h _. ~. f: .
. J: {m It
ﬁnk,“

‘

5
. .
My. ,

 

 

 

 

 

www.mwnﬁ
, m :1... «u.
2.? .l .J

.quvu. . .vaJ...mv.M1Jn.-.

.. if? .
.

bawda

I!) 95.

ut.‘ an. a...
égmmmmis. Jwﬂaauud ,
. :1 z -

a

ll .
lit-.12!
.1416! I.

n~¢z,.
n. 1.!

m- .
.:.l'r3}6 1|
. .i .2. .. .
lurid. w.»
z!

, v.12;

I:
let-V.

31...; 1:;

3.)... I21. )5;
1.1573401: :0
9 .v 7 5.:

‘31!" . . 3

mm? . . _ .. ifﬁéﬁ , ‘ _ 4 $3.»...

 

. y
3..

 

This is to certify that the
dissertation entitled

BEHAVIORAL AND MORPHOLOGICAL DEVELOPMENT IN
A FEMALE-DOMINATED SPECIES, THE SPOTTED HYENA
(CROCUTA CROCUTA)

presented by

Jaime Beth Tanner

 

has been accepted towards fulﬁllment
of the requirements for the

 

LIBRARY
Michigan State

Zoology
Program in Ecology,
Doctoral degree in Evolutionary Biology and
Behavior

y. .w'drSity

 

 

 

 

 

%or%u re
LIX/a;

Date

MS U is an affirmative-action, equal-Opportunity employer

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DaleDue.indd-p.1

BEHAVIORAL AND MORPHOLOGICAL DEVELOPMENT IN A FEMALE-
DOMINATED SPECIES, THE SPOTTED HYENA, CROCUTA CROCUTA

By

Jaime Beth Tanner

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSHOPHY

Department of Zoology
Program in Ecology, Evolutionary Biology and Behavior

2007

ABSTRACT

BEHAVIORAL AND MORPHOLOGICAL DEVELOPMENT IN A FEMALE-
DOMINATED SPECIES, THE SPOTTED HYENA, CROCUTA CROCUTA

By

Jaime Beth Tanner

In this dissertation, I examine changes in skull morphology and feeding
behavior throughout ontogeny in order to determine if there are developmental
constraints operating during the life history of a unique carnivore species, the
spotted hyena (Crocuta crocuta). Spotted hyenas are highly durophagous,
capable of cracking open and consuming bones of large diameter, and their
skulls are highly specialized in order to accomplish such tasks. I use a novel
approach, ﬁnite-element analysis, to investigate adaptations in the skull that
enable hyenas to generate and sustain large bite forces. In particular I test
hypotheses about the functional signiﬁcance of a uniquely elongated fronto-
parietal sinus during bone-cracking. The fronto-parietal sinus, together with a
vaulted forehead, comprise a unique complex in the hyena skull that aids in
dissipating stress away from the facial region and back along the sagittal crest
during bone-cracking bites.

Spotted hyenas, unlike most other carnivores, are highly gregarious and
live in clans that can comprise up to 80 individuals. The prey and carcasses on
which hyenas feed exist as ephemeral and unpredictable patches of food.
Therefore, numerous individuals are likely to be feeding on a carcass

simultaneously and consequently competition can be intense. In an environment

with the dual challenges of intense competition and a durophagous diet, selection
should favor individuals capable of feeding quickly and utilizing all parts of a
carcass. Theory predicts that selection should favor enhanced performance
during life history stages in which the risk of mortality is high yet juveniles are
highly disadvantaged with respect to feeding performance relative to adults
despite a long period of offspring dependence. This disadvantage is likely due to
their small size and inexperience.

I utilized a unique collection of known-age skulls of wild spotted hyenas to
document changes in skull morphology throughout ontogeny. Skull size
continued to increase until 26 months of age shortly after reproductive maturity.
However, skull shape continued to change until 35 months of age, well into
adulthood. Similarly, changes in feeding performance as measured by ingestion
time, continued to change until the third year of life. The results from this work
suggest that there are indeed developmental constraints on skull morphology
and feeding. This provides initial support for a new theory explaining the
evolution of female dominance suggesting that a combination of developmental
constraints and intense feeding competition led to dominant females.

Lastly, in order to investigate whether a form of play often seen in
carnivores, non-nutritive chewing, reﬂect musculo-skeletal development, I
investigated ontogenetic variation in this and other forms of play. Non-nutritive
chewing does not reﬂect morphological changes and may instead allow young

hyenas to investigate novel objects in their environment.

Copyright by
Jaime Beth Tanner

2007

 

ACKNOWLEDGEMENTS

I have been incredibly fortunate during my graduate career and, more
broadly, throughout my life to have been surrounded by many supportive people.
First and foremost, I need to thank my advisor, Dr. Kay Holekamp. I am truly
grateful to have had her as an advisor and mentor. Kay provided me with an
opportunity that I had dreamed about since I was eight years old. More
importantly, she provided me with insightful professional and scientiﬁc training
and always pushed me to put forward my best effort. If I become even one-eighth
of the scientist that she is, I will consider myself to be very successful.

In addition to Kay, my advisory committee has provided useful comments
and support that have improved this dissertation. Dr. Joe Vorro always had kind
words of encouragement when I sought his input on my work. Dr. Laura Smale
provided a great deal of knowledge and advice and has been very supportive
throughout my graduate career. Dr. Barb Lundrigan served as a mentor,
supporter, and friend during my time at MSU and her advice and support were
critical to the success of the work presented here.

The College of Natural Sciences, the Graduate School, the department of
Zoology and the program in Ecology, Evolutionary Biology and Behavior at
Michigan State University have all provided me with funding and logistical
support. In addition, I received funding from the American Society of
Mammalogists and Sigma Xi. The wonderful staff in the Zoology office were
always incredibly helpful and patient with me over the past six years and I am

very grateful to them. My research was part of the Mara Hyena Project, which

has been funded by the National Science Foundation grants IBN9309805,
IBN9630667, |BN9906445, IBNO113170, IBNO343381, and IOBOG18022. I would
also like the thank the Ofﬁce of the President of Kenya for permission to conduct
this research and to the Kenya Wildlife Service, the Narok County Council, and
the Senior Warden of the Masai Mara National Reserve for their cooperation.

This dissertation includes data collected over 18 years and I am indebted
to a number of individuals involved in the data collection, including Nancy Berry,
Chantal Beaudoin, Erin Boydston, Aimee Cokayne, Susan Cooper, Audrey
DeRose-Wilson, Stephanie Dloniak, Martin Durham, Anne Engh, Josh Friedman,
Paula Garrett, lsla Graham, Tyson Harty, Cathy Katona, Karen Kapheim, Kay
Holekamp, Joe Kolowski, Kieth Nelson, Karen Nutt, Gabe Ording, Wiline Pangle,
Laura Smale, Jenn Smith, Micaela Szykman, Kevin Theis, Russ Van Horn, Sofia
Wahaj, Kim Wiebel, Kim Wooten, and Brad White. Additionally, I have had a
small army of undergraduate research assistants that have assisted me on
numerous components of my work, including Stephen Bowman, Suzi Casmer,
Leah DeRose—Wilson, Michelle Engmark, Melissa Gallega, Danielle Harbin,
Rachel Kappler, Leanne Klocke, Lindsay Manuszak, Christine Mikkola, Jami
Pawlowski, Sheri Sanders, and Kim Wooten.

I would like to thank Laura Abrazcinskas and Paula Hildebrand at the
museum for their help in organizing and preparing the specimens used in this
study. I have been very fortunate to collaborate with faculty, both here at MSU
and at other institutions, and these women have provided great training and have

served as role models to me. Dr. Sharleen Sakai in the Psychology Department

vi

at MSU provided me with the opportunity to develop skills in analyzing CT scans
and this work has led to a number of exciting projects. Dr. Miriam Zeldtich at the
University of Michigan has been very generous with her time and advice on using
geometric morphometrics and on statistical analyses. Lastly, Dr. Betsy Dumont at
UMASS opened her lab to me for several weeks and has trained me in an
exciting new technique for analyzing functional morphology. I look fonNard to
future work with her.

I am so lucky to have lived and studied in an amazing country for several
years. I can’t begin to enumerate the many ways in which so many people have
contributed to my success and well-being during my time in Kenya. In particular, I
need to thank amazing friends from ﬁsi camp, including John and Peris Keshe,
Moses Sairowa, James Kerembe, Stephen and Lesingo Nairori. They were not
only amazing people to work with, but also were incredibly loyal friends. In
general, I would like to thank my many friends in Talek and lntipilukwani.
Additional friends in the Mara that helped with car troubles, balloon ﬂights, hot
showers, and most importantly, laughs include Milton and Elly Kirkman, Andrew
Peart, Jakob Bro-Jorgensen, Stephanie Dloniak, Howard Saunders, Bilal Butt,
Nasha (Filis Tasel) and Piers Mason. My dearest friends, the Sinnarys
(Abdulwahab, Amna, Mohammed, Lubaba, Mihad, Luddan),Faith Model, and
Maraka Ewoi are truly like family to me and their hospitality and kindness are
beyond compare.

The Holekamp lab has been a great source of friends throughout my time

here. When I ﬁrst started, Micaela Szykmann, Anne Engh and Russ Van Horn

Vii

 

were great role models and friends. More recently, Katy Califf, Leslie Curren and
Kate Shaw have become great friends and helped to make sure that I laughed in
this past year as I struggled to ﬁnish. Soﬁ Wahaj has been a great friend both in
East Lansing and in Kenya and we have shared many laughs and adventures
together.

I am so grateful to have met Stephen Thomas and in the past two years
his friendship has been incredibly valuable to me. Terri McElhinny and Pat Bills
have been close friends since I started at MSU. Terri was the ﬁrst person I met
that was as big of a dork about skulls and bones as l was and we’ve spent many
hours being geeky together. She has helped me in so many ways, both
professionally and personally. Kevin Theis has been a huge source of support,
friendship, discussions and debates, and working with him in the ﬁeld was a true
pleasure. I was so lucky to start my graduate career with two people who would
become two of my closest friends, Joe Kolowski and Heather Watts. No matter
where we are, the support we provide for one another is unparalleled.
Additionally, some of my greatest support has come from friends afar, including
Amanda Oxenhom, Andrew Engell, Dan Gold and Corinne Sudberg.

I could never have achieved any success without the amazing and
unconditional love and support of my family. Despite not fully understanding what
would possess me to move so far away and live in a tent for years on end, my
family has encouraged me in every new adventure. I can’t begin to express my
gratitude for such a wonderful family. Finally, I would like to thank my third grade

teacher, Mrs. Weigman, for instilling in me a passion for Africa and for wildlife.

viii

TABLE OF CONTENTS

LIST OF TABLES ............................................ x
LIST OF FIGURES ........................................... xi
GENERAL INTRODUCTION .................................... 1
Overview of chapters ..................................... 5
CHAPTER 1
OF ARCS AND VAULTS: THE BIOMECHANICS OF BONE-CRACKING IN THE
SPOTTED HYENA (CROCUTA CROCUTA). ........................ 11
Introduction. ............................................. 1 1
Methods ................................................. 17
Results. ................................................ 22
Discussion ............................................... 29
CHAPTER 2
ONTOGENETIC CHANGE IN SKULL MORPHOLOGY AND MECHANICAL
ADVANTAGE IN THE SPOTTED HYENA (CROCUTA CROCUTA) ......... 34
Introduction. ............................................. 34
Methods ................................................. 39
Results. ................................................ 44
Discussion ............................................... 57
CHAPTER 3

EFFECTS OF AGE, MORPHOLOGY AND SOCIAL RANK ON FEEDING
PERFORMANCE IN FREE-LIVING SPOTTED HYENAS (CROCUTA

CROCUTA). ............................................... 63
Introduction. ............................................ 63
Methods ................................................ 70
Results. ............................................... 75
Discussion .............................................. 85

CHAPTER 4

ONTOGENETIC VARIATION IN THE PLAY BEHAVIOR OF SPOTTED HYENAS

(CROCUTA CROCUTA) ....................................... 95
Introduction. ............................................ 95
Methods ................................................ 101
Results. ............................................... 108
Discussion .............................................. 1 18

APPENDIX ................................................. 132

LITERATURE CITED ......................................... 135

ix

LIST OF TABLES

Table 1.1. Muscle forces applied to each model in order to produce 55OON
of bite force. Values (in Newtons) represent the sum of identical forces applied
on the left and right ........................................... 28

Table 3.1. Variables affecting feeding performance at different subsets of ages.
Signiﬁcant terms (p<0.05)are highlighted in bold font. ................. 84

Table 4.1. Mean rates (min/hr) and SE by categorical predictor of each play type
for Crocuta cubs 2-4 months of age. Significant (p <0.05) predictors from
generalized linear models are highlighted in bold font .................. 116
Table 4.2. Factors associated with rates of play in 2-4 month old spotted hyena
cubs (n=40), identiﬁed in generalized linear models. Only signiﬁcant terms (*),
and terms that indicate trends, are presented ........................ 117

Table 4.3. Sex differences in play behavior in various mammal species. Female-
dominated species are indicated in bold font. ....................... 126

Table A1. List of MSU skull specimens used in Chapter Two ............ 132

Table A2. Description of landmarks for each view used in Chapter Two.. . . . 133

LIST OF FIGURES

Figure 1.1. CT slices of sagittal (A) and coronal (B) views of a Crocuta skull
showing the pneumatized sagittal crest, and (C) a coronal view of a wolf (Canis
lupis) showing the more typical plate-like sagittal crest found in non-hyaenid
mammals .................................................. 15

Figure 1.2. Comparison of STL surface representations with a posterior coronal
cross-section illustrating the different morphologies of the (A) normal, (B) filled,
and (C) ﬂattened models ....................................... 18

Figure 1.3. Von Mises stress during unilateral right-side biting in the three-quarter
lateral view (left) and the lateral view (right) for the (A) normal, (B) ﬁlled, and (C)
ﬂattened models. Stress is measured in megapascals (MP3) with warmer colors
representing higher stress. Arrows indicate the bite point on the third premolar in
each model ................................................. 23

Figure 1.4. Sagittal (left) and coronal (right) slices showing Von Mises stress
distribution (in MP8) in the sagittal crests of the (A) normal, (B) filled, and

(C) ﬂattened models. Note the distribution of stress to the trabeculae within the
frontal sinus as well as along the other bones. The warm colors and white areas
on the left temporomandibular joint illustrate the large stresses on the
contralateral jaw joint during unilateral right-side biting .................. 24

Figure 1.5. The percent of total volume of the facial region (left) and cranial vault
(right) stressed (in MPa) during biting in (A) normal, (B) ﬁlled, and (C) ﬂattened
models. The area under the curve represents 98-99% of the volumes of the
cranial vault and facial region, respectively ........................... 27

Figure 2.1. Photographs of Crocuta skulls illustrating dramatic changes in size
and shape throughout development a) in frontal view (from left to right) at 3
months, 20 months, and 11 years of age, and b) in lateral view at 3 months (left)
and 11 years. .............................................. 35

Figure 2.2. Landmarks digitized in each view used in the study, (*) indicate semi-
landmarks in the lateral cranium and mandible views. See appendix for the
descriptions of each landmark point ............................... 41

Figure 2.3. Deformation grid showing ontogenetic change in skull shape in the
ventral view. The landmarks have been back-reﬂected for ease of interpretation.
Vectors show the direction and magnitude of change from the youngest to the
oldest specimens after centroid size is scaled to the same size for each
specimen .................................................. 45

xi

Figure 2.4. Deformation grid showing ontogenetic change in skull shape in the
lateral view. Vectors show the direction and magnitude of change from the
youngest to the oldest specimens after centroid size is scaled to the same size
for each specimen. .......................................... 46

Figure 2.5. Deformation grid showing ontogenetic change in skull shape in the
lateral mandible view. Vectors show the direction and magnitude of change from
the youngest to the oldest specimens after centroid size is scaled to the same
size for each specimen ........................................ 47

Figure 2.6. PCA plots for each view showing cubs (black circles), sub-adults
(blue crosses) and adults (red stars); a) ventral, b) lateral and c) mandible. . .49

Figure 2.7. Deformation grids showing the two phases of skull development; a)
individuals from two months until weaning (14 months), and b) from weaning until
18 years in the ventral view ..................................... 51

Figure 2.8. Deformation grids showing the two phases of skull development; a)
individuals from two months until weaning (14 months), and b) from weaning until
18 years in the lateral view. .................................... 52

Figure 2.9. Ontogenetic changes in centroid size for each view. Arrows indicate
95% of asymptotic values from the monomolecular equation. Dotted lines
indicate the average ages at weaning (W) and reproductive maturity (RM). . .53

Figure 2.10. Ontogenetic changes in shape (procrustes distance) for each view.
Arrows indicate 95% of the asymptotic values from the monomolecular equation.
Dotted lines indicate the average ages at weaning (W) and reproductive maturity
(RM) ..................................................... 55

Figure 2.11. Ontogenetic changes in a) mechanical advantage of the temporalis,
b) mechanical advantage of the masseter, and c) zygomatic arch breadth. The
arrows indicate 95% of the asymptotic value as calculated from the
monomolecular equation. Dotted lines indicate the average ages at weaning (W)
and reproductive maturity (RM) .................................. 56

Figure 3.1. Timeline illustrating the age at maturity for morphological measures in
relation to life history events (in bold) .............................. 68

Figure 3.2. Measurements of bite force in Newtons (N) obtain with a piezoelectric
transducer in a bite force meter available to wild hyenas ................ 76

Figure 3.3. Ontogenetic change in the time required to consume a 309 dog
biscuit during performance testing. The curved line represents the ﬁt of the
monomolecular growth model, and the straight vertical line indicates 95% of
the asymptotic value achieved at 19 months of age .................... 78

xii

Figure 3.4. Ontogenetic change in the time required to consume a goat femur
during performance testing. The curved line represents the ﬁt of the gompertz
growth model, and the straight vertical line indicates 95% of the asymptotic value
achieved at 20 months of age ................................... 79

Figure 3.5. Linear change in biscuit feeding performance for pre- and post-
asymptotic subsets of data. Age is still a signiﬁcant predictor of feeding
performance in both pre- (a) and post- (b) breakpoint subsets. However, after 34
months (c), age is no longer a signiﬁcant predictor of performance. ....... 80

Figure 3.6. Linear change in bone feeding performance for pre- and post—
asymptotic subsets of data. Age is still a signiﬁcant predictor of feeding
performance in both pre- (3) and post- (b) breakpoint subsets. However, after 36
months (0), age is no longer a signiﬁcant predictor of performance. ....... 82

Figure 3.7. The relationship between social rank and skull centroid size in adult
female spotted hyenas. By convention the highest ranking female is assigned a
rank of 1. ................................................. 86

Figure 3.8. Timeline illustrating the age at maturity for feeding performance (in
italics), for both dog biscuits and goat femurs, relative to life history events and to
the age at maturity for morphological measures. ..................... 90

Figure 4.1. Ontogenetic variation in rates of (a) social play, (b) romping, (c) play
mounting, (d) object play and (e) non-nutritive chewing in spotted hyenas. Age
classes represent two month intervals (eg., 0-2 month interval includes
individuals from 0.1 months to 2 months of age). Because no signiﬁcant sex
differences were found in other forms of play, separate rates for males and
females are shown only in “0”. Horizontal bars indicate the period during which
cubs live at the clan’s communal den on all graphs and on (e) periods of tooth
eruption. Note that y-axis scales vary considerably .................... 109

Figure 4.2. The eruption of the deciduous dentition (shaded) is complete by 2
months of age. Adult teeth (white) begin to erupt around 6 months but are not
fully erupted until 18 months of age ............................... 112

Figure 4.3. Rates of play-mounting when individual mounts were recorded as
events a) among males and females throughout the ﬁrst 8 months of life, b)
among males between 2-4 months of age, as a function of their mothers’ social
ranks ..................................................... 114

Figure 4.4. Rates of social play among cubs 2-4 months old in relation to
maternal social rank category. Letters indicate signiﬁcant difference among
maternal rank categories ....................................... 119

Images in this dissertation are presented in color.

xiii

GENERAL INTRODUCTION

Understanding the relationship between morphology and behavior has
been one of the fundamental questions in the fields of ecology and evolutionary
biology (lrschick, 2002; Lauder, 1981 ). Broadly, morphological traits permit some
behaviors to occur while they prevent others from occurring. More speciﬁcally,
variation in morphology determines the performance abilities of organisms.
Wainwright (1994) deﬁned performance as an organism’s ability to accomplish
"ecologically relevant tasks”; such abilities have profound effects on survival and
reproduction. Thus, by understanding the constraints imposed by morphology,
we may gain insight into the mediation and evolution of behavior.

In this dissertation I focus on relationships between skull morphology and
feeding performance in a carnivore species that faces extreme challenges in its
feeding environment, the spotted hyena (Crocuta crocuta). Members of the Order
Carnivora are particularly interesting for the study of relationships between
morphology and performance as they face major challenges in capturing and
processing food. Indeed there are pronounced specializations in the skulls of
predatory carnivores that facilitate prey capture such as large jaw muscles and
sharp, robust canines (Van Valkenburgh, 2007). Moreover, the diversity among
carnivores in their adaptations for processing prey reﬂects the variation in diet
among members of this family. At one extreme, there are members of the Order
Carnivora such as pandas (Ailuropoda melanoleuca) that feed exclusively on

vegetation and thus have broad, rounded teeth to aid in grinding tough material.

At the other extreme are felids that specialize in eating meat and thus have a
reduced and blade-like dentition.

An unusual suite of adaptations for coping with a challenging diet is found
in the skulls of carnivores that crack and consume bone. Bone-cracking forms
have appeared only a few times in the evolutionary history of carnivorans: among
hyaenids, percrocutids, and borophagine canids. The skulls of animals capable
of cracking open bone need to generate and sustain bite forces of great
magnitude. Morphological traits likely to aid in these abilities and common to the
skulls of bone-cracking forms include large body size, a simplified but robust
dentition, a complex multidimensional structure in the tooth enamel, and a
vaulted forehead (Van Valkenburgh, 2007). One additional characteristic
common only to the bone-cracking members of the Family Hyaenidae, is a
caudally elongated fronto-parietal sinus, which has been hypothesized to aid in
stress distribution during bone-cracking (Joeckel, 1998).

The relationship between morphology and performance changes as the
performance challenges facing individuals shift throughout ontogeny (Carrier,
1996; Herrel & Gibb, 2006). Although adult morphology often represents an
optimal form for maximizing performance capabilities, selection also acts upon
individuals during earlier life history stages. In fact, selection is expected to
reduce the length of life history stages in which mortality risk is greatest and/or
favor enhanced performance during these risky life history stages (Williams,
1966). Since juvenile mortality is high in most species, it would therefore be

expected that selection should act to favor improved performance during this

early phase of the life span (Carrier, 1996). However, juveniles are limited in their
feeding capabilities due to small size, inexperience and lack of mechanical
advantage of the jaws (Herrel & Gibb, 2006; Binder 8 Van Valkenburgh, 2000).
Such limitations suggest that there are constraints acting upon individuals at
early life stages. The extent to which morphology limits performance and
determines the youngest ages at which adult feeding maxima can be achieved, is
unknown for most mammals (Herrel & O'Reilly, 2006; Thompson et al., 2003;
Erickson et al., 2003).

Spotted hyenas are large, long-lived carnivores that are capable of
capturing large antelope as well as consuming almost every part of a carcass.
Spotted hyenas live in social groups called clans that can contain up to 80
individuals (Kruuk, 1972). Despite a rigid linear dominance hierarchy that
determines priority of access to food, Crocuta face challenges during feeding that
are imposed by intense competition at carcasses. Thus the ability to feed quickly
should be favored by natural selection. However, we know from studies of
captive spotted hyenas that juveniles are at a disadvantage during competition
with adults for food due to their weaker bite forces and slower feeding speed
(Binder & Van Valkenburgh, 2000). In such situations it would be expected that
provisioning of young might continue until individuals are capable of competing
with adults. However, although spotted hyenas wean their offspring at a late age
relative to those observed in other carnivores (Watts, 2007), at 14 months, there
is little to no provisioning of young in this species (Holekamp 8 Smale, 1990). If

young Crocuta, at the time of weaning, are not capable of competing with older

conspeciﬁcs, then individual feeding success after weaning is likely to be directly
limited by morphology. Here I investigate ontogenetic changes in morphology
and feeding performance in spotted hyenas.

In order to investigate relationships between morphology and feeding
behavior in free-living spotted hyenas I used a combination of morphological and
performance data. All of the data presented here come from a study population
of Crocuta that has been monitored continuously since 1988 in the Masai Mara
National Reserve in southwestern Kenya. This rich source of information has
provided a magniﬁcent dataset with which to investigate the ontogeny of a long-
lived species. Behavioral data presented here come from archived behavioral
notes collected for over 15 years. Both traditional and geometric morphometrics
were used to quantify changes in morphology from a collection of known-age
hyena skulls housed at the MSU museum. Additionally I used ﬁnite-element (FE)
analysis to investigate the relationship between form and function in the skull of
an adult hyena. Finally, two sets of performance data were used in this study.
The ﬁrst dataset is comprised of feeding performance tests conducted by Dr. Kay
Holekamp in which she presented known-age hyenas with 309 dog biscuits and
measured the time it took for individual hyenas to consume the biscuit. I later
spent two years in the ﬁeld conducting similar performance tests, but this time
presenting de-ﬂeshed goat femurs to individuals and attempting to measure bite
force from wild hyenas. This novel combination of morphological and
performance data provide a comprehensive account of ontogeny in a wild

carnivore.

