed food
stores. In these monkeys, the mucosal lining of the pouch is wired to the brain
bilaterally (Jones et al., 1986; Manger et al., 1996), something that is not true
for other taxonomically related, but pouchless, primates (Manger et al., 1995).
This suggests the need for ne control of the contents of these pouches.
8. There have been other experiments in humans relevant to this sensory threshold
(Owall
& Vorwerk, 1974), but with a technique that should be modied before
being repeated. The swallowing of very similarly sized particles across a wide
range of mammals might be related to the retentive ability of sphincters lower
down in the gut: there is some evidence that very small particles in dogs slide
directly through from stomach to small intestine, thus omitting a digestive step
(cited in Jenkins, 1978). All this may help to explain why foods in the modern
human diet, often made of extremely small particles, are not cleared effectively
from the mouth by natural means. We are barely aware of their presence until
we look.
9. An analogy to mechanical testing is possible where the rate of change of either
force or displacement can be controlled while the other variable is monitored.
The difference between controlling and monitoring change is important because this has a great effect on crack propagation in the test specimen. Stress
control is more likely to result in fast crack growth than displacement control
(Atkins, 1994). However, the stiffer the material under test, the slower the rate
of displacement has to be because the rate of force build-up will be proportional
to specimen stiffness and sensors that could stop or reverse loading so as to avoid
damage may not have time to act.
10. A large amount of research has been done on this, both theoretical and experimental, but it appears inconclusive. Early papers with novel analyses include
Notes to pp. 7189
11.
12.
13.
14.
15.
16.
17.
18.
295
those by Smith & Savage (1959), Barbenel (1972), Greaves (1978), Bramble
(1978) and Smith (1978). Spencer (1998) reviews and tests most of these
hypotheses.
The joints of cercopithecine Old World monkeys have been proved to endure
such forces (Hylander, 1979b; Brehnan et al., 1981).
Nelson et al. (2001) did not test arabinose, which is a pity since this sugar turns
up in appreciable quantities in some fruits and may be sensed by some other,
as yet unknown, taste receptor. It is worth noting that the relative sweetness of
sugars does not seem to tally with their caloric value.
Breslin et al. (1996) introduced the term monogeusia for compounds that
cannot be discriminated from each other by taste. According to these authors,
sucrose, fructose and glucose are like this and are probably perceived via a single
taste mechanism in the human. This conicts with what I have written. However, these authors agree that maltose is different: it is indistinguishable from
these other sugars at low concentrations, but behaves differently at higher concentrations. They suggest this is because maltose activates a second sweetness
receptor.
The order of taste response, from strongest to weakest, appears to be cystine,
alanine, glutamic acid, serine, methionine, aspartic acid, glycine, thyronine,
histidine, arginine and valine. The response is heightened by the presence of
purine nucleotides. Glycine triggers both the amino acid and the sweet receptor
(Nelson et al., 2002).
Critchley & Rolls (1996) have recently found neurons in the taste cortex of
primates (macaques) that respond to tannic acid.
Recently, evidence is accumulating in humans (Hiiemae & Palmer, 1999) that
some foods form a bolus in the oropharynx, behind the mouth, on the upper
surface of the epiglottis.
Ruminants are exceptional in producing vast amounts of saliva. This is because
saliva provides the uid for the fermentation process in the rst compartment
of their stomachs (Kay et al., 1980; Kay, 1987). This fact shows that copious
salivary secretions are entirely feasible were this ow to be required for the
swallowing process.
Cuticles pass through the mammalian gut undigested even in ruminants. However, they are apparently subject to heavy erosion in regions of high rainfall
(C. Neinhuis, pers. comm.). N. J. Dominy (pers. comm.) found that few tropical rainforest leaves show the lotus effect. Micro-roughening of the surface,
a vital part of this effect, may also explain the self-cleaning properties of the
wings of insects (Wagner et al., 1996) and even perhaps the oral surface of the
tongue too.
4 TOOTH SHAPE
1. There seems to be an unjustied assumption in the literature that strength
is OK as a fracture parameter unless massive differences in size are being
considered. This is not so, and it is not something specic to compression
296
2.
3.
4.
5.
6.
Notes to pp. 89113
either, but applies to any loading regime as Grifths (1920) original paper
showed.
In reality, cusps may contact because of the inertia of jaw movement or because
the diet of many mammals is quite varied: some foods are eaten that do not crack
or they could be chewed down to the brittleductile transition. Furthermore,
if mammals swallow like most humans, then their teeth will contact at the
beginning of this process. Thus, it makes sense to develop upper and lower
tooth shapes that can contact each other without damage.
Bond (1952, 1962) based his work index on an enormous amount of data gleaned
from industrial processes supporting his contention that the energy consumed
in comminution (usually measured by calorimetry) is proportional to change in
the reciprocal of the square root of particle size. However, he stressed that this
index really did not refer to particle dimensions, but to cracks. Two nineteenthcentury theories preceded Bonds. Rittinger suggested that the energy consumed
was proportional to the particle volume being fractured, while Kick argued that
that it was proportional to surface area. The ramications of Rittingers and
Kicks theories are best read in Lowrisons (1974) summary of knowledge on
industrial comminution, because the original accounts are nothing more than
vague suggestions. Neither theory has much to back it up.
This secondary tensile stress is the basis of the CookGordon crack-stopping
effect, whereby a crack, driven by a primary stress and that is heading to a weak
interface, can be arrested by a secondary stress that opens the interface before
the crack reaches it (Cook & Gordon, 1964). The result is a blunted crack. The
problem with emphasizing this mechanism as a major defence of foods (or teeth)
against fracture is that all it really achieves is a complication in crack direction.
Generally, the crack may extend along the interface for a short distance but
then be driven on in the original direction as soon as favourable circumstances
permit. This zigzagging crack in a food particle obviously involves a mixed
mode of fracture, but again this is nothing directly to do with the shape of
tooth surface that is attempting to fragment it.
Remember though that the stress level is not sufcient for the crack to grow.
The energy balance must be in favour of it, something that generally requires
that the stress eld (the volume of tissue being stressed) be large enough.
All told, this section gives the reason why surgeons (and animal anatomists)
need to use sharp scalpels. Some human foods show similar behaviours to
animal soft tissues. For example, popiah skins are thin sheets of pastry used to
wrap savoury foods in certain Chinese communities. (When lled and fried,
these are known in the West as spring rolls.) The skins are cooked by being
laid on one side against a hot metal plate for just a few seconds. A study by
Sim et al. (1993) showed that the stressstrain curves of freshly cooked skins
are completely linear up to fracture and highly extensible. They blunt cracks
signicantly when pulled, resulting in a toughness that is about three times
that when cut with scissors. After 24 hours, this extensibility reduces and the
toughness of a free-running crack is then more or less equal to that in cutting
(Appendix B).
Notes to pp. 117126
297
7. Other examples include muscles and tendons in which bres commonly tear
during exercise (but rarely propagate so as to split the whole structure into two,
as athletes know), the periodontal ligament, an unfused mandibular symphysis
and ligaments around joints.
8. There may also be a tertiary cell wall present, but this is generally irrelevant in
mechanical terms.
9. There are many ways to estimate Vc . Microscopy is very effective for homogeneous tissues. When Vc is low, it is approximately equal to the thickness of the
wall divided by the diameter of the cell or the average cell dimension (Ashby &
Gibson, 1999). In tissues where there are elongated cells, all with the same
orientation, Vc can be estimated by measuring the area occupied by cell walls
in end-on view. This area fraction is effectively the same as Vc . If tissues are
too heterogeneous to support such laborious measurement, then as a crude
calculation for living tissues, Vc is approximately equal to the chemical bre
content (neutral detergent bre or NDF; van Soest, 1996), calculated on a wet
(fresh) weight basis, multiplied by 1.5 (Choong, 1996). This multiple allows,
probably too generously, for the extra density of cell wall as compared to cytoplasmic constituents. Alternatively, if the tissues are dead, as in woods or seed
shells, then it is often possible to cut them into a dened shape, measure and
then weigh them. The density relative to that of solid cell wall (taken to be
1500 kg m3 ) is the same as Vc . Lastly, if this is not possible, then as many
pieces as possible can be collected together, weighed, then dropped into water
in a measuring cylinder to nd the displaced volume. A weight of known volume will be needed to help sink most tissue. This last Archimedes procedure
is generally the least accurate.
10. There is just one direct measurement of cell wall toughness in the literature.
This is on potato parenchyma and is 9.55 kJ m2 , of which one-third of
this, i.e. 3.18 kJ m2 , was estimated to be due to elastic mechanisms acting
in the wall (Hiller et al., 1996). This latter value seems compatible with the
3.45 kJ m2 given in the text. The propensity for plastic work is evidenced by
slope/intercept ratios: these are often close to 3.0 for tissues with only primary
cell walls (Appendix B), but average 10.0 or more for woods across-grain (Lucas
et al., 1997).
11. There is actually a negative correlation between the toughness of woods and
lignin content (Lucas et al., 1997).
12. Turgidity tenses the cell wall and so probably lowers the cost of fracturing it by
doing some of the work for teeth. Exactly how much the toughness drops is
not clear because the measurement of turgour pressure is not easy (Tomos &
Leigh, 1999). Flaccid parenchyma can seem much tougher than when it is
turgid, but the change in the shape of the stressstrain curve, analogous to
what happens when such tissue is boiled (see Chapter 7), may mislead.
13. A bat probably fractures such fruit tissue to reduce the weight of tissue being
processed in its gut. These mammals have very fast gut passage rates and the
type of fruit that they concentrate on often has large quantities of nutrients
(simple sugars and free amino acids) that do not require digestion. Chimpanzees
298
14.
15.
16.
17.
Notes to pp. 126149
often wadge foods, i.e they place plant tissues inside their lower lip and
squeeze juice out of them by pressure against their lower incisons. The piths
of monocots are often treated like this, most likely because these contain bres
that, although comprising only a small proportion of the tissue by volume,
prevent fragmentation by the chimpanzees blunt molars (Wrangham et al.,
1992).
J. F. V. Vincent has talked about the resistance of grasses in this way: trampling
it does not result even in fraying of the blades.
The distal edge of a sabre-tooth canine can be sharp, but is not sharpened
against another tooth. Some herbivores, such as some deer and anthropoid
primates, have enlarged canines that they employ mainly in disputes among
males in a species. These have sharpened blades (Zingeser, 1969; Walker, 1984),
being designed for damage, but not death or ingestion.
Tests of any of the dentaldietary relationships suggested in this book require
that a range of foods be included. If the range is too restricted, then it is
likely that the various mechanical properties will be correlated with each other
(Ashby, 1998), leading to spurious conclusions.
I have failed to locate an old note published by G. E. Hutchinson referring to
the devastating effect of the canines of pet cats on hairless human skin. The
note suggested canine reduction in hominins followed hairlessness.
5 TOOTH SIZE
1. One factor driving exponents up is the use of the reduced major axis, which
is identical to the slope obtained from a least-squares regression, divided by
the correlation coefcient. Since most biological analyses involve something
less than a perfect t, this practice always raises body-mass exponents to some
degree.
2. One of these attempts was a mechanical explanation close to the theme of this
book. It supposed that living organisms might be designed such that the elastic
distortion of their tissues under load is everywhere identical (McMahon, 1973).
This seems unlikely, which extensive tests of the hypothesis have reinforced
(Alexander, 1985).
3. If engineers had to face a population of marauding King Kongs that were
chewing up large buildings, then things might be rather different.
4. An exponent of 0.83 might come as a surprise to those used to allometric
arguments. What other theory could result in such an odd exponent? No
simple fraction gives this number.
5. The evidence for this difference in seed transport distances in mammals is not
clear-cut. Many small mammals spit seeds out, e.g. bats (Phua & Corlett, 1989)
and cercopithecine Old World monkeys (Corlett & Lucas, 1990; Lambert,
1999), doing this close to the parent plant. The distances from a plant that
mammals defaecate swallowed seeds depend on the home range of the mammal
and the patchiness of the fruit source that it is exploiting.
Notes to pp. 150177
299
6. These gures are all biased towards the modern human diet. No one eats old
cow or pig because their esh is much tougher than that of younger animals
(and age is the largest factor in determining meat toughness). There is no
similar age bias for sh data: a large sh of any given species seems a good
catch.
7. Flexible jaws would store a lot of energy. After food fracture, this storage
would feed into cracks, propagating them more rapidly. Wright & Vincent
(1996) discuss this in relation to a comparison of fracture tests of foods on
mechanical testing machines, which are typically very stiff, and conditions in
the mammalian mouth.
8. It appears that while large mammals trapped on small islands get smaller,
the small mammals there may enlarge, so reducing the overall size range.
Unfortunately, these rapid enlargements have not been well studied.
9. Gould (1975) and Fortelius (1985) both feel that downsizing the body would
involve different patterns to the more usual upgrade. I do not agree and am unwilling to invoke developmental impediments to dwarng, arguments that are
often shufed out to explain this phenomenon. There appear to be enormous
selective pressures on mammals that reduce their body size very rapidly and I
suspect a functional explanation for the patterns that they show, an explanation
that fracture scaling provides.
10. The actual t of forceelongation curves for mammalian soft tissues on a log
log plot is often poor. Certainly, many plots in Yamada (1970) remain distinctly
curved after logging both stress and elongation (and after converting to true
strain). Some authors claim exponents as high as 9.0 for these tissues.
11. Van Valkenburgh (1990) was concerned with the prediction of body mass
from dental and skeletal measurements and thus plotted body mass on the
vertical axis and carnassial length on the horizontal. The slope in the text was
obtained by taking the reciprocal of the slope given in this paper, divided by
the correlation coefcient.
12. Foods are more complex than this and can show a considerable amount of
viscoelastic behaviour, which would be fatal in most engineering materials.
However, as pointed out by Atkins (1999), this viscoelasticity would not be
apparent until or unless the load was taken off. Which is to say that the shape
of the stressstrain curve is the main concern, not subtle differences in the
deformational processes that underlie this. This is what allows him to treat
yielding as though it were equivalent to non-linear elasticity.
13. As an undergraduate obsessed with teeth, I was irked for a long time by not
being able to sustain an objection to this throwaway comment from a classmate.
14. It should be noted here, for the benet of proponents of allometry, that these
variables are dimensionless and that the gape angle takes size into account.
This relationship is then very close to the exploration of residuals in allometric
analysis, of which examples can be found in this chapter. The gape analysis
appears to me to be a more direct and elegant approach than an allometric one.
15. There is a very slight curvature to y/x lines that cannot be mimicked in the
data analysis.
300
Notes to pp. 182204

6 TOOTH WEAR
1. The analysis of dental microwear was developed independently for studying the
wear of dentures in human subjects (Heath, 1986).
2. The triggering of this sharpening seems to be psychological stress, but is still
debated. Personally, I will never forget the dreadful noise from a caged male
pigtail macaque (Macaca nemestrina) in Singapore when sharpening its canines.
3. If work-hardening, which is not explained in this book anyway, is ignored, then
the multiple would be 3.0, but the accuracy of all the following calculations is
so limited that none of this really matters.
4. The picture is complicated more than a little by scratches often being variable
in depth, starting very shallow, but deepening and ending up in pits (Walker,
1984).
5. The dental wear features that both particles cause are much smaller than their
own dimensions due to only a small part of their surface being in contact with
the tooth tissue.
6. Approximal wear is also, quite logically, referred to as interstitial wear. However, for some reason, the terms got jumbled up and many people talk about
interproximal wear. The editor of one dental journal used to complain rather
strongly about this.
7. The occlusal plane, a plane drawn through the contact points of the postcanine
dentition of mammals can be curved antero-posteriorly. This is called the curve
of Spee in the human dentition. It seems associated with a high jaw joint, as
explained by Osborn (1987).
8. The swallowing model of Prinz & Lucas (1997) is particularly sensitive to the rate
of food breakdown. Slow breakdown is responsible for weak bolus formation
in carrots. This just makes wear even more serious. Not only will wear reduce
the rate of food breakdown, but it jeopardizes ever forming a bolus. If a bolus
does not form, then the mouth is not properly cleared of particles.
7 THE EVOLUTION OF THE MAMMALIAN DENTITION
1. As pointed out in Chapter 4, the esh of some fruits is an obvious exception.
2. This statement reserves the term chewing for extended intra-oral processing. The 260-million-year-old Palaeozoic reptile Suminia (Rybczynski & Reisz,
2001) and the 230-million-year-old Mesozoic reptile Pisanosaurus (Sereno, 1999)
could both fragment plant material to some extent and there is also some evidence that some late Cretaceous dinosaurs developed this ability (Norman &
Weishampel, 1991). However, none appears to have chewed as mammals do.
A major general trend in dinosaur evolution is increase in body size (Sereno,
1997), but some of the later herbivores were as small as 30 kg, possibly consuming a class of angiosperms (owering plants) called monocots (Norman &
Weishampel, 1991; Crane et al., 1995) although Barrett & Willis (2001) nd
no evidence for this. These newly targeted angiosperms may also have defended
themselves by developing thorns and spines.
Notes to pp. 205213
301
3. Young leaves in the tropics tend not to photosynthesize immediately, delaying

their greening while they develop. They can be covered entirely with small
spines or hairs, but these are shed when the leaf starts to function (Coley,
1983). Relatively few plants seem to have this defence, even though herbivory
statistics support its efcacy against herbivores.
4. Many invertebrates bite, but some suck cell contents (Leigh, 1999).
5. This is where woody tissue excels, in fact, because its structural design sacrices
the highest possible modulus for a massive increase in toughness (Jeronimidis,
1980), so maximizing (R/E )0.5 .
6. Young leaves protected by hairs or small spines do not seem to possess much
defensive chemistry (Coley, 1983), as predicted from this stress-limited strategy.
The assumption that I make here is that plant chemical defences are not a
furtive attempt to poison animals without their sensing it, but are intended
to provide clear clues via taste receptors that poisoning would result if feeding
continued. I suspect that, as is the case in living birds, dinosaurs had few
taste buds.
7. These texture maps will eventually have to be converted to non-linear deformation (along the lines of power law relationships between stress and strain)
rather than modulus in order to reect the resistance of many raw animal (and
cooked vegetable) foods in mammalian (and human) diets more accurately.
Chapter 5 and Appendix A give an inkling of how this could be done. Other
important food properties can be mapped (Ashby, 1999), but no single chart
can show all relevant features of either dentitions or diets. The aim here is to
highlight general patterns of adaptation.
8. Taking this argument to its logical conclusion results in a viewpoint which
seems not to have any precedent: some leaves could have been sacriced by
plants to nurture the growth of insect larvae that would subsequently mature to
form part of the population of pollinators. In other words, angiosperms needed
to breed their (guild of ) pollinators. This might explain why many angiosperm
families appear to have very close ecological relationships with those of
Lepidoptera (Ehrlich & Raven, 1965, 1967). This view explains little about
extant ecosystems because butteries and moths are not now the major pollinators of plants, but perhaps they once were. In contrast, ecological accounts
focus purely on angiosperm defences against herbivory (e.g. Coley & Kursar,
1996). Why though are leaf defences so often faulty?
9. Some very small synapsids have recently been discovered, but their overall size
range appears to remain well above that of the earliest mammals.
10. It is possible that birds, which are basically dwarfed dinosaurs (Sereno, 1999),
sacriced the chance of developing mastication purely by enormous weight
reduction during their evolution. This would not allow the space necessary for
fracture-scaled postcanines.
11. The internal tissues of insects are often thought to have little fracture resistance, but Evans & Sanson (1998) provide evidence from experiments that
disagrees with this. There appears to be no information in the literature on the
mechanical properties of the internal parts of insects.
302
Notes to pp. 213227
12. It is unclear whether a heavily tanned cuticle is really so very different a mechanical obstacle to a lightly tanned one. Moore & Sanson (1995) compared the
ability of two marsupial species, the northern quoll (Dasyurus hallacatus) and
the large short-nosed bandicoot (Isoodon macrourus), to break down beetles,
nding that the former, which has the sharper teeth, was more effective.
13. Digestion of the cuticle itself requires the presence of a gut enzyme called
chitinase (Jeuniaux, 1961), the antiquity of which appears unclear.
14. I dene seeds in this section as the unit that separates from the fruit during
oral processing. It is the seed that a layman would recognize. However, it may
actually possess an outer covering that is derived not from seed tissue, but from
the fruit (from the ovary). However, there is no point to get into this. The
embryology of seeds is exceedingly complex (Corner, 1976) and mechanically
protective layers, when found, may actually develop either from the seed coat
(Corner, 1949) or fruit wall.
15. Janzen (1978) suggests in effect that these defences are parsimoniously organized
in that heavily mechanically protected seeds will be free from toxic chemicals
and vice versa.
16. The chimpanzee often compresses the pith of monocots between the lower lips
and incisors, most likely because it contains bres that, although comprising
only a small proportion of the tissue by volume, prevent fragmentation with
the molars (Wrangham et al., 1992).
17. A bat probably juices fruit to reduce the weight of tissue being processed in its
gut. These mammals have very fast gut passage rates and the type of fruit that
they concentrate on often has large quantities of nutrients (simple sugars and
free amino acids) that do not require digestion (Wendeln et al., 2000).
18. From an examination of stomach contents, Gautier-Hion (1980) concluded
that some African cercopithecine monkeys partially damage a considerable
proportion of the seeds that they swallow.
19. There could be differences due to the age of individuals and because this is
learnt behaviour.
20. Some other variables do confound this pattern, but the overall dominance of
seed size is clear in the data (Corlett & Lucas, 1990).
21. Eaglen (1984) demonstrates that New World spider monkeys have smaller
incisors than Old World colobine monkeys compared to body weight. He
argues that this reects phylogeny because otherwise highly frugivorous spider
monkeys must surely be anticipated to possess larger incisors than folivorous
leaf monkeys. The arguments given in this section suggest that this is not
always expected. The incisal size of colobines vary. According to Ungars (1996)
interpretation of Hylander (1975), the Presbytis species, which are heavy seed
destroyers, may have smaller incisors than Trachypithecus species, which are
closely related but which consume more leaves.
22. Dominy et al. (2001) offer a calculation of how few cells a blade breaks, using
a leaf tissue as an example.
23. Some New World monkeys open fruits with their canines (Kinzey & Norconk,
1990, 1993). These teeth have obvious ridges on them.
Notes to pp. 227238
303
24. It has been suggested to me that the crenulated (wrinkled) enamel seen in many
frugivorous mammals can be explained in this way.
25. Seeds are heavily protected, very often by chemistry rather than mechanics. The
choice of defence may be constrained by seed size. A seedling has to break out
of its container. Nearly all seeds display a suture or some other weak point that
allows a germinating seed, turgid from imbibing water, to produce a crack from
such a notch (e.g. Lucas et al., 1991b). Given the type of woody tissue from
which seed coverings are made, it is relatively simple to show that very small
seeds cannot have substantial mechanical protection because the weak element
would have to be so large that a seed predator could easily utilize it, rendering
the investment in the rest of the seed covering redundant (Lucas & Corlett,
1998). The critical size range of seeds appears to be of the order of 12 mm in
diameter (or maximum seed width), below which any protection is likely to
be chemical. This size range coincides with the mucosal sensitivity threshold
described in Chapter 3, which turns out to have important repercussions for
the oral treatment of seeds. So mechanical protection of seeds is probably sizelimited. However, although mechanical defence is permitted in large seeds,
many of them do not have this. Circumstantial evidence links the presence
of hard shells mainly to an oil-based food reserve (Peters, 1987). It is unclear
why seed starch needs to be defended at all if it is generally indigestible to
mammals.
26. The lower incisors here include the most lateral tooth, conventionally termed
a canine.
27. We know that the enamel crests are important because a rare condition called
amelogenesis imperfecta causes the enamel of inicted herbivores to wear as
rapidly as its other dental tissues. These individuals cannot chew properly. The
length of these crests can remain relatively constant despite considerable wear
(Fortelius, 1985).
28. Hominins used to be hominids: the old use of the word hominid is equivalent
to the new term hominin. The change is purely taxonomic and reects recent
agreement that our line from the apes does not deserve familial rank. The
reason to have taxonomy as part of biological science is that it is difcult to
get very far without being able to put a name to an organism. The problem
with hominin taxonomy is that it seems to be professional death to agree with
anyone else. Thus, while sycophancy is rare, nomenclatural stability is not.
Luckily, even though we cant always put a name to a hominin fossil, we can
usually put a face to it.
29. Better said, the second-generation replacement no longer forms to the point
of mineralization because there are many instances in mammals of tooth
germs that initiate but are resorbed very early in their development. Many
insectivores only have one generation of teeth due to this, including shrews
(Kindahl, 1959). Why? I suggest an extension of the same type of explanation
as for early mammals: a very short lifespan. Some modern primates seem to
be dwarfs. The marmosets and tamarins are the clearest example (Ford, 1980),
these primates having lost their third molars.
304
Notes to pp. 241255
30. Taking this further to involve the other premolars is more difcult. As mentioned above, there are only two premolars in catarrhine primates, but three
in cebids. The lower anterior premolar in Old World monkeys is enlarged for
sharpening against the upper canine.
31. To forestall the development of a new insult, no one is descended from a robust
australopithecine.
32. There have been many explanations for canine reduction in males during the
early evolution of hominins. I cannot review them all here, but few have been
refuted. Holloway (1967) felt that canine reduction was related to tool use.
Kinzey (1971) supposed that the canines might never have been large. There is
still no convincing fossil evidence to support this. Nothing has changed since
Delson & Andrews (1975) view that the most parsimonious suggestion is that
the last common ancestor of catarrhines had large canines in males. However,
the presence of small canines in Sahelanthropus certainly means that reduction
was very early.
33. Farrells work seems to be the principal reason why British dentists have little
interest in understanding how teeth work.
34. As an indication of the methods involved in these experiments, two large
Washington baking potatoes were cut into standardized 20-mm thick slices.
All the parts from one potato, except two reserved for raw comparison, were
placed into boiling water and slices recovered at 2-min intervals. The pieces of
the other potato were roasted for 50 min in a 200 C oven. All particles were
cooled rapidly after cooking and then placed in plastic bags to conserve their
moisture prior to testing. Their toughness and stiffness were then measured
by wedge and cylindrical compression tests respectively. It is well known that
potato tissue is not homogeneous, so all test specimens were extracted from the
same subsurface region of the tuber. Tissue that had been in contact with the
boiling water was excluded. Roasting produced a rubbery outer casing, so this
was tested separately from deeper tissue. The degree of reduction in toughness
from both cooking methods is discussed in the text.
35. The term texture here is not restricted to surface feel, as is often the case, but
includes deformational and fracture terms.
References
Abd-el-Malek, S. (1955) The part played by the tongue in mastication and deglutition. Journal of Anatomy 89: 250254.
Abler, W. L. (1992) The serrated teeth of tyrannosaurid dinosaurs, and biting
structures in other animals. Paleobiology 18: 161183.
Aerts, R. J., Barry, T. N. & McNabb, W. C. (1999) Polyphenols and agriculture:
benecial effects of proanthocyanidins in forages. Agriculture, Ecosystems and
Environment 75: 112.
Agrawal, K. R. (1999) The effect of food texture on chewing patterns in human
subjects. Ph.D. thesis. Hong Kong: University of Hong Kong.
Agrawal, K. R. & Lucas, P. W. (2003) The mechanics of the rst bite. Proceedings
of the Royal Society London series B 270: 12771282.
Agrawal, K. R., Lucas, P. W., Prinz, J. F. et al. (1997) Mechanical properties of foods
responsible for resisting food breakdown in the human mouth. Archives of
Oral Biology 42: 19.
Agrawal, K. R., Lucas, P. W., Bruce, I. C. et al. (1998) Food properties that inuence
neuromuscular activity during human mastication. Journal of Dental Research
77: 19311938.
Agrawal, K. R., Lucas, P. W. & Bruce, I. C. (2000) The effect of food fragmentation
index on mandibular closing angle in human mastication. Archives of Oral
Biology 45: 577584.
Aidos, I., Lie, O. & Espe, M. (1999) Collagen content in farmed Atlantic
salmon (Salmo salar L.). Journal of Agricultural and Food Chemistry 47: 1440
1444.
Aiello, L. & Dean, C. (1990) An Introduction to Human Evolutionary Anatomy.
London: Academic Press.
Aitchison, J. (1946) Hinged teeth in mammals: a study of the tusks of muntjacs
(Muntiacus) and Chinese water deer (Hydropotes inermis). Proceedings of the
Zoological Society London 116: 329338.
Akersten, W. A., Lowenstam, H. & Walker, A. (1984) Pigmentation of soricine
teeth: composition, ultrastructure, and function. Abstract, American Society
of Mammalogists, 64th Annual Meeting no. 153, 40.
Akersten, W. A., Lowenstam, H., Walker, A. et al. (2002) How and why do some
shrews have red teeth? Abstract, Society of Vertebrate Paleontology, 62nd Annual
Meeting p. 31A.
305
306
References
Alexander, R. M. (1985) Body support, scaling, and allometry. In: Functional Vertebrate Morphology (eds. M. Hildebrand, D. M. Bramble, K. F. Liem & D. B.
Wake), pp. 2637. Cambridge, MA: The Belknap Press of Harvard University.
(1991) Optimization of gut structure and diet for higher vertebrate herbivores.
Philosophical Transactions of the Royal Society London series B 333: 249255.
(1994) Optimum gut structure for specied diets. In: The Digestive System in
Mammals (eds. D. J. Chivers & P. Langer), pp. 5462. Cambridge: Cambridge
University Press.
(1996) Optima for Animals, 2nd edn. Princeton, NJ: Princeton University Press.
(1998) News of chews: the optimization of mastication. Nature 391: 329.
(2001) Design by numbers. Nature 412: 591.
Alroy, J. (1998) Copes rule and the dynamics of body mass evolution in North
American fossil mammals. Science 280: 731734.
Anapol, F. & Herring, S. W. (2000) Ontogeny of histochemical ber types and
muscle function in the masseter muscle of miniature swine. American Journal
of Physical Anthropology 112: 595613.
Anderson, D. J. (1976) The incidence of tooth contacts in normal mastication and
the part they play in guiding the nal stage of mandibular closure. In: Mastication (eds. D. J. Anderson & B. Matthews), pp. 237241. Bristol: Wright.
Anderson, D. J., Hannam, A. G. & Matthews, B. (1970) Sensory mechanisms in
mammalian teeth and their supporting structures. Physiological Review 50:
171195.
Anderson, D. J., Hector, M. P. & Linden, R. W. A. (1985) The possible relation
between mastication and parotid secretion in the rabbit. Journal of Physiology
364: 1929.
Andreasen, J. O. (1972) Traumatic Injuries of the Teeth. Copenhagen: Munksgaard.
Aranwela, N., Sanson, G. & Read, J. (1999) Methods of assessing leaf-fracture
properties. New Phytologist 144: 369393.
Ardran, G. M., Kemp, F. H. & Ride, W. D. L. (1958) A radiographic analysis
of mastication and swallowing in the domestic rabbit Oryctolagus cuniculus.
Proceedings of the Zoological Society London 130: 257274.
Ardrey, R. (1961) African Genesis. New York: Atheneum.
Ashby, M. F. (1989) On the engineering properties of materials. Acta Metallurgica
37: 12731293.
(1998) Checks and estimates for material properties. I. Ranges and simple correlations. Proceedings of the Royal Society London series A 454: 13011321.
(1999) Materials Selection in Mechanical Design, 2nd edn. Oxford: Butterworth
Heinemann.
Ashby, M. F. & Jones, D. R. H. (1996) Engineering Materials, vol. 1, 2nd edn.
Oxford: Butterworth Heinemann.
(1998) Engineering Materials, vol. 2, 2nd edn. Oxford: Butterworth Heinemann.
Ashby, M. F., Easterling, K. E., Harryson, R. et al. (1985) The fracture and toughness
of woods. Proceedings of the Royal Society London series A 398: 261280.
Atkins, A. G. (1974) Imparting strength and toughness to brittle composites. Nature
252: 116118.
References
307
(1980) On cropping and related processes. International Journal of Mechanical

Science 22: 215231.
(1982) Topics in indentation hardness. Metal Science 16: 127137.
(1994) Scale effects in engineering failures. Engineering Failure Analysis 1: 201
214.
(1999) Scaling laws for elastoplastic fracture. International Journal of Fracture 95:
5165.
Atkins, A. G. & Felbeck, D. K. (1974) Applying mutual indentation hardness
phenomena to service failures. Metal Engineering Quarterly 14: 5561.
Atkins, A. G. & Mai, Y.-W. (1979) On the guillotining of materials. Journal of
Materials Science 14: 27472754.
(1985) Elastic and Plastic Fracture. Chichester: Ellis Horwood.
Atkins, A. G. & Vincent, J. F. V. (1984) An instrumented microtome for improved
histological sections and the measurement of fracture toughness. Journal of
Materials Science Letters 3: 310312.
Baker, G., Jones, L. H. P. & Wardrop, I. D. (1959) Cause of wear in sheeps teeth.
Nature 184: 15831584.
Ball, T. B., Brotherson, J. D. & Gardner, J. S. (1993) A typologic and morphometric
study of variation in phytoliths from einkorn wheat (Triticum monococcum).
Canadian Journal of Botany 71: 11821192.
Balooch, M., Demos, S. G., Kinney, J. H. et al. (2001) Local mechanical and
optical properties of normal and transparent root dentin. Journal of Materials
Science: Materials in Medicine 12: 507514.
Baragar, F. A. & Osborn, J. W. (1984) A model relating patterns of human jaw
movement to biomechanical constraints. Journal of Biomechanics 17: 757
767.
Baragar, F. A., van der Glas, H. W. & van der Bilt, A. (1996) An analytic probability
density for particle-size in human mastication. Journal of Theoretical Biology
181: 169178.
Barbenel, J. C. (1972) The biomechanics of the human temporomandibular joint:
a theoretical study. Journal of Biomechanics 5: 251256.
Barnes, J. (1987) Early Greek Philosophy. London: Penguin.
Barrett, P. M. & Willis, K. J. (2001) Did dinosaurs invent owers? Dinosaur
angiosperm coevolution revisited. Biological Reviews 76: 411447.
Barthlott, W. & Neinhuis, C. (1997) Purity of the sacred lotus or escape from
contamination in biological surfaces. Planta 202: 18.
Basmajian, J. V. & De Luca, C. J. (1985) Muscles Alive, 5th edn. Baltimore, MD:
Williams & Wilkins.
Beard, K. C., Tong, Y., Dawson, M. et al. (1996) Earliest complete dentition of an
anthropoid primate from the late Middle Eocene of Shanxi province, China.
Science 272: 8285.
Beecher, R. M. (1977) Function and fusion at the mandibular symphysis. American
Journal of Physical Anthropology 47: 325336.
(1979) Functional signicance of the mandibular symphysis. Journal of Morphology 159: 117130.
308
References
Bell, E. A. (1984) Toxic compounds in seeds. In: Seed Physiology, vol. 1, Development
(ed. D. R. Murray), pp. 245264. Sydney: Academic Press.
Bell, R. H. V. (1971) A grazing ecosystem in the Serengeti. Scientic American 225:
8689.
Bemis, W. E. (1984) Morphology and growth of lepidoserenid lungsh tooth plates
(Pisces: Dipnoi). Journal of Morphology 179: 7393.
Benedict, F. G. (1938) Vital Energetics: A Study in Comparative Basal Metabolism.
Washington, DC: Carnegie Institute of Washington.
Bennett, E. L. (1983) The banded langur: ecology of a colobine in West Malaysia.
Ph.D. thesis. Cambridge: University of Cambridge.
Berkovitz, B. K. B. (2000) Tooth replacement patterns in non-mammalian vertebrates. In: Development, Function and Evolution of Teeth (eds. M. F. Teaford,
M. M. Smith & M. W. J. Ferguson), pp. 186200. Cambridge: Cambridge
University Press.
Berkovitz, B. K. B. & Poole, D. F. G. (1977) Attrition of the teeth in ferrets. Journal
of Zoology 183: 411418.
Berkovitz, B. K. B., Moxham, B. J. & Newman, H. N. (1995) The Periodontal
Ligament in Health and Disease, 2nd edn. London: Mosby-Wolfe.
Bertram, J. E. A. & Gosline, J. M. (1986) Fracture toughness design in horse hoof
keratin. Journal of Experimental Biology 125: 2947.
Bialek, W. (1987) Physical limits to sensation and perception. Annual Review of
Biophysics and Biophysical Chemistry 16: 455478.
Biewener, A. (1992) Biomechanics. Oxford: IRL Press.
Blackburn, J., Hodgskinson, R., Currey, J. D. et al. (1992) Mechanical properties
of microcallus in human cancellous bone. Journal of Orthopaedic Research 10:
237246.
Blanton, P. L., Biggs, N. L. & Perkins, R. C. (1970) Electromyographic analysis of
the buccinator muscle. Journal of Dental Research 49: 389394.
Bodmer, R. E. (1989) Frugivory in Amazonian Artiodactyla: evidence of the evolution of the ruminant stomach. Journal of Zoology 219: 457467.
Boesch, C. & Boesch, H. (1990) Tool use and tool making in wild chimpanzees.
Folia Primatologica 54: 8699.
Bond, F. C. (1952) Third theory of comminution. American Institute of Mining
Engineers Transactions 193: 484494.
(1962) The laws of rock breakage. In: Zerkleinern Symposion (ed. H. Rumpf ),
pp. 194202. Dusseldorf: Verlag Chemie.
Boughter, J. D. & Gilbertson, T. A. (1999) From channels to behavior: an integrative model of NaCl taste. Neuron 22: 213215.
Bourne, M. C. (1976) Compression rates in the mouth. Journal of Texture Studies
8: 373376.
(2002) Food Texture and Viscosity, 2nd edn. New York: Academic Press.
Bouvier, M. (1986) A biomechanical analysis of mandibular scaling in Old World
monkeys. American Journal of Physical Anthropology 69: 473482.
Bowden, F. P. & Tabor, D. (1950) The Friction and Lubrication of Solids. Oxford:
Oxford University Press.
References
309
Boyde, A. (1964) The structure and development of enamel. Ph.D. thesis. London:
University of London.
Boyde, A. & Fortelius, M. (1986) Development, structure and function of
rhinoceros enamel. Zoological Journal of the Linnean Society 87: 181214.
Brace, C. L. (1963) Structural reduction in evolution. American Naturalist 97:
3949.
(1964) The probable mutation effect. American Naturalist 98: 453455.
Brace, C. L., Rosenberg, K. & Hunt, K. D. (1987) Gradual change in human
tooth size in the late Pleistocene and post-Pleistocene. Evolution 41: 705
720.
Braden, M. (1976) Biophysics of the tooth. In: Frontiers of Oral Physiology, vol. 2
(ed. Y. Kawamura), pp. 137. Basel: Karger.
Bramble, D. M. (1978) Origin of the mammalian feeding complex: models and
mechanisms. Paleobiology 4: 271301.
Brehnan, K., Boyd, R. L., Gibbs, C. H. et al. (1981) Direct measurement of loads at
the temporomandibular joint in Macaca arctoides. Journal of Dental Research
60: 18201824.
Brennan, J. G. (1980) Food texture measurement. In: Food Analysis Techniques (ed.
R. D. King), pp. 118. London: Applied Science.
Breslin, P. A. S., Beauchamp, G. K. & Pugh, E. N. Jr (1996) Monogeusia for
fructose, glucose, sucrose, and maltose. Perception and Psychophysics 58: 327
341.
Brochu, C. (1999) Phylogenetics, taxonomy, and historical biogeography of Alligatoroidea. Journal of Vertebrate Paleontology 19 (supplement to No. 2):
9100.
Brunet, M., Guy, F., Pilbeam, D. et al. (2002) A new hominid from the Upper
Miocene of Chad, Central Africa. Nature 418: 145151.
Buckland-Wright, J. C. (1975) The structure and function of cat skull bones in
relation to the transmission of biting forces. Ph.D. thesis. London: University
of London.
Butler, P. M. (1952) The milk molars of the Perissodactyla, with remarks on molar
occlusion. Proceedings of the Zoological Society London 121: 777817.
(1978) Molar cusp nomenclature and homology. In: Development, Function and
Evolution of Teeth (eds. P. M. Butler & K. A. Joysey), pp. 439453. New York:
Academic Press.
Cachel, S. (1984) Growth and allometry in primate masticatory muscles. Archives
of Oral Biology 29: 287293.
Calcagno, J. M. & Gibson, K. R. (1991) Selective compromise: evolutionary trends
and mechanisms in hominid tooth size. In: Advances in Dental Anthropology
(eds. M. A. Kelley & C. S. Larsen), pp. 5976. New York: Wiley.
Calder, W. A. (1984) Size, Function, and Life History. Cambridge, MA: Harvard
University Press.
Carlson, D. S. (1977) Condylar translation and the function of the supercial
masseter in the rhesus monkey (M. mulatta). American Journal of Physical
Anthropology 47: 5364.
310
References
Carlsson, G. E. (1974) Bite force and masticatory efciency. In: Frontiers of Oral
Physiology, vol. 2 (ed. Y. Kawamura), pp. 265292. Basel: Karger.
Carpita, N. C. & Gibeaut, D. M. (1993) Structural models of primary cell walls in
owering plants: consistency of molecular structure with the physical properties of the walls during growth. Plant Journal 3: 130.
Chaimanee, Y., Jolly, D., Benammi, M. et al. (2003) A Middle Miocene hominoid
from Thailand and orangutan origins. Nature 422: 6165.
Chandrashekar, J., Mueller, K. L., Hoon, M. A. (2000) T2Rs function as bitter
taste receptors. Cell 100: 703711.
Chapman, L. J., Chapman, C. A. & Wrangham, R. W. (1992) Balanites wilsoniana:
elephant dependent dispersal? Journal of Tropical Ecology 8: 275283.
Charalambides, M. N., Williams, J. G. & Chakrabarti, S. (1995) A study of the
inuence of ageing on the mechanical properties of Cheddar cheese. Journal
of Materials Science 30: 39593967.
Chin-Purcell, M. V. & Lewis, J. L. (1996) Fracture of articular cartilage. Journal of
Biomechanical Engineering 118: 545556.
Choong, M. F. (1996) What makes a leaf tough and how this affects the pattern
of Castanopsis ssa leaf consumption by caterpillars. Functional Ecology 10:
668674.
(1997) Patterns of herbivory in tropical Fagaceae. Ph.D. thesis. Hong Kong:
University of Hong Kong.
Choong, M. F., Lucas, P. W., Ong, J. Y. S. et al. (1992) Leaf fracture toughness and
sclerophylly: their correlations and ecological implications. New Phytologist
121: 597610.
Christensen, M., Purslow, P. P. & Larsen, L. M. (2000) The effect of cooking
temperature on mechanical properties of whole meat, single muscle bres
and perimysial connective tissue. Meat Science 55: 301307.
Ciochon, R. L., Piperno, D. R. & Thompson, R. G. (1990) Opal phytoliths
found on the teeth of the extinct ape Gigantopithecus blacki: implications for
paleodietary studies. Proceedings of the National Academy of Science, USA 87:
81208124.
Cipollini, M. L. & Levey, D. J. (1997) Secondary metabolites of eshy vertebratedispersed fruits: adaptive hypotheses and implications for seed dispersal.
American Naturalist 150: 346372.
Cochard, L. R. (1987) Postcanine tooth size in female primates. American Journal
Coley, P. D. (1983) Herbivory and defensive characteristics of tree species in a
lowland tropical rainforest. Ecological Monographs 53: 209233.
Coley, P. D. & Kursar, T. A. (1996) Anti-herbivore defenses of young tropical leaves:
physiological constraints and ecological trade-offs. In: Tropical Forest Plant
Ecophysiology (eds. S. S. Mulkey, R. L. Chazdon & A. P. Smith), pp. 305336.
New York: Chapman & Hall.
Collinson, M. E. & Hooker, J. J. (1991) Fossil evidence of interactions between
plants and plant-eating mammals. Philosophical Transactions of the Royal Society London series B 333: 197208.
References
311
Cook, J. & Gordon, J. E. (1964) A mechanism for the control of crack propagation
in all brittle systems. Proceedings of the Royal Society London series A 282: 508
518.
Corlett, R. T. (1996) Characteristics of vertebrate-dispersed fruits in Hong Kong.
Journal of Tropical Ecology 12: 819833.
Corlett, R. T. & Lucas, P. W. (1990) Alternative seed-handling strategies
in primates: seed-spitting by long-tailed macaques. Oecologia 82: 166
171.
Cornelissen, J. H. C. (1999) A triangular relationship between leaf size and seed size
among woody species: allometry, ontogeny, ecology and taxonomy. Oecologia
118: 248255.
Corner, E. J. H. (1949) The annonaceous seed and its four integuments. New
Phytologist 48: 332364.
(1976) The Seeds of Dicotyledons. Cambridge: Cambridge University Press.
Costa, R. L. & Greaves, W. S. (1983) The pattern of wear responsible for the
formation of enamel ridges on teeth with exposed dentin. American Journal
Cottrell, A. H. (1964) The Mechanical Properties of Matter. New York: Wiley.
Cox, H. L. (1952) The elasticity and strength of paper and other brous materials.
British Journal of Physics 3: 7279.
Crane, P. R., Friis, E. M. & Pederson, K. R. (1995) The origin and early diversication of angiosperms. Nature 363: 342344.
Creighton, G. K. (1980) Static allometry of mammalian teeth and the correlation
of tooth size and body size in contemporary mammals. Journal of Zoology 191:
235243.
Critchley, H. D. & Rolls, E. T. (1996) Responses of primate taste cortex neurons
to the astringent tastant tannic acid. Chemical Senses 21: 135145.
Crompton, A. W. (1963) On the lower jaw of Diarthrognathus and the evolution
of the mammalian lower jaw. Proceedings of the Zoological Society London 140:
697753.
(1971) The origin of the tribosphenic molar. In: Early Mammals (eds. D. M.
Kermack & K. A. Kermack), pp. 6587. Supplement no. 1 to Zoological
Journal of the Linnean Society, vol. 50.
Crompton, A. W. & Hiiemae, K. M. (1970) Molar occlusion and mandibular
movements during occlusion in the American opossum (Didelphis marsupialis). Zoological Journal of the Linnean Society 49: 2147.
Crompton, A. W. & Kielan-Jaworowska, Z. A. (1978) Molar structure and occlusion in Cretaceous therian mammals. In: Development, Function and Evolution
of Teeth (eds. P. M. Butler & K. A. Joysey), pp. 249288. New York: Academic
Press.
Crompton, A. W. & Sita-Lumsden, A. (1970) Functional signicance of the therian
molar pattern. Nature 222: 678679.
Curran, L. M. & Leighton, M. (2000) Vertebrate responses to spatio-temporal
variation in seed production by mast-fruiting Bornean Dipterocarpaceae.
Ecological Monographs 70: 121150.
312
References
Currey, J. D. (1967) The failure of exoskeletons and endoskeletons. Journal of