OVERVIEW OF CHAPTERS

In Chapter One, I investigate the functional signiﬁcance of a unique
characteristic in the skulls of both extinct and extant bone-cracking hyenas, a
caudally elongated frontal sinus (Joeckel, 1998), which extends into the parietal
bone throughout the entire sagittal crest above the braincase. For this study a
cutting edge technique borrowed from the field of mechanical engineering, ﬁnite-
element (FE) analysis, was used to investigate patterns of stress distribution in
the skull of an adult spotted hyena during unilateral biting. During bone-cracking,
the hyena’s skull must simultaneously be able to generate enormous bite-forces,
while withstanding enormous stresses. The FE method allowed me to
experimentally manipulate morphology, and to visualize the effects of these
manipulations on mechanical performance. In addition, I tested alternate
hypotheses about the functional role of the fronto-parietal sinus in stress
dissipation. I constructed and compared three FE models: (1) a “normal” model
of an adult Crocuta skull; (2) a model in which the fronto-parietal sinus was ﬁlled
in with bone; and (3) a model in which we ﬂattened the sagittal crest so that it
resembled the plate-like crests of other mammals. Using this approach I found
that during biting, an arc of stress extends from the bite point up through the
vaulted forehead and along the sagittal crest. These results suggest that
pneumatization of the hyena’s skull does not only enhance the ability to resist
bending, but that the fronto-parietal sinus, together with the vaulted forehead,

play a critical role in evenly dissipating stress away from the facial region. This

chapter was recently accepted with revisions for publication in the Biological
Journal of the Linnean Society.

Although the adult Crocuta skull is highly modiﬁed for dealing with the
stresses incurred during bone-cracking behavior, it appears to take the skulls of
juveniles a long period of time to reach adult size and shape. In Chapter Two I
used both traditional methods and geometric morphometrics to document the
ontogenetic trajectories of skull morphology, mechanical advantage, and muscle
mass in order to predict changes in feeding performance during development
that are described in Chapter Three. I utilized a unique collection of 65 skulls
collected from free-living spotted hyenas that were of known age when they died.
The specimens in this ontogenetic series ranged in age from two months to 18
years. Linear and piecewise regressions were used to visualize the changes in
shape occurring throughout ontogeny in three different views of the skull: the
ventral and lateral views of the cranium and the lateral view of the mandible. A
principle component analysis revealed that there were two distinct phases of
ontogeny. The ﬁrst phase occurred between three months and 14 months; the
latter is the average age of weaning. The second phase of ontogeny occurred
between weaning and 18 years. Additionally, I used nonlinear growth models to
calculate the age at maturity for each skull trait, calculated as 95% of the
asymptotic value. Skull size continued to increase until around or shortly after the
time of sexual maturity, between 24 and 29 months of age. Skull shape reached
maturity at 18 months in the lateral view and 23 months in the mandible but

continued to change until 35 months in the ventral view. Mechanical advantage

reached maturity at 22 months of age but muscle mass (as indicated by
zygomatic arch breadth) continued to increase until 33 months of age. Overall,
these findings suggest that skull morphology continues to change into adulthood
despite the late age of weaning in this species. The age at maturity of a
traditional estimate of performance, mechanical advantage, would suggest that
feeding performance should reach maturity at 22 months of age, but actual
measures of performance are presented in Chapter three where I investigate
ontogenetic changes in feeding.

I used both direct and indirect measures of performance in Chapter Three
to accomplish three main objectives. First, I documented ontogenetic changes in
feeding performance as measured by ingestion time. Second, I examined which
of the several morphological variables described in Chapter Two (dentition, skull
size, skull shape, zygomatic arch breadth, or mechanical advantage of the
temporalis) might act as the main limiting factors with respect to feeding
performance as hyenas mature. Third, I tested the hypothesis that social rank
inﬂuences feeding performance.

Although my attempts at measuring bite force in the ﬁeld were largely
unsuccessful, l was able to collect enough data on changes in feeding speed to
document ontogenetic changes in performance. Feeding performance in wild
hyenas continued to improve until late in the third year of life, between 33 and 36
months of age. Because the eruption of dentition, mechanical advantage of the
temporalis and skull size reach maturity at ages much younger than feeding

performance reached its asymptote, it is unlikely that these are the main factors

limiting performance throughout ontogeny. However, it appears that skull shape
and adductor muscle mass (as measured by zygomatic arch breadth) may be the
main factors limiting performance, as these measures reach maturity at around
the same age as feeding performance reaches asymptotic levels.

In addition to morphology, social factors inﬂuence feeding performance
throughout hyena development. Social rank was a signiﬁcant predictor of
ingestion time among adults with low-ranking individuals feeding faster than high-
ranking hyenas. Detailed investigation of rank-related variation in skull
morphology revealed that lower-ranking females had signiﬁcantly larger skulls
than did higher-ranking females, which might explain the rank-related variation
observed in adult feeding performance. Among juvenile hyenas, however, the
inﬂuence of social rank was in the opposite direction from that observed in adults,
with higher-ranking cubs feeding faster than their lower-ranking peers. Previous
research (Wahaj 8 Holekamp, 2006; Hofer 8 East, 1996) suggests that high-
ranking cubs grow at faster rates than do low-ranking cubs, and the fact that
higher-ranking youngsters are large might explain their surperior feeding
performance.

Lastly, in Chapter Four, I examined the possible functions of a behavior
often observed in young carnivores, a form of object play referred to as non-
nutritive chewing. I investigated whether ontogenetic variation in this behavior
might be related to morphological changes during development in skull size and
shape. Alternatively, non-nutritive chewing might merely reﬂect “teething” or

functions similar to those served by other forms of play. Therefore, I investigated

ontogenetic rates of non-nutritive chewing, and compared these to rates of other
forms of play. Consequently, in this chapter I first documented the ontogenetic
patterns of the various forms of play in order to draw inferences about potential
functions of each type of play. Additionally, given the sex-role reversal in this
species I investigated sexual dimorphism in speciﬁc play behaviors. Finally, I
tested hypotheses suggesting social and energetic inﬂuences on play. With this
comprehensive work I was able to suggest potential functions of each type of
play.

Temporal peaks in non-nutritive chewing did not coincide with either
periods of tooth replacement or signiﬁcant changes in skull morphology. Instead,
rates of this activity peaked quite early in development, between two and four
months of age, and then steadily declined after that. This suggests that non-
nutritive chewing does not reﬂect musculoskeletal development, but rather that it
serves as a way for young hyenas to explore and gain information about their
environments via sensory receptors in and around the mouth. This chapter is
currently in press to be published in the Journal of Developmental Processes.

Collectively, this work provides a great deal of insight into developmental
relationships between morphology and behavior in Crocuta. My results suggest
that development is a slow and prolonged process in spotted hyenas, consistent
with the hypothesis that there are developmental constraints in juvenile spotted
hyenas that affect their performance capabilities and thus their need for
assistance from their mothers. In order to determine whether development is

delayed or prolonged in this species relative to that in other carnivores, and

whether this might be one of the selective forces leading to the evolution of
female dominance, comparable studies on other carnivores are needed.

This dissertation is the result of a great deal of effort and insight from a
number of collaborators, and each chapter was prepared as a manuscript for
publication with some of those collaborators. Therefore, the term “we” is used

throughout.

10

CHAPTER ONE

OF ARCS AND VAULTS: THE BIOMECHANICS OF BONE-CRACKING IN
SPOTTED HYENAS (CROCUTA CROCUTA)

INTRODUCTION

Organisms that cope effectively with extreme challenges in their
environment provide excellent opportunities for exploring the relationship
between form and function (Wainwright 8 Reilly, 1994). The skulls of carnivores
are especially interesting in this context because they allow us to examine the
interplay between forces shaping the highly developed neural and sensory
organs critical for locating and capturing prey, and those shaping a feeding
apparatus designed for both the capture of uncooperative prey and the
processing of their carcasses. The skulls of bone-cracking animals represent
extreme examples. Bone-cracking forms have appeared only a few times in the
evolutionary history of carnivorans: among hyaenids, percrocutids, and
borophagine canids, with only hyaenids surviving to the present. Comparative
analysis reveals some striking morphological similarities among the various
bone-cracking forms, including large body size, a simpliﬁed but robust dentition,
a complex multidimensional structure in the tooth enamel, and a vaulted
forehead (Van Valkenburgh, 2007).

Although extant carnivores provide useful models for the study of skull

form and function, in vivo studies of the forces and strains generated during

11

biting are often logistically and practically impossible to accomplish. Only scant
data are available documenting bite forces generated by carnivores, and almost
all of these are from captive animals (Binder 8 Van Valkenburgh, 2000; Dessem,
1989). Captive studies yield bite force maxima that almost certainly
underestimate the true feeding capabilities of these animals in the wild (Stefen,
1997). Such limitations force us to look elsewhere for methods to more
accurately assess the relationship between form and function in carnivore skulls.
Here we use ﬁnite element (FE) analysis to evaluate this relationship in the skulls
of spotted hyenas (Crocuta crocuta).

Spotted hyenas are highly durophagous animals that are capable of
breaking open and consuming bones of large diameter. These gregarious
carnivores are recently descended from carrion-feeding ancestors (Lewis 8
Werdelin, 2000). Spotted hyenas share with other bone-cracking hyaenids, and
with percrocutids and borophagine canids, a suite of cranio-dental characteristics
that maximize their ability to feed on virtually all components of a carcass. Bone-
cracking carnivorans use their premolar teeth (p3 in hyaenids and percrocutids,
and p4 in borophagine canids) for opening large bones (Werdelin, 1989; Van
Valkenburgh, 2007). As a result, the main bone-cracking premolar is often robust
and pyramidal in shape (Van Valkenburgh, 2007; Werdelin, 1989; Biknevicius 8
Van Valkenburgh, 1996). Likewise in hyaenids, percrocutids, and at least one
borophagine (Borophagus secundus), there is a highly derived enamel
microstructure that aids in preventing fracture (Rensberger 8 Wang, 2005;

Rensberger, 1999; Stefen, 1997). Furthermore, the height of the mandibular

12

corpus and the thickness of the cortical bone in the dentary in bone-cracking
carnivores is greater than that found in carnivores not specialized for bone-
cracking; presumably these mandibular characteristics enhance resistance to
dorsoventral bending (Biknevicius 8 Ruff, 1992; Biknevicius 8 Van Valkenburgh,
1996; Therrien, 2005).

Bone-cracking mammals typically have larger jaw muscles than their non-
durophagous relatives (Werdelin, 1989). Associated with this enhanced jaw
musculature is usually a prominent sagittal crest that provides a relatively large
surface area for attachment of the temporalis muscle, as well as a relatively wide
bi-zygomatic breadth that increases attachment area for the masseter, and
allows more space in which to accommodate the enlarged jaw adductors.
Additionally, bone-cracking hyaenids, percrocutids and borophagine canids have
uniquely vaulted foreheads created by an enlarged frontal sinus. Based on the
work of Buckland-Wright (1978; 1971 ), which suggested that compressive forces
travel in an arc from the bite point dorsally through the region anterior to the eye
orbit, Werdelin (1989) hypothesized that this vaulting creates an arc of bone that
transfers stresses incurred during biting from the facial region caudally along the
sagittal crest.

Although this suite of adaptations for durophagy occurs in all bone-
cracking carnivores, a rare trait shared by extinct and extant bone-cracking
members of the Family Hyaenidae is a caudally-elongated frontal sinus that
invades the parietal bones (from now on referred to as the‘fronto-parietal sinus)

and completely overlies the braincase (Negus, 1958; Buckland-Wright, 1969;

13

Paulli, 1900; Joeckel, 1998). Thus, the skull is not only pneumatized anteriorly
between the postorbital processes (as in borophagines), but also all along the
sagittal crest (Figure 1.1A). There are other species in which the frontal sinus
invades the parietal bone (ie. Brown bears, pigs, elephants, etc.), however, the
sinus does not lie underneath the sagittal crest as it does in hyenas. Thus, the
presence of the elongated frontal sinus within the sagittal crest of bone-cracking
hyenas results in the unique triangular cross section (Figure 1.1 B) that is
dramatically different from the laterally compressed, plate-like sagittal crest of
most other mammals (Figure 1.1 C).

Joeckel (1998) was able to rule out hypotheses suggesting that the
elongated fronto-parietal sinus functions to enhance olfaction, dissipate shock to
the brain, cool the brain, or maximize muscle attachment area. Instead, he
likened the space created by the caudal portion of the elongated fronto-parietal
sinus to an architectural shell or arch, and suggested that the sinus increases
structural support by enhancing resistance to muscular loads applied to the
sagittal crest during bone cracking. Based on principles derived from structural
engineering, he predicted that the uniquely pneumatized sagittal crest in hyenas
should be more resistant to vertical loading by the temporalis muscles than the
plate-like sagittal crests of other mammals. Joeckel concluded that the hollow,
arch-like sagittal crest makes the crest more resistant to bending than a vertical
ﬂat crest, and simultaneously lightens the skull. Although these are interesting
ideas, the relative importance of the elongated fronto-parietal sinus in

strengthening and lightening the skull remains unclear. For example, it is

14

 

Figure 1.1. CT slices of sagittal (A) and coronal (B) views of a Crocuta skull
showing the lpneumatized sagittal crest, and (C) a coronal view of a wolf (Canis
lupis) showing the more typical plate-like sagittal crest found in non-hyaenid
mammals.

possible that it is the expanded base of the crest, rather than the fact that the
crest is hollow that makes it more resistant to large muscle forces during biting.
Here we used the FE method to predict patterns of stress distribution in an adult
Crocuta skull during a bone-cracking bite and test Joeckel’s hypothesis about the
role of the fronto-parietal sinus in bending resistance of the sagittal crest. We
also evaluate more broadly the function of the sinus in the context of stresses
generated by both muscle loading and bite forces.

We built and compared three different FE models for this study: a detailed
model of the skull of an adult female spotted hyena, a model of the same
specimen in which the parietal portion of the pneumatized fronto-parietal sinus
was ﬁlled in with bone, and a third model in which we modiﬁed and ﬂattened the
sagittal crest so that it resembled the more plate-like crests of other mammals. If
Joeckel’s hypothesis is correct that the elongated fronto-parietal sinus functions
to resist bending of the sagittal crest, then the model with the normal sagittal
crest should be less stressed during biting than the model with the ﬂattened
crest. If the function of the sinus is purely to lighten the skull, then normal and
"ﬁlled” models should undergo similar stress regimes. Because the FE method
allows us to experimentally manipulate morphology and visualize the effects of
these manipulations on mechanical performance, we can use it to evaluate the
inﬂuence of competing demands for bite force, bone strength, and skull mass in

the evolution of skull form among these durophagous carnivores.

16

METHODS

The three FE models used in this study were all based on the skull of one
55 month old adult female Crocuta crocuta (Michigan State University Museum —
MSU 36567). This specimen originated from a longitudinal study population of
free-living spotted hyenas in the Masai Mara National Reserve in southwestern
Kenya (Frank et al., 1995). The specimen was scanned on a General Electric
Discovery ST 16 slice scanner at the Department of Radiology at Michigan State
University, with a slice thickness of 0.625mm. Methods for generating a 3D FE
model from CT scans follow Dumont et al. (2005). First, we produced a 3D
surface model from the scans using VGstudioMax software (Volume Graphics
Inc., Heidelberg, Germany). This model was then imported to Geomagic Studio
(Raindrop Geomagic Inc., Research Triangle Park, North Carolina) where we
repaired scanning defects (holes, beam hardening, etc) and simpliﬁed regions
that, in our experience, can be difﬁcult to mesh with ﬁnite elements. These
included very small foramina and thin structures such as the turbinates, which
are unlikely to be load-bearing. Cranial sutures of adult hyenas are completely
fused (Schweikher, 1930) and therefore were not included in the model.

We then used Geomagic to alter the surface model based on the scans
(the “normal” model, Figure 1.2 A) to create the two modiﬁed morphologies. For
the “filled” model (Figure 1.2 B), the sinus cavity was truncated at the post-orbital
processes; the anterior portion of the fronto-parietal sinus and the external
morphology of the sagittal crest were not altered in any way. For the "ﬂattened”

model (Figure 1.2 C) we again removed the sinus cavity caudal to the post—orbital

17

 

Figure 1.2. Comparison of STL surface representations with a posterior coronal
cross-section illustrating the different morphologies of the (A) normal, (B) ﬁlled,
and (C) ﬂattened models.

processes and also ﬂattened the crest to form a thin blade-like crest that was
similar in shape to those found in most other crest-bearing mammals (Figure
1.10). All three surface models (“normal”, “ﬁlled” and “ﬂattened”) were saved
from Geomagic as STL ﬁles and then imported to the FE analysis tool Strand7
(G+D Computing, Sydney, Australia). We used the solid mesh generation
algorithm in Strand7 to create a volumetric mesh composed of 4-noded
tetrahedrals from each surface model. Despite their morphological differences,
the resulting FE models were similar in size (“normal” model = 1,319,307
elements, “ﬁlled” model = 1,332,181 elements, “ﬂattened” model = 1,255,253
elements).

Unfortunately, there are no data available that summarize Young’s
modulus or Poisson’s ratio values for bone in hyena skulls. Therefore, we
assigned our models values for the material properties of cortical bone used by
Verrue et al. (2001) for FE models of dog skulls (E = 1.37 X 104 MPa; v = 0.3).
We modeled bone as homogeneous and isotropic, even though this is not likely
to be the case (Currey, 2002). Therefore the absolute values of stress predicted
by our model may not reﬂect actual values, and should be interpreted with
caution. Past studies have shown that, although varying elastic properties may
cause slight variations in the magnitude of stress distribution, overall gross
patterns of stress distribution are quite robust (Strait et al.2005). Because we
applied the same material properties to each of our models, we can conﬁdently
compare stress distributions among the three models, and attribute any

differences observed to modiﬁcations in the morphology of the sagittal crest.

19

We applied loads to our models that mimic the forces produced by the
three main jaw-adductor muscles: temporalis, masseter, and pterygoid. Data
summarizing the average mass of unilateral temporalis and masseter muscles (E
= 247 i 18.2 g and E = 136 i 9.3 g, respectively) were drawn from four
necropsies of Crocuta conducted in the ﬁeld (JBT). Based on the relative size of
the pterygoids in hyenas, and on data from felids (Gorniak 8 Gans, 1980), we
estimated that the pterygoids contribute approximately 18% to the total mass of
the jaw adductors. Therefore, we modeled the relative contribution of
temporaliszmasseter:pterygoid to adductor muscle force production as 50:32:18.
The areas of attachment for each of the three muscle groups were based on
descriptions of Crocuta and Hyaena myological studies by Buckland-Wright
(1969) and also from our own ﬁeld necropsies. To deﬁne the vectors of muscle
force, we identiﬁed a node on the mandible within the region of attachment of
each muscle. As in previous studies (Strait et al., 2005) we constrained the
models at one node on each articular surface of the temporomandibular joint.
These constraints deﬁned an axis around which the models could rotate in
response to muscle forces. A node on the occlusal surface of the third premolar
was also constrained in order to generate a reaction force that models bite force.
Hyenas, like other carnivores, use unilateral biting during feeding (Biknevicius 8
Van Valkenburgh, 1996) and therefore we loaded our model to reﬂect forces
incurred by one side of the jaw.

After deﬁning muscle attachment sites and constraints, we used the Visual

Basic program BoneLoad to apply muscle forces to FE models (Grosse et al., In

20

Press). This loading algorithm applied muscle forces based on both the traction
muscles generate and the normal forces that accumulate as muscle ﬁbers wrap
around curved bone surfaces. For each FE model an initial analysis was run
using an arbitrary total muscle force. Based on the resulting bite reaction force
(measured perpendicular to the palate at the constrained premolar), we then
adjusted the muscle loads in order to obtain a biologically relevant bite force. In
captivity, measures of bite forces up to 3500 N have been obtained from spotted
hyenas (Binder 8 Van Valkenburgh, 2000). However, Stefen (1997) and others
(Meers, 2002; Therrien, 2005; Erickson et al.1996) estimate that 7000-9000 N
are required to crack open the long bones of antelope, such as wildebeest
(Connochaetes taun'nus), on which wild hyenas commonly feed. Therefore we
applied muscle loads to each FE model here to generate a conservative bite
force of 5500 N.

The resulting FE analyses were compared in two ways. First, we visually
inspected the results to compare and contrast the magnitude and distribution of
von Mises stress, a predictor of failure for ductile materials due to distortion or
shear. Stresses were quantiﬁed in units of megapascals (MPa). These will be
referred to as our “qualitative” results. Second, we constructed histograms to
compare the distribution of stress within two separate regions of the skull: the
face and the cranial vault. The face included the entire skull anterior to the post-
orbital processes, and the vault included the lateral and posterior walls of the

braincase and the sagittal crest. Third, we compared the absolute muscle forces

21

required to generate 5500 N of bite force in each model. The histograms and

muscle forces will be referred to as our “quantitative” results.

RESULTS

The patterns of stress distribution in the normal adult Crocuta model
during unilateral right-sided biting are illustrated in Figure 1.3 A. When sufﬁcient
muscle loads were applied to generate 5500 N of bite force, high areas of stress
appeared on the working side maxilla anterior to the orbit and dorsal to the
infraorbital foramen. Additionally, there was a hotspot of stress just anterior to the
jugal-squamosal suture in the zygomatic arch. As predicted by Buckland-Wright
(1971; 1978) and Werdelin (1989), an arc of stress extended from the bite point
through the region anterior to the orbit, and along the vaulted forehead. This arc
of stress then gradually dissipated posteriorly along the sagittal crest. The
anterior portion of the braincase experienced somewhat higher concentrations of
stress than did the rest of the braincase. Stress along the crest was distributed
around the external surface of the fronto-parietal sinus and along the trabeculae
within it (Figure 1.4 A).

Both the ﬁlled and ﬂattened models deviated from the patterns seen in the
normal model. Overall the ﬁlled model experienced less stress than that
observed in the normal model (Figure 1.3 B). As in the normal model, an arc of
stress extended from the bite point to the vaulted forehead, however the stress
pathway terminated just posterior to the postorbital processes. Therefore, there

was less stress along the sagittal crest and braincase than that observed in the

22

Brlok Stress: VM (MPa) A
50 o '

 

 

Figure 1.3. Von Mises stress during unilateral right-side biting in the three-quarter
lateral view (left) and the lateral view (right) for the (A) normal, (B) ﬁlled, and (C)
ﬂattened models. Stress is measured in megapascals (MPa) with warmer colors
representing higher stress. Arrows indicate the bite point on the third premolar in
each model.

23

Brlck Stress: VM (MPa) A

[rm-.132?
I l

 

 

       

IT- " i
,wij‘rqu: TI

Figure 1.4. Sagittal (left) and coronal (right) slices showing Von Mises stress
distribution (in MPa) in the sagittal crests of the (A) normal, (B) ﬁlled, and

(C) ﬂattened models. Note the distribution of stress to the trabeculae within the
frontal sinus as well as along the other bones. The warm colors and white areas
on the left temporomandibular joint illustrate the large stresses on the
contralateral jaw joint during unilateral right-side biting.

24

normal model (Figure 1.4 B). Although the filled model was generally less
stressed, there were higher concentrations of localized stress on the working
side of the vaulted forehead (as illustrated by the yellow area of stress adjacent
to the postorbital process on the working side), and on the anterior edge of the
orbit, than in the other two models.

The ﬂattened model exhibited by far the greatest stress values along the
sagittal crest, as well as slightly elevated stress in the anterior portion of the
braincase (Figure 1.3 C). There were high values of stress throughout the crest
but the highest stress values were in the dorsal region as indicated In a coronal
cross-section (Figure 1.4 C). Additionally, the stresses on the working side
maxilla were higher than in the other two models.

We observed only slight differences among the three models with respect
to the amount of localized stress in the facial region, with the ﬂattened model
showing the highest stress values in this area. The most obvious differences
among the models were seen in the area between left and right post-orbital
processes, and along the sagittal crest. The ﬂow of stress from the face to the
sagittal crest evident in the normal and ﬂattened models was interrupted in the
ﬁlled model. The normal model, although more heavily stressed than the ﬁlled
model, illustrated an intermediate pattern of stress distribution relative to the
other two models. The stresses incurred during unilateral biting appeared to be
more evenly distributed in the normal model than in either of the other models,

with lower stresses on the edge of the orbit and in the forehead than in the filled

25

model, as well as lower stresses along the sagittal crest and zygomatic arch than
in the ﬂattened model.

Quantitative differences among the three models in the percent of total
skull volume stressed in the face and cranial vault mirrored the results of the
qualitative comparisons (Figure 1.5). For the cranial vault, the filled model was
clearly the least stressed, and the ﬂattened model the most stressed. Although
variation in the proportion of the facial skeleton experiencing stress was very
slight, the areas that were most highly stressed differed among the three models.
The ﬁlled model had higher concentrations of stress along the edge of the orbit
and on the forehead than did the other two models, whereas the ﬂattened model
had higher values in the maxilla.

In considering these results it is important to point out that the muscle
forces required to produce 5500 N of bite force differed among the three models
(Table 1.1). The ﬁlled model required the least amount of muscle force, whereas
the ﬂattened model required the greatest amount of force to produce the same
bite force; the normal model was intermediate. From a slightly different
perspective, the muscle forces required to produce a 5500 N bite force in the
normal model produced a bite force of 5750 N in the ﬁlled model, and a bite force
of 5350 N in the ﬂattened model. Thus the normal model was intermediate in
terms of the efﬁciency of transferring muscle forces to bite force, and the filled

model was most efﬁcient.