Morphology 123: 116.
(1980) Mechanical properties of mollusc shell. In: The Mechanical Properties
of Biological Materials (eds. J. F. V. Vincent & J. D. Currey), pp. 7597.
Cambridge: Cambridge University Press.
Currey, J. D. & Brear, K. (1992) Fractal analysis of compact bone and antler fracture
surfaces. Biomimetics 1: 102118.
Cuy, J. L., Mann, A. B., Livi, K. J. et al. (2002) Nanoindentation mapping of the
mechanical properties of human molar enamel. Archives of Oral Biology 47:
281291.
DArcy Thompson, W. (1961) On Growth and Form. Cambridge: Cambridge
University Press.
Dale, H. E., Shanklin, M. D., Johnson, H. D. et al. (1970) Energy metabolism of
the chimpanzee. In: The Chimpanzee, vol. 2 (ed. G. H. Bourne), pp. 100122.
Basel: Karger.
Damuth, J. (1990) Problems in estimating body masses of archaic ungulates
using dental measurements. In: Body Size in Mammalian Paleobiology (ed.
J. Damuth & B. J. MacFadden), pp. 229253. Cambridge: Cambridge
University Press.
Danielson, D. R. & Reinhard, K. J. (1998) Human dental microwear
caused by calcium oxalate phytoliths in prehistoric diet of the Lower
Pecos region, Texas. American Journal of Physical Anthropology 107: 297
304.
Darvell, B. W. (1990) On uniaxial compression tests and the validity of indirect
tensile strength. Journal of Materials Science 25: 757780.
Darvell, B. W., Lee, P. K. D., Yuen, T. D. B. et al. (1996) A portable fracture
toughness tester for biological materials. Measurement Science and Technolology
7: 954962.
Davies, A. G. & Baillie, I. C. (1988) Soil-eating by red-leaf monkeys (Presbytis
rubicunda) in Sabah, northern Borneo. Biotropica 20: 252258.
Davies, A. G., Bennett, E. L. & Waterman, P. G. (1988) Food selection by
two South-east Asian colobine monkeys (Presbytis rubicunda and Presbytis
melalophos) in relation to plant chemistry. Biological Journal of the Linnean
Society 34: 3356.
de Bruijne, D. W., Hendricks, H. A. C. M., Anderliesten, L. et al. (1993) Mouthfeel
of foods. In: Food Colloids and Polymers: Stability and Mechanical Properties
(eds. E. Dickinson & P. Walstra), pp. 204213. Cambridge: Royal Society of
Chemistry.
Dean, M. C. (2000) Incremental markings in enamel and dentine: what they can
tell us about the way teeth grow. In: Development, Function and Evolution of
Teeth (eds. M. F. Teaford, M. M. Smith & M. W. J. Ferguson), pp. 119130.
Dean, M. C., Jones, M. E. & Pilley, J. R. (1992) The natural history of tooth wear,
continuous eruption and periodontal disease in wild shot great apes. Journal
of Human Evolution 16: 2329.
References
313
Dechow, P. C. & Carlson, D. S. (1990) Occlusal force and craniofacial biomechanics during growth in rhesus monkeys. American Journal of Physical Anthropology 83: 219237.
Dechow, P. C. & Hylander, W. L. (2000) Elastic properties and masticatory bone
stress in the macaque mandible. American Journal of Physical Anthropology
112: 553574.
DeGusta, D., Everett, M. A. & Milton, K. (2003) Natural selection on molar size
in a wild population of howler monkeys (Alouatta palliata). Proceedings of the
Royal Society London series B (Supplement: Biological Letters). 0b10005.S1.
Delgado, S., Casane, D., Bonnaud, L. et al. (2001) Molecular evidence for preCambrian origin of amelogenin, the major protein of enamel. Molecular
Biology and Evolution 18: 21462153.
DeLong, R. & Douglas, W. H. (1983) Development of an articial oral environment
for the testings of dental restorations: bi-axial and movement control. Journal
of Dental Research 62: 3236.
Delson, E. G. & Andrews, P. J. (1975) Evolution and interrelationships of the
catarrhine primates. In: Phylogeny of the Primates: A Multidisciplinary Approach
(eds. W. P. Luckett & F. S. Szalay), pp. 405446. New York: Plenum.
Delson, E., Terranova, C. J., Jungers, W. J. et al. (2000) Body mass in Cercopithecidae (Primates, Mammalia): estimation and scaling in extinct and extant taxa. American Museum of Natural History, Anthropological Papers 83: 1
159.
Demes, B. & Creel, N. (1988) Bite force, diet, and cranial morphology of fossil
hominids. Journal of Human Evolution 17: 657670.
Demment, M. W. & Van Soest, P. J. (1985) A nutritional explanation for body-size
patterns of ruminant and nonruminant herbivores. American Naturalist 125:
641672.
Deutsch, D., Palmon, A., Dafni, L. et al. (1995) The enamelin (tuftelin) gene.
International Journal of Developmental Biology 39: 135143.
Diekwisch, T. G. H. (1998) Subunit compartments of secretory stage enamel matrix. Connective Tissue Research 38: 101111.
Diekwisch, T. G. H., Berman, B. S., Gentner, S. et al. (1995) Initial enamel
crystals are spatially not associated with mineralised dentine. Cell and Tissue
279: 149167.
Dixon, A. D. (1963) Nerve plexuses in the oral mucosae. Archives of Oral Biology
8: 435447.
Dobrin, P. B. & Doyle, J. M. (1970) Vascular smooth muscle and the anisotropy
of dog carotid artery. Circulation Research 27: 105119.
Dominy, N. J. (2001) Trichromacy and the ecology of food selection in four African
primates. Ph.D. thesis. Hong Kong: University of Hong Kong.
(2003) Color as an indicator of food quality to anthropoid primates: ecological
evidence and an evolutionary scenario. In: Anthropoid Origins: New Visions
(eds. C. Ross & R. F. Kay), pp. 599628. New York: Kluwer.
Dominy, N. J., Lucas, P. W., Osorio, D. et al. (2001) The sensory ecology of primate
food perception. Evolutionary Anthropology 10: 171186.
314
References
Dorrington, K. L. (1980) The theory of viscoelasticity in biomaterials. In: The

Mechanical Properties of Biomaterials (eds. J. F. V. Vincent & J. D. Currey),
pp. 289314. Cambridge: Cambridge University Press.
Douglas, W. H., Sakaguchi, R. L. & DeLong, R. (1985) Frictional effects between
natural teeth in an articial mouth. Dental Materials 1: 115119.
Drake, B. K. (1963) Food crushing sounds: an introductory study. Journal of Food
Science 28: 233241.
(1965) Food crushing sounds: comparisons of objective and subjective data.
Journal of Food Science 30: 556559.
Druzinsky, R. E. (1993) The time allometry of mammalian chewing movements:
chewing frequency scales with body mass in mammals. Journal of Theoretical
Biology 160: 427440.
Dunbar, R. I. M. (1977) Feeding ecology of gelada baboons: a preliminary report. In: Primate Ecology (ed. T. H. Clutton-Brock), pp. 251273. London:
Academic Press.
Dyment, M. L. & Synge, J. L. (1935) The elasticity of the periodontal membrane.
Oral Health 25: 105109.
Eaglen, R. H. (1984) Incisor size and diet revised: the view from a platyrrhine
perspective. American Journal of Physical Anthropology 69: 262275.
Edgar, W. M. & OMullane, D. M. (1996) Saliva and Dental Health, 2nd edn.
London: British Dental Association.
Edwards, C., Read, J. & Sanson, G. (2000) Characterising sclerophylly: some
mechanical properties of leaves from heath and forest. Oecologia 123: 158
167.
Ehrlich, P. & Raven, P. H. (1965) Butteries and plants: a study in coevolution.
Evolution 18: 586608.
(1967) Butteries and plants. Scientic American 216: 104113.
Emerson, S. B. & Radinsky, L. (1980) Functional analysis of sabertooth cranial
morphology. Paleobiology 6: 295312.
Emmons, L. H. (1991) Frugivory in treeshrews (Tupaia). American Naturalist 138:
642649.
(2000) Tupai: A Field Study of Bornean Treeshrews. Berkeley, CA: University of
California Press.
Emmons, L. H., Nais, J. & Biun, A. (1991) The fruit and dispersers of Rafesia
keithii (Rafesiaceae). Biotropica 23: 197199.
Epstein, B. (1947) The mathematical description of certain breakage functions leading to the logarithmico-normal distribution. Journal of the Franklin Institute
244: 471477.
Eriksson, O., Friis, E. M. & Lofgren, P. (2000) Seed size, fruit size, and dispersal systems in angiosperms from the early Cretaceous to the late Tertiary.
Evans, A. R. & Sanson, G. D. (1998) The effect of tooth shape on the breakdown
of insects. Journal of Zoology 246: 391400.
(2003) The tooth of perfection: functional and spatial constraints on mammalian
tooth shape. Biological Journal of the Linnean Society 78: 173191.
References
315
Every, R. F. (1970) Sharpness of teeth in man and other primates. Postilla 143:
120.
Farrell, J. (1956) The effect of mastication on the digestion of food. British Dental
Journal 100: 149155.
Fengel, D. & Wegener, G. (1989) Wood: Chemistry, Ultrastructure, Reactions. Berlin:
Walter de Gruyter.
Fincham, A. G., Luo, W., Morodian-Oldak, J. et al. (2000) Enamel biomineralization: the assembly and disassembly of the extracellular organic
matrix. In: Development, Function and Evolution of Teeth (eds. M. F. Teaford,
M. M. Smith & M. W. J. Ferguson), pp. 3761. Cambridge: Cambridge
University Press.
Finney, D. (1971) Probit Analysis, 3rd edn. Cambridge: Cambridge University Press.
Fish, D. R. & Mendel, F. C. (1982) Mandibular movement patterns relative to
food types in common tree shrews (Tupaia glis). American Journal of Physical
Fleagle, J. G. (1999) Primate Adaptation and Evolution, 2nd edn. San Diego, CA:
Academic Press.
Fooden, J. & Albrecht, G. H. (1993) Latitudinal and insular variation of skull
size in crab-eating macaques (Primates, Cercopithecidae: Macaca fascicularis).
American Journal of Physical Anthropology 92: 521538.
Ford, S. M. (1980) Callithricids as phyletic dwarfs, and the place of Callithricidae
in Platyrrhini. Primates 21: 3143.
Fortelius, M. (1985) Ungulate cheek teeth: developmental, functional and evolutionary interrelationships. Acta Zoologica Fennica 180: 176.
(1990) Problems with using fossil teeth to estimate body sizes of extinct
mammals. In: Body Size in Mammalian Paleobiology (eds. J. Damuth
& B. J. MacFadden), pp. 207228. Cambridge: Cambridge University
Press.
Fox, C. L., Juan, J. & Albert, R. M. (1996) Phytolith analysis on dental calculus,
enamel surface, and burial soil: information about diet and paleoenvironment.
Fox, P. G. (1980) The toughness of tooth enamel, a natural brous composite.
Journal of Materials Science 15: 31133121.
Frank, F. C. & Lawn, B. R. (1967) On the theory of Hertzian fracture. Proceedings
of the Royal Society London series A 299: 291316.
Freeman, P. W. (1979) Specialized insectivory: beetle-eating and moth-eating
molossid bats. Journal of Mammalogy 60: 467479.
(1981a) A multivariate study of the family Molossidae (Mammalia, Chiroptera):
morphology, ecology and evolution. Fieldiana Zoology (new series) 7: 1173.
(1981b) Correspondence of food habits and morphology in insectivorous bats.
Journal of Mammalogy 62: 166173.
(1988) Frugivorous and animalivorous bats (Microchiroptera): dental and cranial
adaptations. Biological Journal of the Linnean Society 33: 249272.
(1992) Canine teeth of bats (Microchiroptera): size, shape and role in crack
propagation. Biological Journal of the Linnean Society 45: 97115.
316
References
(2000) Macroevolution in Microchiroptera: recoupling morphology and ecology with phylogeny. Evolutionary Ecology Research 2: 317335.
Freeman, P. W. & Weins, W. N. (1997) Puncturing ability of bat canine teeth:
the tip. In: Life among the Muses: Papers in Honor of James S. Findley (eds.
T. L. Yates, W. L. Gannon & D. E. Wilson), pp. 225232. Albuquerque, NM:
Museum of Southwestern Biology, University of New Mexico.
Frisch, J. E. (1963) Sexual dimorphism in canines of Hylobates lar. Primates 4:
110.
Frolich, L. M., LaBarbera, M. & Stevens, W. P. (1994) Poissons ratio of a crossed
ber sheath: the skin of aquatic salamanders. Journal of Zoology 232: 231252.
Gabbott, S. E., Aldridge, R. J. & Theron, J. N. (1995) A giant conodont with
preserved muscle tissue from the Upper Ordovician of South Africa. Nature
374: 800803.
Galileo, G. (1638) Dialogues concerning the Two Sciences. Translated (1914) by H.
Crew & A. de Salvico. Evanston, IL: Northwestern University Press.
Gantt, D. G. & Rafter, J. A. (1998) Evolutionary and functional signicance of
hominoid tooth enamel. Connective Tissue Research 39: 195206.
Gardner, R. P. & Austin, L. G. (1962) A chemical engineering treatment of batch
grinding. In: Zerkleinern Symposion (ed. H. Rumpf ), pp. 217248. Dusseldorf:
Verlag Chemie.
Garn, S. M. & Lewis, A. B. (1962) The relationship between third molar agenesis
and a reduction in tooth number. Angle Orthodontist 33: 1418.
Gaudin, A. M. & Meloy, T. P. (1962) Model and a comminution distribution equation for single fracture. American Institute of Mining Engineers Transactions
233: 4143.
Gautier-Hion, A. (1980) Seasonal variations of diet related to species and sex in
a community of Cercopithecus monkeys. Journal of Animal Ecology 49: 237
269.
Gautier-Hion, A., Duplantier, J. M., Quris, R. et al. (1985) Fruit characters as a basis
of fruit choice and seed dispersal in a tropical forest vertebrate community.
Oecologia 65: 324337.
Gibbs, C. H., Mahan, P. E., Lundeen, H. C. et al. (1981) Occlusal forces during chewing and swallowing as measured by sound transmission. Journal of
Prosthetic Dentistry 46: 443449.
Gibbs, S., Collard, M. & Wood, B. (2000) Soft tissue characters in higher primate
phylogenetics. Proceedings of the National Academy of Sciences, USA 97: 11130
11132.
Gibson, L. J. (1985) The mechanical behaviour of cancellous bone. Journal of
Biomechanics 18: 317328.
Gibson, L. J. & Ashby, M. F. (1999) Cellular Solids Structure and Properties, 2nd
edn. Cambridge: Cambridge University Press.
Gibson, L. J., Easterling, K. E. & Ashby, M. F. (1981) The structure and mechanics
of cork. Proceedings of the Royal Society London series A 377: 99117.
Gibson, L. J., Ashby, M. F. & Easterling, K. E. (1988) The structure and mechanics
of the iris leaf. Journal of Materials Science 23: 30413048.
References
317
Gilbertson, T. A. (1998) Gustatory mechanisms for the detection of fat. Current

Opinion in Neurobiology 8: 447452.
Gilbertson, T. A., Liu, L., York, D. A. et al. (1998) Dietary fat preferences are
inversely correlated with peripheral gustatory fatty acid sensitivity. Annals of
the New York Academy of Sciences 855: 165168.
Gipps, J. M. & Sanson, G. D. (1984) Mastication and digestion in Pseudocheirus.
In: Possums and Gliders (eds. A. P. Smith & I. D. Hume), pp. 237246. Sydney:
Australian Mammal Society.
Glantz, P.-O. (1970) The surface tension of saliva. Odontologisk Revy 21: 119127.
Gordon, I. J. & Illius, A. W. (1994) The nutritional ecology of African ruminants:
a reinterpretation. Journal of Animal Ecology 65: 1828.
Gordon, J. E. (1978) Structures. London: Penguin.
(1980) Biomechanics: the last stronghold of vitalism. In: The Mechanical Properties of Biomaterials (eds. J. F. V. Vincent & J. D. Currey), pp. 111. Cambridge:
Cambridge University Press.
(1991) The New Science of Strong Materials, 2nd edn. London: Penguin.
Gordon, J. E. & Jeronimidis, G. (1980) Composites with high work of fracture.
Philosophical Transactions of the Royal Society London series A 294: 545550.
Gordon, K. D. (1984) A study of microwear on chimpanzee molars: implications
for dental microwear analysis. Journal of Dental Research 63: 195215.
Gosline, J. M., Lillie, M., Carrington, E. et al. (2002) Elastic proteins: biological
roles and mechanical properties. Philosophical Transactions of the Royal Society
London Series B 357: 121132.
Gould, S. J. (1966) Allometry and size in ontogeny and phylogeny. Biological
Reviews 41: 587640.
(1975) On the scaling of tooth size in mammals. American Zoologist 15: 351
362.
Grajal, A., Strahl, S. D., Parra, R. et al. (1989) Foregut fermentation in the hoatzin,
a neotropical leaf-eating bird. Science 245: 12361238.
Greaves, W. S. (1973) The inference of jaw motion from tooth wear facets. Journal
of Paleontology 47: 10001001.
(1978) The jaw-lever system in ungulates: a new model. Journal of Zoology 184:
271285.
Green, B. G. (1993) Oral astringency: a tactile component of avor. Acta Psychologica 84: 119125.
Greenberg, A. R., Mehling, A., Lee, M. et al. (1989) Tensile behaviour of grass.
Journal of Materials Science 24: 25492554.
Grifth, A. A. (1920) Phenomena of rupture and ow in solids. Philosophical
Transactions of the Royal Society London series A 221: 163198.
Grine, F. E. (1981) Trophic differences between gracile and robust
australopithecines: a scanning electron microscope analysis of occlusal events.
South African Journal of Science 77: 203230.
Groves, C. P. & Napier, J. R. (1968) Dental dimensions and diet in australopithecines. In: Proceedings of the 8th Congress of Anthropological and Ethnological
Sciences, vol. 3, pp. 273276. Tokyo: Science Council of Japan.
318
References
Gurney, C. & Hunt, J. (1967) Quasi-static crack propagation. Proceedings of the

Royal Society London series A 299: 508524.
Habelitz, S., Marshall, S. J., Marshall, G. W. Jr et al. (2001) The functional width
of the dentino-enamel junction determined by AFM-based nanoscratching.
Journal of Structural Biology 135: 294301.
Habelitz, S., Marshall, G. W. Jr, Balooch, M. et al. (2002) Nanoindentation and
storage of teeth. Journal of Biomechanics 35: 995998.
Haines, D. J., Berry, D. C. & Poole, D. F. G. (1963) Behavior of tooth enamel
under load. Journal of Dental Research 42: 885888.
Hamilton, G. M. & Goodman, L. E. (1966) The stress eld created by a circular
sliding contact. Journal of Applied Mechanics 33: 371376.
Hankins, G. A. (1925) Report on the effects of adhesion between the indenting tool and the material in ball and cone indentation hardness tests. Proceedings of the Institution of Mechanical Engineers 1: 611
645.
Harris, B. (1980) The mechanical behaviour of composite materials. In: The
Mechanical Properties of Biological Materials (eds. J. F. V. Vincent & J. D.
Currey), pp. 3774. Cambridge: Cambridge University Press.
Harrison, M. J. S. (1986) Feeding ecology of black colobus (Colobus satanus) in
central Gabon. In: Primate Ecology and Conservation (eds. J. G. Else & P. C.
Lee), pp. 3137, Cambridge: Cambridge University Press.
Harvey, P. H., Kavanagh, M. & Clutton-Brock, T. H. (1978) Sexual dimorphism
of primate teeth. Journal of Zoology 186: 475485.
Hatley, T. & Kappelman, J. (1980) Bears, pigs and Plio-Pleistocene hominids: a
case for the exploitation of belowground resources. Human Ecology 8: 371
390.
Hayes, V. J., Freedman, L. & Oxnard, C. E. (1995) The differential expression
of dental sexual dimorphism in subspecies of Colobus guereza. International
Journal of Primatology 16: 971996.
Heath, M. R. (1982) The effect of maximum biting force and bone loss upon
masticatory function and dietary selection of the elderly. International Dental
Journal 32: 345356.
(1986) Functional interpretation of patterns of occlusal wear on acrylic teeth.
Restorative Dentistry 2: 100107.
Hector, M. P. & Linden, R. W. A. (1999) Reexes of salivary secretion. In: Neural
Mechanism of Salivary Gland Secretion (eds. J. R. Garrett, J. Ekstrom & L. C.
Anderson), pp. 196217. Basel: Karger.
Helkimo, E., Carlson, G. E. & Helkimo, M. (1978) Chewing efciency and
state of dentition: a methodological study. Acta Odontologica Scandinavica
36: 3341.
Hellekant, G., Ninomiya, Y., DuBois, G. E. et al. (1996) Taste in chimpanzee.
I. The summated response to sweeteners and the effect of gymnemic acid.
Physiology and Behavior 60: 469479.
Hendrichs, H. (1965) Vergleichende Untersuchung des Wiederkauverhalten. Biologisches Zentralblatt 84: 651721.
References
319
Herrera, C. M. (2002) Seed dispersal by vertebrates. In: PlantAnimal Interactions:

An Evolutionary Approach (eds. C. M. Herrera & O. Pellmyr), pp. 185208.
Oxford: Blackwell.
Herring, S. W. (1972) The role of canine morphology in the evolutionary divergence
of pigs and peccaries. Journal of Mammalogy 53: 500512.
(1975) Adaptations for gape in the hippopotamus and its relatives. Forma e
Functio 8: 85100.
Herring, S. W. & Herring, S. E. (1974) The supercial masseter and gape in
mammals. American Naturalist 108: 561575.
Heyes, J. A. & Sealey, D. F. (1996) Textural changes during nectarine (Prunus
persica) development and ripening. Scientia Horticulturae 65: 4958.
Hiiemae, K. M. (1978) Mammalian mastication: a review of the activity of the
jaw muscles and the movements they produce in chewing. In: Development,
Function and Evolution of Teeth (eds. P. M. Butler & K. A. Joysey), pp. 359398.
New York: Academic Press.
Hiiemae, K. M. & Ardran, G. M. (1968) A cineradiographic study of feeding in
Rattus norvegicus. Journal of Zoology 154: 139154.
Hiiemae, K. M. & Crompton, A. W. (1985) Mastication, food transport, and
swallowing. In: Functional Vertebrate Morphology (eds. M. Hildebrand, D. M.
Bramble, K. F. Liem & D. B. Wake), pp. 262290. Cambridge, MA: Belknap
Press of Harvard University.
Hiiemae, K. M. & Palmer, J. B. (1999) Food transport and bolus formation during
complete sequences on foods of different initial consistency. Dysphagia 14:
3142.
Hill, A. V. (1950) The dimensions of animals and their muscular dynamics. Scientic
Progress 38: 209230.
Hill, D. A. & Lucas, P. W. (1996) Toughness and ber content of major leaf foods
of wild Japanese macaques (Macaca fuscata yakui) in Yakushima. American
Journal of Primatology 38: 221231.
Hill, D. A., Lucas, P. W. & Cheng, P. Y. (1995) Bite forces used by Japanese
macaques (Macaca fuscata yakui) on Yakushima Island, Japan to open aphidinduced galls on Distylium racemosum (Hamamelidaceae). Journal of Zoology
237: 5763.
Hiller, S., Bruce, D. M. & Jeronimidis, G. (1996) A micropenetration technique for
mechanical testing of plant cell walls. Journal of Texture Studies 27: 559587.
Hillerton, J. E., Reynolds, S. E. & Vincent, J. F. V. (1982) On the indentation
hardness of insect cuticle. Journal of Experimental Biology 96: 4552.
Hills, M., Graham, S. H. & Wood, B. A. (1983) The allometry of relative cusp size
in hominoid mandibular molars. American Journal of Physical Anthropology
62: 311316.
Hoaglund, R. G., Rosenfeld, A. R. & Hahn, G. T. (1972) Mechanisms of fast
fracture and arrest in steels. Metallurgical Transactions 3: 123136.
Holland, G. R. (1994). Morphological features of dentine and pulp related to dentine sensitivity. Archives of Oral Biology 39 (supplement):
3S11S.
320
References
Holloway, R. L. (1967) Tools and teeth: some speculations regarding canine reduction. American Anthropologist 69: 6367.
Hopson, J. A. (2001) Origin of mammals. In Palaeobiology, vol. 2 (eds. D. E. G.
Briggs & P. R. Crowther), pp. 8894. Oxford: Blackwell.
Howe, H. F. (1989) Scatter- and clump-dispersal and seedling demography: hypothesis and implications. Oecologia 79: 417426.
Hrycyshyn, A. W. & Basmajian, J. V. (1972) Electromyography of the oral stage
of swallowing in man. American Journal of Anatomy 133: 333340.
Hume, I. D. (1994) Gut morphology, body size and digestive performance in
rodents. In: The Digestive System in Mammals (eds. D. J. Chivers & P. Langer),
Hunter, J. P. & Jernvall, J. (1995) The hypocone as a key innovation in mammalian
evolution. Proceedings of the National Academy of Sciences, USA 92: 10718
10722.
Hutchings, J. B. & Lillford, P. J. (1988) The perception of food texture: the philosophy of the breakdown path. Journal of Texture Studies 19: 103115.
Hylander, W. L. (1975) Incisor size and diet in anthropoids with special reference
to Cercopithecidae. Science 189: 10951097.
(1977) Morphological changes in human teeth and jaws in a high-attrition
environment. In: Orofacial Growth and Development (eds. A. A. Dahlberg &
T. M. Graber), pp. 301330. The Hague: Mouton.
(1978) Incisal bite force direction in humans and the functional signicance of
mammalian mandibular translation. American Journal of Physical Anthropology 48: 18.
(1979a) Mandibular function in Galago crassicaudatus and Macaca fascicularis:
an in vivo approach to stress analysis of the mandible. Journal of Morphology
159: 253296.
(1979b) An experimental analysis of temporomandibular joint reaction force in
macaques. American Journal of Physical Anthropology 51: 433456.
(1985) Mandibular function and biomechanical stress and scaling. American
Zoologist 25: 315330.
(1992) Functional anatomy. In: The Temporomandibular Joint: A Biological Basis
for Clinical Practice. (eds. B. G. Sarnat & W. B. Laskin), pp. 6092. Philadelphia, PA: W. B. Saunders.
Hylander, W. L., Ravosa, M. J., Ross, C. F. et al. (2000) Symphyseal fusion and
jaw-adductor muscle force: an EMG study. American Journal of Physical Anthropology 112: 469492.
Illius, A. W. & Gordon, I. J. (1987) The allometry of food intake in grazing
ruminants. Journal of Animal Ecology 56: 989999.
Inglis, C. E. (1913) Stresses in a plate due to the presence of cracks and sharp
corners. Transactions of the Institution of Naval Architects 55: 219230.
Iwamoto, T. (1979) Feeding ecology. In: Ecological and Social Studies of Gelada
Baboons (ed. M. Kawai), pp. 279335. Basel: Karger.
Jablonski, N. G. (1993) Theropithecus: The Life and Death of a Primate Genus.
References
321
Jablonski, N. G. & Crompton, R. H. (1994) Feeding behavior, mastication and

tooth wear in the Western tarsier (Tarsius bancanus). International Journal of
Primatology 15: 2959.
Jackson, A. P., Vincent, J. F. V. & Turner, R. M. (1988) The mechanical design of
nacre. Proceedings of the Royal Society London series B 234: 415440.
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),
Janis, C. M. & Ehrhardt, D. (1988) Correlation of relative muzzle width and relative
incisor width with dietary preference in ungulates. Zoological Journal of the
Linnean Society 92: 267284.
Janis, C. M. & Fortelius, M. (1988) On the means whereby mammals achieve
increased functional durability of their dentitions, with special reference to
limiting factors. Biological Reviews 63: 197230.
Janson, C. H. (1983) Adaptation of fruit morphology to dispersal agents in a
neotropical forest. Science 219: 187189.
Janvier, P. (1996). Fishy fragments tip the scales. Nature 383: 757758.
Janzen, D. H. (1978) The ecology and evolutionary biology of seed chemistry as
relates to seed predation. In: Biochemical Aspects of Plant and Animal Evolution
(ed. J. B. Harborne), pp. 163206. London: Academic Press.
Janzen, D. H. & Martin, P. S. (1982) Neotropical anachronisms: the fruits the
gomphotheres ate. Science 215: 1927.
Jarman, P. J. (1974) The social organization of antelope in relation to their ecology.
Behaviour 48: 215266.
Jeannorod, M. (1997) The Cognitive Neuroscience of Action. Cambridge, MA: Blackwell.
Jenkins, G. N. (1978) The Physiology and Biochemistry of the Mouth, 4th edn.
Oxford: Blackwell.
Jennings, J. S. & Macmillan, N. H. (1986) A tough nut to crack. Journal of Materials
Science 21: 15171524.
Jensen, J. L., Lamkin, M. S. & Oppenheim, F. G. (1992) Adsorption of human
salivary proteins to hydroxyapatite: a comparison between whole saliva and
glandular salivary secretions. Journal of Dental Research 71: 15691576.
Jernvall, J. & Selanne, L. (1999) Laser confocal microscopy and geographical
information systems in the study of dental morphology. Palaeontologia
Electronica 2(1): 18 pp. http://www-odp.tamu.edu/paleo/1999 1/confocal/
issue1 99.htm.
Jernvall, J. & Thesleff, I. (2000) Reiterative signaling and patterning during mammalian tooth morphogenesis. Mechanisms of Development 92: 1929.
Jernvall, J., Hunter, J. P. & Fortelius, M. (2000) Trends in the evolution of molar
crown types in ungulate mammals: evidence from the northern hemisphere.
In: Development, Function and Evolution of Teeth (eds. M. F. Teaford, M. M.
Smith & M. Ferguson), pp. 269281. Cambridge: Cambridge University
Press.
322
References
Jeronimidis, G. (1980) The fracture behaviour of wood and the relations between
toughness and morphology. Proceedings of the Royal Society London series B
208: 447460.
Jeuniaux, C. (1961) Chitinase: an addition to the list of hydrolases in the digestive
tract of vertebrates. Nature 192: 135136.
Ji, Q., Luo, Z.-X., Wible, J. R. et al. (2002) The earliest known eutherian mammal.
Nature 416: 816822.
Jolly, C. J. (1970) The seed-eaters: a new model of hominid differentiation. Man
(new series) 5: 126.
Jones, E. G., Schwark, H. D. & Callahan, P. A. (1986) Extent of the ipsilateral representation in the ventral posterior medial nucleus of the monkey thalamus.
Experimental Brain Research 63: 310320.
Jones, S. J. (1981). Cement. In: A Companion to Dental Studies (eds. A. H. R. Rowe
& R. B. Johns), vol. 2, Dental Anatomy and Embryology (ed. J. W. Osborn),
pp. 193205. Oxford: Blackwell.
Jones, S. J. & Boyde, A. (1974) Coronal cementogenesis in the horse. Archives of
Jordano, P. (1995) Angiosperm eshy fruits and seed dispersers: a comparative
analysis of adaptation and constraints in plant-animal interaction. American
Naturalist 145: 163191.
Jungers, W. E. (1978) On canine reduction in early hominids. Current Anthropology
19: 155156.
Junqueira, L. C., Toledo, A. M. & Doine, A. L. (1973) Digestive enzymes in the
parotid and submandibular glands of mammals. Anais da Academia Brasileira
de Ciencias 45: 629643.
Kasapi, M. D. & Gosline, J. M. (1997) Complexity and fracture control
in the equine hoof wall. Journal of Experimental Biology 200: 1639
1659.
Kastelic, J. & Baer, E. (1980) Deformation in tendon collagen. In: The Mechanical
Properties of Biological Materials (eds. J. F. V. Vincent & J. D. Currey), pp. 397
435. Cambridge: Cambridge University Press.
Kawamura, Y. & Yamamoto, T. (1978) Studies on neural mechanisms
of the gustatorysalivary reex in rabbits. Journal of Physiology 285:
3547.
Kawasaki, K. & Weiss, K. M. (2003) Mineralized tissue and vertebrate evolution:
the secretory calcium-binding phosphoprotein gene cluster. Proceedings of the
National Academy of Sciences, USA 100: 40604065.
Kay, R. F. (1975a) Allometry and early hominids. Science 189: 63.
(1975b) The functional adaptations of primate molar teeth. American Journal of
Physical Anthropology 42: 195215.
(1978) Molar structure and diet in extant Cercopithecidae. In: Development,
Function and Evolution of Teeth (eds. P. M. Butler & K. A. Joysey), pp. 309
340. New York: Academic Press.
(1981) The nut crackers: a new theory of the adaptations of the Ramapthecinae.
References
323
Kay, R. F. & Covert, H. H. (1984) Anatomy and behavior of extinct primates. In:
Food Acquisition and Processing in Primates (eds. D. J. Chivers, B. A. Wood
& A. Bilsborough), pp. 467508. New York: Plenum.
Kay, R. F. & Hylander, W. L. (1978) The dental structure of mammalian folivores with special reference to Primates and Phalangeroidea (Marsupialia).
In: The Biology of Arboreal Folivores (ed. G. G. Montgomery), pp. 173192.
Washington, DC: Smithsonian Institution Press.
Kay, R. N. B. (1987) Weights of salivary glands in ruminant animals. Journal of
Zoology 211: 431436.
Kay, R. N. B., Engelhardt, W. V. & White, R. G. (1980) Digestive physiology
of wild ruminants. In: Digestive Physiology and Metabolism in Ruminants:
Proceedings of the 5th International Symposium on Ruminant Physiology
(eds. Y. Ruckebusch & P. Thivend), pp. 743761. Lancaster, PA: MTP
Press.
Kayser, A. (1980) Shortened dental arches and oral function. Journal of Oral Rehabilitation 8: 457462.
Kemp, A. (1995) The Hornbills. Oxford: Oxford University Press.
Kendall, K. (1978a) Complexities of compression failure. Proceedings of the Royal
Society London series A 361: 245263.
(1978b) The impossibility of comminuting small particles by compression. Nature 272: 710711.
(2001) Molecular Adhesion. New York: Kluwer.
Kermack, K. A. & Haldane, J. B. S. (1950) Organic correlation and allometry.
Biometrika 37: 3041.
Keyser, A. W. (2000) The Drimolen skull: the most complete australopithecine
cranium and mandible to date. South African Journal of Science 96: 189193.
Kier, W. M. & Smith, K. K. (1985) Tongues, tentacles and trunks: the biomechanics of movement in muscular-hydrostats. Zoological Journal of the Linnaean
Society 83: 307324.
Kiltie, R. A. (1982) Bite force as a basis for niche differentiation between rain forest
peccaries. Biotropica 14: 188195.
Kindahl, M. (1959) Some aspects of the tooth development in the Soricidae. Acta
Odontologica Scandinavica 17: 203237.
Kinney, J. H., Balooch, M., Marshall, S. J. et al. (1996) Hardness and Youngs
modulus of human peritubular and intertubular dentine. Archives of Oral
Biology 41: 913.
Kinzey, W. G. (1971) Evolution of the human canine tooth. American Anthropologist
73: 680694.
(1972) Canine teeth and the monkey, Callicebus moloch: lack of sexual dimorphism. American Journal of Physical Anthropology 35: 91100.
Kinzey, W. G. & Norconk, M. A. (1990) Hardness as a basis of fruit choice in two
sympatric primates. American Journal of Physical Anthropology 81: 515.
(1993) Physical and chemical properties of fruit and seeds eaten by Pithecia and
Chiropotes in Surinam and Venezuela. International Journal of Primatology 14:
207227.
324
References
Kirk, E. C. & Simons, E. L. (2001) Diets of fossil primates from the Fayum
Depression of Egypt: a quantitative analysis of molar shearing. Journal of
Human Evolution 40: 203229.
Kirkham, J. & Robinson, C. (1995) The biochemistry of the bres of the periodontal ligament. In: The Periodontal Ligament in Health and Disease, 2nd
edn. (eds. B. K. B. Berkovitz, B. J. Moxham & H. N. Newman), pp. 5581.
London: Mosby-Wolfe.
Kleiber, M. (1932) Body size and metabolism. Hilgardia 6: 315353.
(1961) The Fire of Life: An Introduction to Animal Energetics. New York: Wiley.
Klineberg, I. (1980) Inuences of temporomandibular articular receptors on functional jaw movements. Journal of Oral Rehabilitation 7: 307317.
Klineberg, I. & Wyke, B. O. (1983) Articular reex control of mastication. In: Oral
Surgery, vol. 4 (ed. L. W. Kay), pp. 253258. Copenhagen: Munksgaard.
Korhonen, R. K., Laasanen, M. S., Toyras, J. (2002) Comparison of the equilibrium response of articular cartilage in unconned compression, conned
compression and indentation. Journal of Biomechanics 35: 903909.
Krishnamani, R. & Mahaney, W. C. (2000) Geophagy among primates: adaptive
signicance and ecological consequences. Animal Behaviour 59: 899915.
LaBarbera, M. (1989) Analyzing body size as a factor in ecology and evolution.
Annual Reviews of Ecology and Systematics 20: 97117.
Lacey, R. W. (1994) Hard to Swallow. Cambridge: Cambridge University Press.
Laine, P. & Siirila, H. S. (1971) Oral and manual stereognosis and two-point
tactile discrimination of the tongue. Acta Odontologica Scandinavica 29: 197
204.
Lake, G. J. & Yeoh, O. H. (1978) Measurement of rubber cutting resistance in the
absence of friction. International Journal of Fracture 14: 509526.
Lakes, R. S. (1987) Foam structures with a negative Poissons ratio. Science 235:
10381040.
Lambert, J. E. (1999) Seed handling in chimpanzees (Pan troglodytes) and redtail
monkeys (Cercopithecus ascanius): implications for understanding hominoid
and cercopithecine fruit-processing strategies and seed dispersal. American
Lambrecht, J. R. (1965) The inuence of occlusal contact area on chewing performance. Journal of Prosthetic Dentistry 15: 444450.
Landry, S. O. Jr (1970) The Rodentia as omnivores. Quarterly Review of Biology
45: 351372.
Lanyon, J. M. & Sanson, G. D. (1986) Koala (Phascolarctos cinereus) dentition and
nutrition. II. Implications of tooth wear in nutrition. Journal of Zoology 209:
169181.
Lanyon, L. E. & Rubin, C. T. (1985) Functional adaptation in skeletal structures.
In: Functional Vertebrate Morphology (eds. M. Hildebrand, D. M. Bramble,
K. F. Liem & D. B. Wake), pp. 125. Cambridge, MA: Belknap Press of
Harvard University.
Laska, M., Kohlmann, S., Hernandez-Salazar, L. T. et al. (2001) Gustatory responses to polycose in four species of non-human primates. Folia Primatologica 72: 171172.
References
325
Lauer, C. (1975) A comparison of sexual dimorphism and range of variation in

Papio cynocephalus and Gorilla gorilla dentition. Primates 16: 17.
Lavigne, G. J., Kato, T., Kolta, A. et al. (2003) Neurobiological mechanisms
involved in sleep bruxism. Critical Reviews in Oral Biology and Medicine 14:
3046.
Lawn, B. R. (1993) Fracture of Brittle Solids, 2nd edn. Cambridge: Cambridge
University Press.
Lee, C.-S., Lawn, B. R. & Kim, D. K. (2001) Effect of tangential loading on critical
conditions for radial cracking in brittle coatings. Journal of the American
Ceramic Society 84: 27192721.
Lee, V. M. & Linden, R. W. A. (1992) An olfactory-submandibular salivary reex
in humans. Experimental Physiology 77: 221224.
Lees, C., Vincent, J. F. V. & Hillerton, J. E. (1991) Poissons ratio in skin. BioMedical Materials and Engineering 1: 1923.
Leigh, E. G. Jr (1999) Tropical Forest Ecology: A View from Barro Colorado Island.
New York: Oxford University Press.
Leighton, M. (1993) Modeling dietary selectivity by Bornean orangutans: evidence
for integration of multiple criteria in fruit selection. International Journal of
Primatology 14: 257313.
Leighton, M. & Leighton, D. R. (1983) Vertebrate responses to fruiting seasonality
within a Bornean rain forest. In: Tropical Rain Forest: Ecology and Management
(eds. S. L. Sutton, T. C. Whitmore & A. C. Chadwick), pp. 181196. Oxford:
Blackwell.
Lermer, C. M. & Mattes, R. D. (1999) Perception of dietary fat: ingestive and
metabolic implications. Progress in Lipid Research 38: 117128.
Levine, A. S. & Silva, S. E. (1980) Absorption of whole peanuts, peanut oil and
peanut butter. New England Journal of Medicine 303: 917918.
Lieberman, D. E. (1993) Life history variables preserved in dental cementum
microstructure. Science 261: 11621164.
(1996) How and why recent humans grow thin skulls: experimental data on
systemic cortical robusticity. American Journal of Physical Anthropology 101:
217236.
Lieberman, D. E. & Crompton, A. W. (2000) Why fuse the mandibular symphysis?
A comparative analysis. American Journal of Physical Anthropology 112: 517
540.
Lightoller, G. H. S. (1925) Facial muscles. The modiolus and muscles surrounding
the rima oris with some remarks about the panniculus adiposus. Journal of
Anatomy 60: 185.
Lillford, P. J. (1991) Texture and acceptability of human foods. In: Feeding and the
Texture of Food (eds. J. F. V. Vincent & P. J. Lillford), pp. 93121. Cambridge:
(2000) The materials science of eating and food breakdown. MRS (Materials
Research Society) Bulletin (December): 3843.
Linden, R. W. A. (1990) Periodontal mechanoreceptors and their functions. In:
Neurophysiology of the Jaws and Teeth (ed. A. Taylor), pp. 5295. London:
Macmillan.
326
References
Logan, M. & Sanson, G. D. (2002) The association of tooth wear with sociality of
free-ranging male koalas (Phascolarctos cinereus Goldfuss). Australian Journal
of Zoology 50: 621626.
Lovegrove, B. G. (2000) The zoogeography of mammalian basal metabolic rate.
Lowrison, G. C. (1974) Crushing and Grinding. London: Butterworth.
Lucas, P. W. (1980) Adaptation and form of the mammalian dentition with special reference to primates and the evolution of Man. Ph.D. thesis. London:
University of London.
(1981) An analysis of canine size and jaw shape in some Old and New World
non-human primates. Journal of Zoology 195: 437448.
(1982a) Basic principles of tooth design. In: Teeth: Form, Function and Evolution
(ed. B. Kurten), pp. 154162. New York: Columbia University Press.
(1982b) An analysis of the canine tooth size of Old World higher primates in
relation to facial length and body weight. Archives of Oral Biology 27: 493496.
(1989) A new theory relating seed processing by primates to their relative
tooth sizes. In: The Growing Scope of Human Biology (eds. L. H. Schmitt,
L. Freedman & N. W. Bruce), pp. 3749. Perth: Centre for Human Biology,
University of Western Australia.
(1991) Fundamental physical properties of fruits and seeds in the diet of Southeast
Asian primates. In: Primatology Today (eds. A. Ehara, T. Kimura, O. Takenaka
& M. Iwamoto), pp. 125128. Amsterdam: Elsevier.
Lucas, P. W. & Corlett, R. T. (1991) Quantitative aspects of the relationship between
dentitions and diets. In: Feeding and the Texture of Food (eds. J. F. V. Vincent
& P. J. Lillford), pp. 93121. Cambridge: Cambridge University Press.
(1998) Seed dispersal by long-tailed macaques. American Journal of Primatology
45: 2944.
Lucas, P. W. & Luke, D. A. (1983) Methods for analysing the breakdown of food
during human mastication. Archives of Oral Biology 28: 813819.
(1984a) Optimal mouthful for food comminution in human mastication.
Archives of Oral Biology 29: 205210.
(1984b) Chewing it over: basic principles of food breakdown. In: Food Acquisition
and Processing in Primates (eds. D. J. Chivers, B. A. Wood & A. Bilsborough),
pp. 283302. New York: Plenum.
(1986) Is food particle size a criterion for the initiation of swallowing? Journal of
Oral Rehabilitation 13: 127136.
Lucas, P. W. & Peters, C. R. (2000) Function of postcanine tooth shape in mammals. In: Development, Function and Evolution of Teeth (eds. M. F. Teaford,
M. M. Smith & M. Ferguson), pp. 282289. Cambridge: Cambridge University Press.
Lucas, P. W. & Teaford, M. F. (1994) Functional morphology of colobine teeth.
In: Colobine Monkeys: Their Ecology, Behaviour and Evolution (eds. A. G.
Davies & J. F. Oates), pp. 171203. Cambridge: Cambridge University
Press.
(1995) Signicance of silica in leaves eaten by long-tailed macaques (Macaca
fascicularis). Folia Primatologica 64: 3036.
References
327
Lucas, P. W., Corlett, R. T. & Luke, D. A. (1985) Plio-Pleistocene hominids: an