26

Cranial Vault

   

 

   

Facial Region
18
Normal 15 Normal
14
0 0
E E
2 2
§ 2
g 39
0 5 10 15 2O 25 I!) 35 40 45 50 55 60 O 5 1o 15 20 25 30 35
Stress (MPa) Stress (MPa)
13
Filled Filled
I) U
E E
3 3
B '6
> >
35 5°
0 5 10 15 20 25 3) 35 40 45 50 55 60 0 5 10 15 20 25 30 35
Stress (MPa) Stress (MPa)
18
Flattened 16 Flattened
14
0 O
E E
2 3
‘>’ >°
38 a?

   

0
O 51015202530354045505560
Stress(MPa)

Stress (MPa)

Figure 1.5. The percent of total volume of the facial region (left) and cranial vault
(right) stressed (in MPa) during biting in (A) normal, (B) filled, and (C) ﬂattened
models. The area under the curve represents 98-99% of the volumes of the
cranial vault and facial region, respectively.

27

Table 1.1. Muscle forces applied to each model in order to produce 5500N of
bite force. Values (in Newtons) represent the sum of identical forces applied on
the left and dim.

Model

 

Temporalrs Masseter Pterygord Total (N)

 

(N) (N) ('9
Normal 101.6 64.6 45.2 211.4
Filled 97.2 61.8 43.2 202.2
Flattened 104.5 66.4 46.5 217.4

 

28

DISCUSSION

Here, for the ﬁrst time, we examined patterns of stress distribution during
biting in the skull of a bone-cracking hyaenid. The FE method allowed us to
experiment with alternative morphologies and take a holistic approach to
predicting patterns of force distribution during biting behavior. This kind of
detailed “snapshot” of the stress state throughout the entire skull is impossible to
achieve using any other technique, and highlights the utility of this method in
comparative biology, especially when investigating areas of the skull, such as the
elongated fronto-parietal sinus, from which it would othenlvise be impossible to
collect data (Ross, 2005). Although FE analysis is a modeling technique that
necessarily requires certain assumptions, it greatly enhances our ability to
generate theories and predictions about relationships between form and function
in the vertebrate skull. Only through the use of FE analysis were we able to
understand the interplay of biting forces and visualize what this might look like.
Thus this study illustrates the usefulness of the FE method in hypothesis-testing,
and its potential for addressing questions about form-function relationships in
organisms that cope with extreme conditions.

Given the unique challenges of a highly durophagous diet and the unique
morphology seen in the feeding apparatus of the hyena, this study provides
compelling evidence that the form of the fronto-parietal sinus, and consequently
the pneumatized sagittal crest, represent adaptations to cope effectively with the
large stresses experienced and forces imposed during bone-cracking. If the

sagittal crest of hyenas functioned simply to increase muscle attachment area,

29

then we would expect to find a morphology similar to that in our ﬂattened model.
On the other hand, if the main selective pressure was simply for resisting higher
muscle forces in the cranial vault, we might expect the skull of a hyaenid to more
closely resemble that of our ﬁlled model. Consequently, the pneumatized sagittal
crest appears to represent the most efﬁcient structure for meeting the concurrent
demands of the massive jaw musculature necessary for generating large bite
forces, and skull strength sufﬁcient for resisting large stresses, without increasing
skull weight.

The general pattern of stress distribution observed in our normal model
conforms to predictions made by Buckland-Wright (1978; 1971) and Werdelin
(1989), who concluded that forces generated in hyenas during premolar biting
must pass through the face anterior to the orbit, and then continue along the
vaulted forehead to the sagittal crest. The greater stress in the sagittal crest of
the ﬂattened model than in the normal model also supports Joeckel’s (1998)
hypothesis that the elongated fronto-parietal sinus helps to resist loads by
dissipating stress along its curvature (Figures 1.3 8 1.4). Speciﬁcally, the
stresses are dissipating laterally along the edges of the sinus (Figure 1.4 A) in
addition to the pathway of stress dissipating caudally (Figure 1.3 A). However,
although Joeckel predicted that the highest areas of stress in a plate-like crest
would be found at the base, our ﬂattened model revealed higher stress along the
top of the crest (Figure 1.4 C). This may be due to the tensile stresses caused by
the simultaneous pull of the muscles on either side of the plate-like crest in our

FE model. If temporalis activity were asymmetrical, it is likely that there would

30

indeed be bending stress at the base of the crest, but stresses were applied
symmetrically in our models. In any case, Joeckel was correct in his
interpretation that the vaulted sagittal crest provides more structural support
during biting than does a ﬂattened crest. Certainly, the greatest variation in stress
among our three models occurred in the cranial vault. It is noteworthy that,
despite the limited technology available to these earlier researchers, their
insights into these unique adaptations have proven to be remarkably accurate.

Including the ﬁlled model in our analysis allowed us to extend Joeckel’s
work by teasing apart the relative importance of the fronto-parietal sinus in
resisting muscle loads versus lightening the skull. The volume of the fronto-
parietal sinus in the specimen used here was 89.7 cm3 (Sakai, Tanner,
Lundrigan 8 Holekamp, In Prep). Therefore, assuming the density of bone is
similar to that found in primate and canid skulls (around 1.62 g/cm3) (Novecosky
8 Popkin, 2005; Wang et al., 2006), our ﬁlled model would result in an addition of
1459 of bone. It is difﬁcult to estimate the extent to which this small additional
mass might affect the animal’s ability to hunt and capture prey. We can, however,
predict the effects that a solid sagittal crest would have on patterns of stress in
the skull during biting.

Our quantitative results showed that, overall, the presence of a sinus
actually increased the amount of stress in the skull during biting. That is, in
comparison to the ﬁlled model, the normal model exhibited signiﬁcantly higher
stresses in the cranial vault (Figure 1.5). The presence of the sinus also

decreased the efﬁciency of transferring muscle force into bite force. The ﬁlled

31

model was more efﬁcient, requiring less muscle force to produce a bite force of
5500 N. This suggests that a ﬁlled arch-shaped crest would allow for the
generation of higher bite forces than a pneumatized crest, which might be
advantageous for an animal that makes its living cracking open bones. However,
we have observed extant spotted hyenas cracking open giraffe long bones up to
7 cm in diameter, so their pneumatized skulls allow them to cope effectively even
with bones much larger than those of their usual prey. Furthermore, the ﬁlled
model exhibited slightly higher stress concentrations in the facial region. The
sinus, therefore, appears to reﬂect a trade-off between maximizing efﬁciency in
force production and minimizing regions of stress concentration.

Although the normal model experienced more stress than the ﬁlled model,
the qualitative comparisons indicate that the skull is more evenly stressed in the
normal model. This is particularly striking in light of the fact that the normal model
required higher muscle forces to achieve the same bite force than the ﬁlled
model. Therefore, even under larger loads, the normal model has fewer areas
with very high concentrations of stress. For instance, the areas of highly
concentrated stress in the orbit and forehead of the ﬁlled model predict that the
bone in these areas would reach the point of failure under lower muscle loads
than the same areas in the normal model. This highlights the signiﬁcant role
played by the fronto-parietal sinus in moving forces away from the face, through
the forehead to the cranial vault. Thus, there appears to be a unique complex
comprised of the elongated fronto-parietal sinus and the vaulted forehead that is

critical for stress dissipation in the skulls of these bone-cracking hyenas. When

32

the sinus in the sagittal crest is absent, as in our ﬁlled model, the force
transmission pathway from the face to the crest is broken. A critical next step in
understanding the biomechanical adaptations for durophagy would be to
compare the arc of stress dissipation we have demonstrated in hyaenids with
that in percrocutids or borophagine canids, which both have similar vaulting in
the forehead but lack an elongated fronto-parietal sinus beneath the sagittal
crest. We would expect to find a pattern similar to that in our ﬁlled model, lacking
the advantages of an elongated fronto-parietal sinus to evenly dissipate stress

away from the facial region.

33

CHAPTER 'IWO
ONTOGENETIC CHANGE IN SKULL MORPHOLOGY AND MECHANICAL
ADVANTAGE IN THE SPOTTED HYENA (CROCUTA CROCUTA)
INTRODUCTION

The ability to obtain and process food is essential for the survival of all
mammals after weaning, yet an individual’s functional abilities to perform such
tasks may still be quite immature at the time of nutritional independence
(Monteiro et al., 1999; Wainwright 8 Reilly, 1994; Herrel 8 Gibb, 2006; Herrel 8
O'Reilly, 2006). The functional challenges associated with weaning may be
especially demanding for carnivorous mammals, as they must feed on
uncooperative and unpredictable prey, and they must also often cope with ﬁerce
competitors. Morphological changes during development are rarely isometric, so
juvenile carnivores are not miniature replicas of their adult counterparts (Figure
2.1). Instead their adult form is the result of changes in both size and shape, and
these changes have functional signiﬁcance with respect to feeding performance.
Wainright (1994) deﬁnes performance as an organism’s ability to carry out
ecologically relevant tasks. Both morphology and behavior delimit the boundaries
of performance, which in turn constrain the environmental resources that an
organism can utilize, and thereby affect ﬁtness (Wainwright, 1994).
One approach adopted by ecomorphologists interested in constraints is to study
the functional capacities of organisms that live in extreme environments

(Wainwright, 1994). Here, we investigate developmental changes in skull

34

 

Figure 2.1. Photographs of Crocuta skulls illustrating dramatic changes in size
and shape throughout development a) in frontal view (from left to right) at 3
months, 20 months, and 11 years of age, and b) in lateral view at 3 months (left)
and 11 years.

35

morphology of a carnivore, the spotted hyena (Crocuta crocuta), that must
simultaneously cope effectively with two types of extreme conditions. First, the
intense feeding competition characteristic of spotted hyenas represents an
extreme environment in which selection has apparently favored traits, including
female ‘masculinization,’ that set these animals apart from other gregarious
mammals (Frank, 1997). Second, spotted hyenas have a highly durophagous
diet that demands cracking and consumption of bones of large diameter (Kruuk,
1972; Werdelin, 1989). Enhanced feeding performance in such an extreme
environment may have profound effects on an individual’s ﬁtness. Mortality is
high among juvenile spotted hyenas, with less than half of all cubs surviving to
adulthood (Frank, 1997; Holekamp et al., 1996), so selection should favor rapid
improvement in feeding performance among juveniles after weaning. However,
past studies indicate that juvenile spotted hyenas may in fact be at a
disadvantage for rapid feeding long after weaning, possibly due to their relatively
small size and poor mechanical advantage of the feeding apparatus (Biknevicius,
1996; Binder, 1998; Binder 8 Van Valkenburgh, 2000).

Adult spotted hyenas can consume 1.3 kg of meat and bone per minute
(Kruuk, 1972), so they are capable of very rapid feeding. Furthermore, the skull
of an adult spotted hyena is highly modiﬁed to deal with the performance
demands of bone-cracking. Like other extant and extinct bone-cracking members
of the Family Hyaenidae, as well as extinct percrocutids and borophagine canids,
adult spotted hyenas possess a number of morphological characteristics that

enhance their ability to generate and sustain bite forces of great magnitude.

36

These adaptations for durophagy include a robust and simpliﬁed dentition, large
jaw adductors, a pronounced sagittal crest, a vaulted forehead, wide bi-
zygomatic arch breadth, increased cortical thickness of the dentary bone and
large body size (Van Valkenburgh, 2007). Furthermore, all extant bone-cracking
hyaenids have an unusually long lactation period relative to other carnivores,
lasting over one year and extending to 24 months in some cases (Kruuk, 1972;
Holekamp et al., 1996; Watts, 2007), suggesting that development of the
specialized morphology of bone-cracking forms might require a great deal of
time.

In contrast to most other carnivores, including the other extant bone-
cracking hyaenids, spotted hyenas are highly gregarious. The size of their social
groups, called clans, can reach 80 individuals (Kruuk, 1972). Each clan is
structured by a rigid linear dominance hierarchy, and social rank determines
priority of access to food. The challenges posed by weaning in spotted hyenas
may thus be exacerbated both by slow development of the specialized traits
needed to cope with a durophagous diet, and also by intensive feeding
competition with conspeciﬁcs. It is noteworthy that there is little or no provisioning
of young by adult spotted hyenas (Holekamp 8 Smale, 1990). Therefore, aside
from potential maternal assistance during feeding competition, individual feeding
success after weaning in this species is likely to be directly limited by
morphology.

One method traditionally used to inquire about changes in feeding ability

during mammalian development has been to model the jaw as a lever in order to

37

draw inferences about mechanical advantage of the feeding apparatus
(Radinsky, 1981b; Greaves, 1985; Greaves, 1983; Thomason, 1991; Smith,
1993). In lever-arm models, relative bite force is inferred by estimating size and
mechanical advantage of the primary masticatory muscles, the temporalis and
masseter (Binder, 1998; Hildebrand, 1984; Radinsky, 1981a). The traditional
indicator of adductor muscle mass on Mammalian skulls is maximal width across
the zygomatic arches, in-lever arm lengths are indicated by distances separating
muscle insertion points, and out-lever (moment arm of resistance) length by the
distance from the jaw joint to the bite point on the mandible (Binder, 1998;
Gittleman 8 Van Valkenburgh, 1997; Radinsky, 1981b). As the length of the
temporalis and masseteric moment arms increase relative to the length of the
resistance arm, mechanical advantage increases (Hurov et al., 1988; Gittleman 8
Van Valkenburgh, 1997). These linear measurements have been used in myriad
studies of functional jaw morphology.

One shortcoming of this traditional approach is that it is difﬁcult to
integrate the results into an accurate visualization of overall shape change in the
feeding apparatus as it matures. A more effective method for describing such
change is geometric morphometric methods (Bookstein, 1991; Rohlf 8 Marcus,
1993; Zelditch et al., 2004), which combines traditional multivariate statistics with
the graphical tradition introduced by D'Arcy Thompson (1917), and allows us to
investigate changes in skull size and shape independently. Here we use both

traditional methods and geometric morphometrics to document ontogenetic

38

variation in skull size, skull shape, and mechanical advantage in spotted hyenas

in order to predict changes in feeding performance during development.

MATERIALS AND METHODS

Samples

In this study we utilized an ontogenetic series of 68 Crocuta skulls,
representing individuals ranging in age from two months to 18 years (Figure 2.1).
All specimens are catalogued into the Michigan State University Museum (Table
A1). Most originated from a study population of spotted hyenas that has been
monitored since 1979 (Frank, 1986) in the Masai Mara Game Reserve in
southwestern Kenya, but the youngest specimen, a two month old, was obtained
from a captive colony of spotted hyenas at UC Berkeley. For 37 specimens,
including all subadults, exact age at death was known because these individuals
were born into the study population, and their birth dates were estimated (to i 7
days) based on the appearance of cubs when ﬁrst observed above ground
(Holekamp et al., 1996). For all other adult individuals we estimated age (i 6
months) based on canine height and occlusal wear on the surface of the third

lower premolar as described in Van Horn et al. (2003).

Geometn'c Morphometrics
To quantify changes in skull size and shape we used 2D geometric
morphometrics. Homologous landmarks were chosen from photographs of all

specimens in the ventral (27 landmarks) and lateral (19 landmarks) views of

39

thecranium. Whenever available, mandibles (n=62) were also photographed in
lateral view (39 landmarks). Landmarks at suture points and homologous
structures were digitized using tpsDig (tpsDig; SUNY, Stony Brook, NY). Sutures
on the cranium are completely fused in adult hyenas (Schweikher, 1930), and
there are few suitable landmarks on the mandible. Therefore, semi-landmarks,
points evenly spaced along curves, were used in lateral views of the cranium and
mandible to capture overall shape. The landmarks and semi-landmarks in each
view are shown in Figure 2.2 and described in Table A2. Each specimen was
photographed and digitized three times on separate days by the same observer
(JBT) in order to estimate measurement error. The data for each individual were
then averaged so that each specimen had one mean value for every landmark in
each view.

For the ventral view specimens were arranged so that the palate was
parallel to the photographic plane. Landmarks were digitized on both sides, but to
avoid inﬂating degrees of freedom the coordinates for bilaterally homologous
landmarks were subsequently reﬂected and averaged to yield 16 landmarks. In
the lateral views of the cranium and mandible, the camera was oriented
perpendicular to the sagittal plane of the specimen. A 1cm scale was used in all
photographs for all views.

A Generalized Least-Squares Procustes superimposition algorithim was
used to align the landmark and semi-landmark data. This removes information
not related to shape, i.e. scale, position and orientation (Zelditch et al., 2004;

Rohlf 8 Slice, 1990). Skull size was calculated using centroid size, which is the

40

 

Figure 2.2. Landmarks digitized in each view used in the study, (*) indicate semi-
landmarks in the lateral cranium and mandible views. See appendix for the
descriptions of each landmark point.

41

square root of the summed squared distances from each landmark to the
geometric center of the object. Procrustes distance was used to quantify the
magnitude of differences in shape, and the thin-plate spline function was used to
draw the pictures (Bookstein, 1996; Zelditch et al., 2004). Superimposition was
done using CoordGen for the ventral and lateral views of the cranium and
SemiLand for the lateral view of the mandible (Sheets, Canisius College, Buffalo,
NY).

For each view of the skull we performed a linear regression of shape
against centroid size to illustrate the dominant linear changes in shape
throughout ontogeny and generate deformation grids. In order to determine
whether a linear model was appropriate for our ontogenetic data, we performed

an exploratory analysis using principal component analysis (PCA).

Mechanical Advantage and muscle mass

To estimate mechanical advantage of the jaw adductors we used
traditional linear measurements. Following Radinsky (1981b), mechanical
advantage was calculated as the in-Iever divided by the out-lever, with the in-
lever (or moment arm) of the temporalis measured as the distance from the
dorsal tip of the coronoid process to the mandibular condyle, and the in-lever for
the masseter measured as the distance from the mandibular condyle to the
middle of the ventral masseteric depression. The out-levers for both the
temporalis and the masseter were measured as the distance from the mandibular

condyle to the bite point. The third lower premolar was used as the bite point in

42

this study because that is the main tooth used for bone-cracking in this species
(Werdelin, 1989; Van Valkenburgh, 1988). Zygomatic arch breadth at the widest

point on the skull was used as an estimate of jaw adductor mass.

Age at Maturity

To determine the age at which skull size and shape reached adult values,
we used a nonlinear growth model to calculate asymptotic values. A number of
different growth models were ﬁtted to the data using an IMP program,
GrowChoice (Sheets, Canisius College, Buffalo, NY), and Aikaike’s Information
Criterion (AIC) weights were calculated for each model. For each measure (size,
shape, mechanical advantage and zygomatic arch breadth) the monomolecular
model (following Gaillard et al., 1997) had the highest AIC weight; we therefore
used this model to estimate asymptotic values. The monomolecular model is
appropriate when initial growth is rapid, but then levels off (Karkach, 2006). The

formula used was

X(t)= All -eXpK“o‘t)}

where x(t) is the measurement of interest at time t, A is the asymptotic adult
value, K is the rate of approach to adult value, and to is the age at which growth
begins (Gaillard et al., 1997; Zelditch et al., 2003). We report age at maturity as
95% of the asymptotic value. After 95% of the asymptote was determined, this

age (in months) was used as a breakpoint and a regression was performed on

43

the subset of data after this age to ensure that age was no longer a signiﬁcant

predictor of the measurement of interest.

RESULTS
Ontogenetic changes in Size and Shape

The linear regression of change in shape in the ventral view from two
months to 18 years of age (Figure 2.3) shows that ontogeny is characterized by a
relative increase in the bi-zygomatic arch breadth and lengthening of the
zygomatic arches. Additionally there is a relative narrowing of the basicranium
and palate. Although skull length increases relative to breadth, this change is
less than the changes in the zygomatic arches, which dominate the deformation
grid in this view.

In the lateral view (Figure 2.4), we observed the typical mammalian
pattern in which the braincase becomes less bulbous during ontogeny, as
illustrated by the downward direction of the vectors from semi-landmarks along
the mid-dorsal surface of the cranium. This ﬂattening of the cranial proﬁle is
further accentuated by a relative increase in height at the posterior end of the
skull, representing expansion of the nuchal and occipital crests. The lengthening
of the zygomatic'arches is also evident in this view, as is the lengthening of the
rostrum compared to the rest of the skull.

Ontogenetic change in the mandible (Figure 2.5) is dominated by
expansion of the masseteric fossa and a marked anterior-dorsal reorientation of

the coronoid process resulting in a more vertical orientation in adults than in

44

 

Figure 2.3. Deformation grid showing ontogenetic change in skull shape in the
ventral view. The landmarks have been back-reﬂected for ease of interpretation.
Vectors show the direction and magnitude of change from the youngest to the
oldest specimens after centroid size is scaled to the same size for each
specimen.

45

 

Figure 2.4. Deformation grid showing ontogenetic change in skull shape in the
lateral view. Vectors show the direction and magnitude of change from the
youngest to the oldest specimens after centroid size is scaled to the same size

for each specimen.

46

 

“ {5* .“ .‘ﬁ’

Figure 2.5. Deformation grid showing ontogenetic change in skull shape in the
lateral mandible view. Vectors show the direction and magnitude of change from
the youngest to the oldest specimens after centroid size is scaled to the same
size for each specimen.

47

juveniles. We were unable to document changes in the tooth row since the teeth
in juveniles and adults differ due to replacement of the deciduous dentition.

Although the regression analyzed shape change based on the assumption
that it was linear throughout ontogeny, the trajectories of ontogenetic change in
skull shape generated by the PCAs were actually curved in all three views
(Figure 2.6 a, b, and c), with curvature in ventral and lateral views being
particularly dramatic. For labeling purposes, specimens were divided into three
age classes: cubs (two to 13 months of age), subadults(14 to 24 months of age),
and adults (older than 24 months since this is when spotted hyenas are
physiologically capable of breeding). In each PCA cubs differ from the other age
groups.

The dominant linear trend was described in each view by PC1, which
explained almost half of the variation in skull shape (45.9% in ventral view,
41.6% in lateral view, and 47.8% in the mandible). P02 accounted for 12.4% of
the variation in the ventral view, 12.3% in the lateral view and 14% in the
mandible, which suggest substantial deviations from a linear trend. These
components had a nonlinear relationship with age but were correlated with age
such that PC2 increased from two months to 14 months and subsequently

decreased from 14 months to 18 years of age indicating two distinct phases of

48

0- x *t*’

O. x‘,‘ g t
x

 

.c'>
o
N
.
.0

 

 

. I I A J L— A I—
-0.06 -0.04 -0.02 0 0.02 0.04 0.06
P01

0. x

P02
X
1»
no
to

0.030 *

 

 

—‘— ‘ ‘— ‘ I —|_
0.06 0.04 0.02 0 0.02 0.04

Figure 2.6. PCA plots for each view showing cubs (black circles), sub-adults
(blue crosses) and adults (red stars); a) ventral, b) lateral and c) mandible.

49

skull ontogeny: two months to around the time of weaning (14 months), and
weaning to adulthood.

To illustrate the differences in shape change between these two phases of
ontogeny, we generated deformation grids for each phase for ventral (Figure 2.7)
and lateral (Figure 2.8) views. Early in ontogeny (Figure 2.7a), changes in the
ventral view were characterized by lengthening of the zygomatic arches and
palate, and narrowing of the basicranium. In the second phase of ontogeny
(Figure 2.7b) the zygomatic arches continue to lengthen, but here we also see a
marked increase in bi-zygomatic arch breadth and a slight lengthening of the
basisphenoid.

The differences in shape change between early and late phases of
development were particularly interesting in the lateral view. Early in ontogeny
(Figure 2.8a) there was a striking dorso-ventral ﬂattening of the cranial region
and a dramatic increase in the length of the rostrum relative to the other skull
dimensions. Both are characteristic of ontogenetic patterns in many other
mammals. Later ontogeny (Figure 2.8b) was characterized in hyenas by an
increase in height and length of the caudal skull region, with development of the

sagittal and nuchal crests.

Age at Maturity
Skull size (Figure 2.9), as indicated by centroid size, reached 95% of its
asymptotic value around the time of sexual maturity or shortly thereafter in the

ventral, lateral, and mandible views (at 26, 29, and 24 months respectively). The

50

 

Figure 2.7. Deformation grids showing the two phases of skull development; a)
individuals from two months until weaning (14 months), and b) from weaning until
18 years in the ventral view.

51

 

Figure 2.8. Deformation grids showing the two phases of skull development; a)
individuals from two months until weaning (14 months), and b) from weaning until
18 years in the lateral view.

52

Centroid Size

Centroid Size

Procrustes Distance

320 r
300 r
280 .
260 r
240 .
220 r
200 r

180

160 -
140 ~

 

Ventral

 

120
0

400'

380

360 r
340 .
320 ~

280

260 ’
240 -
220 -
200 >
180 >

 

0...

'5V""”'
0

 

100 120 140 160 180 200

Age (Months)

40 60 80

Lateral

 

 

160

0.10 *
0.09 r
0.08 '
0.07 -
0.06 .
0.05 -
0.04 r
0.03 .
0.02 '
0.01 0

 

20

100 120 140 160 180 200

Age (Months)

40 60 80

Ventral

 

4L

 

20

100 120 140 160 180 200

Age (Months)

40 60 80

Figure 2.9. Ontogenetic changes in centroid size for each view. Arrows indicate
95% of asymptotic values from the monomolecular equation. Dotted lines
indicate the average ages at weaning (W) and reproductive maturity (RM).

53

95% asymptotes for shape, as measured by Procrustes distance (of each
individual from the average for the youngest age class), were much more
variable (Figure 2.10). In all three views shape change was initially rapid, but
change in the ventral view leveled off more gradually than did change in the other
two, not reaching its 95% asymptote until 35 months of age (compared to 18 and
23 for the lateral view and mandible, respectively). Interestingly, there was a
considerable individual variation observed even after adult size and shape were

reached in all of the views examined (Figure 2.9 and 2.10).

Mechanical Advantage and Muscle Mass

Mechanical advantage of the temporalis (Figure 2.11a) increased rapidly
before leveling off at 22 months of age, shortly before sexual maturity. In
contrast, mechanical advantage of the masseter (Figure 2.11b) showed very high
inter-individual variation and no clear pattern of ontogenetic change. Size of the
jaw musculature, as reﬂected in zygomatic arch breadth, continued to increase
until 33 months of age (Figure 2110), well past the age at which mechanical
advantage of the temporalis reached an asymptote.

Regressions were performed to inquire whether age was a signiﬁcant
predictor beyond 95% of the asymptotic value for each measurement (centroid
size, shape, mechanical advantage and muscle mass) and in each case, age

was not a signiﬁcant predictor.

54

Procrustes Distance

Procrustes Distance

Procrustes Distance

Ventral

 

 

 

 

 

 

 

 

 

0.10» WRM
| u . . C Q
0.09» :: . .o. °.° . '
‘ ' ’.o..o 00:000’ 0
, o o
008 .3=".,:°' "- ’
0.07' 5.6
0.06» . 9‘ i
g u o
°°°5' \ : :
0.04» ' '
0m”.
0
0.02-
‘ l i
0.01 . A . - L _ ‘ - g
0 20 40 60 80 100 120 140 160 180 200
Age(Months)
Lateral
0.12»
w M
. R . .
I I . .
, h i O . O Q . . .0 .0
0.10 .": o. :o.’ .0...o. 0. o . .
008- ° 2 i
J: '
o i
0.06l
O
.0
0.04’
O
0.02)!
0 20 40 60 80 100 120 140 160 180 200
Age(Months)
Mandible
0.12- w RM 0
010 E E . . o
i 50.5 e T. ' 0. ‘ ' 0 o. o
I C.
008 If). , .3 z, 0 ~ 0°
I I 0 0° ' ' .
006 if .
004 "i
O
0.02-3
0.00 A . A A A A A A A A A A
0 20 40 60 80 100 120 140 160 180 200
Age(Months)

Figure 2.10. Ontogenetic changes in shape (procrustes distance) for each view.
Arrows indicate 95% of the asymptotic values from the monomolecular equation.
Dotted lines indicate the average ages at weaning (W) and reproductive maturity

(RM).

55

Mechanical Advantage (Tern

Mechanical AdVantage (Masseter)

Zygomatic Arch Breadth (cm)

0.35 '
0.34 .
0.33 .
0.32 >
0.31 ~
0.30 '
0.29 .
0.28 -
0.27 >

0.26

18

16*

14,

12’

10»

8 .

 

6

.. . ‘..o o
P-------.00-v‘

 

E4 . l A . A

 

 

 

100 120 140 160 180

Age (Months)

40 60 80

200

 

33'0“""5

80 100 120 140 160 180

Age (Months)

I!
'Z

_---------2------
A;
v

 

.A

200

 

0

100 120 140 160 180

Age (months)

20 40 60 80

200

Figure 2.11. Ontogenetic changes in a) mechanical advantage of the temporalis,
b) mechanical advantage of the masseter, and c) zygomatic arch breadth. The
arrows indicate 95% of the asymptotic value as calculated from the
monomolecular equation. Dotted lines indicate the average ages at weaning (W)
and reproductive maturity (RM).

56

DISCUSSION

The ontogenetic changes in hyena skull morphology documented here
reﬂect developmental changes in the associated musculature and thus also most
likely in feeding performance, which is contingent upon morphology. The
increase zygomatic arch breadth creates a larger space through which both the
masseter and temporalis muscles can pass, and presumably corresponds to the
increasing size of these jaw adductor muscles. Lengthening of the zygomatic
arches (Figure 2.3) provides more muscle attachment area for the masseter.
Similarly, the increase in height of the sagittal and nuchal crests, as shown in
lateral view (Figure 2.4), results in increased muscle attachment area for the
temporalis muscle, which is relatively massive in this species. Both the
temporalis and masseter are importantly involved in jaw closure and bite force
generation in all mammals, but the temporalis appears to play a
disproportionately large role in bone-cracking hyaenids.

The increase in length of the rostrum during development also has
implications for feeding performance. Jaw length dictates gape, which will
determine the size of bones that individuals are able to crack open (Binder, 1998;
Binder 8 Van Valkenburgh, 2000). As the upper jaw and mandible elongate (as
seen in all three views), the out-lever increases in length yet the in-levers also
lengthen with the increase in height of the coronoid process (Figure 2.4) and the
increase in depth of the dentary, where the masseter inserts (Figure 2.5). The
increase in the length of the in-levers relative to the out-lever results in improved

mechanical advantage of the jaw musculature.

57

The age-related increase in mechanical advantage of the temporalis
(Figure 2.11a) suggests that the length of the temporalis in-lever is growing with
positive allometry relative to the out-lever. In contrast, the lack of a clear
developmental pattern in mechanical advantage of the masseter (Figure 2.11b)
suggests that the out-lever and the in-lever of the masseter are growing
isometrically therefore there is no distinct pattern of change in the ratio of the in-
lever to the out-lever. As the temporalis is the much larger of the two main
adductor muscles in this species, it is not surprising that mechanical advantage is
tied closely to the size of this muscle. However, although the masseter is not as
prominent as the temporalis, the size of the masseter also changes during
development.

The ﬁndings suggest that the traditional measurement of mechanical
advantage of the masseter will not accurately predict its contribution to changes
in feeding performance occurring during ontogeny. The estimate of muscle mass
(zygomatic arch breadth) continued to increase until 33 months, well beyond the
22 months for the mechanical advantage of the temporalis (Figure 2.11).
Therefore, by taking changes in muscle mass into account we may have a more
realistic indicator of performance than the estimates of mechanical advantage
alone. Performance is expected to increase with age as juveniles shift from a
relatively easy diet comprised mainly of milk, to one that is more challenging after
weaning. Perhaps the most interesting result here is that the transition from a

young juvenile skull, with a feeding apparatus ill-suited for a durophagous diet, to

58

an adult skull highly specialized for durophagy, continues long after nutritional
independence from the mother.

On average, spotted hyenas in our free-living study population are
weaned at 14 months of age (Holekamp et al., 1996). Thus, like other bone-
cracking hyaenids, these animals are weaned quite late relative to other
carnivores (Gittleman, 1989; Watts, 2007). Nevertheless, our data show clearly
that the skulls of newly-weaned spotted hyenas are still far from adult size and
shape (Figure 2.9, 2.10). Centroid size continues to increase past sexual maturity
in each of the views examined in this study (Figure 2.9). Although centroid size
reaches maturity at 26 months in the ventral view, shape in this view continues to
change until 35 months of age, well into adulthood. This disparity may be due to
the increase in breadth of the zygomatic arches along with the narrowing of the
braincase.

In contrast shape change in the lateral view reaches maturity at only 18
months of age and centroid size continues to increase for another year after
shape has reached maturity. This disparity is interesting, and might indicate that
growth continues isometrically in the lateral view past 18 months of age.
Alternatively, the early termination of change in shape in the lateral view relative
to that in other views might be due to distortion from using 2D images to sample
3D shapes. Changes in shape beyond 18 months are apparent, particularly in the
sagittal crest, when visually inspecting skulls side by side. Therefore, SD

morphometrics might better capture development of the sagittal crest.

59

In both the ventral and lateral skull views, we observed two distinct phases
of ontogeny. The shape changes occurring in early ontogeny differed from the
changes occurring later in ontogeny, and the shift from one phase to the other
occurred around the time of weaning at 14 months. This bi-phasic growth might
reﬂect changes in shape due to bone remodeling in response to the forces
placed upon the feeding apparatus as young Crocuta start to eat more
challenging foods after weaning. In any case, it is clearly inappropriate to treat
ontogenetic development as a linear process in this species.

Our results show that maturation of skull size and shape in the spotted
hyena is a complex process that continues well beyond the age of weaning.
Based on our calculations of mechanical advantage, which derive strictly from
skull size parameters, we would expect measures of feeding performance in this
species to reach asymptotic values by around 22 months of age, when centroid
size approaches maturity. However, our data show clearly that shape change
continues in ventral views of the skull of the spotted hyena until 35 mo of age.
This is not only long after weaning, at around 14 months, but is also long after
spotted hyenas reach reproductive maturity at 24 months. Thus these animals
are potentially capable of producing offspring nearly a year before development
of the feeding apparatus is complete.

If feeding performance in these animals is affected by skull shape as well
as by skull size, we would expect measures of feeding performance to reach
asymptotic values at 30-35 months of age rather than at around 22 months, as

suggested by size alone. In Chapter Three we investigate the relationship

60

between the morphological measures presented here with direct measures of
feeding performance in wild spotted hyenas.

Maternal investment in the spotted hyena is enormous compared to that
seen in most carnivores (Hofer 8 East, 1993; Hofer 8 East, 1995; Holekamp et
al., 1996; Holekamp 8 Smale, 1998a), and ontogenetic development of the
physical and behavioral traits required for rapid ingestion of food seems
unusually protracted in this species. Adult spotted hyenas are remarkably well
equipped for the difﬁcult tasks of capturing, killing, consuming, and defending
their prey, but juveniles are handicapped during hunting and feeding by their
small body size, weak chewing muscles, poor mechanical advantage of the jaws,
and inexperience (Biknevicius, 1996; Binder, 1998; Binder 8 Van Valkenburgh,
2000). It would be surprising if these same traits did not handicap all young
mammalian carnivores. However, extant Crocuta descended only 200,000 to
990,000 years ago from ancestors specialized for scavenging and durophagy
(Lewis 8 Werdelin, 2000). Constrained ontogenetic development of a
craniodental morphology competent for durophagy may have generated strong
selection pressures during Crocuta’s recent evolutionary shift to feeding in large
groups on live prey, favoring not only rapid development of the behavioral and
morphological tools needed to cope with difﬁcult foods, but also favoring mothers
able to help their cubs during competitive feeding.

Spotted hyenas are unique among carnivores, and highly unusual among
mammals, in that they exhibit female dominance over males. In light of the highly

specialized bone-cracking morphology in this species, its prolonged period of

61

offspring dependence, and the intensive feeding competition characteristic of this
species, we hypothesize that developmental constraints on juvenile feeding
capabilities, occurring in an environment characterized by intensive feeding
competition, may have shaped the evolution of female dominance. Evidence of
slow maturation of the feeding apparatus in Crocuta relative to that in other
carnivores would suggest the operation of developmental constraints in young
spotted hyenas.

Although our current results indicate that the feeding apparatus continues
to change well into adulthood in spotted hyenas, further work comparing
developmental patterns in skull size and shape among a large array of carnivore
species is needed to determine whether an extended development is unique to
bone-cracking hyaenids. Such a ﬁnding would provide evidence for
developmental constraints on the feeding apparatus, which, together with
intensive feeding competition, may have led to the evolution of female

dominance in spotted hyenas.

62

CHAPTER THREE
EFFECTS OF AGE, MORPHOLOGY AND SOCIAL RANK ON FEEDING
PERFORMANCE IN FREE-LIVING SPOTTED HYENAS (CROCUTA

CROCUTA)

INTRODUCTION

Selection is likely to act upon individuals throughout their lives, until death
or reproductive senescence. Therefore, selection should either reduce the length
of the life history stages in which mortality risk is high, favor enhanced
performance to improve chances of survival during risky life history stages, or
both (Williams, 1966). Since juvenile mortality is high in most animals, selection
for improved performance should be expected during early life stages (Carrier,
1996; Williams, 1966). Carrier (1996) predicted that juveniles, constrained by
their small size, should compensate for their limitations by improving their
performance relative to adults. Indeed, studies on reptiles and mammals have
shown this to be the case for certain aspects of locomotor performance, in that
juveniles are capable of anning or jumping at higher speeds than adults (Carrier,
1996; Irschick, 2000; Herrel 8 Gibb, 2006). However, the few studies that have
investigated ontogenetic variation in feeding performance have found that
juveniles are limited in their feeding capabilities relative to adults, suggesting the
operation of constraints (Binder, 1998; Binder 8 Van Valkenburgh, 2000;
Thompson et al., 2003; Herrel 8 O'Reilly, 2006; Vincent et al., 2006). Schwenk

and Wagner (2004) deﬁned a constraint as any mechanism that limits the the

63

response of characters to the selective pressures acting during particular life
stages. In the case of feeding performance, pressures from competition might
lead to the selection for rapid maturation of the feeding apparatus, and
subsequently feeding performance. However, a limitation on this response such
as delayed development of a highly specialized skull would represent a
developmental constraint.

Bite force is a common measure of feeding performance (Binder 8 Van
Valkenburgh, 2000; Herrel et al., 2001; Thomason et al., 1990; Wroe et al., 2004;
Thomason, 1991) as it may determine the types of food that can be utilized by an
individual. However, due to the difﬁculty of obtaining direct measures of bite force
in large animals, many studies use indirect estimates derived from skull
measurements (Wroe et al., 2004; Christiansen 8 Adolfssen, 2005; Christiansen,
2007; Thomason, 1991; Radinsky, 1981b; Radinsky, 19813; Radinsky, 1982). In
most mammals, bite force is likely to be largely determined by jaw muscle mass
and mechanical advantage. Traditional approaches outlined by Maynard Smith
and Savage (1959), and further developed by Greaves (1985) and Radinsky
(1981b), model the jaw as a third class lever and use linear measurements from
the dentary to calculate mechanical advantage of the jaw muscles, and maximum
zygomatic arch breadth as an estimate of muscle mass. Calculation of
mechanical advantage involves measuring the in-Iever of the masseter and
temporalis as the distance from the muscle insertion points on the mandible to

the jaw joint. These distances are then divided by the out-lever, measured from

64

the jaw joint to the bite point. Larger values of mechanical advantage and muscle
mass are expected to reﬂect the ability to achieve greater bite forces.

Although these indirect measures have provided a great deal of insight
into the morphological determinants of feeding performance, it would be
preferable to obtain, when possible, direct measures of performance from live
animals (Herring, 1985). The few studies directly quantifying feeding
performance of wild animals have been primarily limited to reptiles (lizards and
turtles: Herrel 8 O'Reilly, 2006; snakes: Vincent et al., 2006; alligators: Erickson
et al., 2003). To date, all such studies of mammals have utilized captive
populations (gray short-tailed opossum Thompson et al., 2003; Virginia opossum:
Thomason et al., 1990; spotted hyena: Binder 8 Van Valkenburgh, 2000;
Herring, 1985).

Here, we use both direct and indirect measures to investigate changes in
feeding performance throughout development in free-living spotted hyenas
(Crocuta crocuta). For large carnivores like hyenas, both capture and processing
of food is extremely demanding, and juveniles may be particularly strongly
disadvantaged during competition with adults. Where juvenile performance is
poor relative to that of adults, provisioning or assistance from parents or other
relatives might be expected until individuals have reached adult performance
capabilities. Spotted hyenas are particularly interesting in this respect because
they face extreme challenges during feeding, yet provisioning of young by

parents is very rare (Holekamp 8 Smale, 1990).

65

Adult spotted hyenas have a highly durophagous diet and are capable of
cracking open and consuming bones of large diameter. Associated with bone-
cracking is a suite of traits in the feeding apparatus that allow hyenas to sustain
and generate large bite forces. These include robust dentition, a vaulted
forehead, a pneumatized sagittal crest, and enlarged jaw musculature (Joeckel,
1998; Werdelin, 1989; Van Valkenburgh, 2007). In addition to extreme
durophagy, Crocuta also experience intensive feeding competition (Kruuk, 1972).
Unlike most carnivores, spotted hyenas live in large social groups called clans
that contain up to 80 individuals (Kruuk, 1972). Clans are structured by rigid
linear dominance hierarchies, with social rank determining priority of access to
food resources. In intensely competitive environments, like that characteristic of a
spotted hyena clan, the ability to consume large quantities of food in a short
period of time can be highly advantageous. Therefore juveniles should be
disadvantaged during competition over food if they are unable to feed as quickly
as adults.

The period of offspring dependence is unusually long for Crocuta and
other bone-cracking members of the Family Hyaenidae relative to this period in
other carnivores (Watts, 2007). Spotted hyena cubs are weaned, on average, at
14 months, but this can range up to 24 months of age (Holekamp et al., 1996;
Kruuk, 1972). Yet, despite the late age at which Crocuta are weaned, it appears
that individuals have still not reached adult performance levels by the time they
attain nutritional independence from their mother. That is, in captive spotted

hyenas, Binder and Van Valkenburgh (2000) found that bite force continued to

66

increase well past puberty, which occurs in this species at around 24 months.
Furthermore, the feeding tests they conducted using de-ﬂeshed sheep femurs
revealed that juveniles were at a disadvantage compared to adults when eating
bone, and that there was a marked increase in feeding speed around 22 months
of age. However, performance in this test was still improving in the oldest captive
hyenas measured (30 months), so these earlier investigations did not establish
when during ontogeny asymptotic performance was achieved. External linear
measures of skull size from live captive hyenas leveled off at 20 months of age,
shortly before the age of sexual maturity (Binder 8 Van Valkenburgh, 2000).
They therefore attributed any improvements in performance after 20 months to
increased size of the jaw musculature. In Chapter Two we used geometric
morphometrics to quantify changes in skull size and shape in an ontogenetic
series of Crocuta skulls from the wild (Figure 3.1 ), and found that skull size
continued to increase until 26 months, but that skull shape continued to change
until 35 months of age. In contrast to the pattern of change in skull shape, we
found that mechanical advantage of the temporalis, estimated from traditional
lever-arm models, reached maturity at 22 months of age. Using a traditional
measurement of zygomatic arch breadth to estimate jaw muscle mass, we found
that this continued to increase until 33 months of age. However, it remains to be
determined which, if any, of these traditional estimates accurately reﬂect actual
feeding performance.

Biknevicius (Biknevicius, 1996) proposed that the eruption of adult

dentition, which is fully complete in spotted hyenas by 18 months

67

—Birth

14 mo. “Weaning
18 mo. -Adult dentition fully erupted

22 mo. —Adult mechanical advantage of the Jaw
24 mo. —Reproductive maturity
26 mo. —Adult skull size

33 mo. i—Adult zygomatic arch breadth
35 mo. —Adult skull shape

 

\/

Figure 3.1. Timeline illustrating the age at maturity for morphological measures in
relation to life history events (in bold).

68

(Van Horn et al., 2003), might be the main limiting factor affecting feeding
performance in juveniles, and that weaning is delayed until tooth eruption is
complete because this is when juveniles are capable of competing with adults at
kills. If either mechanical advantage or skull size was the main factor limiting
performance in young hyenas, then feeding performance should reach maturity
at around 22 or 26 months of age, respectively. If skull shape or muscle mass
was the main limiting factor, then feeding performance should continue to
improve until 33 or 35 months of age, respectively when these morphological
measures reach asymptotic values. Each of these morphological variables
probably has important consequences for performance, and all of these variables
are changing throughout early ontogeny. However, if feeding performance
continues to improve beyond the time at which a particular morphological
variable has ceased changing, this would suggest that other variables must
subsequently be contributing to improved feeding.

Another factor likely to affect feeding performance in spotted hyenas is
social rank, which is maternally inherited and therefore persists throughout life in
females, and until dispersal in males. Higher ranking individuals in a clan feed
ﬁrst at a kill and have preferential access to the soft parts of a carcass such as
viscera and muscle (Kruuk, 1972; Bearder, 1977). Consequently, the ability to
generate large bite forces should be especially advantageous for lower-ranking
hyenas as they are more likely to feed on relatively large amounts of bones.
Furthermore the ability to feed quickly should be particularly helpful for lower-

ranking hyenas as they may be displaced from food shortly after prey capture by

69

higher-ranking conspeciﬁcs. If social rank inﬂuences feeding performance then
lower-ranking individuals should feed faster than higher-ranking individuals of the
same age.

The objectives of this study were threefold. First, we hoped to document
ontogenetic changes in feeding performance as directly measured by both bite
force generation and ingestion time. Second, we examined the relative roles of
each of several morphological variables (dentition, skull size, skull shape,
zygomatic arch breadth, or mechanical advantage of the temporalis) in limiting
feeding performance during ontogeny. Finally, we tested the hypothesis that

social rank inﬂuences feeding performance in this species.

MATERIALS AND METHODS

Study population

We conducted performance testing on known-age hyenas from a study
population (Frank, 1986) that has been continuously monitored since 1988 in the
Masai Mara National Reserve in southwestern Kenya. All individuals in the
population were identiﬁed by their unique spot patterns and birth dates were
estimated (to i 7 days) based on the appearance of cubs when they were ﬁrst
observed above ground (Holekamp et al., 1996). We evaluated dominance
relationships among mothers based on the direction of submissive behaviors
(Holekamp 8 Smale, 1993); by convention, the highest ranking female was
assigned a rank of one. Cubs were assigned the same social ranks as their

mothers.

70

Direct Measurements of Bite Force

We obtained direct measurements of bite force using a bite force
transducer (Kistler, Amherst, NY) presented to known-age hyenas in the study
population between August 2003 and April 2005. The transducer was baited
either with powdered milk for younger individuals, or meat for older individuals,
and attached to our study vehicle. When an individual hyena voluntarily bit down
on the probe, a piezoelectric plate embedded in the probe measured the force
generated in Newtons, and this was recorded from a handheld charge amplifier.
Similar devices have been used with captive hyenas at UC Berkeley (Binder,
1998; Binder 8 Van Valkenburgh, 2000), captive short-tailed opossums
(Thompson et al., 2003), and with both captive and wild American alligators

(Erickson et al., 2003).

Feeding Speed

Since both the ability to feed quickly and the ability to generate large bite
forces are critical for feeding success in the competitive environment of spotted
hyenas, we used feeding speed as another indicator of feeding performance. We
measured ingestion time for two different types of food objects. First, from June
1989 until February 1996 we conducted 174 feeding tests by presenting 309 dog
biscuits to 84 spotted hyenas of known age. Although, the dog biscuit is not an

object that hyenas would naturally encounter in their environment, it allowed us

71

to present a food item of identical size, shape and consistency to each individual
tested, and to conduct tests on cubs as young as three months of age.

From July 2003 until June 2005 we also conducted 203 feeding tests on
43 individuals using de-ﬂeshed goat femurs that were purchased from a local
butchery. The bones used in performance tests varied in mass (70-118 grams),
so the weight of each bone was recorded prior to testing. All femurs used in this
study were fresh and presented to individuals within three or four days of
preparation at the butchery. Although we did not know the last time at which each
tested hyena had eaten, we never presented bones to individuals that had
distended bellies, indicating recent ingestion of a substantial meal. Furthermore
all individuals were tested when they were away from other individuals.

For both types of feeding tests, we measured ingestion time with a
stopwatch. We only included the seconds during which each individual was
actively processing the object; i.e., if they ceased chewing for any reason while
food still remained (eg., to engage in vigilance), we did not include those

seconds in recorded ingestion time.

Analyses

In order to identify the age at which feeding performance ceased
improving, we used nonlinear growth models to calculate asymptotic values.
Using an IMP program, GrowChoice (Sheets, Canisius College, Buffalo, NY),
eight different growth models were ﬁtted to the data and Aikaike’s Information

Criterion (AIC) weights were calculated for each model. The monomolecular

72

model (following Gaillard et al., 1997) had the highest AIC weight for the biscuit

feeding data. The formula used was

x(t)= A{1-eprIto'tl}

where x(t) is ingestion time at time t, A is the asymptotic adult value, K is the rate
of approach to the adult value, and to is the age at which growth begins (Gaillard
et al., 1997; Zelditch et al., 2003). For the bone test data we used a Gompertz

equation as parameterized by Fiorello 8 German (1997):

x(t) = Aexp'ke'bt

Here, k is the initial growth rate and b is the decay of the growth rate (Fiorello 8
German, 1997; Zelditch et al., 2003). We have reported age at maturity for each
measure as 95% of its asymptotic value.

Although these models ﬁt the curves well, there was a great deal of
variation in both sets of performance data. Therefore, to further elucidate the
causes of this variation, we used the breakpoint (the age at which 95% of the
asymptotic value was achieved) to divide the data, and then constructed
generalized linear mixed models (GLMM) that included both ﬁxed and random
effects for analysis of each set of linear data separately. Here age and social
rank were modeled as ﬁxed factors and the random factor of individual identity

took into consideration repeated testing of some individuals at different ages to

73

avoid pseuoreplication. For the bone tests, we also included the weight of the
bone as a ﬁxed factor since this varied among tests. Additionally, in order to
determine the age at which feeding performance ceased to improve while taking
the above predictors into consideration, we iteratively increased the lowest age
(in months) represented in each subset of data until age was no longer a
signiﬁcant (p 2 0.05) predictor of feeding performance in the model.