approach combining masticatory and ecological analysis. Journal of Human
Evolution 14: 187202.
(1986a) Postcanine tooth size and diet in anthropoids. Zeitschrift fur Morphologie
und Anthropologie 76: 253276.
(1986b) Sexual dimorphism of teeth in anthropoid primates. Human Evolution
1: 2339.
(1986c) New approach to postcanine tooth size applied to Plio-Pleistocene hominids. In: Primate Evolution (eds. J. G. Else & P. C. Lee), pp. 191201.
Lucas, P. W., Luke, D. A., Voon, F. C. T. et al. (1986d) Breakdown patterns
produced by human subjects possessing articial and natural teeth. Journal
of Oral Rehabilitation 13: 205214.
Lucas, P. W., Ow, R. K. K., Ritchie, G. M. et al. (1986e) Relationship between jaw
movement and food comminution in human mastication. Journal of Dental
Research 65: 400404.
Lucas, P. W., Choong, M. F., Tan, H. T. W. et al. (1991a) Fracture toughness
of the leaf of the dicotyledonous angiosperm, Calophyllum inophyllum L.
Lucas, P. W., Lowrey, T. K., Pereira, B. et al. (1991b) The ecology of Mezzettia
leptopoda Hk. f. et Thoms. (Annonaceae) seeds as viewed from a mechanical
perspective. Functional Ecology 5: 345353.
Lucas, P. W., Oates, C. G. & Lee, W. P. (1993) Fracture toughness of mung bean
gels. Journal of Materials Science 28: 11371142.
Lucas, P. W., Peters, C. R. & Arrandale, S. (1994) Seed-breaking forces exerted by
orangutans with their teeth in captivity and a new technique for estimating
forces produced in the wild. American Journal of Physical Anthropology 94:
365378.
Lucas, P. W., Darvell, B. W., Lee, P. K. D. et al. (1995) The toughness of plant
cell walls. Philosophical Transactions of the Royal Society London series B 348:
363372.
Lucas, P. W., Tan, H. T. W. & Cheng, P. Y. (1997) The toughness of secondary cell
wall and woody tissue. Philosophical Transactions of the Royal Society London
series B 352: 341352.
Lucas, P. W., Turner, I. M., Dominy, N. J. et al. (2000) Mechanical defences to
herbivory. Annals of Botany 86: 913920.
Lucas, P. W., Beta, T., Darvell, B. W. et al. (2001) Field kit to characterize physical,
chemical and spatial aspects of potential foods of primates. Folia Primatologica
72: 1115.
Lucas, P. W., Prinz, J. F., Agrawal, K. R. et al. (2002) Food physics and oral
physiology. Food Quality and Preference 13: 203213.
Lucas, P. W., Osorio, D., Yamashita, N. et al. (2003) Dietary analysis. I. Physics.
In: Field and Laboratory Methods in Primatology (eds. J. Setchell & D. Curtis),
Luke, D. A. & Lucas, P. W. (1983) The signicance of cusps. Journal of Oral
Rehabilitation 10: 197210.
328
References
(1985) Chewing efciency in relation to occlusal and other variations in the

dentition. British Dental Journal 159: 401403.
Lumsden, A. G. S. & Osborn, J. W. (1977) The evolution of chewing: a dentists
view of palaeontology. Journal of Dentistry 5: 269287.
Lund, J. P. (1991) Mastication and its control by the brainstem. Critical Reviews in
Oral Biology and Medicine 2: 3364.
Lundberg, M., Nord, P. G. & strand, P. (1974) Changes in masticatory function
after surgical treatment of mandibular prognathism. (Cineradiographic study
of bolus position). Acta Odontologica Scandinavica 32: 3949.
Luo, Z.-X., Cifelli, R. L. & Kielan-Jaworowska, Z. (2001) Dual origin of tribosphenic mammals. Nature 409: 5357.
Luschei, E. S. & Goldberg, L. J. (1981) Neural mechanisms of mandibular control:
mastication and voluntary biting. In: Handbook of Physiology, section I, The
Nervous System, vol. 2, Motor Control, part 2 (eds. J. M. Brookhart, V. B.
Mountcastle, V. B. Brooks & S. R. Geiger), pp. 12371273. Bethesda: American
Physiological Society.
Macfadden, B. J., Solounias, N. & Cerling, T. E. (1999) Ancient diets, ecology and
extinction of 5-million-year-old horses from Florida. Science 283: 824827.
Maglio, V. J. (1972) Evolution of mastication in the Elephantidae. Evolution 26:
638658.
(1973) Origin and evolution of the Elephantidae. Transactions of the American
Philosophical Society of Philadelphia (new series) 63: 1149.
Magnusson, W. E. & Saniotti, T. M. (1987) Dispersal of Miconia seeds by the rat
Bolomys lasiurus. Journal of Tropical Ecology 3: 277278.
Mai, Y.-W. & Atkins, A. G. (1975) Scale effects and crack propagation in non-linear
elastic structures. International Journal of Mechanical Sciences 17: 673675.
Mai, Y.-W. & Cotterell, B. (1989) On mixed-mode plane stress ductile fracture.
In: Advances in Fracture Research (eds. K. Salama, K. Ravi-Chandar, D. M. R.
Taplin & P. Ramo Rao), pp. 22692278. Oxford: Pergamon Press.
Manger, P. R., Woods, T. M. & Jones, E. G. (1995) Representation of the face
and intra-oral structures in area 3b of the squirrel monkey (Saimiri sciureus)
somatosensory cortex, with special reference to the ipsilateral representation.
Journal of Comparative Neurology 362: 597607.
(1996) Representation of face and intra-oral structures in area 3b of macaque
monkey somatosensory cortex. Journal of Comparative Neurology 371: 513
521.
Manly, R. S. & Braley, L. C. (1950) Masticatory performance and efciency. Journal
Mao, J. J. (2002) Mechanobiology of craniofacial sutures. Journal of Dental Research
81: 810816.
Mao, J., Stein, R. B. & Osborn, J. W. (1992) The size and distribution of ber
types in jaw muscles: a review. Journal of Craniomandibular Disorders: Facial
and Oral Pain 6: 192201.
Marshall, D. B. & Lawn, B. R. (1986) Indentation of brittle materials. In:
Microindentation Techniques in Science and Engineering ASTMMSTP 889
References
329
(eds. P. J. Blau & B. R. Lawn), pp. 2646. Philadelphia, PA: American Society
for Testing and Materials.
Marshall, G. W. Jr, Balooch, M., Gallagher, R. R. et al. (2001) Mechanical
properties of the dentinoenamel junction: AFM studies of nanohardness,
elastic modulus, and fracture. Journal of Biomedical Materials Research 54:
8795.
Martin, L. B. (1985) Signicance of enamel thickness in hominoid evolution. Nature
314: 260263.
Martin, R. D. (1979) Phylogenetic aspects of prosimian behaviour. In: The Study of
Prosimian Behaviour (eds. G. A. Doyle & R. D. Martin), pp. 4577. London:
Academic Press.
(1990) Primate Origins and Evolution. Princeton, NJ: Princeton University Press.
Martin, R. D., Chivers, D. J., MacLarnon, A. M. et al. (1985) Gastrointestinal
allometry in primates and other mammals. In: Size and Scaling in Primate
Biology (ed. W. L. Jungers), pp. 6189. New York: Plenum.
Martinez del Rio, C. (1990) Dietary and phylogenetic correlates of intestinal sucrase
and maltase activity in birds. Physiological Zoology 63: 9871011.
Mattes, R. D. (2001) Oral exposure to butter, but not fat replacers elevates postprandial triaglycerol concentration in humans. Journal of Nutrition 131: 1491
1496.
Matthews, P. B. C. (1972) Mammalian Muscle Receptors and their Central Actions.
London: Edward Arnold.
Maynard Smith, J. (1979) Game theory and the evolution of behaviour. Proceedings
of the Royal Society London series B 205: 475488.
Maynard Smith, J. & Price, G. R. (1973) The logic of animal conict. Nature 146:
1518.
McCann, M. C. & Roberts, K. (1991) Architecture of the primary cell wall. In: The
Cytoskeletal Basis of Plant Growth (ed. C. W. Lloyd), pp. 109129. London:
Academic Press.
McKey, D. B., Gartlan, J. S., Waterman, P. G. et al. (1981) Food selection by black
colobus monkeys (Colobus satanus) in relation to plant chemistry. Biological
Journal of the Linnean Society 16: 114146.
McMahon, T. A. (1973) Size and scaling in biology. Science 179: 12011204.
(1975) Using body size to understand the structural design of animals:
quadrapedal locomotion. Journal of Applied Physiology 39: 619627.
McMahon, T. A. & Greene, P. R. (1979) The inuence of track compliance on
running. Journal of Biomechanics 12: 893904.
McMillan, A. S. & Hannam, A. G. (1992) Task-related behaviour of motor units
in different regions of the human masseter muscle. Archives of Oral Biology
37: 849857.
Mendoza, M., Janis, C. M. & Palmqvist, P. (2002) Characterizing complex craniodental patterns related to feeding behaviour in ungulates: a multivariate
approach. Journal of Zoology 258: 223246.
Mercader, J., Panger, M. & Boesch, C. (2002) Excavation of a chimpanzee stone
tool site in the African Rainforest. Science 296: 14521455.
330
References
Metcalfe, C. R. & Chalk, L. (1950) Anatomy of Dicotyledons, vols. 1 and 2. Oxford:

Clarendon Press.
Mills, J. R. E. (1955) Ideal dental occlusion in the Primates. Dental Practitioner 6:
4761.
Milton, K. (1979) Factors inuencing leaf choice by howler monkeys: a test of some
hypotheses of food selection by generalist herbivores. American Naturalist 114:
362378.
Miyazato, M., Chen, J. Y., Ishiguro, E. et al. (1994) Studies on the non-destructive
measurement of Poissons ration of watermelon. I. Model development.
Memoirs of the Faculty of Agriculture, Kagoshima University 30: 97104.
Mohsenin, N. N. (1977) Characterization and failure in solid foods with particular
reference to fruits and vegetables. Journal of Texture Studies 8: 169193.
Mongini, F., Tempia-Valenti, G. & Benvegnu, G. (1986) Computer-based assessment of habitual mastication. Journal of Prosthetic Dentistry 55: 638649.
Moore, S. J. & Sanson, G. D. (1995) A comparison of the molar efciency of two
insect-eating mammals. Journal of Zoology 235: 175192.
Morgan, E. (1972) The Descent of Woman. London: Souvenir.
Mowlana, F. & Heath, M. R. (1993) Assessment of masticatory efciency: new
methods appropriate for clinical research in dental practice. European Journal
of Prosthodontics and Restorative Dentistry 1: 121125.
Moyà Solà, S. & Kohler, M. (1997) The phylogenetic relationships of Oreopithecus
bambolii Gervais 1872. Comptes Rendus de lAcademie des Sciences Paris (serie II)
324 : 141148.
Muhlemann, H. R. (1951) Periodontometry: a method for measuring tooth mobility. Oral Surgery Oral Medicine Oral Pathology 4: 12201233.
Murphy, T. (1959) Compensatory mechanisms in facial height adjustment to functional tooth attrition. Australian Dental Journal 4: 312323.
Murray, C. G. & Sanson, G. D. (1998) Thegosis: a critical review. Australian Dental
Journal 43: 192198.
Murray, P. (1975) The role of cheek pouches in cercopithecine monkey adaptive
strategy. In: Primate Functional Morphology and Evolution (ed. R. H. Tuttle),
pp. 151194. The Hague: Mouton.
Mysterud, A. (1998) The relative roles of body size and feeding type on activity
time of temperate ruminants. Oecologia 113: 442446.
Neinhuis, C. & Barthlott, W. (1997) Characterization and distribution of waterrepellent, self-cleaning plant surfaces. Annals of Botany 79: 667677.
Nelson, G., Hoon, M. A., Chandrashekar, J. et al. (2001) Mammalian sweet taste
receptors. Cell 106: 381390.
Nelson, G., Chandrashekar, J., Hoon, M. A. et al. (2002) An amino acid taste
receptor. Nature 416: 199202.
Nguyen, Q. & Zarkadas, C. G. (1989) Comparison of the amino acid composition
and connective tissue protein contents of selected bovine skeletal muscles.
Journal of Agricultural and Food Chemistry 32: 12791286.
Norman, D. B. & Weishampel, D. B. (1991) Feeding mechanisms in some small
herbivorous dinosaurs: processes and patterns. In: Biomechanics in Evolution
References
331
(eds. J. M. V. Rayner & R. J. Wootton), pp. 161181. Cambridge: Cambridge