We compared the age at which feeding performance reached maturity in
each test to the ages at maturity for mechanical advantage of the temporalis,
skull size, skull shape, zygomatic arch breadth and eruption of the adult dentition.
Mechanical advantage of the temporalis was measured as the distance from the
dorsal tip of the coronoid process to the jaw joint, divided by the distance from
the jaw joint to the third lower premolar. Zygomatic arch breadth was measured
as the widest distance across the zygomatic arches. These measurements were
obtained from 65 Crocuta skulls ranging in age from 2 months to 18 years
(housed at the MSU Museum). Similarly, using geometric morphometrics, as
described previously (Chapter Two), we calculated centroid size for each of the
65 specimens and examined shape variation (measured using Procruste’s
distances) among them. The monomolecular growth model described above was
used to calculate 95% of the asymptotic value for each of these morphological
measures.

Lastly, to determine if there were any effects of social rank on skull size and
shape among adults, we performed a regression on skull centroid size and on

skull shape variables (Procruste’s distances) for adult females, using age and

74

social rank as predictor variables. For these analyses we used centroid size and
shape data from the ventral view calculated in the previous chapter from 11 adult
females of known age and rank. All of the analyses were conducted with R
software version 2.4.1 (R Core Development Team, 2005). The GLMMs were

performed using the lme package in R.

RESULTS

Bite Force

Although we recorded 65 bites on the transducer from 31 hyenas, for a
number of reasons we were unable to collect sufﬁcient reliable bite force data to
document ontogenetic changes in bite force. First, it was impossible to obtain
measures from adult hyenas. Although juveniles would often become highly
engaged in biting the transducer, the adults that approached it would usually
remove the meat covering, test the transducer gently with their teeth, and then
cease interacting with the device altogether. Second, it was clear that most
juveniles biting the transducer were not biting down as hard as they could. In
Figure 3.2, we present the bite force data collected from 31 individuals between
the ages of 3.5 and 24 months. Interestingly, our largest bite force readings were
from one individual at 12.2 and 13.1 months of age. At 13.1 months, this hyena
had a bite force of 2682 N. Given how much greater this value is than all other
values save one, it is unlikely that the other bites recorded reﬂected bite force

maxima for those individuals.

75

2800;
2600} °
2400;
2200?
2000}
1800i
1600}
1400;
1200}
1000} '
800; .
600}
400:— . t. '00.. 0:.
2°33 b.4353 .
24 81012141618202224
Age(Months)

Bite Force (N)

 

6

Figure 3.2. Measurements of bite force in Newtons (N) obtain with a piezoelectric
transducer in a bite force meter available to wild hyenas.

76

Ingestion Time

Ingestion time measured in biscuit tests rapidly decreased during early
ontogeny (Figure 3.3) and reached 95% of the asymptotic value at 19 months of
age. Ingestion time measured from tests using goat femurs (Figure 3.4) followed
a similar trajectory, reaching 95% of the asymptotic value at 20 months of age.
However, when the ontogenetic data were subdivided into pre- and post-
breakpoint subsets, age continued to be a signiﬁcant predictor of performance
well beyond 20 months (Figures 3.5 and 3.6). Speciﬁcally, ingestion time in the
biscuit tests continued to decrease until 34 months of age (Figure 3.5c) and
ingestion time in bone tests until 36 months of age. Thus, even though the most
rapid and dramatic ontogenetic change in feeding performance in both tests
occurred during the ﬁrst 20 months of life, performance during ingestion of both
soft (biscuits) and hard (bones) foods continued to improve signiﬁcantly until late
in the third year of life.

Table 3.1 shows the results of the mixed models for both biscuit and bone
tests when social rank and individual identity were considered in addition to age.
The GLMM for the biscuit tests revealed that, prior to 19 months, age was a
signiﬁcant predictor of ingestion time and that there was a trend for higher-
ranking cubs to feed more quickly than lower-ranking cubs. After 19 months, age
continued to be a signiﬁcant predictor but there was no indication that rank
inﬂuenced feeding speed. By 34 months, age was no longer signiﬁcant, and

social rank was highly significant. However, the direction of the inﬂuence of rank

77

Biscuits
1600 -

1400

A 1200 -
1000 ~
800 »

600 »

Ingestion Time (sec

400 .

 

200»
0....0. 0" .0 .u... o 0 ‘

80 100 120 140 160
Age(Months)

 

 

Figure 3.3. Ontogenetic change in the time required to consume a 309 dog
biscuit during performance testing. The curved line represents the ﬁt of the
monomolecular growth model, and the straight vertical line indicates 95% of
the asymptotic value achieved at 19 months of age.

78

Bones
1600(

1400 * '
1200 ~
1000 -

800 ~

600 -

Ingestion Time (sec)

400 I

  

 

d 0 O 0

ﬂ.. An.-”..n’. . .. . O 1.
0 20 40 60 80 100 120 140 160
Age (Months)

200 >

 

 

Figure 3.4. Ontogenetic change in the time required to consume a goat femur
during performance testing. The curved line represents the fit of the gompertz
growth model, and the straight vertical line indicates 95% of the asymptotic value
achieved at 20 months of age.

79

Figure 3.5. Linear change in biscuit feeding performance for pre- and post-
asymptotic subsets of data. Age is still a significant predictor of feeding
performance in both pre- (a) and post- (b) breakpoint subsets. However, after 34
months (0), age is no longer a significant predictor of performance.

80

Ingestion Time (sec) Ingestion Time (sec)

Ingestion Time (sec)

1400'

1200~

100

3

140-

120

100'

3

40’

20»

Biscuits <19 months

' r=-0.88, p<.0001

 

 

 

2 4 6 8 10 12 14 16 18 20
Age(Months)

Biscuits >19 months

L =-0.56, p<.0001

 

 

 

20 40 60 80 100 120 140 160
Age (Months)

Biscuit >34 months

r=-.26, p= 0.09

 

 

40 60 80 100 120 140 160
Age (Months)

81

Figure 3.6. Linear change in bone feeding performance for pre- and post-
asymptotic subsets of data. Age is still a signiﬁcant predictor of feeding
performance in both pre- (a) and post— (b) breakpoint subsets. However, after 36
months (0), age is no longer a signiﬁcant predictor of performance.

82

Ingestion Time (sec)

Ingestion Time (sec)
-I —I N N 0.) A
o e s e 8 8 § 8 s

Ingestion Time (sec)

3

Bones <20 months
1600 -

1400 . ‘ r=-0.69, p<.0001

1200-

 

 

 

Age (Months)

Bones > 20 months

r=—0.26, p= 0.001

 

 

20 40 60 80 100 120 140 160
Age (Months)

Bones >36 months

=-0.21, p= 0.08

3

U!
o

 

O

 

 

Age (Months)

83

Table 3.1. Variables affecting feeding performance at different subsets of ages.
Signiﬁcant terms (p<0.05) are highlighted in bold font.

 

 

Feeding Age
Test subset Terms b :1- SE t df p
Biscuits < 19 Age -50.24 i 6.02 -8.33 26 <.00001
Rank 10.61 i 5.71 1.86 26 .07
> 19 Age -1.05 i 0.19 -5.60 39 <.OOOO1
Rank -1.06 11.09 -0.97 39 0.34
> 34 Age -0.18 $0.11 -1.75 11 .1085
Rank -2.12 i 0.59 -3.60 11 .0042
Bones < 20 Bone wt. 8.5 i 2.6 3.28 35 .0024
Age -80.18 i 10.0 -8.01 35 <.00001
Rank -10.41 i 10.0 -1.04 35 .3052
> 20 Bone wt. 1.61 i 0.38 4.17 111 .0001
Age -0.90 i 0.32 -2.77 111 .006
Rank -4.23 i 2.19 -1.93 111 .055
> 36 Bone wt. 1.84 i 0.49 3.74 58 .0004
Age -0.45 i 0.25 -1.82 58 .07
Rank -O.87 i 2.40 -0.36 58 .72

 

84

had changed from the pattern seen before 19 months, such that lower—ranking
adult hyenas now had shorter ingestion times than did higher-ranking adults.

Bone test models included an additional variable, the weight of the bone,
which was highly significant at all ages. Not surprisingly, heavier bones took
longer to ingest. Furthermore, age continued to be a signiﬁcant predictor of
performance in this test until 36 months of age. Social rank was not signiﬁcant in
any of the GLMMs for bones.

Neither age nor social rank were signiﬁcant predictors of skull shape in
adult female hyenas. However, separate regressions of skull centroid size on age
and rank for females older than 26 months of age (which is when skull size
reached maturity), showed that age was not, but that social rank was, a
significant predictor of centroid size (Figure 3.7). Lower-ranking females had

larger skulls than higher-ranking females (n=11, r=0.69, p = .02).

DISCUSSION
The difﬁculty we experienced obtaining bite force measures from adult
hyenas in the wild highlights the impracticality of collecting such measures from
wild carnivores, even those very well habituated to the presence of observers.
Although, wild alligators can be stimulated to bite down with great force by
touching their teeth with a transducer while the mouth is open (G. M. Erickson,
Pers. Comm.), there is no way to elicit a comparable response from spotted

hyenas. Future endeavors to collect such data will require greater creativity in

85

304.
302 _’ o ' r=0.5309, p=0.0283
300:
298:
296:
294:
292-
290:
288:
286:
284:
282:
280:
278: . .
2761 0
274 91-1 -
0 2 4 6 8 10 12 14 16 18 20 22 24
Absolute Social Rank

Centroid Size

 

 

 

Figure 3.7. The relationship between social rank and skull centroid size in adult
female spotted hyenas. By convention the highest ranking female is assigned a
rank of 1.

86

obtaining bite force measures. The challenge of presenting an expensive and
sensitive piece of equipment to a large carnivore in a way that might elicit a
motivated response is difﬁcult, to say the least. Even when individuals appear
highly motivated to bite the transducer, it is difficult to determine whether values
that have been recorded approach their maximum possible bite force. This
difficulty is illustrated by the variation in bite force evident among young Crocuta
(Figure 3.2). The strongest bite measured from a young hyena, 2682 N, was
substantially stronger than most of the bite force readings obtained from adult
hyenas in a captive colony (Binder 8 Van Valkenburgh, 2000). This suggests that
previous bite force measures obtained from captive animals may have similarly
failed to capture bite force maxima, and that they probably greatly underestimate
the bite force capabilities of wild hyenas. It has been estimated that forces of
between 7000-9000 N are required to crack open the limb bones of a wildebeest-
sized ungulate (Stefen, 1997; Erickson et al., 1996; Meers, 2002; Therrien,
2005), a task spotted hyenas perform daily in many parts of Africa. Furthermore,
spotted hyenas have been observed cracking open giraffe femurs and various
bones from hippopotamus and rhinoceros (Deane, 1962), suggesting that they
are capable of generating substantially more than 9000N of force during bone-
cracking. All of these facts suggest that adult free-living Crocuta are capable of
bite forces much greater than those reported here or earlier.

Our results from feeding performance tests measuring ingestion time were
more comprehensive than those obtained with our bite force meter. These

feeding performance tests allowed us to examine ontogenetic variation in feeding

87

speed, and investigate factors limiting performance. Although the ﬁtted growth
models of raw performance data suggest that ingestion time reaches 95% of ﬁnal
adult values at around 20 months of age in spotted hyenas, a more accurate
pattern emerges when additional variables inﬂuencing feeding performance are
considered. Changes in feeding speed appear instead to continue changing well
beyond 20 months and into adulthood, until between 33 and 36 months of age.
This means that feeding performance in wild hyenas continues to improve, not
just beyond weaning, but also for roughly a year beyond reproductive maturity.
Our performance data indicate that eruption of the adult dentition is not a
primary contributor to improvement in feeding performance later in ontogeny.
Dentition certainly determines the types of foods individuals can utilize,
particularly among bone-cracking animals. The ability to crack open bone is
expected to improve once the robust and specialized adult dentition is in place,
yet there are clearly other aspects of development that contribute to improvement
in feeding performance well beyond the age (18 months) at which tooth eruption
is complete. Biknevicius (1996) suggested that, since the timing of weaning
roughly coincides with the eruption of adult dentition, juveniles must be capable
of effectively competing for food with conspeciﬁcs at this time. However, neither
the results presented here nor the data from captive hyenas (Binder, 1998;
Binder 8 Van Valkenburgh, 2000) support this hypothesis. Instead, it appears
that hyenas are not able to compete effectively for food with older individuals via
rapid feeding until 33 to 36 months of age, when feeding performance stops

improving.

88

Mechanical advantage of the temporalis reached maturity at an earlier age
than ingestion time (Figure 3.8). Our results suggest either that the traditional
measures of mechanical advantage may not accurately reﬂect developmental
changes in three-dimensional morphology, or that any improvements in feeding
performance beyond the age at which mechanical advantage has reached 95%
of the asymptotic value are due instead to changes in skull size, shape or muscle
mass. Skull size reached maturity at 26 months of age, but increases in muscle
mass (indicated by zygomatic arch breadth) and skull shape continued until 33
and 35 months of age, respectively. Therefore, it is most likely changes in these
two aspects of the feeding apparatus that limit feeding speed until late in the third
year of life. Additionally, experience gained by ingesting bones and other foods
might also increase feeding speed in the third year.

One limitation of our data that prevents us from adequately testing
hypotheses about the inﬂuences of morphological variables on feeding
performance is that our performance measures and morphological measures
were taken from different individuals. Ideally, in order to address these questions,
both sets of measures would be taken together. However this was logistically
impossible in our case, as behavioral measures could not be obtained from the
individuals from which we obtained skulls.

In addition to the inﬂuences of morphology, social factors appear to
inﬂuence variation in feeding performance during ontogeny. The results from the
GLMMs on the biscuit data support our prediction that lower-ranking individuals

feed faster than their higher-ranking conspeciﬁcs as adults. Rank-related

89

-BIrth

14 mo. —-Weaning

18 mo. -—Adult dentition fully erupted

22 mo. —Adult mechanical advantage of the Jaw

24 mo. —Reproductive maturity

26 mo. —Adult skull size

33 mo. —Adult zygomatic arch breadth
34 mo. —Adult feeding performance (biscuits)
35 mo. —Adult skull shape

36 mo. -Adult feeding performance (goat femurs)

 

\/

Figure 3.8. Timeline illustrating the age at maturity for feeding performance (in
italics), for both dog biscuits and goat femurs, relative to life history events and to
the age at maturity for morphological measures.

90

variation in feeding speed might reﬂect variation with respect to motivation level,
morphology, or both. When we investigated rank-related variation in skull size
(Figure 3.7), we found that lower-ranking females had larger skulls than higher-
ranking females, which might contribute to the observed differences in feeding
performance. One other hypothesis to explain the rank-related variation in skull
size is that this reﬂects differential experience in feeding on bones. Lower-
ranking hyenas are expected to feed on greater proportions of bone in their diet
than are high-ranking individuals due to their relatively low priority of access to
food. Skull and other bones remodel themselves in response to the stresses
placed upon them (McGowan, 1998), and the greater stresses experienced by
individuals that must feed on more bones might be reﬂected in increased skull
size. Another possibility is that larger body size enhances survival among low-
ranking female hyenas such that we are more likely to acquire skulls of larger
than smaller individuals in this rank category among adults.

Interestingly, the trend seen in effects of social rank on feeding
performance among hyenas younger than 19 months was in the opposite
direction from that seen among adults. Higher-ranking sub-adults fed faster on
biscuits than did lower-ranking sub-adults. Previous studies have shown that
intra-litter rank and maternal rank affect growth rates among juveniles, with
higher-ranked individuals growing faster than their lower-ranking peers (Wahaj 8
Holekamp, 2006; Hofer 8 East, 1996). The relatively large size of high-ranking

juveniles might explain their ability to feed more rapidly.

91

The effect of social rank on performance was not as apparent from
feeding tests using goat femurs as those using biscuits, perhaps due to the
variation introduced by bone weight. There was a trend for lower-ranking hyenas
to feed faster than higher-ranking individuals in the post-breakpoint linear dataset
after 20 months of age, but this trend was not evident in the smaller subset of
data encompassing individuals older than 36 months. Although goat femurs are
undoubtedly more biologically meaningful objects than dog biscuits for
performance testing, the variation in bone size and weight contributes additional
error to the model. The goal of our study was to document ontogenetic changes
in feeding performance, so we chose to use objects that even young hyenas
could process. The trade-off, however, is that biscuits might not have been
sufﬁciently challenging for adults, and thus not adequate for reﬂecting
ontogenetic and social inﬂuences on feeding performance at older ages. Future
work might utilize larger bones to address this issue. Presenting larger bones to
adult hyenas might also provide a more accurate measure of their maximum bite
force, as the force needed to crack open bones of particular diameters and
material properties could be straightforwardly calculated. Nevertheless, spotted
hyenas eat soft as well as hard foods and the data presented here clearly show
that juveniles are handicapped relative to adults at eating both types of food.

Both the morphological variables measured in the previous chapter and
the performance variables measured here indicate that juvenile spotted hyenas
are at a disadvantage relative to adults with respect to feeding performance at

the time of nutritional independence. In wild spotted hyenas, the phase of the life

92

history during which risk of mortality is greatest occurs immediately after weaning
(Watts, 2007). This likely reﬂects the fact that individuals at the time of weaning
are neither equipped for, nor capable of, the levels of feeding performance
attained by adults. Even after the age when individuals are ﬁrst capable of
reproducing, at 24 months of age, skull shape continues to change and ingestion
time to increase, suggesting that there are developmental constraints on the
feeding apparatus that limit performance well into adulthood. Although the period
of offspring dependence is longer in spotted hyenas than in other carnivores,
young hyenas after weaning are still poorly equipped to compete with adults.
Given the limited feeding capabilities of young hyenas, we expect that females
aid their offspring in other ways after weaning, perhaps by defending them while
they feed. Indeed, individuals whose mothers survive past the individual’s
weaning age survive better than those whose mothers die shortly after weaning
(Watts, 2007). This suggests that maternal presence aids in cub survival after
weaning.

Spotted hyenas are unique among carnivores in that adult females are
socially dominant to adult immigrant males in each clan. One explanation for the
evolution of female dominance is that intensive feeding competition has selected
for large, aggressive females who are better able than other females to provide
resources to their dependent offspring (Frank, 1997). Although there is intense
feeding competition in this species, there are other large carnivores, such as wild
dogs and wolves, that experience similar levels of competition at kills (Creel 8

Creel, 2002; Schmidt 8 Mech, 1997). Thus selection pressures imposed by

93

feeding competition alone do not seem sufﬁcient to explain the evolution of
female dominance. Alternatively, it may be that developmental constraints on
juvenile feeding performance, in addition to intensive feeding competition, led to
the evolution of dominant mothers capable of assisting their handicapped
offspring. The data presented here, which indicate that development is a very
slow and prolonged process in Crocuta, support this hypothesis. However,
comparable studies on other carnivores are needed in order to determine

whether the slow ontogeny of the spotted hyena is unique among carnivores.

94

CHAPTER FOUR
ONTOGENETIC VARIATION IN THE PLAY BEHAVIOR OF SPOTTED HYENAS
(CROCUTA CROCUTA)
INTRODUCTION
Play is one of the most ubiquitous developmental processes known to
occur in mammals, yet it also remains one of the most poorly understood.
Because little is currently known about its functions, we adopt the structural

deﬁnition of play suggested by Bekoff and Byers (1981), as "any post-natal motor

 

VA!” . -

activity that appears to be purposeless, in which motor patterns from other
contexts may often be used in modiﬁed forms and temporal sequencing.”
Although this deﬁnition implies that play may serve no purpose, the frequency
with which play occurs among juveniles, its wide taxonomic distribution (Fagen,
1981), and its costs with respect to energy expenditure and predation risk
(Harcourt, 1991) suggest that play probably has important ﬁtness consequences
in the lives of mammals. Bekoff (1984) and Byers (1998) have argued that
careful documentation of the occurrence of play during ontogenetic development
in a wide array of species is a necessary prerequisite for evaluating alternative
hypotheses that suggest speciﬁc functions of play. Among mammals, rates of
play are frequently highest, and the forms of play most diverse, in primates and
carnivores (Fagen, 1981). Thus study of play in members of these orders may
provide useful insight into the adaptive signiﬁcance of these behaviors. Here we
report on play behavior in a highly gregarious carnivore, the spotted hyena

( Crocuta crocuta).

95

Spotted hyenas live in complex social groups, called clans, which may
contain up to 90 individuals (Kruuk, 1972). Every clan is structured by a strict
linear dominance hierarchy, and an individual’s position in the hierarchy
determines its priority of access to food (Kruuk, 1972; Engh et al., 2002).
Juvenile hyenas ‘inherit’ the social ranks of their mothers (Smale et al., 1993;
Holekamp 8 Smale, 1993) via the same mechanisms as those operating in
cercopithecine primates (Engh et al., 2000).

Female Crocuta give birth in an isolated natal den to litters of one, two, or
rarely three cubs. Cubs remain at the natal den for the ﬁrst 2-5 weeks of life (East
et al., 1989; Kruuk, 1972). They are then brought to the clan’s communal den
where the cubs of multiple females are sheltered together in an underground
burrow system. Cubs live at a communal den until 8 to 12 months of age, after
which they cease using the den for shelter, and begin traveling around the clan’s
territory with their mother. The communal den serves as the social center of the
clan and is periodically visited by all clan members. When older hyenas are
present at the den, cubs emerge to nurse, rest and play.

As in other mammals, the play behavior exhibited by hyenas and other
carnivores takes a variety of forms. These include play-ﬁghting, sexual play,
object play, and solitary locomotor play. Although it is widely recognized that
each form of play behavior might have its own function, and thus offer unique
beneﬁts to the individual, it is not always clear whether such benefits are
immediate for the developing animal or deferred until adulthood (Pellegrini 8

Smith, 1998). One of the goals of this study was to assess the various forms of

96

play behavior exhibited by spotted hyenas in their natural habitat, and to
document the relative frequency at which each form of play occurs throughout
ontogeny. Like Byers (1998), we view this as a critical ﬁrst step toward
understanding the functional significance of each type of play.

Another goal of our study was to evaluate sexual dimorphism in play
behavior. In many mammals (Bekoff, 1974), including humans (Hines, 2003),
some forms of play appear strongly sexually dimorphic whereas others do not. In
those species for which the physiological substrates of play have been
examined, sexually dimorphic forms of play are mediated by pre- or early post-
natal exposure to androgenic hormones or their metabolites (monkeys, Goy,
1996; Goy et al., 1988; Goy 8 Wallen, 1979; ground squirrels, Nunes et al.,
1999; humans, Hines, 2003).

The spotted hyena is a particularly interesting species in which to examine
sexual dimorphism in play behavior because many aspects of the behavior and
morphology of female hyenas are heavily masculinized. For example, females
are more aggressive than males (Szykman et al., 2003), and as adults, females
are also socially dominant to males (Kruuk, 1972). In these respects, spotted
hyenas are thus “sex-role reversed” relative to mammalian norms. Furthermore,
the external genitalia of female hyenas are strongly virilized; the hypertrophied
clitoris forms a large and fully erectile pseudopenis, and the fused vaginal labia
form a structure resembling the scrotal sac of the male. Recent experimental
work with captive hyenas has shown that, although androgens later modify the

ﬁne-structure of the phallus, these odd genitalia develop before androgens are

97

present and therefore formation of the genitalia must be largely androgen-
independent (Drea et al., 1998; Drea et al., 2002; Glickman et al., 2005;
Glickman et al., 2006). However, the maternal ovary produces high
concentrations of a weak androgen called androstenedione, which is transformed
by placental enzymes into testosterone that reaches developing fetuses via the
umbilical vein (Licht et al., 1998; Yalcinkaya et al., 1993). Fetal hyenas of both
sexes are thus apparently bathed in androgens late in gestation.

Work with primates and rodents has shown that perinatal androgen
exposure masculinizes social and sexual play behavior in juveniles, even in the
absence of any circulating hormone during the period of ontogeny when play
occurs (Goy 8 Wallen, 1979; Goy et al., 1988; Nunes et al., 1999; Pellis, 2002).
Furthermore, the masculinizing effect of androgen on juvenile play behavior is
separated in time from its masculinizing effect on genital differentiation (Wallen 8
Baum, 2002). In rhesus monkeys, for example, androgen exposure early in fetal
development masculinizes the genitalia of the female but has no effect on
juvenile play behavior; by contrast, androgen exposure late in gestation produces
no genital masculinization, but strongly affects juvenile rough-and-tumble play
and play-mounting (Goy et al., 1988). Similarly, in humans, girls affected with
congenital adrenal hyperplasia (CAH), and consequently exposed to heightened
levels of testosterone prenatally, exhibit more male-typical toy preferences and
more masculine play behaviors than unaffected girls (Hines, 2003).

Interestingly, in both humans (Hines, 2003) and spotted hyenas (Dloniak

et al., 2006), naturally-occurring variations in concentrations of maternal

98

androgens during pregnancy are associated with variations in the sexually
dimorphic play behavior of the offspring resulting from those pregnancies. That
is, in both species, higher androgen concentrations in pregnant females are later
associated with higher rates of some forms of male-typical play in female
offspring. However, it is not currently known how many different forms of play are
sexually dimorphic among young spotted hyenas, nor which forms of play might
be inﬂuenced by prenatal androgen exposure. If androgens inﬂuence play in
spotted hyenas as they do in primates and other mammals, then patterns of both
social play and play-mounting in female hyenas should be similar to those of
males. Pedersen et al. (1990) found that, among captive hyenas in same-sex
groups, females engaged in more social play than did males. In the current
study, we evaluated this possibility in wild spotted hyenas.

Our ﬁnal goal was to test two hypotheses about particular classes of
environmental variables known to inﬂuence rates of play in other mammals:
energy availability and social context. Numerous studies on juvenile animals
ranging from birds to primates have shown that rates of play decrease when
energy supplies decline (Sharpe et al., 2002; Bateson et al., 1990; Nunes et al.,
1999; Nunes et al., 2004; Pozis-Francois et al., 2004; Rajpurohit et al., 1995;
Barrett et al., 1992; Baldwin 8 Baldwin, 1976). However, in some species the
opposite relationship has been shown. For instance, by reducing energy
availability, inducing an early onset of weaning increases rates of social and
locomotor play in domestic cats (Martin 8 Bateson, 1985). Similarly, underfed

rats play more than well-fed individuals (Loranca et al., 1999), and captive

99

chimpanzees and bonobos engage in more social play before than after feeding;
in these species play might function to reduce the intensity of aggression and
competition around food (Palagi et al., 2004; Palagi et al., 2006).

Here we tested several predictions of an hypothesis suggesting that low
levels of available energy are associated with relatively low rates of play in hyena
cubs. Because intra-Iitter rank has a signiﬁcant effect on cub growth rates (Hofer
8 East, 1996; Wahaj 8 Holekamp, 2006), we inquired whether rates of play
behavior also varied here with intralitter rank. If so, we expected dominant cubs
within twin litters to play at higher rates than their subordinate siblings, as they
enjoy superior access to the mother’s milk. Social dominance determines priority
of access to food resources in this species, with high-ranking individuals enjoying
better food access than low-ranking ones (Kruuk, 1972). Therefore, if the
hypothesis is correct that energy affects hyena play, we expected that rates of
play would be elevated in cubs with high-ranking mothers relative to those with
lower-ranking mothers. Furthermore, after controlling for numbers of potential
play partners, we expected singleton cubs, which have sole access to their
mother’s milk and therefore grow faster than twins (Hofer 8 East, 1996), to
exhibit more play than cubs from twin litters. Additionally, if food availability
affects rates of play, then play rates should be lower in all cubs during periods
when food is scarce than during periods of prey abundance.

Finally, one important feature of the immediate social context that might
inﬂuence rates of play is the number of potential play partners available to an

animal (Poirier 8 Smith, 1974). The number of available partners can be quite

100

variable for spotted hyena cubs, who live for several months at a communal den
where offspring of two to several adult females may reside. If availability of play
partners affects rates of play, we expected that individuals from large cohorts of
cubs would play at higher rates than individuals from small cohorts. We
emphasize that hypotheses suggesting energetic and social variables as
determinants of play rates are not mutually exclusive, and that multiple variables

might be operating concurrently to affect rates of play.

MATERIALS AND METHODS

Study Animals

We collected data from members of one large clan of spotted hyenas
inhabiting the Talek region of the Masai Mara National Reserve in southwestern
Kenya (Frank, 1986; Boydston, 2001 ). Adult and sub-adult hyenas were
identiﬁed individually by their unique spot patterns. Young cubs, which have a
solid black Ianugo, were identiﬁed on the basis of distinctive scars, nicks on the
ears and patterns of molt. Relationships between mothers and offspring were
determined based on nursing associations, and birth dates were estimated (i 7
days) based on the appearance of cubs when they were ﬁrst observed
(Holekamp et al., 1996). Cub sex was determined from the dimorphic shape of

the glans of the erect phallus (Frank et al., 1990).

101

Data collection

Repeated focal animal samples (Altmann, 1974), lasting on average 27 _+.
2.11 minutes, were conducted on individual hyenas ranging in age from less than
1 month to adulthood. All occurrences of play were recorded during focal
sampling. Ontogenetic data were collected between June 15, 1988 and August
14, 1998 from 24 males and 20 females, including 13 singletons and 31 animals
from twin litters. Most cubs were sampled biweekly between birth and 30 months
of age, with an average of 2.05 :l: 0.21 hours of focal animal sampling for each
individual during every two-month age interval. A total of 691 focal sampling
hours were included in this component of the study. Den-dwelling cubs spend
much of the day underground, typically entering the den two to three hours after
sunrise and emerging again in the late afternoon, approximately two to three
hours before sunset. Behavioral data were therefore collected mainly during
daylight hours between 0600h and 0900h, and again between 1600h and 1930h.

The same two observers (K.E.H. and LS.) collected all data included in
the 2-4 month and play-mounting analyses. However, the larger ontogenetic
study, which spanned a decade, included observations made by 16 observers
who usually worked in pairs. All observers were blind with respect to the
hypotheses tested here, and all were trained individually by K.E.H. for several
months. Observers were allowed to collect data independently only once they
could consistently and accurately identify all of the study animals as well as all

play behaviors described below.

102

Play was consistently recognized by observers based on exaggerated
movements, apparent purposelessness, speciﬁc initiation postures, and the
absence of either aggressive posturing or vocalizations indicative of pain inﬂicted
when bitten by another individual. We observed ﬁve types of play behavior: social
play, romping, object play, play-mounting and non-nutritive chewing. Social play
involved multiple individuals engaged together in chasing, wrestling, jumping or
chewing on one another. Romping was identiﬁed as exaggerated running and
jumping behavior exhibited by solitary individuals. Object play occurred when we
observed an animal running with or tugging on sticks, bones, feathers or old
scraps of skin. Play-mounting involved one cub approaching another from behind
and placing its forepaws on the back of another cub to assume a posture like that
exhibited by adult males during copulation; the phallus was often erect during
play-mounting, but was not a necessary condition for us to record the occurrence
of this form of play. If the animal being mounted did not pull away immediately,
then the mounting animal typically also exhibited thrusting behavior, but no
intromission. Lastly, one additional type of object play recorded here, seen
mainly in carnivores, is non-nutritive chewing. This is deﬁned as biting or
gnawing on objects from which little or no energy or nutrients could be derived,
including plants, sticks, or old dried pieces of bone (Power, 2000; Fagen, 1981).
To determine whether non-nutritive chewing might simply reﬂect a process
similar to the teething behavior seen in human infants, we examined non-nutritive

chewing here in relation to patterns of tooth eruption and replacement in hyenas.

103

All types of play behavior occurred in bouts lasting from a few seconds to
several minutes. In order to compare rates of play and examine effects of
energetic and social variables on play rates, we recorded all play behaviors as
states rather than discrete events. That is, if any play of a particular type
occurred within a given minute of observation, then that 60-second interval was
counted as a minute of play of that type. If a particular type of play occurred in
multiple shorter bouts within a single minute, this was nevertheless counted as
having occurred only once. Hourly rates of play behavior of each type were
calculated for every individual in each two-month age interval from birth to 30
months of age. Rates were calculated by dividing the number of minutes in which
each type of play occurred by the number of minutes in which that animal was
observed as a focal animal during each age interval.

Although rates of play-mounting were recorded as states for purposes of
comparison with other types of play, for more detailed analyses of play-mounting
we also recorded each mount by a particular individual as a discrete event, such
that it could be recorded as occurring multiple times within a given minute of
observation. Each occurrence of play-mounting was recorded as a critical
incident (Altmann, 1974) from 21 male cubs and 15 female cubs monitored
throughout the ﬁrst 8 months of life during all observation sessions at dens. An
observation session began when we arrived at the den, and ended when
observers left the den. Observation sessions at dens ranged in duration from five

minutes to several hours. Cubs included in this analysis of play mounts as critical

104

incidents were each observed at dens for an average of 117 i 4 hours. These
data were collected during 4212 hours of observation of all cubs individually.

In order to examine potential effects on play rates of social (cohort size)
and energetic variables (prey density, litter size, intralitter rank and maternal
social rank) we focused exclusively on the 2-4 month age interval because this
proved to be the period during which the highest rates of play were observed. Of
the original 44 cubs observed in the overall ontogenetic analysis, 40 (23 males
and 17 females) were included in analyses focused exclusively on the 2-4 month
age interval. We used a mean of 7.00i 0.48 focal animal samples per individual
in this analysis (3.58: 0.23 focal samples per individual per month), such that
each individual was observed for an average of 3.36 i 0.22 hours during the 2-4
month age interval. The total observation time used for focal animal sampling in
this component of the study was 134 hours.

We evaluated dominance relationships among mothers on the basis of the
direction of submissive behaviors (Holekamp 8 Smale, 1993); by convention, the
highest ranking female was assigned a rank of one. Cubs were assigned the
same social ranks as their mothers. The hierarchy was then divided into equal
thirds and individuals were assigned to one of three categories; high-, mid- and
low-ranking. Throughout the study period bi-weekly prey censuses were
conducted by counting all ungulates within 100m of each of two 4 km transects to
monitor local prey availability in the Talek area. Prey availability was categorized
as either high or low at each census based on whether the count was above or

below the mean prey density for the entire study period. For each individual, its

105

cohort size was calculated as the number of cubs born within six months of its
birth date. Rank relations within litters were determined both by outcomes of
dyadic aggressive interactions between littermates and by nursing positions. The
dominant cub within a litter usually assumes a preferred nursing position against
the mother’s belly whereas the subordinate cub typically nurses while lying

between the mother’s hind legs (Wahaj 8 Holekamp, 2006).

Statistical Analyses

For the overall ontogenetic analysis we used non-parametric statistics to
evaluate effects of age and sex on play rates because these data were not
normally distributed. Within each age interval we compared rates for males and
females, using the Mann-Whitney-U test statistic, and we examined changes
across all age intervals with a KruskaI-Wallis ANOVA, followed by multiple
comparisons among speciﬁc intervals using Bonferroni corrections. We used a
Friedman test statistic to analyze variation in play rates throughout ontogeny for
individuals for which we had data from all age intervals. For each sex we
examined variation in the number of play mounts exhibited during the ﬁrst eight
months of life using the Friedman test statistic, followed by comparisons among
speciﬁc intervals using Wilcoxon Signed Ranks tests. A Spean'nan’s correlation
coefﬁcient was calculated to examine relationships between maternal social rank
and rates of play-mounting. These statistical analyses were performed in

STATISTICA 6.1 (Statsoft, 2002).

106

To test hypotheses about the effects of social and energetic factors on
rates of play during the 2-4 month age interval, we used a generalized linear
model (GLIM) with binomial error distributions and a logit link function to analyze
each play type. A quasi-binomial distribution was used in order to correct for
overdispersion in the model, and the signiﬁcance of each term was evaluated
using an F test (Crawley, 1993). The number of minutes during which a cub was
observed engaging in a particular type of play throughout the 2-4 month age
interval, and the number of minutes observed in which that type of play did not
occur, comprised the binomial response variable in each model. This approach
analyzes the proportion of minutes in which a speciﬁc type of play was observed,
but weights the analysis based on the number of minutes observed (Venables 8
Ripley, 1999). The same predictor variables were initially used in each model,
with cohort size as a continuous predictor and sex, litter size (singleton or twin),
social rank (high, mid, low) and prey availability (high, low) as categorical
predictors. Each model was then simpliﬁed by eliminating any variables that did
not contribute to its explanatory power. Only individuals observed for at least 30
minutes during the 2-4 month age interval were included in this analysis. The
GLIM analysis was conducted using R software version 2.1.1 (R Core
Development Team, 2005).

Finally, to evaluate the effect of intra-Iitter dominance on rates of play
during the 2-4 month age interval we performed Wilcoxon matched pairs

analyses between dominant and subordinate littermates. Differences between

107

groups were considered signiﬁcant when a<0.05 after Bonferroni correction, and

mean values are presented as t SE throughout.

RESULTS

Ontogenetic variation in rates of play

Social play was the type of play behavior most frequently observed,
occurring on average during more than 12 minutes in each hour observed during
the 2-4 month age interval (Figure 4.1 a). The overall mean rate of social play
between birth and 30 months of age was 4.50 :l: 0.38 min/hr, but rates peaked
during the second two-month age interval, and then declined (Figure 4.1a,
Kruskal-Wallis H (18,268), F = 121.6, p < 0.00001). As we did not manage to
sample every cub in every age interval, we performed a repeated-measures
ANOVA on only the subset of cubs (n =13) sampled in every interval from 0-20
months, and we observed a pattern of results identical to that in our original
analysis (Friedman test statistic = 41.46, Kendall coefﬁcient of concordance =
0.39, df = 8, p < 0.00001). We found no sex difference in rates of social play
(Mann-Whitney U = 204, p = 0.40). Interestingly, social play continued into
adulthood at hourly rates much higher than those observed for any other form of
play (Figure 4.1).

Romping occurred, on average, during 1.61 :t 0.20 min/hr between birth
and 30 months of age. The highest rates of romping occurred during the ﬁrst two
months of life (Y = 7.83 1: 1.97 min/hr), and dropped to near-zero values after 6

months of age (Figure 4.1 b, KruskaI-Wallis H (18,268), F = 155.04, p < 0.00001).

108

m
P'

Communal Den Communal Den

.L
O

      

  

 

 

 

E A 8
E 10 E
E 8 g 6
o. a.
w 4
2 -' ‘ 7 . 2
°gvggee-esaaaaas °
“ éésseégssse
Age(Months)
c d.
1.6 Communal Den 1-0 Communal Den
IMaIes
1.4 -Females
A08
E12 §
210.8 E
.—.. 0.
gas $0.4
o
20.4 g
>
(u 0.2
E 0.2
°'°gvsgeezesaax 0-0
" " «5 9 s s .4 e a 4
Age(Months)
e.
10 CommunalDen
l————l
MilkTeeIh
1: Emptlng AduttTeethEnptIng
,5: 8 l———l
E
8’
‘ 6
E
.C
O
.3 4
‘3
‘F
5 2
z
0
"T ENZSE’RN'ma’B
“Hgéésiesé §§§e
Age(Months)

Figure 4.1. Ontogenetic variation in rates of (a) social play, (b) romping, (c) play
mounting, (d) object play and (e) non-nutritive chewing in spotted hyenas. Age
classes represent two month intervals (eg., 0-2 month interval includes
individuals from 0.1 months to two months of age). Because no signiﬁcant sex
differences were found in other forms of play, separate rates for males and
females are shown only in “c". Horizontal bars indicate the period during which
cubs live at the clan’s communal den on all graphs and on (e) periods of tooth
eruption. Note that y-axis scales vary considerably.

109

Analysis of the subset of cubs sampled in every age interval produced the same
pattern (Friedman test statistic = 46.98, Kendall coefﬁcient of concordance =
0.45, df = 8, p < 0.00001). We found no sex difference in rates of romping (Mann-
Whitney U = 190.5, p = 0.24).

Play-mounting was the only type of play for which we found a significant
sex difference (Figure 4.1c, Mann-Whitney U = 94, p = 0.0005), with males
engaging in this activity at much higher rates than females. Play-mounting
occurred at lower rates throughout development than did either social play or
romping (males: ‘x‘ = 0.30 :I: 0.07 min/hr; females: 2 = 0.05 :l: 0.02), but there
was signiﬁcant variation among age classes (Figure 4.1c, Kruskal-Wallis H (18,
268), F= 66.78, p < 0.00001). For males, a peak in mounting rates occurred
during the second age interval, at 2-4 months of age (Figure 4.1c). Mounting
behavior vanished from the play repertoire much earlier than did other forms of
play; it was never observed after 12 months of age among females, nor after 16
months of age among males. Male spotted hyenas pass through puberty at
around 24 months of age (Holekamp 8 Smale, 1998b).

Object play occurred at the lowest rates of any play type observed (2 =
0.25 i 0.03 min/hr). Although we observed significant variation among age
classes in rates of object play for all individuals, and also for those individuals
sampled in every age interval (Figure 4.1d, Kruskal-Wallis H (18, 268), F:
148.34, p < 0.00001; Friedman test statistic = 25.96, Kendall coefﬁcient of
concordance = 0.25, df = 8, p < 0.0011), the ontogenetic pattern in object play

was less clear than for other forms of play. Rates of object play appeared to be

110

largely determined by availability of suitable objects in the immediate
environment of each focal cub at the time it was sampled. We found no
signiﬁcant difference between males and females in rates of object play (Mann-
Whitney U = 204, p = 0.39).

After social play, the mean rate of non-nutritive chewing was the highest of
the various types of play observed between birth and adulthood (x: 2.61 :I: 0.24
min/hr), with cubs 2-4 months old spending a mean of seven minutes of every
hour observed engaged in this activity. We documented signiﬁcant ontogenetic
variation in rates of non-nutritive chewing (Figure 4.1e, Kruskal-Wallis H (18,
268), F = 148.34, p <0.00001; Friedman test statistic = 39.53, Kendall coefﬁcient
of concordance = 0.38, df = 8, p < 0.00001). Chewing rates were highest
between two and four months of age (E: 6.98 :I: 0.68 min/hr), and then declined
to nil by 24 months of age.

To inquire whether higher rates of non-nutritive chewing reﬂected teething,
we examined rates of this form of play in temporal relation to eruption of both
deciduous (milk) teeth and adult dentition. Although Crocuta cubs are born with
fully erupted deciduous incisors and canines (Frank et al., 1991 ), the deciduous
cheek teeth erupt during the ﬁrst several weeks after birth, and this process is
complete by two months of age (Figure 4.2). There is then a four-month hiatus
before the adult teeth start to erupt at six to seven months of age, followed by a
period during which both milk and adult teeth are present concurrently. The adult

teeth are fully erupted by 18 months of age (Van Horn et al., 2003).

111

 

 

 

2 months

 

 

11 months

 

 

18 months

 

 

 

Figure 4.2. The eruption of the deciduous dentition (shaded) is complete by two
months of age. Adult teeth (white) begin to erupt around six months but are not
fully erupted until 18 months of age.

112

The highest rates of non-nutritive chewing occurred during the period (2-4
months of age) between eruptions of deciduous and adult teeth; rates declined
signiﬁcantly before initial eruption of the adult teeth (Figure 4.1e, multiple
comparisons with corrected p—values: 2-4 months vs. 1-2 months, p < 0.000001;
2-4 months vs. 4-6 months, p < 0.00001; 24 months vs. 6-8 months p < 0.0001 ).
This pattern was largely inconsistent with the hypothesis that this behavior
represents teething. Furthermore, non-nutritive chewing was observed, albeit at
low rates, at 18-22 months of age, after all adult teeth have fully erupted (Figure

4.1e).

Analysis of play mounts as events

Mean rates of play-mounting were significantly higher among males than
females during all but the youngest age interval (Figure 4.3a), during which 14 of
15 females exhibited no mounting at all but one female engaged in this behavior
at an extremely high rate. Interestingly, as an adult, this female (RV) never
reproduced and appeared to be infertile, so may have been exposed in utero to
an unusually high concentration of androgens relative to other females.
Yalcinkaya et al. (1993) suggested that early exposure of female spotted hyenas
to androgens appears to severely modify ovarian histology, and it may be that
this occurred in RV and impaired such physiological processes essential for
female reproduction as oogenesis or ovulation. Among females, rates of play-

mounting were generally low, and did not vary signiﬁcantly with age (Friedman

113

 

- Males
- Females

 

 

O
O
E
2
3
O
2
Age (months)
b.
120 -
100 - °
0
80 - o
8
‘8
«‘2
2

 

 

 

Figure 4.3. Rates of play-mounting when individual mounts were recorded as
events a) among males and females throughout the ﬁrst eight months of life, b)

among males between 2-4 months of age, as a function of their mothers’ social
ranks.

114

test statistic= 4.220, Kendall coefficient of concordance= 0.094, df=3, p > 0.05;
Figure 4.3a).

Among males, rates of play-mounting varied significantly with age
(Friedman test statistic = 15.857, Kendall coefﬁcient of concordance = 0.252, p <
0.001; Figure 4.3a). Speciﬁcally, play-mounting rates were relatively low from 0-2
months, peaked during the 2-4 month age interval, decreased between four and
six months, and subsequently remained relatively low between six and eight
months of age. Finally, mounting rates among males 2-4 months of age were
positively correlated with maternal rank such that sons of high-ranking females
engaged in this behavior at signiﬁcantly higher rates than did sons of low-ranking

females (Figure 4.3b; Spearrnan’s R= 0.389, F= 4.306, p < 0.05).

Modulating inﬂuences on play at 2-4 months of age

Table 4.1 presents mean rates of play of each type during the 2-4 month
age interval, and shows how mean values for each play type were affected by
each categorical predictor variable. Table 4.2 summarizes results from analyses
using generalized linear models, indicating which independent variables had
signiﬁcant or near-signiﬁcant effects on rates of play of each type. Because play-
mounting was strongly sexually dimorphic, we included only males in the 2-4
month GLIM of this type of play. There was a trend for more romping to occur
among males than females, but sex was not a signiﬁcant predictor of romping
rates (Table 4.2, GLIM, F 1,38 = 5.29, p= 0.09), nor was any other independent

variable included in the ﬁnal model for this form of play.

115

 

 

 

 

 

 

 

116

 

7090.0 0.0H00 No.30 No.30 0000.0 V0000 3.000;. N.0H0.0 N.0H0.0 v.0H00 .5030

n.0HN... N.0H0.0 N.0H0.0 0090.0 0.03.? 0.0000 0v.0HN._. 0.09 _. .0 N090... Noﬁmd ESQ—2

5000.0 0. _.H0.v 530% 00.10% N. 30.0 0.0HFO 0_.. 710.0 0.00; 0.0H00 0,090.0 QEom

5090K 0.NH0.0 N930 N930 meﬂms 0.0HN.0 N0._.H0.0 00410.0 0.30.5 503.0 30:0
«.30.? 0.NH0.0 v.39: #7100 00210.3 veHNNP 0.3040? @302. 5:60? w.FH0.mF _m_oow

2,3 :9: 25.. Es. BE as; 85.96 mason. was.

M 0%... SE
>20 xcmm 3000 050 .00: x00

 

 

 

 

 

 

.th Eon E 0020:sz 9m £009: 50:: 03:90:00 E00 92200.5 30.0 v 3 0:85:90
0% .6 ch0E YN mono 3:080 .2 003 BE some he 500605 3260008 >0 00 new 05:25 mega. cam—2 ._..v 03m...

 

Table 4.2. Factors associated with rates of play in 2-4 month old spotted hyena
cubs (n=40), identified in generalized linear models. Only signiﬁcant terms (*),
and terms that indicate trends, are presented.

 

 

Play Type Effect Model Term F df p

Social Low > High Prey Category 4.79 1,32 0.04 *
High > Mid & Low Rank category 3.27 2,36 0.05 *
Larger > smaller Cohort size 3.23 1,38 0.08

Romp M > F Sex 5.29 1,38 0.09

Play Mount Low > High Prey Category 5.29 1,18 003*

(males only)

Object Singleton > Twin Litter Size 4.71 1,38 0.04*

Non- All N/S

nutritive

Chewing

 

117

Singleton cubs engaged in object play at higher hourly rates than did cubs
from twin litters (Tables 4.1 & 4.2, GLIM, F 1, 38 = 4.71, p = 0.04), but litter size did
not appear to affect rates of any other play type. None of the independent
variables examined here had any appreciable inﬂuence on rates of non-nutritive
chewing during the 2-4 month age interval. Similarly, intralitter rank appeared not
to affect rates of any form of play, as we found no signiﬁcant effects of this F
variable (Wilcoxon Matched Pairs, social play. n = 15, T = 55, p = 0.78; romp:

n=15, T: 53, p = 0.69; object play. n=15, T = 19, p = 0.12; non-nutritive chewing:

 

n=15, T = 47, p = 0.73). Maternal dominance status had a signiﬁcant effect on
rates of social play, with high-ranking cubs engaging in more play than either
mid- or low-ranking animals (Figure 4.4; Table 4.2, GLIM, F 2, 36 = 3.27, p=0.049).
Unexpectedly, social play occurred more frequently during periods of low prey
availability than during periods when prey were abundant (Tables 4.1 & 4.2,
GLIM, F132 = 4.79, p=0.04). Similarly, rates of play-mounting among males 2-4
months of age were higher during low than high prey periods (Table 4.2, GLIM, F
1,18 = 5.29, p = 0.03). We observed a trend towards higher rates of social play in

cohorts of large size (Table 4.2, GLIM, F 1, 38 = 3.23, p = 0.08).

DISCUSSION
Play appears to be a signiﬁcant component of the behavioral repertoire of
wild spotted hyenas, particularly during early development. Drea et al. (1996)

examined the emergence of play behavior in captive Crocuta cubs from twin

118

Social Play (min/hr)

 

 

High Mid Low
Social Rank

Figure 4.4. Rates of social play among cubs 2-4 months old in relation to
maternal social rank category. Letters indicate signiﬁcant difference among
maternal rank categories.

119

litters in the ﬁrst four weeks of life, a period during which wild cubs are seldom
seen above ground, and found that rates of social play increased over these first
few weeks. These authors hypothesized that early social play promotes the
development of a repertoire of social behavior that prepares young cubs to
become integrated into the clan when they are moved to the communal den
(Drea et al., 1997). When these results from captivity are combined with the data
presented here from free-living Crocuta, we have, for the first time, a
comprehensive description of the ontogenetic patterns of play in this gregarious

carnivore.