University Press.
Norman, T. L., Vashishth, D. & Burr, D. B. (1992) Effect of groove on bone
fracture toughness. Journal of Biomechanics 25: 14891492.
Nose, K. (1961) Study on the hardness of human and animal teeth. Journal of the
Kyoto Prefectural University of Medicine 69: 19251945.
Oates, J. F. (1978) Water-plant and soil consumption by guereza monkeys (Colobus
guereza): a relationship with minerals and toxins in the diet? Biotropica 10:
241253.
Olthoff, L. W., van der Bilt, A., Bosman, F. et al. (1984) Distribution of particle
sizes in food comminuted by human mastication. Archives of Oral Biology 29:
899903.
Orchardson, R. & Cadden, S. W. (1998) Mastication. In: The Scientic Basis of
Eating (ed. R. W. A. Linden), pp, 76121. Basel: Karger.
Oron, U. & Crompton, A. W. (1985) A cineradiographic and electromyographic
study of mastication in Tenrec ecaudatus. Journal of Morphology 185: 155
182.
Osborn, H. F. (1888) The nomenclature of the mammalian molar cusps. American
Naturalist 22: 926928.
Osborn, J. W. (1961) An investigation into the interdental forces occurring between
the teeth of the same arch during clenching of the jaws. Archives of Oral Biology
5: 202211.
(1969) Dentine hardness and incisor wear in the beaver (Castor ber). Acta
Anatomica 72: 123132.
(1971a) A relationship between the striae of Retzius and prism directions
in the transverse plane of the tooth. Archives of Oral Biology 16: 461
470.
(1971b) The ontogeny of tooth succession in Lacerta vivipara Jacquin (1787).
Proceedings of the Royal Society London series B 179: 261289.
(1974) Variation in structure and development of enamel. In: Oral Sciences
Reviews, vol. 3, Dental Enamel. Copenhagen: Munksgaard.
(1978) Morphogenetic gradients: elds versus clones. In: Development, Function
and Evolution of Teeth (eds. P. M. Butler & K. A. Joysey), pp. 171213. New
York: Academic Press.
(1981) Companion to Dental Studies (eds. A. H. R. Rowe & R. B. Johns), vol. 2A,
Dental Anatomy and Embryology. Oxford: Blackwell.
(1982) Helicoidal plane of occlusion. American Journal of Physical Anthropology
57: 273281.
(1985) The design of the human TMJ: design, function, and failure. Journal of
Oral Rehabilitation 12: 279293.
(1987) Relationship between the mandibular condyle and the occlusal plane
during hominid evolution: some of its effects on jaw mechanics. American
(1993) A model to describe how ligaments may control symmetrical jaw opening
movements in man. Journal of Oral Rehabilitation 20: 585604.
332
References
(1995a) Internal derangement and the accessory ligaments around the temporomandibular joint. Journal of Oral Rehabilitation 22: 731740.
(1995b) Biomechanical implications of lateral pterygoid contribution to biting
and jaw opening in humans. Archives of Oral Biology 40: 10991108.
Osborn, J. W. & Baragar, F. A. (1985) Predicted pattern of human muscle activity during clenching derived from a computer-assisted model. Journal of
Biomechanics 18: 599612.
Osborn, J. W. & Lumsden, A. G. S. (1978) An alternative to thegosis and a reexamination of the ways in which mammalian molars work. Neues Jahrbuch
fur Geologie und Palaontologie Abhandlungen 156: 371396.
Osborn, J. W., Baragar, F. A. & Grey, P. (1987) The functional advantage of
proclined incisors in man. In: Teeth Revisited: Proceedings of 7th International Symposium on Dental Morphology (eds. D. E. Russell, J. P. Santoro
& D. Sigognean-Russell). Memoirs du Musee National dHistoire Naturelle,
Paris (Serie C) 53: 445458.
Ostry, D. J., Gribble, P. L., Levin, M. F. et al. (1997) Phasic and tonic stretch
reexes in muscles with few muscle spindles: human jaw-opener muscles.
Experimental Brain Research 116: 299308.
Otani, T. & Shibata, E. (2000) Seed dispersal and predation by Yakushima
macaques, Macaca fuscata yakui, in a warm temperate forest of Yakushima
Island, southern Japan. Ecological Research 15: 133144.
Ottenhoff, F. A., van der Bilt, A., van der Glas, H. W. et al. (1996) The relationship
between jaw elevator muscle surface electromyogram and simulated food
resistance during dynamic condition in humans. Journal of Oral Rehabilitation
23: 270279.
Overdorff, D. J. & Strait, S. G. (1998) Seed handling by three prosimian primates
in Southeastern Madagascar: implications for seed dispersal. American Journal
of Primatology 45: 6982.
Owall,
B. (1978) Interocclusal perception with anaesthetized and unanaesthetized
temporomandibular joints. Swedish Dental Journal 2: 199208.
Owall,
B. & Vorwerk, P. (1974) Analysis of a method for testing oral tactility during
chewing. Odontologisk Revy 25: 110.
Owen-Smith, N. (1988) Megaherbivores: The Inuence of Very Large Body Size on
Ecology. Cambridge: Cambridge University Press.
Owen-Smith, N., Robbins, C. T. & Hagerman, A. E. (1993) Browse and browsers:
interactions between woody plants and mammalian herbivores. Trends in
Ecology and Evolution 8: 158160.
Page, D. H., El-Hosseiny, F. & Winkler, K. (1971) Behaviour of single wood bres
under axial tensile strain. Nature 229: 252253.
Panger, M.A, Brooks, A. G., Richmond, B. G. et al. (2002) Older than the
Olduwan? Rethinking the emergence of hominin tool use. Evolutionary
Paphangkorakit, J. & Osborn, J. W. (1997) Effect of jaw opening on the direction
and magnitude of human incisal bite forces. Journal of Dental Research 76:
561567.
References
333
(2000) The effect of normal occlusal forces on uid movement through human
dentine in vitro. Archives of Oral Biology 45: 10331041.
Partt, G. J. (1960) Measurement of the physiological mobility of individual teeth
in an axial direction. Journal of Dental Research 39: 608618.
Peleg, M., Gomez-Brito, L. & Malevski, Y. (1976) Compressive failure patterns of
some juicy fruits. Journal of Food Science 41: 13201324.
Peleg, M. & Gomez-Brito, L. (1977) Textural changes in ripening plantains. Journal
of Texture Studies 7: 457463.
Pereira, B. P., Lucas, P. W. & Teoh, S. H. (1997) Ranking the fracture toughness of
mammalian soft tissues using the scissors cutting test. Journal of Biomechanics
30: 9194.
Perez-Barberia, F. J. & Gordon, I. J. (1998a) Factors affecting food comminution
during chewing in ruminants: a review. Biological Journal of the Linnean
Society 63: 233256.
(1998b) The inuence of molar occlusal surface area on the assimilation efciency, chewing behaviour and diet selection of red deer. Journal of Zoology
245: 307316.
(2001) Relationship between oral morphology and feeding style in the Ungulata:
a phylogenetically controlled evaluation. Proceedings of the Royal Society
London series B 268: 10211030.
Peters, C. R. (1979) Towards an ecological model of African Plio-Pleistocene hominid adaptations. American Anthropologist 81: 261278.
(1981) Australopithecus vs. Homo dietary capabilities: the natural competitive
advantage of the megadonts. In: Perceptions of Human Evolution, vol. 7 (eds.
L. L. Mai, E. Shanklin & R. W. Sussman), pp. 161181. Los Angeles: UCLA
Anthropology.
(1982) Electron-optical microscopic study of incipient dental microdamage from
experimental seed and bone crushing. American Journal of Physical Anthropology 57: 283301.
(1987) Nut-like oil seeds: food for monkeys, chimpanzees, humans and
probably ape-men. American Journal of Physical Anthropology 73: 333
363.
Peters, C. R. & Maguire, B. (1981) Wild plant foods of the Makapansgat area: a
modern ecosystems analogue for Australopithecus africanus adaptations. Journal of Human Evolution 10: 565583.
Peters, C. R., OBrien, E. M. & Drummond, R. B. (1992) Edible Wild Plants of
Sub-Saharan Africa. Kew: Royal Botanical Gardens.
Peters, R. H. (1983) The Ecological Implications of Body Size. Cambridge: Cambridge
University Press.
Pettifor, E. (2000) From the teeth of the dragon: Gigantopithecus blacki. In: Selected
Readings in Physical Anthropology (ed. P. Scully), pp. 143149. Dubuque, IA:
Kendall/Hunt.
Phua, P. B. & Corlett, R. T. (1989) Seed dispersal by the lesser short-nosed fruit
bat (Cynopterus brachyotis, Pteropodidae, Megachiroptera). Malayan Nature
Journal 42: 251256.
334
References
Picton, D. C. A. (1965) On the part played by the socket in tooth support. Archives
of Oral Biology 10: 945955.
Pilbeam, D. R. & Gould, S. J. (1974) Size and scaling in human evolution. Science
186: 892901.
(1975) Allometry and early hominids. Science 189: 64.
Plavcan, J. M., van Schaik, C. P. & Kappeler, P. M. (1995) Competition, coalitions
and canine size in primates. Journal of Human Evolution 28: 245276.
Poon, T. F. (1974) Physiological studies on fruits of Nephelium lappaceum L. B.Sc.
dissertation. Singapore: Department of Botany, National University of
Singapore.
Popovics, T. E. & Fortelius, M. (1997) On the cutting edge: tooth blade sharpness in
herbivorous and faunivorous mammals. Annales Zoologici Fennici 34: 7388.
Popovics, T. E., Remsberger, J. M. & Herring, S. W. (2002) The fracture behaviour
of human and pig molar cusps. Archives of Oral Biology 46: 112.
Preston, C. M. & Sayer, B. G. (1992) Whats in a nutshell: an investigation of structure by carbon-13 cross-polarization magic-angle spinning nuclear magnetic
resonance spectroscopy. Journal of Agricultural Food Chemistry 40: 206220.
Preston, R. D. (1974) The Physical Biology of Plant Cell Walls. London: Chapman
& Hall.
Prinz, J. F. (in press) Abrasives in foods and their effect on intra-oral processing: a
two-colour chewing gum study. Journal of Oral Rehabilitation.
Prinz, J. F. & Lucas, P. W. (1995) Swallow thresholds in humans. Archives of Oral
Biology 40: 401403.
(1997). An optimization model for mastication and swallowing in mammals.
Proceedings of the Royal Society London series B 264: 17151721.
(2000) Saliva tannin interactions. Journal of Oral Rehabilitation 27: 991994.
(2001) The rst bite of the cherry: intra-oral manipulation prior to the rst bite
in humans. Journal of Oral Rehabilitation 28: 614617.
Prinz, J. F., Silwood, C. J. L., Claxson, A. W. D. et al. (2003) Simulated digestion status of intact and exoskeletally punctured insects and insect larvae: a
spectroscopic investigation. Folia Primatologica 74: 1226.
Prothero, D. R. & Sereno, P. C. (1982) Allometry and paleoecology of medial
Miocene dwarf rhinoceroses from the Texas Gulf Coastal Plain. Paleobiology
8: 1630.
Purnell, M. A. (1995) Microwear on conodont elements and macrophagy in the
rst vertebrates. Nature 374: 798800.
Purslow, P. P. (1983) Measurement of the fracture toughness of extensible connective
tissues. Journal of Materials Science 18: 35913598.
(1985) The physical basis of meat texture: observations on the fracture behaviour
of cooked bovine M. semitendinosus during heating. Meat Science 12: 3960.
(1991a) Notch-sensitivity of nonlinear materials. Journal of Materials Science 26:
44684476.
(1991b) Measuring meat texture and understanding its structural basis. In: Feeding and the Texture of Food (eds. J. F. V. Vincent & P. J. Lillford), pp. 3556.
References
335
Purslow, P. P., Bigi, A., Ripamonti, A. et al. (1984) Collagen bre orientation around
a crack in biaxially stretched aortic media. International Journal of Biological
Macromolecules 6: 2125.
Rajaram, A. (1986) Tensile properties and fracture of ivory. Journal of Materials
Science Letters 5: 10771080.
Ralls, K. (1977) Mammals in which females are larger than males. Quarterly Review
of Biology 51: 245275.
Rasmussen, S. T., Patchin, R. E., Scott, D. B. et al. (1976) Fracture properties of
human enamel and dentine. Journal of Dental Research 55: 154164.
Raven, J. A. (1983) The transport and function of silicon in plants. Biological
Reviews 58: 179207.
Ravosa, M. J. (1991) Structural allometry of the prosimian mandibular corpus and
symphysis. Journal of Human Evolution 20: 320.
Rees, J. S. & Jacobsen, P. H. (1997) Elastic modulus of the periodontal ligament.
Biomaterials 18: 995999.
Rees, L. A. (1954) The structure and function of the mandibular joint. British
Dental Journal 96: 125133.
Rensberger, J. M. (1973) An occlusion model for mastication and dental wear in
herbivorous mammals. Journal of Palaeontology 47: 515528.
(2000) Pathways to functional differentiation in mammalian enamel. In:
Development, Function and Evolution of Teeth (eds. M. F. Teaford, M. M.
Smith & M. W. J. Ferguson), pp. 252268. Cambridge: Cambridge University
Press.
Rensberger, J. M. & von Koenigswald, W. (1980) Functional phylogenetic interpretation of enamel microstructure in rhinoceroses. Paleobiology 6: 447495.
Renson, C. E. & Braden, M. (1971) The experimental deformation of human
dentine by indenters. Archives of Oral Biology 16: 563572.
(1975) Experimental determination of the rigidity modulus, Poissons ratio and
elastic limit of shear of human dentine. Archives of Oral Biology 20: 4347.
Rich, T. H., Flanner, T. F., Trusler, P. et al. (2002) Evidence that monotremes and
ausktribosphenids are not sistergroups. Journal of Vertebrate Paleontology 22:
466469.
Ridley, H. N. (1930) The Dispersal of Plants around the World. Ashford: Reeve.
Ringel, R. L. & Ewanowski, S. J. (1965) Oral perception. I. Two-point discrimination. Journal of Speech and Hearing Research 8: 389397.
Robbins, C. T., Spalinger, D. E. & Van Hoven, W. (1995) Adaptation of ruminants
to browse and grass diets: are anatomical-based browsergrazer interpretations
valid? Oecologia 103: 208213.
Robbins, M. W. (1977) Biting loads generated in the laboratory rat. Archives of
Robinson, C., Brookes, S. J., Bonass, W. A. et al. (1997) Enamel maturation. In:
Dental Enamel (Ciba Foundation Symposium 205), pp. 156174. Chichester:
Wiley.
Robinson, J. T. (1956) The dentition of the Australopithecinae. Memoirs of the
Transvaal Museum 9: 1179.
336
References
Rogers, M. E., Maisels, F., Williamson, E. A. et al. (1990) Gorilla diet in the Lope
Reserve, Gabon: a nutritional analysis. Oecologia 84: 326339.
Rolls, E. T., Critchley, H. D., Browning, A. S. et al. (1999) Responses to the sensory
properties of fat of neurons in the primate orbitofrontal cortex. Journal of
Neuroscience 19: 15321540.
Romer, A. S. (1966) Vertebrate Paleontology, 3rd edn. Chicago, IL: University of
Chicago Press.
Rose, H. E. & Sullivan, R. M. E. (1961) Vibration Mills and Vibration Milling.
London: Constable.
Rose, K. D., Walker, A. & Jacobs, L. (1981) Function of the mandibular tooth
comb in living and extinct mammals. Nature 289: 583585.
Rosenberger, A. L. & Kinzey, W. G. (1976) Functional patterns of molar occlusion
in platyrrhine primates. American Journal of Physical Anthropology 45: 281
298.
Roth, V. L. (1990) Insular dwarf elephants: a case study in body mass estimation and
ecological inference. In: Body Size in Mammalian Paleobiology (ed. J. Damuth
& B. J. MacFadden), pp. 151179. Cambridge: Cambridge University
Press.
Rubin, C., Turner, S., Bain, S. et al. (2001) Extremely low level mechanical signals
are anabolic to trabecular bone. Nature 412: 603604.
Runham, N. W., Thornton, P. R., Shaw, D. A. et al. (1969) Mineralization and
hardness of the radular teeth of the limpet Patella vulgata. Zeitschrift fur
Zellforschung und Mikroscopische Anatomie Abteilung Histochemie 99: 608
626.
Ryan, J. M. (1986) Comparative morphology and evolution of cheek pouches in
rodents. Journal of Morphology 190: 2741.
Rybczynski, N. & Reisz, R. R. (2001) Earliest evidence for efcient oral processing
in a terrestrial herbivore. Nature 411: 684687.
St Hoyme, L. E. & Horitzer, R. T. (1971) Signicance of canine wear in pongid
evolution. American Journal of Physical Anthropology 35: 145147.
Sato, K., Yoshinaka, R., Sato, M. et al. (1986) Collagen content in the muscles of
sh in association with their swimming movement and meat texture. Bulletin
of the Japanese Society of Scientic Fisheries 52: 15951600.
Savage, R. J. G. (1977) Evolution in carnivorous mammals. Palaeontology 20: 237
271.
Scapino, R. P. (1965) The third joint of the canine jaw. Journal of Morphology 116:
2350.
Schoeld, R. M. S., Nesson, M. H. & Richardson, K. A. (2002) Tooth-hardness
increases with zinc-content in mandibles of young adult leaf-cutter ants.
Naturwissenschaften 89: 579583.
Schmidt-Neilsen, K. (1972) How Animals Work. Cambridge: Cambridge University
Press.
Schwartz, J. H. (1974) Premolar loss in the primates: a theoretical reinvestigation.
In: Prosimian Biology (eds. R. D. Martin, G. A. Doyle & A. C. Walker),
pp. 621640. Pittsburgh, PA: University of Pittsburgh Press.
References
337
Sclafani, A. (1991) Starch and sugar tastes in rodents: an update. Brain Research
Bulletin 27: 383386.
Semaw, S., Renne, P., Harris, J. W. K. et al. (1997) 2.5 million-year-old stone tools
from Gona, Ethiopia. Nature 385: 333336.
Sereno, P. C. (1997) Origin and evolution of dinosaurs. Annual Review of Earth
and Planetary Sciences 25: 435489.
(1999) The evolution of dinosaurs. Science 284: 21372147.
Shama, F. & Sherman, P. (1973) Evaluation of some textural properties of foods
with the Instron universal testing machine. Journal of Texture Studies 4: 344
353.
Sharp, S. J., Ashby, M. F. & Fleck, N. A. (1993) Material response under static and
sliding indentation loads. Acta Metallurgica et Materialia 41: 685692.
Sharpe, P. T. (2000) Homeobox genes in initiation and shape of teeth during development in mammalian embryos. In: Development, Function and Evolution
of Teeth (eds. M. F. Teaford, M. M. Smith & M. W. J. Ferguson), pp. 312.
Shaw, D. M. (1917) Form and function of teeth: a theory of maximum shear.
Journal of Anatomy 52: 97106.
Sheikh-Ahmad, J. Y. & McKenzie, W. M. (1997) Measurement of tool wear and
dulling in the machining of particleboard. In: Proceedings of the 13th International Wood Machining Seminar, Vancouver, Canada, pp. 659670.
Sheine, W. S. (1979) The effect of variations in molar morphology on masticatory
effectiveness and digestion of cellulose in prosimian primates. PhD thesis.
Duke University, Durham, NC.
Sheine, W. S. & Kay, R. F. (1977) An analysis of chewed food particle size and
its relationship to molar structure in the primates Cheirogaleus medius and
Galago senegalensis and the insectivoran Tupaia glis. American Journal of Physical Anthropology 47: 1529.
Shellis, R. P. (1981) Comparative histology of dental tissues. In: Companion to
Dental Studies (eds. A. H. R. Rowe & R. B. Johns), Dental Anatomy and
Embryology (ed. J. W. Osborn), vol. 2A, pp. 158165. Oxford: Blackwell.
Shellis, R. P. & Dibden, G. H. (2000) Enamel microporosity and its functional
implications. In: Development, Function and Evolution of Teeth (eds. M. F.
Teaford, M. M. Smith & M. W. J. Ferguson), pp. 242268. Cambridge:
Shipley, L. A. & Spalinger, D. E. (1992) Mechanics of browsing in dense food
patches: effects of plant and animal morphology on intake rate. Canadian
Journal of Zoology 70: 17431753.
Shipley, L. A., Gross, J. E., Spalinger, D. E. et al. (1994) The scaling of intake rate
of mammalian herbivores. American Naturalist 143: 10551082.
Sim, B. J., Lucas, P. W., Pereira, B. P. et al. (1993) Mechanical and sensory assessment
of the texture of refrigerator-stored spring roll pastry. Journal of Texture Studies
24: 2744.
Simpson, G. G. (1936) Studies of the earliest mammalian dentitions. Dental Cosmos
78: 940953.
338
References
(1941) The Function of Saber-Like Canines in Carnivorous Mammals. American

Museum Novitates no. 1130.
(1953) The Major Features of Evolution. New York: Columbia University Press.
Smith, J. M. & Savage, R. J. G. (1959) The mechanics of mammalian jaws. School
Science Review 141: 289301.
Smith, K. K. (1992) The evolution of the mammalian pharynx. Zoological Journal
of the Linnean Society 104: 313349.
Smith, M. M. and Coates, M. I. (2000) Evolutionary origins of teeth and jaws:
developmental models and phylogenetic patterns. In: Development, Function
and Evolution of Teeth (eds. M. F. Teaford, M. M. Smith & M. W. J. Ferguson),
Smith, M. M. & Johanson, Z. (2003) Separate evolutionary origins of teeth from
evidence in fossil jawed vertebrates. Science 29: 12351236.
Smith, M. M. and Sansom, I. J. (2000) Evolutionary origins of dentine in the
fossil record of early vertebrates: diversity, development and function. In:
Development, Function and Evolution of Teeth (eds. M. F. Teaford, M. M.
Smith & M. W. J. Ferguson), pp. 6581. Cambridge: Cambridge University
Press.
Smith, R. J. (1978) Mandibular biomechanics and temporomandibular joint function in primates. American Journal of Physical Anthropology 49: 341350.
(1983) The mandibular corpus of female primates: taxonomic, dietary and allometric correlates of interspecic variations in size and shape. American Journal
(1984) Comparative functional morphology of maximum mandibular opening
(gape) in primates. In: Food Acquisition and Processing in Primates (eds. D. J.
Chivers, B. A. Wood & A. Bilsborough), pp. 231255. New York: Plenum.
Sofaer, J. A. (1973) A model relating developmental interaction and differential
evolutionary reduction of tooth size. Evolution 27: 427434.
(1977) Co-ordinated growth of successively initiated tooth germs in the mouse.
Archives of Oral Biology 22: 7172.
Sofaer, J. A., Chung, C. S., Niswander, J. D. et al. (1971a) Developmental
interaction, size and agenesis among maxillary incisors. Human Biology 43:
3645.
Sofaer, J. A., Bailit, H. L. & MacLean, J. (1971b) A developmental basis for differential tooth reduction during hominid evolution. Evolution 25: 509517.
Solounias, N. & Moelleken, S. M. C. (1993) Dietary adaptation of some extinct
ruminants determined by premaxillary shape. Journal of Mammalogy 74:
10591071.
Solounias, N., Teaford, M. F. & Walker, A. (1988) Interpreting the diet of extinct
ruminants: the case of a grazing giraffe. Paleobiology 14: 287300.
Spears, I. R., van Noort, R., Crompton, R. H. (1993) The effects of enamel
anisotropy on the distribution of stress in a tooth. Journal of Dental Research
72: 15261531.
Spencer, M. (1998) Force production in the primate masticatory system: electromyographic tests of biomechanical hypotheses. Journal of Human Evolution
34: 2554.
References
339
Spielman, A. I. (1990) Interaction of saliva and taste. Journal of Dental Research

69: 838843.
Stirton, R. A. (1947) Observations on evolutionary rates in hypsodonty. Evolution
1: 3241.
Strait, S. G. (1993a) Molar morphology and food texture among small-bodied
insectivorous mammals. Journal of Mammalogy 74: 391402.
(1993b) Differences in occlusal morphology and molar size in frugivores and
faunivores. Journal of Human Evolution 25: 471484.
(1993c) Molar microwear in extant small-bodied faunivorous mammals: an analysis of feature density and pit frequency. American Journal of Physical Anthropology 92: 16379.
Strait, S. G. & Overdorff, D. J. (1996) Physical properties of fruits eaten by
Malagasy primates. American Journal of Physical Anthropology, Supplement
22: 224.
Strait, S. G. & Vincent, J. F. V. (1998) Primate faunivores: physical properties of
prey items. International Journal of Primatology 19: 867878.
Su, H. H. & Lee, L. L. (2001) Food habits of Formosan rock macaques
(Macaca cyclopsis) in Jentse, northeastern Taiwan, assessed by fecal analysis and behavioral observation. International Journal of Primatology 22: 359
377.
Swanson, S. A. V. (1980) Articular cartilage. In: The Mechanical Properties of Biological Materials (eds. J. F. V. Vincent & J. D. Currey), pp. 377395. Cambridge:
Synge, J. L. (1933) The tightness of the teeth, considered as a problem concerning
the equilibrium of a thin incompressible membrane. Philosophical Transactions
of the Royal Society London series A 231: 435477.
Szalay, F. S. (1975) Huntingscavenging protohominids. Man (new series) 10: 420
429.
Szczesniak, A. S. (1963) Classication of textural characteristics. Journal of Food
Science 28: 385389.
Tabor, D. (1951) The Hardness of Metals. Oxford: Clarendon Press.
Tanne, K., Tanaka, E. & Sakuda, M. (1991) The elastic modulus of the temporomandibular joint disc from adult dogs. Journal of Dental Research 70:
15451548.
Tavare, S., Marshall, C. R., Will, O. et al. (2002) Using the fossil record to estimate
the age of the last common ancestor of extant primates. Nature 416: 726729.
Teaford, M. F. (1988) A review of dental microwear and diet in modern mammals.
Scanning Microscopy 2: 11491166.
(1994) Dental microwear and dental function. Evolutionary Anthropology 3: 17
30.
Teaford, M. F. & Walker, A. (1983) Dental microwear in adult and still-born guinea
pigs (Cavia porcellus). Archives of Oral Biology 28: 10771081.
Teng, S., Xu, Y., Cheng, M. et al. (1991) Biomechanical properties and collagen ber orientation of temporomandibular joint discs in dogs. II. Tensile
mechanical properties of discs. Journal of Craniomandibular Disorders: Facial
and Oral Pain 5: 107114.
340
References
Terborgh, J. (1983) New World Primates: A Study in Comparative Ecology. Princeton,

NJ: Princeton University Press.
(1986) Keystone plant resources in the tropical forest. In: Conservation Biology:
Science of Scarcity and Diversity, (ed. M. Soule), pp. 330344. Sunderland,
MA: Sinauer.
Theimer, T. C. (2003) Intraspecic variation in seed size affects scatterhoarding
behaviour of an Australian tropical rain-forest rodent. Journal of Tropical
Ecology 19: 9598.
Thexton, A. J. & Crompton, A. W. (1998) The control of swallowing. In:
The Scientic Basis of Eating (ed. R. W. A. Linden), pp. 168222. Basel:
Karger.
Tittelbach, T. J. & Mattes, R. D. (2001) Oral stimulation inuences postprandial triaglycerol concentrations in humans: nutrient specicity. Journal of the
American College of Nutrition 20: 485493.
Tomos, A. D. & Leigh, R. A. (1999) The pressure probe: a versatile tool in plant
cell physiology. Annual Review of Plant Physiology and Plant Molecular Biology
50: 447472.
Trulsson, M. & Essick, G. K. (1997) Low-threshold mechanoreceptive afferents in
human lingual nerve. Journal of Neurophysiology 77: 737748.
Turner, I. M. (2001) The Ecology of Trees in the Tropical Rain Forest. Cambridge:
Uden, P. & Van Soest, P. J. (1982) The determination of digesta particles in some
herbivores. Animal Feed Science and Technology 7: 3544.
Ungar, P. S. (1992) Incisal microwear and feeding behavior of four Sumatran anthropoids. Ph.D. thesis. State University of New York at Stony
Brook.
(1994) Patterns of ingestive behavior and anterior tooth use differences in sympatric anthropoid primates. American Journal of Physical Anthropology 95:
197219.
(1996) Relationship of incisor size to diet and anterior tooth use in sympatric
Sumatran anthropoids. American Journal of Primatology 38: 145146.
Ungar, P. S. & Grine, F. E. (1991) Incisor size and wear in Australopithecus africanus
and Paranthropus robustus. Journal of Human Evolution 20: 313340.
Ungar, P. S. & Teaford, M. F. (1996) A preliminary examination of non-occlusal
dental microwear in anthropoids: implications for the study of fossil primates.
(2002) Human Diet: Its Origin and Evolution. Westport, CT: Bergin & Garvey.
Ungar, P. S. & Williamson, M. D. (2000) Exploring the effects of tooth wear
on functional morphology: a preliminary study using dental topographic
analysis. Palaeontologia Electronica 3: 18 pp. http://www-odp.tamu.edu/paleo/
2000 1/gorilla/main.htm.
Ungar, P. S., Teaford, M. F., Glander, K. E. et al. (1995) Dust accumulation in the
canopy: implications for the study of dental microwear in primates. American
Ungar, P. S., Fennell, K. J., Gordon, K. et al. (1997) Neanderthal incisor beveling.
Journal of Human Evolution 32: 407421.
References
341
Utz, K. H. (1986) Untersuchungen u ber die interokklusale tactile Feinsensibilitat naturischer Zahne mit Hilfe von Aluminium-oxid-teilchen. Deutsch
Zahnarztliche Zeitschrift 41: 313315.
van den Braber, W., van der Glas, H. W., van der Bilt, A. et al. (2001) Chewing
efciency of pre-orthognathic surgery patients: selection and breakage of food
particles. European Journal of Oral Science 109: 306311.
van der Bilt, A., Olthoff, L. W., van der Glas, H. W. et al. (1987) A mathematical
description of the comminution of food in human mastication. Archives of
van der Glas, H. W., van der Bilt, A., Olthoff, L. W. et al. (1987) Measurement
of selection chances and breakage functions during chewing in man. Journal
(1992) A selection model to estimate the interaction between food particles and
the postcanine teeth in human mastication. Journal of Theoretical Biology 155:
103120.
van Reenen, J. F. (1982) The effects of attrition on tooth dimensions of San (Bushmen). In: Teeth: Form, Function and Evolution (ed. B. Kurten), pp. 182203.
New York: Columbia University Press.
van Roosmalen, M. G. M. (1980) Habitat preferences, diet, feeding strategy and
social organization of the black spider monkey (Ateles paniscus paniscus Linnaeus 1958) in Surinam. Ph.D. thesis. Rijksuniversiteit voor Natuurbeheer,
Leersum.
van Schaik, C. P., Fox, E. A. & Sitompul, A. F. (1996) Manufacture and use of
tools in wild Sumatran orangutans. Naturwissenschaften 83: 186188.
Van Soest, P. J. (1994) Nutritional Ecology of the Ruminant, 2nd edn. Ithaca, NY:
Cornell University Press.
(1996) Allometry and ecology of feeding behavior and digestive capacity in
herbivores: a review. Zoo Biology 15: 455479.
Van Valen, L. (1960) A functional analysis of hypsodonty. Evolution 14: 531
532.
Van Valkenburgh, B. (1988) Incidence of tooth breakage among large, predatory
mammals. American Naturalist 131: 291300.
(1990) Skeletal and dental predictors of body mass in carnivores. In: Body Size in
Mammalian Paleobiology (eds. J. Damuth & B. J. MacFadden), pp. 181205.
(1996) Feeding behavior in free-ranging, large African carnivores. Journal of
Mammalogy 77: 240254.
Van Valkenburgh, B. & Hertel, F. (1993) Tough times at La Brea: tooth breakage
in large carnivores of the late Pleistocene. Science 261: 456459.
Van Valkenburgh, B. & Ruff, C. B. (1987) Canine tooth strength and killing
behaviour in large carnivores. Journal of Zoology, London 212: 379397.
Van Valkenburgh, B., Teaford, M. F. & Walker, A. (1990) Molar microwear and
diet in large carnivores. Journal of Zoology 22: 319340.
van Vliet, T. (2002) On the relation between texture perception and fundamental
mechanical properties of liquids and time dependent solids. Food Quality and
Preference 13: 227236.
342
References
Vickers, Z. (1981) Relationships of chewing sounds to judgements of crispness,