Factors modulating rates of play

Our results were mixed with respect to the potential effects of availability
of energy and social partners on rates of play by hyena cubs. We documented a
trend toward higher rates of social play among cubs that grew up with the largest
cohorts of peers. This suggests, but cannot conﬁrm, that partner availability
inﬂuences patterns of play in spotted hyenas. Similarly, although the inﬂuence of
available energy on play rates was suggested by our ﬁnding that rates of all ﬁve
play types were higher in singletons than in twins, none of these differences were
statistically signiﬁcant. Cubs of the highest-ranking females engaged in social
play at significantly higher rates than did their lower-ranking peers (Figure 4.4).
This result is consistent with the notion that energy availability affects play rates,
but an alternative explanation is that it reﬂects a preference among clan

members for high-ranking play partners. Indeed, although partner preferences

120

 

have not yet been examined in young hyenas, adult females prefer to associate
with high-ranking females (Holekamp et al., 1997). More detailed analysis of play
initiation will be needed to evaluate kin and rank-related partner preferences
during play among free-living Crocuta.

We found that prey density was actually inversely related to rates of play
among hyenas. Although this was surprising, there are at least two possible
explanations. First, prey availability is relatively high throughout the year in the
Masai Mara National Reserve (Ogutu & Dublin, 2002), where our Talek study site
was situated. It is therefore possible that, even during periods of low prey
abundance, energy resources are sufﬁcient to support high rates of play among
Talek cubs. Indeed, ﬂuctuations in prey abundance do not appear to affect cub
survival in this study population (Watts, 2007). Alternatively, our time frame for
looking at energy availability may have been too broad, and may therefore have
obscured relationships between food intake and play. A finer-grained analysis of
play in temporal relation to nursing bouts for individual cubs might reveal a
different pattern from that observed here. Nevertheless, neither of these
explanations could account for our observation of signiﬁcantly higher rates of
play during low than high prey periods.

Some studies of other species have reported higher rates of play during
periods of low than high food availability (Martin & Bateson, 1985; Loranca et al.,
1999; Palagi et al., 2004; Palagi et al., 2006). It has been suggested that play
under these conditions may function to reduce stress associated with competition

for limited food resources which might conceivably be the case among Talek

121

 

hyenas. However, this seems unlikely among hyenas because, when we observe
den-dwelling cubs who are energetically stressed after their mothers die, play is
one of the ﬁrst activities to disappear from their behavioral repertoire. Lastly, it is
possible that female Crocuta spend more time in search of food during periods of
prey scarcity, and less time at the den, than when prey are abundant. Therefore,
increased rates of social play during the former periods might reﬂect more “free”
time available to cubs lacking opportunities to nurse. This pattern appears to
occur in juvenile vervet monkeys (Cercopithecus aethiops) in which play rates
are elevated when mothers are absent (Govindarajulu et al., 1993). We are
currently examining the relationship between prey availability, den attendance

and other aspects of maternal care (Stricker et al., unpublished data).

Possible functions of play in spotted hyenas

Although our goal was not to test speciﬁc hypotheses about the functions
of each type of play, the ontogenetic patterns we observed here suggest some
ways in which play might enhance ﬁtness in spotted hyenas. In the wild, social
play occurred more frequently and continued until much later in life than did any
other form of play (Figure 4.1). The highest rates of social play occurred when
cubs were 2-4 months old, during which time individuals engaged in vigorous
social play for roughly 20% of each hour observed. This pattern suggests that the
beneﬁts derived from this form of play must be considerable in order to outweigh
the energetic costs associated with it. These high rates of social play were

observed shortly after cubs were brought to the clan’s communal den, when

122

Crocuta cubs must learn not only the identities of their group members, but also
their own places in the social hierarchy (Holekamp & Smale, 1993).

The ontogenetic pattern observed here suggests that social play facilitates
the acquisition of social knowledge and integration into the clan (Gomendio,

1988). It seems unlikely that adults would continue to engage in this activity if no

 

beneﬁts were accruing from it. Indeed, the persistence of social play behavior It"
from the natal den into adulthood suggests that the beneﬁts from this form of play

probably continue to accrue long after cubs become independent of the den.

Although adults were observed playing away from the den, their play rates may J

largely reﬂect interactions with den-dwelling cubs when older focal individuals
returned to the den to visit. It may be that the older individuals are gaining
information about new members of the clan by playing with them, or that they are
honing their own parental skills during their extensive playful interactions with

young cubs.

Rates of romping were highest during the ﬁrst two months of life, but
subsequently decreased. The ﬁrst months of life are critical for developing
locomotor skills, and it is likely that romping enhances the development of
coordination and movement in very young cubs, as predicted by the ‘motor
training hypothesis’(Spinka et al., 2001). The gradual decrease in romping after
two months of age may also reﬂect a transition from individual play to social play
that accompanies transfer from the natal den to the communal den between two

and ﬁve weeks of age. That is, to the extent that play facilitates development of

123

coordination and other motor skills, these beneﬁts may be attained mainly from

vigorous social play, rather than solitary romping, after the ﬁrst few weeks of life.

Social play, play-mounting and non-nutritive chewing all reached their
peaks during the 2-4 month age interval, shortly after cubs were moved to the
clan’s communal den. However, these rates of play were not merely a response
to living at the communal den, as they dropped to adult levels before cubs
became independent of the den. Furthermore, captive cubs play at high rates

during the ﬁrst few weeks of life, even in the absence of a communal den (Drea

 

et al., 1996). If these forms of play facilitate acquisition by cubs of new
information about their social and physical environments, the fact that play rates
drop even before wild cubs leave the den suggests that the ability to capture new
information via playful interactions is declining. We found that solitary object play
occurred quite infrequently throughout ontogeny, but the rarity of this play type
may merely reﬂect a shortage of suitable objects (“toys”) at the den, or the
abundance of playmates there.

Non-nutritive chewing is not likely to be primarily associated with tooth
eruption or alveolar development in spotted hyenas, as the highest rates of this
form of play were observed during an age interval when no tooth eruption is
occurring (Figure 4.1e). Additionally, the skull and jaw muscles of a spotted
hyena continue to develop even after 32 months of age (Tanner, unpublished
data), whereas rates of non-nutritive chewing drop dramatically after 18 months,
and this behavior is no longer observed after 22 months of age. Although we are

unable to rule out the possibility that non-nutritive chewing might stimulate

124

musculo-skeletal development early in life, its occurrence certainly does not
correspond well with the period of most pronounced change in the morphology of
the feeding apparatus. Another possible function of non-nutritive chewing is that
it allows individuals to explore their immediate environment. Spotted hyenas, like
other cursorial carnivores, are limited in their abilities to manipulate objects with
their forelimbs when compared with primates or even with more dexterous
carnivores such as bears or raccoons (Glickman & Sroges, 1966). Hyenas may
therefore rely heavily on their mouths for gaining information about objects
encountered in their environments. Thus it is not surprising that the rates of this
activity are highest after individuals have moved to a new environment when they

are transferred to the clan’s communal den.

Sex differences in play and their implications

The variation observed here among the different types of play suggests
that the mechanisms underlying their development might also vary. Although we
found that rates of play-mounting were higher in male than female hyena cubs,
rates of social play did not differ between the sexes. By contrast, in many other
mammalian species including humans, males typically engage in higher rates of
both of these forms of play than do females (Table 4.3). In a study of captive
spotted hyenas, Pedersen et al. (1990) found that females engaged in more
social play than males when they were monitored in same-sex groups, but these
investigators observed no sex differences when youngsters were in mixed-sex

groups (Pedersen et al., 1990). The results from the mixed-

125

Table 4.3. Sex differences in play behavior in various mammal species. Female-

domlnated species are indicated in bold font.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Social l Rough- Sex
Sp°°'°s and-Tumble Pla Pla s°”’°°
Spotted hyena _
(wild) M - F M > F Present study
M < F
Spotted hyena (same sex groups) (Pedersen,J., et. al
(captive) = F ---- 1990)
3 mixed sex groups)
3 Domestic cat M = F ____ (Barrett, P. &
E (Fe/is cattus) Bateson, P. 1978)
E Coyotes. wolves and
o Beagles M = F M > F (Bekoff, M., 1974)
(Canis sap. )
Biben M. 1982'
Ferret _ ( . . ,
(Mustela furo) M _ F M > F Sgcgcgliman, et. al
Galapagos fur seal
(Arctocephalus M > F --- $512223: 8‘1 985)
galapagoensis) ’ "
Belding’s ground squirrel M = F M > F (Nunes, S. et. al,
(Spermophilus beldingi) 1999; 2004)
Richardson’s ground
g SQUirrel (Spermophilus M = F M > F (5)::sz T“ J" et' al,
g n’chardsonil)
a: Golden Hamsters M > F ____ (Vieira, M. L. et. al,
Mesocn'cetus auratus) 2005)
Rats -
(Rattus norve icus) M > F ---- (Pellis, S. M., 2002)
10 ' .
3 Cuvner's gazelle (Gaze/Ia (Gomendio, M.,
g cuvien) M > F M > F 1988)
2’
Lambs
:
Ovis aries) M > F M > F (Orgeur, P., 1995)
Japanese monkey (Macaca
M > F ---— (Koyama, N., 1985)
fuscata)
(Goy, R. W. et. al
Rhesus Macaque (Macaca M > F M > F 1988' Wallen K
W F I ‘l
2 mulatta) 1996)
E Ring-tailed lemur (Lemur M = F M > F (Gould L 1990)
E catta ' "
(Pellegrini A. D. &
Humans M > F ____ Smith, P. K., 1998;
(Homo sapiens) Scott & Panksepp,
2003)

 

126

 

sex groups in captivity are most easily comparable to those from the current
study, as both male and female cubs were present at communal dens in the wild.
In a study on another female-dominated species, the ring-tailed lemur (Lemur
catta), Gould (1990) similarly found no sex differences in social play, but did
observe dimorphic sex play (Table 4.3).

The lack of sexual dimorphism in social play is intriguing given the sex-
role reversal seen in both lemurs and spotted hyenas with regard to social
dominance. A popular hypothesis for the function of social play in general, and
rough-and-tumble play in particular, is that it serves as practice for aggressive
encounters in adulthood (Maestripieri & Ross, 2004; Spinka et al., 2001; Smith,
1982). Therefore, the higher rates of rough-and-tumble play observed among the
males of many other mammalian species have been attributed to the male’s
greater need to practice for male-male combat in adulthood (Pellis et al., 1997;
Smith, 1982).

Although overall rates of aggressive behavior are higher among adult
female than male Crocuta (Szykman et al., 2003), both sexes engage in
considerable aggression, particularly in the context of intrasexual interactions
during feeding competition, and when members of neighboring clans encounter
each other. Therefore, the absence of a sex difference in social play within
female-dominated species such as spotted hyenas and ring-tailed lemurs is
consistent with the hypothesis that social play provides practice for agonistic
encounters in adulthood. However, sexual monorphism in social play has also

been found in some other carnivore species studied in captivity, including cats,

127

 

ferrets and various canids, all of which hunt terrestrial vertebrate prey (Table
4.3). Therefore, rough-and-tumble play among young carnivores may provide
practice for skills necessary in pursuing and capturing prey as well as for adult
combat; presumably, development of these skills would be just as important for
females as for males in predatory species (Fagen, 1981).

Copulation represents an unusually difﬁcult challenge for the adult male
spotted hyena (Drea et al., 2002). First, he must be tolerated by the large,
socially dominant and highly aggressive adult female. Then he must insert his
erect phallus into her ﬂaccid one, which is positioned far more anteriorly than is
the vaginal opening in other carnivore species. To achieve successful mating in
adulthood, the male hyena may thus need practice earlier in life, and the
communal den may offer the male his only opportunity for such practice. We
found that male hyena cubs engaged in play-mounting at rates considerably
higher than those exhibited by females (Figure 4.3a). Dimorphic sex play is seen
among juveniles of a variety of mammalian species (Table 4.3), and these sex
differences are promoted by differential exposure to testosterone earlier in
development (Goy, 1996).

Both male and female Crocuta cubs are exposed to androgens late in
gestation (Yalcinkaya et al., 1993; Licht et al., 1998). Although prenatal treatment
with anti-androgens can de-masculinize certain aspects of the genitalia and
nervous system of the spotted hyena (Drea et al., 1998; Drea et al., 2002; Forger
et al., 1996; Fenstemaker et al., 1999), it is not currently known whether, or to

what extent, such treatment also affects the sexually dimorphic patterns of play

128

 

observed here among young hyenas. Three mechanistic hypotheses might
account for the patterns observed here. First, sex differences in play may be
androgen-independent in this species, although this seems unlikely because it
would make spotted hyenas unique among mammals in this regard. Second, it
may be that the critical period for sensitivity of key neural substrates to
testosterone does not occur during prenatal development in this species, but a
instead occurs during the ﬁrst month after birth, when testosterone I
concentrations are higher in male than female cubs (Frank et al., 1991). Third,

there may be a sex difference in prenatal testosterone exposure or receptor

 

Q19
density in this species that has not yet been detected.

The rates of play-mounting observed here were affected by maternal rank
as well as by offspring sex. Specifically, at 2-4 months of age, sons of high-
ranking mothers mounted other cubs at higher rates than did sons of lower-
ranking mothers. Interestingly, this pattern was apparent before the ages at
which maternal rank begins to inﬂuence dominance relations among cubs
(Holekamp & Smale, 1993). Mounting rates among young male Crocuta are
correlated with maternal androgen concentrations late in gestation, and these, in
turn, vary with maternal rank such that sons of high-ranking females are exposed
to higher androgen concentrations in utero than are sons of low-ranking females

(Dloniak et al., 2006).

Conclusions

The patterns of development described here are consistent with the idea

129

that some forms of play confer immediate beneﬁts whereas other forms confer
beneﬁts later in life. Here it appeared that most beneﬁts conferred to hyenas by
romping and non-nutritive chewing were probably largely immediate, whereas
those conferred by play-mounting were most likely delayed until later in life;
furthermore, it appeared that social play might confer both immediate and
delayed beneﬁts. The ﬁve types of play we documented in hyenas pale in
comparison to the broad array of human behaviors that ﬁt the definition of play
adopted here (from Bekoff & Byers 1981). Nevertheless, as appears to be true of
hyena play, some forms of human play may help children cope with speciﬁc
challenges encountered during one or more early stages of development,
whereas others may have long-lasting consequences for individual welfare in
adulthood. Just as social play may help young Crocuta cubs become integrated
into the clan, there also appears to be a relationship between rough-and-tumble
play in male children and measures of social competence (Pellegrini, 1995).
Although it is not known to what extent play reorganizes the individual’s
behavioral phenotype, anecdotal evidence from humans indicates that early play
deprivation is often closely associated with pathological behavior in adulthood
(Brown, 1998). This suggests that play importantly affects structural or functional
aspects of brain development that enhance the individual’s ability to cope
effectively with environmental complexity in general, and with speciﬁc stressors
in particular (Siviy, 1998).

Finally, the patterns observed here highlight the fact that mammalian play

must be heterogeneous in its regulatory mechanisms as well as in its form and

130

 

function. We observed a great deal of variability among the different forms of
hyena play in regard to their patterns of occurrence during development. This
suggests that the genetic and neural substrates regulating expression of one
form of play are not necessarily the same as those regulating expression of other
forms. For example, we found that patterns of temporal variation differed
strikingly between non-sexual social play and play mounting. Furthermore, we
found that one of these two forms of social play, both of which are sexually
dimorphic in many other mammals, is ‘masculinized’ in female spotted hyenas

(non-sexual social play) but the other (play mounting) is not. These data suggest

 

0:"

that genetic mechanisms and neural circuits regulating developmental patterns
differ between these two forms of social play. Although the speciﬁc motor
patterns involved and the exact patterns of sex differences in hyena and human
play are distinctly different, overall it appears that the general principles
concerning the functions of play, as well as the heterogeneous nature of the
mechanisms regulating its development and its expression, may be surprisingly

similar in these two distantly related species.

131

APPENDIX

Table A1. List of MSU skull specimens used in Chapter Two.

 

Catalog Number Age (mos) Catalog Number Age (mos)

 

132

36093 2.00 36558 57.19
36089 4.37 35857 61 .83
36088 4.40 36074 62.40
36087 5.13 36550 63.13
36166 5.47 36168 64.39
36579 6.13 36553 65.27
36086 7.00 36160 70.04
36159 7.00 36570 80.87
36157 7.20 36163 82.97
36075 7.23 36079 83.09
36155 7.53 36569 87.80
36085 8.00 36009 91 .22
35863 8.57 35854 96.57
35861 10.00 35853 96.71
36010 11.13 36165 97.40
36549 13.80 36161 102.50
36090 14.37 35856 107
36578 15.00 897BFT 107.17
36167 15.27 486ECO 113.98
36162 16.73 36083 115.20
36081 16.87 36551 117.50
36577 18.00 225VGS 1 19.85
799HUM 20.53 36566 120.65
35859 22.23 36580 123.65
36576 31 .60 36084 125.23
36556 32.93 36008 128.87
36156 38.58 35852 133.16
36082 39.85 36094 138.36
36164 40.35 36552 140.35
35858 47.33 3601 1 143.83
36581 47.97 36080 148.58
36567 50.83 36571 153.00
36568 53.47 36078 157.53
35855 55.90 36077 199.73

 

Table A2. Description of landmarks for each view used in Chapter Two.

Ventral Landmarks

 

1 Juncture between incisors on the premaxilla

2 Premaxilla-maxilla suture there it intersects the medial edge of the
canine

3 Posterior-most point of the incisive foramen

4 Posterior-most premaxilla-maxialla suture on the palate

5 Anterior-most point on the foramen magnum

6 Palatal foramen

7 Maxilla-palatine midline suture

8 Midline suture between the left and right palatine

9 Maxilla-palatine suture on the posterior edge of the palate

1O Medial edge of the maxilla-jugal suture

11 Lateral edge of the jugal-squamosal suture

12 Anterior lateral edge of the glenoid fossa

13 Medial edge of the glenoid process

14 Medial edge of the jugular foramen

15 Anterior edge of the external auditory meatus

16 Lateral and posterior edge of the basisphenoid

 

Lateral Landmarks

 

Anterior-most upper edge of the l3

Anterior-most upper edge of the canine
Posterior-most upper edge of the canine

Upper edge of the infraorbital foramen

Upper edge of the lacrimal foramen

Uppermost tip of the zygomatic process

Upper edge of the jugal-squamosal suture

Ventral edge of the jugal — squamosal suture
Poster-most edge of the pterygoid

1O Anterior-most upper edge of the external auditory meatus
11 Anterior-most upper edge of the occipital condyle
12 Posterior edge of the nuchal crest

13 *Ray on the dorsal edge of the cranium

14 *Ray on the dorsal edge of the cranium

15 *Ray on the dorsal edge of the cranium

16 *Ray on the dorsal edge of the cranium

17 Tip of the post-orbital process

18 Anterior-most edge of the nasal-premaxilla suture
19 Posterior-most edge of the premaxilla-jugal suture

CDCDNODU'I-bOON—K

Mandible Landmarks

 

1 Anterior edge of l4
2 Anterior edge of the canine

133

Posterior edge of the canine

Upper edge of the mental foramen
Dorsal apex of the curve on the coronoid process?
Posterior edge of the coronoid process
Anterior edge of the mandibular condyle
Posterior edge of the mandibular condyle
Dorsal tip of the articular process
Posterior edge of the mandibular symphysis
Dorsal edge of the masseteric fossa
Ventral apex of the curve of the dentary
* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

* Semi-landmark

134

 

 

LITERATURE CITED

Altmann, J. 1974. Observational study of behavior: sampling methods.
Behaviour, 49, 227-267.

Baldwin, J. D. & Baldwin, J. I. 1976. Effects of Food Ecology on Social Play: A
Laboratory Simulation. 2. Tierpsychol., 40, 1-14.

Barrett, L., Dunbar, R. I. & Dunbar, P. 1992. Environmental inﬂuences on play
behaviour in immature gelada baboons. Animal Behaviour, 44, 111-115.

Bateson, P., Mendl, M. & Feaver, J. 1990. Play in the domestic cat is enhanced
by rationing of the mother during lactation. Animal Behaviour, 40, 514-525.

Bearder, S. K. 1977. Feeding habits of spotted hyaenas in a woodland habitat.
East African Wildlife Journal, 15, 263-280.

Bekoff, M. 1974. Social Play and Play-Soliciting by Infant Canids. American
Zoologist, 14, 323-340.

Bekoff, M. 1984. Social Play Behavior. BioScience, 34, 228-233.

Bekoff, M. & Byers, J. A. 1981. A critical reanalysis of the ontogeny and
phylogeny of mammalian social and locomotor play: an ethological
hornet's nest. ln: Behavioral Development: The Bielefeld Interdisciplinary
Project (Ed. by lmmelmann, K., Barlow, G. W., Petrinovich, L. & Main, M.),
pp. 296-337. Cambridge: Cambridge University Press.

Biknevicius, A. R. 1996. Functional discrimination in the masticatory apparatus of
juvenile and adult cougars (Puma concolor) and spotted hyenas (Crocuta
cmcuta). Canadian Journal of Zoology. 74, 1934-1942.

Biknevicius, A. R. & Ruff, C. B. 1992. The structure of the mandibular corpus and
its relationship to feeding behaviours in extant carnivores. Journal of
Zoology, 228, 479-507.

Biknevicius, A. R. & Van Valkenburgh, B. 1996. Design for killing: craniodental
adaptations of predators. In: Carnivore behavior, ecology, and evolution
(Ed. by Gittleman, J. L.), pp. 393-428. Ithaca: Cornell University Press.

Binder, W. J. 1998. Functional aspects of tooth and jaw development in large

carnivores. ln: Department of Biology. Los Angeles: University of
California.

135

 

Binder, W. J. & Van Valkenburgh, B. 2000. Development of bite strength and
feeding behaviour in juvenile spotted hyenas (Crocuta crocuta). Journal of
Zoology, 252, 273-283.

Bookstein, F. K. 1991. Morphometn'c Tools for Landmark Data. Cambridge:
Cambridge University Press.

Bookstein, F. L. 1996. Combining the tools of geometric morphometrics. In:
Advances in Morphometn'cs (Ed. by Marcus, L. F., Corti, M., Loy, A.,
Naylor, G. J. & Slice, D. E.), pp. 131-151. New York: Plenum Press.

Boydston, E. E. 2001. Socio-ecological determinants of space utilization in the
spotted hyena, Crocuta crocuta. In: Department of Zoology, pp. 172. East
Lansing: Michigan State University.

Brown, S. 1998. Play as an organizing principle: clinical evidence and personal
observations. In: Animal Play (Ed. by Bekoff, M. & Byers, J. A.), pp. 243-
259. Cambridge: Cambridge University Press.

Buckland-Wright, J. C. 1969. Craniological observations on Hyaena and Crocuta
(Mammalia). Journal of Zoology, London, 159, 17-29.

Buckland-Wright, J. C. 1971. The distribution of biting forces in the skulls of dogs
and cats. Journal of Dental Research, 50, 1168-1169.

Buckland-Wright, J. C. 1978. Bone Structure and the Patterns of Force
Transmission in the Cat Skull (Felis catus). Journal of Morphology, 155,
35-62.

Byers, J. A. 1998. Biological effects of locomotor play: Getting into shape, or
something more speciﬁc? In: In Animal Play (Ed. by Bekoff, M. & Byers,
J.), pp. 205-220. Cambridge: Cambridge University Press.

Carrier, D. R. 1996. Ontogenetic limits on locomotion performance. Physiological
Zoology, 69, 467-488.

Christiansen, P. 2007. Evolutionary implications of bite mechanics and feeding
ecology in bears. Journal of Zoology.

Christiansen, P. & Adolfssen, J. S. 2005. Bite forces, canine strength and skull
allometry in carnivores (Mammalia, Carnivora). J. Zool. Land, 266, 133-
151.

Crawley, M. J. 1993. GLIM for Ecologists. Oxford: Blackwell Scientific
Publications.

136

 

Creel, S. & Creel, N. M. 2002. The African wild dog: behavior, ecology, and
conservation. Princeton, New Jersey: Princeton University Press.

Currey, J. D. 2002. Bones: structures and mechanics. Princeton, NJ: Princeton
University Press.

Deane, N. N. 1962. The spotted hyaena Crocuta crocuta crocuta. Lammergeyer,
2, 26-43.

Dessem, D. 1989. Interaction between jaw-muscle recruitment and jaw-joint
forces in Canis familiaris. J. Anat., 164, 101-121.

Dloniak, S. M., French, J. A. & Holekamp, K. E. 2006. Rank-related maternal
effects of androgens on behaviour in wild spotted hyaenas. Nature, 440,
1190-1193.

 

 

Drea, C. M., Hawk, J. E. & Glickman, S. E. 1996. Aggression decreases as play : '
emerges in infant spotted hyaenas: preparation for joining the clan. Animal ”j
Behaviour, 51 , 1323-1336.

Drea, C. M., Hawk, J. E. & Glickman, S. E. 1997. The emergence of afﬁliative
behavior in infant spotted hyenas (Crocuta crocuta). Annals of the New
York Academy of Sciences, 807, 498-500.

Drea, C. M., Place, N. J., Weldele, M. L., Coscia, E. M., Licht, P. & Glickman, S.
E. 2002. Exposure to naturally circulating androgens during foetal life
incurs direct reproductive costs in female spotted hyenas, but is
prerequisite for male mating. Proceedings of the Royal Society of London
Series B, 269, 1981-1987.

Drea, C. M., Weldele, M. L., Forger, N. G., Coscia, E. M., Frank, L. G., Licht, P. 8.
Glickman, S. E. 1998. Androgens and masculinization of genitalia in the
spotted hyaena (Crocuta crocuta). 2 Effects of prenatal anti-androgens.
Journal of Reproduction and Fertility, 11 3, 1 17-127.