crunchiness and hardness. Journal of Food Science 47: 121124.
Vincent, J. F. V. (1980) Insect cuticle: a paradigm for natural composites. In: The
Mechanical Properties of Biomaterials (eds. J. F. V. Vincent & J. D. Currey),
(1981) Morphology and design of the extensible intersegmental membrane of
the female migratory locust. Tissue and Cell 13: 831853.
(1982) The mechanical design of grass. Journal of Materials Science 17: 856860.
(1990) Fracture properties of plants. Advances in Botanical Research 17: 235287.
(1991) Strength and fracture of grasses. Journal of Materials Science 26: 1947
1950.
(1992) Biomaterials. Oxford: IRL Press.
Vincent, J. F. V. & Hillerton, J. E. (1979) The tanning of insect cuticle: a
critical review and a revised mechanism. Journal of Insect Physiology 25: 653
658.
Vincent, J. F. V. & Khan, A. A. (1993) Anisotropy in the fracture properties of
apple esh as investigated by crack-opening tests. Journal of Materials Science
28: 4551.
Vincent, J. F. V. & Sibbing, F. A. (1992) How the grass carp (Ctenopharnygodon
idella) chooses and chews its food: some clues. Journal of Zoology 226: 435
444.
Vincent, J. F. V., Jeronimidis, G., Khan, A. A. & Luyten, H. (1991) The wedge
fracture test: a new method for measurement of food texture. Journal of Texture
Studies 22: 4557.
Vincent, J. F. V., Saunders, D. E. J. & Beyts, P. (2002) The use of stress intensity
factor to quantify hardness and crunchiness objectively. Journal of Texture
Studies 33: 149159.
Vinyard, C. J., Wall, C. E., Williams, S. H. et al. (2003) Comparative functional
analysis of skull morphology of tree-gouging primates. American Journal of
Visser, M. (1991) The Rituals of Dinner. London: Penguin.
von Koenigswald, W. (1982) Enamel structure in the molars of Arvicolidae
(Rodentia, Mammalia), a key to functional morphology and phylogeny. In:
Teeth: Form, Function and Evolution (ed. B. Kurten), pp. 109122. New York:
Columbia University Press.
Voon, F. C. T., Lucas, P. W., Luke, D. A. et al. (1986) A simulation approach
to understanding the masticatory process. Journal of Theoretical Biology 119:
251262.
Wagner, T., Neinhuis, C. & Barthlott, W. (1996) Wettability and contaminability
of insect wings as a function of their surface sculptures. Acta Zoologica 77:
213225.
Walker, A. C. (1981) Diet and teeth: dietary hypotheses and human evolution.
(1984) Mechanisms of honing in the male baboon canine. American Journal of
References
343
Walker, A. C., Hoech, H. N. & Perez, L. (1978) Microwear of mammalian teeth

as an indicator of diet. Science 201: 908910.
Walker, P. L. (1979) The adaptive signicance of pongid lip mobility. Ossa 6:
277284.
Walker, P. & Murray, P. (1975) An assessment of masticatory efciency in a series
of anthropoid primates with special reference to the Colobinae and Cercopithecinae. In: Primate Functional Morphology and Evolution (ed. R. H. Tuttle),
pp. 135150. The Hague: Mouton.
Waterman, P. G. (1984) Food acquisition and processing as a function of plant
chemistry. In: Food Acquisition and Processing in Primates (eds. D. J. Chivers,
B. A. Wood & A. Bilsborough), pp. 177211. New York: Plenum.
Waterman, P. G. & Kool, K. M. (1994) Colobine food selection and plant chemistry.
In: Colobine Monkeys: Their Ecology, Behaviour and Evolution (eds. A. G.
Davies & J. F. Oates), pp. 251284. Cambridge: Cambridge University Press.
Waterman, P. G. & Mole, S. (1994) Analysis of Phenolic Plant Metabolites. Oxford:
Blackwell.
Waters, N. E. (1975) Aspects of dental biomechanics. In: Scientic Aspects of Dental
Materials (ed. J. A. von Fraunhofer), pp. 147. London: Butterworth.
(1980) Some mechanical and physical properties of teeth. In: The Mechanical
Properties of Biological Materials (eds. J. F. V. Vincent & J. D. Currey), pp. 99
135. Cambridge: Cambridge University Press.
Weijs, W. (1975) Mandibular movements of the albino rat during feeding. Journal
of Morphology 154: 107124.
Weijs, W. & Dantuma, R. (1981) Functional anatomy of the masticatory apparatus
in the rabbit. Netherlands Journal of Zoology 31: 99147.
Wendeln, M. C., Runkle, J. R. & Kalko, E. K. V. (2000) Nutritional values of
14 g species and bat feeding preferences in Panama. Biotropica 32: 489501.
White, S. N., Luo, W., Paine, M. L. et al. (2001) Biological organization of hydroxyapatite crystallites into a brous continuum toughens and controls
anisotropy in human enamel. Journal of Dental Research 80: 321326.
Whitten, A. J. (1982) Diet and feeding behaviour of kloss gibbons on Siberut
Island, Indonesia. Folia Primatologica 37: 177208.
Wictorin, L., Hedegard, B. & Lundberg, M. (1971) Cineradiographic studies of
bolus position during chewing. Journal of Prosthetic Dentistry 26: 236246.
Wilding, R. J. (1993) The association between chewing efciency and occlusal
contact area in man. Archives of Oral Biology 38: 589596.
Williams, G. (1956) The relationship between the length of the jaw and the length
of the molar series in some eutherian mammals. Journal of Zoology 126: 51
56.
Williamson, L. & Lucas, P. W. (1995) The effect of moisture content on the mechanical properties of a seed shell. Journal of Materials Science 30: 162166.
Wilsea, M., Johnson, K. L. & Ashby, M. F. (1975) Indentation of foamed plastics.
International Journal of Mechanical Sciences 17: 457460.
Wilson, E. O. (1975) Sociobiology: The New Synthesis. Cambridge, MA: Belknap
Press of Harvard University.
344
References
Wing, S. L., Hickey, L. J. & Swisher, C. C. (1993) Implications of an exceptional

fossil ora for Late Cretaceous vegetation. Nature 363: 342344.
Wong, K. (2003) An ancestor to call our own. Scientic American (January): 4251.
Wood, B. A. (1984) Tooth size and shape and their relevances to studies of human
evolution. Philosophical Transactions of the Royal Society London series B 292:
5764.
Wood, B. A. & Abbott, S. A. (1983) Analysis of the dental morphology of PlioPleistocene hominids. I. Mandibular molars: crown area measurements and
morphological traits. Journal of Anatomy 136: 197219.
Wood, B. A. & Collard, M. (1999a) The human genus. Science 284: 6571.
(1999b) Is Homo dened by culture? Proceedings of the British Academy 99: 1123.
Wood, B. A., Abbott, S. A. & Graham, S. H. (1983) Analysis of the dental
morphology of Plio-Pleistocene hominids. II. Mandibular molars: study of
cusp areas, ssure pattern and cross-sectional shape of the crown. Journal of
Anatomy 137: 287314.
Wood, B. A., Abbott, S. A. & Uytterschaut, H. (1988) Analysis of the dental
morphology of Plio-Pleistocene hominids. IV. Mandibular postcanine root
morphology. Journal of Anatomy 156: 107139.
Wrangham, R. W., Conklin, N. L., Chapman, C. A. et al. (1992) The signicance of brous foods for Kibale Forest chimpanzees. In: Foraging Strategies
and Natural Diets of Monkeys, Apes, and Humans (eds. A. Whiten & E. M.
Widdowson), pp. 1118. Oxford: Clarendon Press.
Wrangham, R. W., Jones, J. H., Laden, G. et al. (1999) The raw and the stolen:
cooking and the ecology of human origins. Current Anthropology 40: 567
594.
Wright, W. & Illius, A. (1995) A comparative study of the fracture properties of
ve grasses. Functional Ecology 9: 269278.
Wright, W. & Vincent, J. F. V. (1996) Herbivory and the mechanics of fracture in
plants. Biological Reviews 71: 401413.
Yamada, H. (1970) The Strength of Biological Materials (ed. F. G. Evans). Baltimore,
MD: Williams & Wilkins.
Yamashita, N. (1998) Functional dental correlates of food properties in ve
Malagasy lemur species. American Journal of Physical Anthropology 106:
169188.
(2000) Mechanical thresholds as a criterion for food selection in two prosimian
primate species. In: Plant Biomechanics 2000 (eds. H.-C. Spatz & T. Speck),
pp. 590595. Stuttgart: Georg Thieme.
(2003) Food procurement and tooth use in two sympatric lemur species.
Young, G. C., Karatajute, V. N. & Smith, M. M. (1996) A possible Late Cambrian
vertebrate from Australia. Nature 383: 810812.
Yucker, A. S., Matthews, K. L. & Sharpe, P. T. (1998) Transformation of tooth type
induced by inhibition of BMP signaling. Science 282: 11361138.
Yurkstas, A. A. (1965) The masticatory act: a review. Journal of Prosthetic Dentistry
15: 248260.
References
345
Yurkstas, A. A. & Manly, R. S. (1949) Measurement of occlusal contact area effective

in mastication. American Journal of Orthodontics 35: 185195.
Zahradnik, R. T. & Moreno, E. C. (1975) Structural features of human dental
enamel as revealed by isothermal water vapour sorption. Archives of Oral
Biology 20: 317325.
Zarkadas, C. G., Karatzas, C. N., Khalili, A. D. et al. (1988) Quantitative
determination of the myobrillar proteins and connective tissue content in
selected porcine skeletal muscles. Journal of Agricultural and Food Chemistry
36: 11311146.
Zhang, X., Gubbels, G. H. M., Terpstra, R. A. et al. (1997) Toughening of
calcium hydroxyapatite with silver particles. Journal of Materials Science 32:
235243.
Zingeser, M. R. (1969) Cercopithecoid honing mechanisms. American Journal of
(1973) Dentition of Brachyteles arachnoides with reference to alouattine and ateline afnities. Folia Primatologica 20: 351390.
Zuccotti, L. F., Williamson, M. D., Limp, W. F. et al. (1998) Modeling primate occlusal topography using geographical information systems technology. American Journal of Physical Anthropology 107: 137142.
Zwell, M. & Pilbeam, D. R. (1972) The single-species hypothesis, sexual dimorphism and variability in early hominids. Yearbook of Physical Anthropology 16:
6979.
Index
Abbott, S. H. 194
Abler, W. L. 131
abrasiveness 168, 183200
acid
from bacteria 6
from food 74, 182
Aerts, R. J. 77, 192
Agrawal, K. R. 108, 110, 130, 235
Aidos, I. 150
Aiello, L. 47
Aitchison, J. 35, 184
Akersten, W. A. 215216
Albrecht, G. H. 156
Alexander, R. M. 12, 78, 86, 147, 151
Alexanders caveat 144145, 158
allometry 133
Anapol, F. 46, 73
Anderson, D. J. 37, 70, 73
Andreasen, J. O. 31
angiosperms 210
anisotropy 30, 151, 265
antlers 95
Aranwela, N. 279
Ardran, G. M. 49, 55, 155
Ardrey, R. 233
Artiodactyla 154, 228
Ashby, M. F. 96, 117, 123124, 161, 201, 203, 207,
264, 281
Atkins, A. G. 94, 95, 107, 110, 117119, 124, 138,
139, 142, 159160, 163, 186189, 193, 252,
266, 275276, 279, 281
Austin, L. G. 60, 63
australopithecines, robust 198, 200, 242243
aye-aye 16
baboons 194, 226, 228
Baker, G. 70, 186
Ball, T. B. 186, 193
Balooch, M. 271
Baragar, F. A. 63, 72, 179, 244
Barnes, J. 12
Barthlott, W. 85, 171

Basmajian, J. V. 43, 54
bats 69, 126, 174, 186, 226
Beard, K. C. 238
bears 174
beavers 95, 194
Beecher, R. M. 33
beetles 186
Bell, E. A. 218
Bell, R. H. V. 150, 231
Bemis, W. E. 30
Benedict, F. G. 135, 165
Bennett, E. L. 225, 228
Berkovitz, B. K. B. 37, 196, 210
Bertram, J. E. A. 131
Bialek, W. 67
Biewener, A. 33, 282
birds 36
bite forces 33, 3739, 152153
blades 13, 105, 124, 126, 129, 206
Blanton, P. L. 51, 66
Bodmer, R. E. 242
Boesch, C. 246
Boesch, H. 246
bolus formation 8284
Bond, F. C. 108
Bonds work index 108
bone 27, 95, 129, 186
of attachment 34
jaw 3234
remodelling 33, 37
strain 33
Youngs modulus 34
Boughter, J. D. 74
Bourne, M. C. 68, 281
Bouvier, M. 155
bovids 151
Bowden, F. P. 182
Boyde, A. 23, 35, 131, 193
Brace, C. L. 233, 245
Braden, M. 31
346
Index
brain 9
Braley, L. C. 166
breakage function 60
breakage sites 169170, 198
Brear, K. 95
Brennan, J. G. 281
Brochu, C. 19
browsers 152, 192, 229
Brunet, M. 234, 243
bruxism 184185
bunodonty 198
Butler, P. M. 183
Cachel, S. 151
Cadden, S. W. 55
Calcagno, J. M. 245
Calder, W. A. 135
canids 174, 197
canines
crown height 198
denition 16
shape 129, 131, 216
size 173176
carbohydrate (starch) 217, 253
Carlson, D. S. 178
carnivores 56, 65, 129, 159163, 173174, 179,
196, 209217
canines 129, 216
carnassials 113, 161, 173, 196, 216
Carpita, N. C. 118
cats 174
cement 14, 35
cetaceans 17
Chaimanee, Y. 234
Chalk, L. 186
Chandrashekar, J. 76
Chapman, C. A. 224
Charalambides, M. N. 265
cheek pouches 52, 155, 195,
228
cheeks 51, 84, 174, 195, 243
cheeses 265
chewing see also mastication
cycle 22, 6467
rate 136, 147148
chimpanzees 126, 226
Chinese water deer 35, 179
Choong, M. F. 128, 254
Christensen, M. 250
Ciochon, R. L. 185
Cipollini, M. L. 94
Coates, M. I. 2
Cochard, L. R. 164
Coley, P. D. 76, 149, 205207
collagen bres 36
347
Collard, M. 242, 246

Collinson, M. E. 217
colobine monkeys 228
communication 9
smell 9, 73
sound 10
vocalization 79
composites 117
conodont 3
cooking 5, 245253
cork 281
Corlett, R. T. 69, 218, 220223,
254
Cornelissen, J. H. C. 149
Costa, R. L. 192
Cotterell, B. 92
Cottrell, A. H. 82, 281
Covert, H. H. 146
Cox, H. L. 123
crack
bluntness 111
extension 92
free-running 123
Creel, N. 152
Creighton, G. K. 153
crocodilians 19
Crompton, A. W. 32, 47, 55, 66, 68, 79, 115, 183,
213, 215
Crompton, R. H. 215
crystals see hydroxyapatite
crystal-free region in enamel 22
Curran, L. M. 218, 228
Currey, J. D. 2, 29, 95
cuticle (leaf ) 85
DArcy Thompson, W. 138
Dale, H. E. 165
Damuth, J. 145
Danielson, D. R. 186
Dantuma, R. 55
Darvell, B. W. 89, 116, 272, 273
Davies, A. G. 225, 228
de Bruijne, D. W. 45
De Luca, C. J. 54
Dean, M. C. 24, 47, 198
Dechow, P. C. 34
deformation 257
elastic 261
elastoplastic 261
plastic 89, 261
DeGusta, D. 166, 198, 200
Delgado, S. 2
DeLong, R. 193
Demes, B. 152
Demment, M. W. 147
348
dental decay
effect of acids 6
effect of sugars 56
dentaldietary adaptations 205209
dentine
calcospherites 27
description 2, 13, 2729
interglobular dentine 31
isotropy 31
mantle dentine 27
mechanical properties 3031
microwear 189190
odontoblasts 27
peritubular dentine 30
petrodentine 30
Deutsch, D. 26
diastema 37, 154, 157
Dibden, G. H. 26
Diekwisch, T. G. H. 26
digestive enzymes
gut 10
saliva 4345
dinosaurs 202205
Dixon, A. D. 69
dogs 174
dolphins 17
Dominy, N. J. 7678, 186, 223224, 226, 255,
281
Dorrington, K. L. 281
Douglas, W. H. 193
Drake, B. 70
Druzinsky, R. E. 147148
Dunbar, R. I. M. 226, 228, 242
dwarsm 156, 237238
Dyment, M. L. 36
Edgar, W. M. 6, 77, 83
Edwards, C. 279
Ehrhardt, D. 229, 230
elephants 21, 65, 96, 133, 156, 196, 224, 228
Emerson, S. B. 174
Emmons, L. H. 69, 211
enamel
ameloblasts 24
anisotropy 30, 265
cross-striations 24
crystal-free region 22
decussation 131
demineralization 77
description 2, 13, 2226
lamellae 31
maturation 26
mechanical properties 3031
microwear 187189
pellicle 78
Index
perikymata 181
rod/prism 22
as a sponge 26
striae of Retzius 24
tufts 31
Epstein, B. 59, 60
Eriksson, O. 149, 217
Essick, G. K. 86
Evans, A. R. 110
Every, R. G. 184185, 203
Ewanowski, S. J. 69
Farrell, J. 245
Felbeck, D. K. 189
felids 161, 174
Fengel, D. 118
Ferrets 196
Fincham, A. G. 26
ngernails 131
Fish, D. R. 65
Fleagle, J. G. 238
eshy fruits 126, 217
seed treatment 220226
with skin 218
stone cells 186
with peel 218, 224
uid
ow through enamel 21, 26
ow through dentine 21, 27, 70
in mouth 45, 80, 83
food, properties/qualities of
comminution 60, 99105, 107
energy density 150
external attributes 10, 70, 168169
homogeneity 151
internal (mechanical) properties 1011
stickiness 171173
structural connectivity 161
Youngs modulus 102
food particles
fracture, chance (probability) of 11, 169
fragmentation 11
lubrication by intra-oral uid 80
shape 128, 170171
size/rate of size reduction 5764, 99105,
149150, 166
size at swallowing 40, 69, 80
Fooden, J. 156
foods
beef 82, 84
brazil nuts 60, 81
carrots (raw) 45, 60, 84
owers 149
galls 152
grasses 56, 161, 198, 228
Index
leaves 56, 149, 150, 171, 228, 241, 279
meat 250251
onion 249
potato 248
proteins 7576
seeds 56, 117, 124, 127, 161, 173, 217, 227
turnip 248
yogurt 81
force
bite 23, 3739, 152153
muscle 37
Fortelius, M. 23, 110, 131, 135136, 142, 147, 153,
155156, 166, 193, 197
Fox, C. L. 181, 186
Fox, P. G. 26
fracture 88
elastic 9195, 137
elastoplastic 261
geometry 9195
limits 9697
mechanisms 124125
modes 93, 267268
plastic 95
Frank, F. C. 99
Freeman, P. W. 110, 150151, 174, 186, 213, 220,
226
friction 39, 40, 44, 86, 97, 101, 117, 131, 182, 188,
189, 192193, 230, 232, 280281
Frisch, J. E. 198
frugivores 217220, 226228
Gabbott, S. E. 3
Galileo, G. 88, 117, 138, 161
Gantt, D. G. 198, 234
Gardner, R. P. 60, 63
Garn, S. M. 238
Gautier-Hion, A. 149, 218, 219
gel 26, 265
genes to form teeth 7
geometry, of loading 9091
geophagy 186
Gibbs, C. H. 21, 70
Gibbs, S. 7
Gibeaut, D. M. 118
Gibson, K. R. 245
Gibson, L. J. 34, 116117, 123, 161, 204, 281
Gilbertson, T. A. 7677
Gipps, J. M. 195
Glantz, P. O. 45
Goldberg, L. J. 65
Gomez-Brito, L. 126
gomphotheres 200
Goodman, L. E. 192
Gordon, I. J. 153, 166, 229, 230
Gordon, J. E. 22, 88, 117, 121, 204
349
Gordon, K. D. 189
Gosline, J. M. 131
Gould, S. J. 135, 139, 146, 150, 156
Grajal, A. 146
grazers 152, 229
Greaves, W. S. 73, 155, 192, 243
Green, B. G. 78
Grifth, A. A. 88
Grine, F. E. 198, 235
Groves, C. P. 242
guinea pigs 184, 190
Gurney, C. 139
Habelitz, S. 30, 77
Haines, D. J. 21
Haldane, J. B. S. 145
Hamilton, G. M. 192
Hankins, G. E. 101
hardness 184, 186, 205, 216, 269271
Harris, B. 118
Harrison, M. J. S. 225
Harvey, P. H. 164
Hatley, T. 233, 236
Hayes, V. J. 173
Heath, M. R. 33, 108
Hector, M. P. 74
Helkimo, E. 167
Hellekant, G. 75
Hendrichs, H. 147
Heraclitus 12
herbivores 64, 69, 151152, 156, 204, 228232
tooth size 163164, 230232
tooth wear 190193
Herrera, C. M. 218
Herring, S. E. 46, 73, 147, 174, 176, 244
Herring, S. W. 147, 174
Hertel, F. 198
Heyes, J. A. 126
Hiiemae, K. 21, 47, 55, 65, 66, 68, 79
Hill, A. V. 139, 143
Hill, D. A. 152, 254
Hillerton, J. E. 213, 254
Hills, M. 235
hippopotamus 156, 174, 190
Hoaglund, R. G. 104
Holland, G. R. 29, 70
hominins 198, 234
hoof, of a horse 131
Hooker, J. J. 217
Hopson, J. A. 209, 211, 212
Howe, H. F. 149
humans, diet and teeth 233253
Hume, I. D. 146
Hunt, J. 139
Hunter, J. P. 7
350
Index
Hutchings, J. B. 80
Hrycyshyn, A. W. 43
hyenas 217
hydroxyapatite 21
in dentine, bone and cement 27
in enamel 22
Hylander, W. L. 33, 39, 40, 121, 153, 155, 198, 219,
225
hyoid bone 50
hypsodonty 197
Illius, A. W. 153, 229, 230
incisors
crown height 198
denition 16
size 153
spatulate 129132
indentation 30, 99, 186190, 193,
269
ingestion 97, 128132, 208
Inglis, C. E. 110
insectivores 173, 186, 209217
Insects
cuticle 113, 186, 192
as pollinators 210
Iwamoto, T. 226, 228
Jablonski, N. G. 179, 215, 241
Janis, C. M. 193, 197, 229, 230
Janson, C. H. 149, 218, 219
Janvier, P. 2
Janzen, D. H. 218, 229
Jarman, P. J. 150, 231
jaw
bones (maxilla/mandible) 39, 155
gape 73, 174179, 244, 247
joint see temporomandibular joint
mechanics 7173
movements 21, 32, 39, 64, 232, 235
muscles see muscles
see also mandible
Jeannorod, M. 255
Jennings, J. S. 275
Jensen, J. L. 77, 192
Ji, Q. 209
Johanson, Z. 3
Jolly, C. J. 226, 228, 233, 235, 241,
244
Jones, D. R. H. 101, 201, 264, 281
Jones, S. J. 35
Jordano, P. 218
Jernvall, J. 7, 182, 228, 232
Jeronimidis, G. 121, 124
Jungers, W. E. 243
Junqueira, L. C. 253
KIC (critical stress intensity factor) 103,

267268
Kappelman, J. 233, 236
Kastelic, J. 36
Kawamura, Y. 74
Kawasaki, K. 3
Kay, R. F. 59, 121, 135, 146, 164, 173, 179, 198,
220, 225, 227
Kayser, A. 167
Kemp, A. 4
Kendall, K. 89, 105, 110, 184, 201, 270, 276, 280,
281
Kermack, K. A. 145
Keyser, A. W. 235
Khan, A. A. 276
Kielan-Jaworowska, Z. 115
Kier, W. M. 42, 269
Kiltie, R. A. 228
Kinney, J. H. 30
Kinzey, W. G. 173, 219, 228, 243
Kirk, E. C. 220, 227
Kirkham, J. 34
Kleiber, M. 134136
Klineberg, I. 69
koalas 195
Kohler, M. 179
Kool, K. M. 229
Koritzer, R. T. 198
Krishnamami, R. 186
Kursar, T. 76, 207
LaBarbera, M. 135, 145
lactation 165
lagomorphs 16, 154
Laine, P. 69
Lake, G. J. 113
Lakes, R. S. 269
Lambert, J. E. 226
Lambrecht, J. R. 167
Landry, S. O. 155
language 253
Lanyon, J. M. 195
Lanyon, L. E. 33
larynx 10
Laska, M. 75
Lauer, C. 164
Lavigne, G. J. 184
Lawn, B. R. 92, 95, 99, 104, 119, 123, 187188,
193, 266268, 270, 276, 281
Lee, C.-S. 193
Lee, L. L. 69
Lee, V. M. 73
Leigh, E. G. 205
Leigh, R. A. 118
Leighton, D. R. 217, 219
Index
Leighton, M. 217219, 226, 228
lemurs 225, 281
Lermer, C. M. 77
Levey, D. J. 94
Levine, D. S. 245
Lewis, A. B. 238
Lieberman, D. E. 3334, 36
ligamentum nuchae 54
Lightoller, G. H. S. 51
Lillford, P. J. 69, 80, 82, 86, 107, 250
Linden, R. W. A. 35, 69, 7273
lizards 19
loading
bending 139, 275
compressive 90, 91, 275
shear 90, 91
tensile 90, 91, 275
locomotion, compared to mastication 66
Logan, M. 195
lophs 232
Lovegrove, B. G. 135
Lowrison, G. C. 57, 97, 246
Lucas, P. W. 45, 60, 66, 6970, 7882, 84,
94, 96, 108, 110, 116117, 119, 122, 127,
130, 152, 164, 167, 169, 170, 174175, 186,
190, 198, 203, 218, 220223, 227228,
234, 235, 240, 243244, 253255, 271, 272,
279
Luke, D. A. 60, 79, 80, 167170, 190,
227
Lumsden, A. G. S. 185, 254
Lund, J. P. 65
Lundberg, M. 244
lungsh 30
Luo, Z.-X. 6, 211
Luschei, E. S. 65
macaques 16, 152, 178, 220, 224
Macfadden, B. J. 197
Macmillan, N. H. 275
mandible 178
condyle 39, 178
description 39
Maglio, V. J. 21, 156, 200
Magnusson, W. E. 69
Maguire, B. 248
Mahaney, W. C. 186
Mai, Y.-W. 92, 9495, 107, 110, 119, 124, 139,
142, 159160, 193, 209, 252, 266, 276,
281
malocclusions 20
Mammaliaformes 209
mammals, why they evolved mastication 34,
5657
Manly, R. S. 166167
351
Mao, J. J. 33, 73
marmosets 174
Marshall, D. B. 104
Martin, L. B. 234
Martin, P. S. 218
Martin, R. D. 17, 135, 145, 147, 154, 233
Martinez del Rio, C. 6
mastication 208
description of process 78, 97
masticatory sequence 6467
reasons why it evolved in mammals 34
scaling arguments 133, 135, 145146,
163164
Mattes, R. D. 76, 245
Matthews, P. B. C. 69
matrix, of hard tissues 21
Maynard Smith, J. 173
McCann, M. C. 118
McKenzie, W. M. 110
McKey, D. B. 225, 228
McMahon, T. 144
Mendel, F. C. 65
Mendoza, M. 152, 154, 197, 229
Mercader, J. 246
metabolic rate 56, 134137
Metcalfe, C. R. 186
Mesozoic vegetation 205207
mice 196
Mills, J. R. E. 20, 183
Milton, K. 229
modiolus 51, 155, 243
Moelleken, S. M. C. 230
Mohsenin, N. N. 126
molars
action 91
deciduous vs. permanent 17
shape 99105
size 133, 153154, 163164
tribosphenic 6, 113, 215
wear 184
Mole, M. 77
Mongini, F. 65
monkeys 16, 69, 77, 129, 152, 157158, 166, 178,
189, 219220, 228, 243
Moreno, E. C. 26
Morgan, E. 233
moths 186
mouse deer 179
mouth see oral cavity
mouthslit 51, 243
movements see jaw movements
Mowlana, F. 108
Moyà Solà, S. 179
Muhlemann, H. R. 37
Murphy, T. 194
352
Index
Murray, C. G. 185
Murray, P. 52, 59
muscles 4551
architecture 73
buccinator 51
description of jaw muscles 7
digastric 47, 151
of face 5153
jaw-closing 21, 33, 46, 235
jaw-opening 47
lateral pterygoid 40, 47
masseter 46, 151, 174
medial pterygoid 46, 174
of neck 5354
spindles 69
sternocleidomastoid 54
temporalis 45, 151
of tongue 5051
mustelids 161
Mysterud, A. 147
nanomaterials 31
Napier, J. R. 242
nasal cavity 9
Nelson, G. 75, 76
Neinhuis, C. 85, 171
nervous system 9, 29
control of jaw muscles 7173
control of mastication 6770
inside tooth 69
Nguyen, Q. 150
noise 70
non-linear elasticity 265
Norconk, M. A. 173, 228
Nose, K. 283
notch sensitivity 161, 271
Oates, J. F. 186
Olthoff, L. W. 60
OMullane, D. M. 6, 45, 77, 83
opossum, American 55
oral cavity
description 40
mucosal lining 40, 69
size 146148
soft tissues 4043
orang-utans 126
Orchardson, R. 55
Oron, U. 32
Osborn, J. W. 1819, 2224, 27, 31, 39, 40, 69,
72, 129, 179, 185, 193195, 210, 236238, 244,
254
Ostry, D. J. 68
Otani, T. 69
Ottenhoff, F. A. 67
ovarian cyst 19
Overdorff, D. J. 225, 227, 281
Owall,
B. 69
Owen-Smith, N. 147, 192, 229
Page, D. H. 121
Paphangkorakit, J. 22, 27, 39, 69
Panger, M. 246
Partt, G. J. 37
peccaries 174
peeling 280
Peleg, M. 126, 219
Pereira, B. P. 113, 131
Perez-Barberia, F. J. 166, 229
periodontal ligament 14, 3437
collagen bres 36
mechanoreceptors 69
Perissodactyla 154, 228
Peters, C. R. 195, 208, 228, 235, 243, 248
Peters, R. H. 56, 136, 146, 147, 149
Pettifor, E. 200, 241
pharynx 8, 43, 79
Phua, P. B. 69
Picton, D. C. A. 36
pigs 174, 228
Pilbeam, D. R. 135, 146, 164
piscivores 150
plant tissues 115
abrasives in 193
bark 228, 279
cell wall 117119
cellulose 117
crack paths in (fracture of ) 117128
owers 149
fruits 149, 163, 173, 182, 217218
leaves 56, 149, 171, 228, 241, 279
lignin 117
middle lamella 124
phytoliths 186
plant storage organs 206, 228
plastic buckling 121
pods 114, 117, 161
seeds 56, 117, 124, 127, 161, 173, 220228
woody tissue and its toughness 121, 122, 208
platypus 52
Plavcan, J. M. 173
Poissons ratio 21, 90, 95, 110, 115116, 131, 213,
268269
Poole, D. F. G. 196
Popovics, T. E. 110
postcanines 16, 19, 153
pouches, in cheeks 52, 155, 195, 228
pregnancy 165
premolars 17
molarized 235, 243
Index
Preston, C. M. 121
Price, G. R. 173
primates 16, 173, 174, 179
apes 16, 173, 179, 200, 224
aye-aye 16
baboons 194, 226, 228
chimpanzees 126
colobine monkeys 228
lemurs 281
macaques 16, 152, 178, 220, 224
marmosets 174
molar size gradient in 238242
monkeys 16, 69, 77, 129, 152, 157, 158, 166,
178, 189, 219, 220, 228, 243
orang-utans 126
tamarins 219
tarsiers 215
Prinz, J. F. 45, 66, 69, 74, 78, 8182, 131, 196, 213
Prothero, D. R. 156
pulp 2729
Purnell, M. A. 3
Purslow, P. P. 95, 112, 113, 117, 207, 250, 265, 268
quartz 188
rabbits 55, 155
Radinsky, L. 174
Rafter, J. A. 198, 234
Ralls, K. 164
Rasmussen, S. T. 30
rat skin 113, 207
Raven, J. A. 205
Ravosa, M. J. 148
Rees, L. A. 39
Reinhard, K. J. 186
Rensberger, J. M. 23, 31, 59, 72, 131, 193, 216, 227
Renson, C. E. 31
rhinoceros 193
Ridley, H. N. 218, 226
Riesz, R. R. 56
Ringel, R. L. 69
Ringtailed possum 195
Robbins, C. T. 147
Roberts, K. 118
Robinson, C. 26, 34
Robinson, J. T. 235
rodents 16, 21, 52, 65, 69, 95, 154155, 194, 228
Rogers, M. E. 146
Rolls, E. T. 77, 86
Romer, A. S. 173
Rose, H. E. 108, 227
Rose, K. D. 184
Rosenberger, A. L. 219
RosinRammler equation 59
Roth, V. L. 156
Rubin, C. T. 33
Ruff, C. B. 129, 216
Runham, N. W. 216
Ryan, J. M. 52
Rybczynski, N. 56
St Hoyme, L. E. 198
saliva
amylase 76
functions 4, 6, 9, 40, 4345, 7378
proline-rich proteins 77
viscosity 45, 84
salivary glands 43
Saniotti, T. M. 69
Sansom, I. J. 2
Sanson, G. D. 110, 185, 195
Sato, K. 150
Savage, R. J. G. 56, 115, 216
Sayer, B. G. 121
scaling
fracture (negative allometry) 137144
geometric (isometry) 136137
Scapino, R. P. 32
Schmidt-Neilsen, K. 56
Schoeld, R. M. S. 186, 213, 254
scissors 277280
Sclafani, A. 75
sea cows 179
Sealey, D. F. 126
Selanne, L. 182
selection function 60, 169, 244
Semaw, S. 246
sensation
fats (lipids) 77, 217
oral sensation 45
sensory feedback (and its limits) 6770
smell 9, 73
sound 70
special senses 7
taste 9, 73, 75
Sereno, P. C. 156, 204
sexual dimorphism 164166
Shama, F. 264
Sharp, S. J. 187188
Sharpe, P. T. 19
Shaw, D. M. 94
Sheikh-Ahmad, J. Y. 110
Sheine, W. S. 59
Shellis, R. P. 26
Sherman, P. 264
Shibata, E. 69
Shipley, L. A. 146, 229
shrews 196, 215
Sifakas 227
353
354
Siirila, H. S. 69
silica 188, 193, 198, 205, 279
Silva, S. E. 245
Simons, E. L. 220, 227
Simpson, G. G. 96, 129, 197, 216
Sita-Lumsden, A. G. 213, 215
Smith, J. M. 129, 216
Smith, K. K. 9, 40, 4243, 50, 210,
269
Smith, M. M. 2, 3
Smith, R. J. 155, 177, 244
Sofaer, J. A. 238239
Solounias, N. 230
Spalinger, D. E. 229
Spears, I. R. 30, 70
Spencer, M. 72
Spielman, A. R. 73
stickiness 168, 171173
stiffness 137, 258
Stirton, R. A. 197
strain
denition 259
strain energy 89, 137
true strain 264
Strait, S. G. 186, 215, 220, 225, 227,
254, 281
strength
compressive 89
fracture 260
yield 186, 261, 265
stress
denition 259, 265
fracture 102
hydrostatic 265
tensile 103
stressstrain curve 40
Su, H. H. 69
suids 174
Sullivan, R. M. E. 108, 227
surface roughness/smoothness 11, 181195
swallowing 78, 43, 7879
bolus model 8284
dual threshold 8082
particle size threshold 80
Swanson, S. A. V. 39
symphysis 32
synapsids 209, 213
Synge, J. L. 3637
Szalay, F. S. 233, 235
Szczesniak, A. S. 10
Tabor, D. 182
tamarins 219
Tanne, K. 40
tannins 77, 192, 232
Index
tarsiers 215
taste 7375
amino acids 7576
bitterness 76
fatty acids 76
starch receptor 76
sugars 75
taste buds 42
umami 76
Tavare, S. 236
tayussuids 174
Teaford, M.F. 70, 181184, 187189, 195,
227, 233
teeth see tooth
temporomandibular joint 3940
articular disc 40
articular eminence 39
ligaments 39, 72
Teng, S. 40
tenrec 32
Terborgh, J. 219, 228
thegosis 185
Theimer, T. C. 228
Thesleff, I. 7
Thexton, A. J. 55, 79
Tittelbach, T. J. 76, 245
Tomos, A. D. 118
tongue 4042, 67, 195
tools 246
tooth
alignment 1921
classes 16
cusps 99105, 125126
decay 56
dermal armour 23
description 1319
drift 37
features 13, 15
homology 18
number 211, 236238
origins 25
in ovarian cyst 19
roles of teeth 97
sharpness 31, 110116, 124, 134, 184,
185
socket 3437
structure 1316, 2129
tissues 2131
wear 181201
toughness 262263, 266268
tests 276, 277
tree shrews 69, 211
trouser-tear 277
Trulsson, M. 86
Turner, I. M. 149, 218, 228
Index
Uden, P. 69
Ungar, P. S. 129, 149, 181182, 185, 189, 195, 198,
224, 233, 234
ursids 174
Utz, K. H. 70
van den Braber, W. 244
van der Bilt, A. 60
van der Glas, H. W. 60, 169170
van Reenen, J. F. 194, 198
van Roosmalen, M. G. M. 226
van Schaik, C. P. 246
Van Soest, P. J. 69, 119, 147, 229230
van Valen, L. 197
Van Valkenburgh, B. 129, 161, 186, 196, 198,
216217
van Vliet, T. 45, 84, 264
vertebrates
birds 34
collagen content, of soft tissues 150
conodonts 3
description 13
exoskeleton 2
lungsh 30
soft tissues 112, 131
Vickers, Z. 70
Vincent, J. F. V. 70, 113, 117, 126, 131, 161, 213,
215, 232, 275, 276, 279, 280, 282
Vinyard, C. J. 131, 174
viscoelasticity 264
von Koenigswald, W. 193
Voon, F. C. T. 59
wadging 195
Walker, A. 70, 182184, 186, 192, 194, 201,
227228
Walker, P. L. 59, 126, 174179, 195
walruses 179
Waterman, P. G. 77, 218, 229
wear 181201
abrasion 183, 198
approximal 194
and dental efciency 195196
erosion 182
pits 182, 184, 186, 187

scratches 182, 184, 188
terminology 183184
wedges
vs. blades 104, 126
tests 276277
Weiss, K. M. 3
Wegener, G. 118
Weijs, W. A. 49, 55
Weins, W. N. 110
Wendeln, M. C. 149
whales 17
White, S. N. 30
Whitten, A. 226
Wictorin, L. 244
Wilding, R. J. 167
Williams, G. 154
Williamson, L. 123, 127
Williamson, M. D. 182
Wilsea, M. 115, 270
Wilson, E. O. 10
Wing, S. L. 217
Wong, K. 234
Wood, B. A. 194, 235, 242243, 245246
Wrangham, R. W. 5, 236, 246, 248, 251,
253
Wright, W. 232, 279
Wyke, B. O. 69
Yamada, H. 160, 250
Yamamoto, T. 74
Yamashita, N. 129, 225, 281
Yeoh, O. H. 113
Young, G. C. 2
Youngs (elastic) modulus 260, 263264
Yucker, A. S. 18
Yurkstas, A. A. 80, 167
Zahradnik, R. T. 26
Zarkadas, C. G. 150
zebras 196
Zingeser, M. R. 158, 185, 226, 243
Zuccotti, L. F. 182
Zwell, M. 164
355

LUCAS How Teeth Work 1st Ed 2004 PDF

Uploaded by

Copyright:

Available Formats

LUCAS How Teeth Work 1st Ed 2004 PDF

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

LUCAS How Teeth Work 1st Ed 2004 PDF

Uploaded by

Copyright:

Available Formats

DENTAL FUNCTIONAL MORPHOLOGY

How Teeth Work

Dental Functional Morphology offers an innovative alternative to the

CAMBRIDGE UNIVERSITY PRESS

To my wife Mariati and my daughters, Katherine and Diana,

1 How to get excited about teeth

2 The basic structure of the mammalian mouth

3 How the mouth operates

7 The evolution of the mammalian dentition

Teeth cause such dreadful problems in humans that interest in them by

sciences. Well-cited journals here include the American Journal of Physical

Dental Anatomy and Embryology (Volume 2 of A Companion to Dental

A ick-page animation starts here, illustrating the evolution of the human

tribosphenic molar from which those of all living therian mammals

How to get excited about teeth

1 How to get excited about teeth

to explain about just about anything and to be credible that explanation

The origins of teeth

1 How to get excited about teeth

would simply slow it down. The alternative solution, to which mammals

Are teeth just hangers on?

on several food items within a 2-hour period. How could an individual

1 How to get excited about teeth

Basic functional considerations

of upper molars called the hypocone, not present in the tribosphenic

1 How to get excited about teeth

Basic functional considerations

uniquely mammalian possessing a musculature that appears to have no

1 How to get excited about teeth

these animals seem to communicate better by smell and body language

The basic model

of surface quality such as surface roughness and stickiness. These surface

1 How to get excited about teeth

The basic structure of the mammalian mouth

2 The structure of the mammalian mouth

What are teeth?

2 The structure of the mammalian mouth

surrounding jaw bone all seem to have a common developmental origin

Types of teeth and their denitions

generally given to postcanine teeth: molars and premolars. Molars develop

2 The structure of the mammalian mouth

identify canines by their position in that jaw, either in relation to a bony

2 The structure of the mammalian mouth

Individual tooth tissues

Spreading the force of tooth contacts across the dentition by precisely

2 The structure of the mammalian mouth

part, this ow must be through enamel, but Paphangkorakit & Osborn

Individual tooth tissues

2 The structure of the mammalian mouth

Enamel formation by ameloblasts is continuous, but not completely

Individual tooth tissues

2 The structure of the mammalian mouth

Individual tooth tissues

Dentine and pulp

2 The structure of the mammalian mouth

Individual tooth tissues

2 The structure of the mammalian mouth

Appendix B lists the principal mechanical properties of these crucial tissues.

Individual tooth tissues