Dumont, E. R., Piccirillo, J. & Grosse, l. R. 2005. Finite-Element Analysis of
Biting Behavior and Bone Stress in the Facial Skeletons of Bats. The
Anatomical Record, 283, 319-330.

East, M., Hofer, H. & Turk, A. 1989. Functions of birth dens in spotted hyaenas
(Crocuta crocuta). Journal of Zoology, London, 219, 690-697.

Engh, A. L., Esch, K., Smale, L. & Holekamp, K. E. 2000. Mechanisms of
maternal rank 'inheritance' in the spotted hyaena, Crocuta crocuta. Animal
Behaviour, 60, 323-332.

 

137

 

Engh, A. L., Funk, 8. M., Van Horn, R. C., Scribner, K. T., Bruford, M. W.,
Libants, S., Szykman, M., Smale, L. & Holekamp, K. E. 2002.
Reproductive skew among males in a female-dominated mammalian
society. Behavioral Ecology, 13, 193-200.

Erickson, G. M., Lappin, A. K. & Vliet, K. A. 2003. The ontogeny of bite-force
performance in American alligator (Alligator mississippiensis). J Zool,
Land, 260, 317-327.

Erickson, G. M., Van Kirk, S. D., Su, J., Levenston, M. E., Caler, W. E. & Carter,
D. R. 1996. Bite-force estimation for Tyrannosaurus rex from tooth-marked
bones. Nature, 382, 706-708.

Fagen, R. 1981. Animal Play Behavior. New York: Oxford University Press.

Fenstemaker, S. B., Zup, S. L., Frank, L. G., Glickman, S. E. & Forger, N. G.
1999. A sex difference in the hypothalamus of the spotted hyena. Nature
Neuroscience, 2, 943-945.

Fiorello, C. V. & German, R. Z. 1997. Heterochrony within species: Craniofacial
growth in giant, standard, and dwarf rabbbits. Evolution, 51, 250-261.

Forger, N., Frank, L., Breedlove, M. & Glickman, S. E. 1996. Sexual dimorphism
in perineal muscles and motorneurons in spotted hyenas. . The Journal of
Comparative NeuroIOQY, 375, 333-343.

Frank, L. G. 1986. Social organization of the spotted hyaena (Crocuta crocuta). l.
Demography. Animal Behaviour, 34, 1500-1509.

Frank, L. G. 1997. Evolution of genital masculinization: Why do females have
such a large 'penis'? Trends in Ecology and Evolution, 12, 58-62.

Frank, L. G., Glickman, S. E. & Licht, P. 1991. Fatal sibling aggression, precocial
development, and androgens in neonatal spotted hyaenas. Science, 252,
702-704.

Frank, L. G., Glickman, S. E. & Powch, I. 1990. Sexual dimorphism in the spotted
hyaena (Crocuta crocuta). Joumal of Zoology, London, 221, 308-313.

Frank, L. G., Holekamp, K. E. & Smale, L. 1995. Dominance, demography, and
reproductive success of female spotted hyenas. In: Serengeti II: dynamics,
management, and conservation of an ecosystem (Ed. by Sinclair, A. R. E.
& Arcese, P.), pp. 364-384. Chicago, Illinois: University of Chicago Press.

138

 

 

 

Gaillard, J. M., Pontier, D., Allainé, D., Loison, A., Herve, J. C. & Heizmann, A.
1997. Variation in growth form and precocity at birth in eutherian
mammals. Proc. R. Soc. Lond. B, 264, 859-868.

Gittleman, J. L. 1989. Carnivore Behavior; Ecology and Evolution. Ithaca, New
York: Cornell University Press.

Gittleman, J. L. & Van Valkenburgh, B. 1997. Sexual dimorphism in the canines
and skulls of carnivores: effects of size, phylogeny, and behavioural
ecology. Journal of Zoology, London, 242, 97-117.

Glickman, S. E., Cunha, G. R., Drea, C. M., Conley, A. J. & Place, N. J. 2006.
Mammalian sexual differentiation: lessons from the spotted hyena.
Trends in Endocrinology Metabolism, 17, 349-356.

Glickman, S. E., Short, R. V. & Renfree, M. B. 2005. Sexual differentiation in
three unconventional mammals: spotted hyenas, elephants and tammar
wallabies. Hormones and Behavior, 48, 403-417.

Glickman, S. E. & Sroges, R. W. 1966. Curiosity in 200 Animals. Behaviour, 26,
151 .

Gomendio, M. 1988. The development of different types ofplay in gazelles:
implications for the nature and functions of play. Animal Behaviour, 36,
825-836.

Gorniak, G. C. & Gans, C. 1980. Quantitative Assay of Electromyograms During
Mastication in Domestic Cats (Fe/is catus). Journal of Morphology, 163,
253-281.

Gould, L. 1990. The Social Development of Free-Ranging Infant Lemur catta at
Berenty Reserve, Madagascar. International Journal of Primatology, 11,
297-31 8.

Govindarajulu, P., Hunte, W., Vermeer, L. A. & Horrocks, J. A. 1993. The
Ontogeny of Social Play in a Feral Troop of Vervet Monkeys
(Cercopithecus aethiops sabaeus): The Function of Early Play.
International Journal of Primatology, 14, 701-719.

Goy, R. W. 1996. Patterns of juvenile behavior following early hormonal
interventions. In: The Lifespan Development of Individuals Behavioral,
Neurobiological, and Psychosocial Perspecitives A Synthesis (Ed. by
Magnusson, 0.), pp. 296-314: Cambridge University Press.

139

 

 

 

Goy, R. W., Bercovitch, F. B. 8 McBrair, M. C. 1988. Behavioral Masculinization
ls Independent of Genital Masculinization in Prenatally Androgenized
Female Rhesus Macaques. Hormones and Behavior, 22, 552-571.

Goy, R. W. 8 Wallen, K. 1979. Experimental variables inﬂuencing play, foot-clasp
mounting and adult sexual competence in male rhesus monkeys.
PsychoneuroendocrinologY. 4, 1-12.

Greaves, W. S. 1983. A functional analysis of carnassial biting. Biological Journal
of the Linnean Society, 20, 353-363. .

Greaves, W. S. 1985. The generalized carnivore jaw. Zoological Journal of the
Linnean Society, 85, 267-274.

Grosse, I. R., Dumont, E. R., Coletta, C. 8 Tolleson, A. In Press. Techniques for
modeling muscle-induced forces in ﬁnite element models of skeletal
structures. The Anatomical Record.

Harcourt, R. 1991. Survivorship costs of play in the South American fur seal.
Animal Behaviour, 42, 509-511.

Herrel, A., De Grauw, E. 8 Lemos-Espinal, J. A. 2001. Head Shape and Bite
Performance in Xenosaurid Lizards. Journal of Experimental Biology, 290,
101-107.

Herrel, A. 8 Gibb, A. C. 2006. Ontogeny of performance in vertebrates.
Physiological and Biochemical Zoology, 79, 1-6.

Herrel, A. 8 O'Reilly, J. C. 2006. Ontogenetic scaling of bite force in lizards and
turtles. Physiological and Biochemical Zoology, 79, 31-42.

Herring, S. W. 1985. The ontogeny of mammalian mastication. American
Zoologist, 25, 339-349.

Hildebrand, M. 1984. Analysis of Vertebrate Structure. New York: John Wiley 8
Sons.

Hines, M. 2003. Sex Steroids and Human Behavior: Prenatal Androgen
Exposure and Sex-Typical Play Behavior in Children. Annals of the New
York Academy of Sciences, 1007, 272-282.

Hofer, H. 8 East, M. 1995. Population dynamics, population size, and the
commuting system in Serengeti spotted hyenas. In: Serengeti II:
dynamics, management, and conservation of an ecosystem (Ed. by
Sinclair, A. R. E. 8 Arcese, P.), pp. 332-363. Chicago: University of
Chicago Press.

140

 

 

 

Hofer, H. 8 East, M. L. 1993. The commuting system of Serengeti spotted
hyaenas: how a predator copes with migratory prey. lll. Attendance and
maternal care. Animal Behaviour, 46, 575-589.

Hofer, H. 8 East, M. L. 1996. The components of parental care and their ﬁtness
consequences: a life history perspective. Verhandlungen der Deutschen
Zoologischen Gesellschaﬁ, 89, 149-164.

Holekamp, K. E., Cooper, S. M., Katona, C. l., Berry, N. A., Frank, L. G. 8 Smale,
L. 1997. Patterns of association among female spotted hyenas (Crocuta

crocuta). Journal of Mammalogy. 78, 55-74. '3‘
Holekamp, K. E. 8 Smale, L. 1990. Provisioning and food sharing by lactating

spotted hyenas, Crocuta crocuta (Mammalia: Hyaenidae). Ethology, 86,

191-202.
Holekamp, K. E. 8 Smale, L. 1993. Ontogeny of dominance in free-living spotted J

 

hyaenas: juvenile rank relations with other immature individuals. Animal
Behaviour, 46, 451 -466.

Holekamp, K. E. 8 Smale, L. 1998a. Behavioral development in the spotted
hyena. BioScience, 48, 997-1005.

Holekamp, K. E. 8 Smale, L. 1998b. Dispersal status inﬂuences hormones and
behavior in the male spotted hyena. Hormones and Behavior, 33, 205-
216.

Holekamp, K. E., Smale, L. 8 Szykman, M. 1996. Rank and reproduction in the
female spotted hyaena. Journal of Reproduction and Fertility, 108, 229-
237.

Hurov, J., Henry-Ward, W., Phillips, L. 8 German, R. 1988. Growth Allometry of
Craniomandibular Muscles, Tendons, and Bones in the Laboratory Rat
(Rattus norvegicus): Relationships to Oromotor Maturation and
Biomechanics of feeding. The American Journal of Anatomy, 182, 381-
394.

Irschick, D. J. 2000. Effects of Behaviour and Ontogeny on the Locomotor
Performance of a West Indian Lizard, Anolis lineatopus. Functional
Ecology, 14, 438-444.

Irschick, D. J. 2002. Evolutionary Approaches for Studying Functional

Morphology: Examples from Studies of Performance Capacity. Integrative
and Comparative Bio/093’, 42, 278-290.

141

Joeckel, R. M. 1998. Unique frontal sinuses in fossil and living Hyaenidae
(Mammalia, Carnivora): description and interpretation. Journal of
Vertebrate Paleontology, 18, 627-639.

Karkach, A. S. 2006. Trajectories and models of individual growth. Demographic
Research, 15, 347-400.

Kruuk, H. 1972. The spotted hyena: a study of predation and social behavior.
Chicago, Illinois: University of Chicago Press.

Lauder, G. V. 1981. Form and function: structural analysis in evolutionary
morphology. Paleobiology, 7, 430-442.

Lewis, M. E. 8 Werdelin, L. 2000. The evolution of spotted hyenas (Crocuta
crocuta). I. U. C. N. Hyaena Specialist Group Newsletter, 7, 34-36.

Licht, P., Hayes, T., Tsai, P., Cunha, 6., Kim, H., Golbus, M., Hayward, 8.,
Martin, M. C., Jaffe, R. B. 8 Glickman, S. E. 1998. Androgens and
masculinization of genitalia in the spotted hyaena (Crocuta crocuta). 1.
Urogenital morphology and placental androgen production during fetal life.
Joumal of Reproduction and Fertility, 113, 105-116.

Loranca, A., Torrero, C. 8 Salas, M. 1999. Development of Play Behavior in
Neonatally Undernourished Rats. Physiology & Behavior, 66, 3-10.

Maestripieri, D. 8 Ross, S. R. 2004. Sex differences in play among western
lowland gorilla (Gorilla gorilla gorilla) infants: implications for adult
behavior and social structure. American Journal of Physical Anthropology,
123, 52-61.

Martin, P. 8 Bateson, P. 1985. The inﬂuence of experimentally manipulating a
component of weaning on the development of play in domestic cats.
Animal Behaviour, 33, 511-518.

Maynard Smith, J. 8 Savage, J. G. 1959. The mechanics of mammalian jaws.
School Science Review, 141, 289-301.

McGowan, C. 1998. A Practical Guide to Vertebrate Biomechanics. Cambridge:
Cambridge University Press.

Meers, M. B. 2002. Maximum bite force and prey size of Tyrannosaurus rex and

their relationships to the inference of feeding behavior. Historical Biology,
16, 1-1 2.

142

 

 

Monteiro, L. R., Lessa, L. G. 8 Abe, A. S. 1999. Ontogenetic variation in skull
shape of Thrichomys apereoides (Rodentia: Echimyidae). Journal of
Mammalogy, 80, 102-111.

Negus, V. 1958. The Comparative Anatomy and Physiology of the Nose and
Paranasal Sinuses. Edinburgh: E 8 S Livingstone Ltd.

Novecosky, B. J. 8 Popkin, P. R. W. 2005. Canidae volume bone mineral density
values: an application to sites in western Canada. Journal of
Archaeological Science, 32, 1677-1 690.

Nunes, S., Muecke, E. M., Anthony, J. A. 8 Batterbee, A. S. 1999. Endocrine and
energetic mediation of play behavior in free-living Belding's ground
squirrels. Hormones and Behavior, 36, 153-165.

Nunes, S., Mueke, E. M., Lancaster, L. T., Miller, N. A., Mueller, M. A., Muelhaus,
J. 8 Castro, L. 2004. Functions and consequences of play behaviour in
juvenile Belding's ground squirrels. Animal Behaviour, 68, 27-37.

Ogutu, J. O. 8 Dublin, H. T. 2002. Demography of lions in relation to prey and
habitat in the Maasai Mara National Reserve, Kenya. African Journal of
Ecology, 40, 120-129.

Palagi, E., Cordoni, G. 8 Borgognini Tarli, S. M. 2004. Immediate and Delayed
Beneﬁts of Play Behaviour: new Evidence from Chimpanzees (Pan
troglodytes). Ethology, 1 10, 949-962.

Palagi, E., Paoli, T. 8 Borgognini Tarli, S. M. 2006. Short-Term Beneﬁts of Play
Behavior and Conﬂict Prevention in Pan paniscus. International Journal of
Primatology, 27, 1257-1270.

Paulli, S. 1900. Uber die Pneumaticitat des Schaedels bei den Séugethieren. Ill.
Uber die Morphologie des Siebbeins und die der Pneumaticitat bei den
Insectivoren, Hyracoideen, Chiroptera, Camivoren, Pinnipedien,
Edentaten, Prosimiem und Primaten. Morphologisches Jahrbuch, 28, 483-
564.

Pedersen, J. M., Glickman, S. E., Frank, L. G. 8 Beach, F. A. 1990. Sex
differences in the play behavior of immature spotted hyenas, Crocuta
crocuta. Hormones and Behavior, 24, 403-420.

Pellegrini, A. D. 1995. Boys' rough-and-tumble play and social competence:
Contemporaneous and longitudinal relations. In: The future of play theory:
A multidisciplinary inquiry into the contributions of Brian Sutton-Smith (Ed.
by Pellegrini, A. D.), pp. 107-126. Albany: State University of New York
Press.

143

 

 

Pellegrini, A. D. 8 Smith, P. K. 1998. Physical Activity Play: The Nature and
Function of a Neglected Aspect of Play. Child Development, 69, 577-598.

Pellis, S. M. 2002. Sex Differences in Play Fighting Revisited: Traditional and
Nontraditional Mechanisms of Sexual Differentiation in Rats. Archives of
Sexual Behavior, 31, 17-26.

Pellis, S. M., Field, E. F., Smith, L. K. 8 Pellis, V. C. 1997. Multiple Differences in
the Play Fighting of Male and Female Rats. Implications for the Causes
and Functions of Play. Neuroscience and Biobehavioral Reviews, 21, 105-

 

120. "a
Poirier, F. E. 8 Smith, E. O. 1974. Socializing Functions of Primate Play.
American Zoologist, 14, 275-287.
Power, T. G. 2000. Play and exploration in children and animals. Mahwah: 5" .,
Erlbaum. E j

Pozis-Francois, O., Zahavi, A. 8 Zahavi, A. 2004. Social Play in Arabian
Babblers. Behaviour, 141, 425-450.

Radinsky, L. B. 1981a. Evolution of skull shape in carnivores 2. Additional
modern carnivores. Biological Journal of the Linnean Society, 16, 337-
355.

Radinsky, L. B. 1981b. Evolution of skull shape in carnivores. I. Representative
modern carnivores. Biological Journal of the Linnean Society, 15, 369-
388.

Radinsky, L. B. 1982. Evolution of skull shape in carnivores. 3. The origin and
early radiation of the modern carnivore families. Paleobiology, 8, 177-195.

Rajpurohit, L. S., Sommer, V. 8 Mohnot, S. M. 1995. Wanderers between
harems and bachelor bands: male hanuman langurs (Presbytis entellus)
at Jodhpur in Rajasthan. Behaviour, 132, 255-299.

Rensberger, J. M. 1999. Enamel microstructural specialization in the canine of
the spotted hyena (Crocuta crocuta). Scanning Microscopy, 13, 343-361.

Rensberger, J. M. 8 Wang, X. 2005. Microstructural Reinforcement in the Canine
Enamel of the Hyaenid Crocuta crocuta, the Felid Puma concolor and the
Late Miocene Canid Borophagus secundus. Journal of Mammalian
Evolution, 12, 379-403.

Rohlf, F. J. tpsDig. (Ed. by Department of Ecology and Evolution, S. U. o. N. Y.,
Stony Brook). Stony Brook, NY.

144

Rohlf, F. J. 8 Marcus, L. F. 1993. A revolution in morphometrics. Trends in
Ecology and Evolution, 8, 129-132.

Rohlf, F. J. 8 Slice, D. E. 1990. Extensions of the Procrustes method for the
optimal superimposition of landmarks. Systematic Zoology, 39, 40-59.

Ross, C. F. 2005. Finite Element Analysis in Vertebrate Biomechanics. The
Anatomical Record, 283A, 253-258.

Schmidt, P. A. 8 Mech, L. D. 1997. Wolf pack size and food acquisition. The
American Naturalist, 150, 513-517.

Schweikher, F. P. 1930. Ectocranial suture closure in the hyaenas. American
Journal of Anatomy, 45, 443-460.

Schwenk, K. 8 Wagner, G. P. 2004. The relativism of constraints on phenotypic
evolution. In: Phenotypic Integration: Studying the Ecology and Evolution
of Complex Phenotypes (Ed. by Pigliucci, M. 8 Preston, K.), pp. 390-408.
Oxford: Oxford University Press.

Sharpe, L. L., Clutton-Brock, T. H., Brotherton, N. M., Cameron, E. Z. 8 Cherry,
M. I. 2002. Experimental provisioning increases play in free-ranging
meerkats. Animal Behaviour, 64, 113-121.

Siviy, S. M. 1998. Neurobiological substrates of play behavior: glimpses into the
structure and function of mammalian playfulness. In: Animal Play (Ed. by
Bekoff, M. 8 Byers, J. A.), pp. 221-242. Cambridge: Cambridge University
Press.

Smale, L., Frank, L. G. 8 Holekamp, K. E. 1993. Ontogeny of dominance in free-
living spotted hyaenas: juvenile rank relations with adult females and
immigrant males. Animal Behaviour, 46, 467-477.

Smith, K. K. 1993. The Form of the Feeding Apparatus in Terrestrial Vertebrates:
Studies of Adaptation and Constraint. In: The Skull (Ed. by Hanken, J. 8
Hall, 8. K.), pp. 150-185. Chicago: University of Chicago Press.

Smith, P. K. 1982. Does play matter? Functional and evolutionary aspects of
animal and human play. The Behavioural and Brain Sciences, 5, 139-184.

Spinka, M., Newberry, R. C. 8 Bekoff, M. 2001. Mammlian Play: Training for the
Unexpected. The Quateriy Review of Biology, 76, 141-168.

Statsoft. 2002. STATISTICA (data analysis software system). In: StatSoft. Tulsa,
OK.

145

 

 

 

 

Stefen, C. 1997. Differences in Hunter-Schreger bands of carnivores. In: Tooth
Enamel Microstructure (Ed. by Koenigswald, W. V. 8 Sander, P. M.), pp.
123-136. Rotterdam: Balkema Press.

Strait, D. S., Wang, Q., Dechow, P. C., Ross, C. F., Richmond, B. G., Spencer,
M. A. 8 Patel, 8. A. 2005. Modeling Elastic Properties in Finite-Element
Analysis: How Much Precision ls Needed to Produce an Accurate Model?
The Anatomical Record, 283A, 275—287.

Szykman, M., Engh, A. L., Van Horn, R. C., Boydston, E. E., Scribner, K. T.,
Smale, L. 8 Holekamp, K. E. 2003. Rare male aggression directed toward
females in a female-dominated society: baiting behavior in the spotted
hyena. Aggressive Behavior, 29, 457—474.

Team, R. C. D. 2005. R: A language and environment for statistical computing.
(Ed. by Computing, R. F. f. 8.). Vienna, Austria.

Therrien, F. 2005. Mandibular force proﬁles of extant carnivorans and
implications for the feeding behaviour of extinct predators. J. Zool. Lond.,
267, 249-270.

Thomason, J. J. 1991. Cranial strength in relation to estimated biting forces in
some mammals. Can. J. Zool, 69, 2326-2333.

Thomason, J. J., Russell, A. P. 8 Morgeli, M. 1990. Forces of biting, body size,
and masticatory muscle tension in the opossum Didelphis virginiana. Can.
J. Zool, 68, 318-324.

Thompson, D. A. 1917. On Growth and Form (abridged edition). Cambridge:
Cambridge University Press.

Thompson, E. N., Biknevicius, A. R. 8 German, R. 2003. Ontogeny of feeding
function in the gray short-tailed opossum Monodelphis domestica:
empirical support for the constrained model of jaw biomechanics. The
Joumal of Experimental Biology, 206, 923-932.

Van Horn, R. C., McElhinny, T. L. 8 Holekamp, K. E. 2003. Age estimation and
dispersal in the spotted hyena (Crocuta crocuta). Journal of Mammalogy,
84, 1019-1030.

Van Valkenburgh, B. 1988. Incidence of tooth breakage among large, predatory
mammals. The American Naturalist, 131, 291-302.

Van Valkenburgh, B. 2007. Deja vu: the evolution of feeding morphologies in the
Carnivora. Integrative and Comparative Biology.

146

 

 

Venables, W. N. 8 Ripley, B. D. 1999. Modern Applied Statistics with S. NY, NY:
Spnngen

Verrue, V., Dennaut, L. 8 Verhegghe, B. 2001. Three-dimensional ﬁnite element
modelling of a dog skull for the simulation of initial orthopaedic
displacements. European Journal of Orthodontics, 23, 517-527.

Vincent, S. E., Vincent, P. D., Irschick, D. J. 8 Rossell, J. M. 2006. Do juvenile
gape-limited predators compensate for their small size when feeding?
Journal of Zoology, 268, 279-284.

Wahaj, S. A. 8 Holekamp, K. E. 2006. Functions of sibling aggression in the
spotted hyaena, Crocuta crocuta. Animal Behaviour, 71, 1401-1409.

Wainwright, P. C. 1994. Functional morphology as a tool in ecological research.
In: Ecological Morphology (Ed. by Wainwright, P. C. 8 Reilly, S. M.), pp.
42-59. Chicago: University of Chicago Press.

Wainwright, P. C. 8 Reilly, S. M. 1994. Ecological Morphology. Chicago:
University of Chicago Press.

Wallen, K. 8 Baum, M. J. 2002. Masculinization and Defeminization in Altricial
and Precocial Mammals: Comparative Aspects of Steroid Hormone
Action. Hormones, Brain and Behavior, 4, 385-423.

Wang, Q., Strait, D. S. 8 Dechow, P. C. 2006. A comparison of cortical elastic
properties in the craniofacial skeletons of three primate species and its
relevance to the study of human evolution. Journal of Human Evolution,
51, 375-382.

Watts, H. E. 2007. Social and ecological inﬂuences on survival and reproduction
in the spotted hyena, Crocuta crocuta. East Lansing, MI: Michigan State
University.

Werdelin, L. 1989. Constraint and adaptation in the bone-cracking canid
Osteoborus (Mammalia: Canidae). Paleobiology, 15, 387-401.

Williams, G. C. 1966. Adaptation and Natural Selection. Princeton, NJ: Princeton
University Press.

Wroe, S., McHenry, C. 8 Thomason, J. J. 2004. Bite club: comparative bite force

in big biting mammals and the prediction of predatory behaviour in fossil
taxa. Proc. R. Soc. Lond. B, 1-7.

147

 

 

Yalcinkaya, T. M., Siiteri, P. K., Vigne, J., Licht, P., Pavgi, 8., Frank, L. G. 8
Glickman, S. E. 1993. A mechanism for virilization of female spotted
hyenas in utero. Science, 260, 1929-1931.

Zelditch, M. L., Lundrigan, B. L., Sheets, H. D. 8 Garland, T., Jr. 2003. Do
precocial mammals develop at a faster rate? A comparison of rates of

skull development in Sigmodon fulviventer and Mus musculus domesticus.
Journal of Evolutionary Biology, 16, 708-720.

Zelditch, M. L., Swiderski, D. L., Sheets, H. D. 8 Fink, W. L. 2004. Geometric

Morphometrics for Biologists A Primer. San Diego: Elsevier Academic
Press.

148

mm]

 

1..
.3...
. .
. .
z ..