mmflfu .7 l K. . uvrvh «V Lug
. L a 2"... v
. Rama... 3.3M ‘ . 4... $1...

.1 ..

 

LIBRARY
Michigan State
Universuy

 

 

 

This is to certify that the
dissertation entitled

MORPHOLOGICAL VARIATION IN A DUROPHAGOUS
CARNIVORE, THE SPOTTED HYENA,
CROCUTA CROCUTA

presented by

TERESA LYNN MCELHINNY

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in ZOOLOGY

 

 

[fl 441W W

Major Professor’s Signature

25%/ ’2007
E f

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

-c—0-0-0-o-I-n-v-o-o-u—u—I-.-.-o—u—-—--_.-—o-—o—u—c—~—

. _..-

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IPrq1Aoc&Pres/CIRC/DateDue.indd

MORPHOLOGICAL VARIATION IN A DUROPHAGOUS CARNIVORE,
THE SPOTTED HYENA, CROCUTA CROCUTA

By

Teresa Lynn McElhinny

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Zoology
Program in Ecology, Evolutionary Biology, and Behavior

2009

ABSTRACT

MORPHOLOGICAL VARIATION IN A DUROPHAGOUS CARNIVORE,
THE SPOTTED HYENA, CROCUTA CROCUTA

By

Teresa Lynn McElhinny

Spotted hyenas (Crocuta crocuta) are wide ranging African carnivores.
Across their range, they display high levels of morphological variation and
behavioral lability. In an effort to provide a more accurate description of the
patterns of variation in this species, I examined several aspects of morphology in
spotted hyenas, primarily in the skull. Dental attrition levels were high, as is
expected of a durophagous carnivore, but lOss of the bone-crushing teeth
occurred at lower levels than that reported for wolves.

This species is often cited as exhibiting female-biased sexual size
dimorphism, but empirical data have not supported this assertion. Using large
samples of animals measured in the field, I observed slight but significant female-
biased sexual size dimorphism. Further, power analyses indicated that large
sample sizes are required to detect the phenomenon, but that the differences are
maintained in animals bred in captivity and fed uniform diets. I observed no
sexual dimorphism in skull size or shape, examined using geometric
morphometric methods.

Using geometric a large sample of skulls from across the African
continent, I examined patterns of variation in skull size and shape. Crocuta has

long been held as an example of a species that conforms to Bergmann’s rule.

However, spotted hyena skull centroid size, instead of increasing with distance
from the equator, shows a pattern in which the smallest animals are clustered in
eastern/northeastern Africa. Additionally, the relationship with temperature is
opposite that predicted by Bergmann’s rule; skull centroid size increases with
increasing minimum temperature. Data presented here indicate that prey base
affects body size distribution.

Allometric shape variation is highlighted by localized changes in structures
that serve as origin or insertion points for head and neck muscles. Non-allometric
shape variation exhibits a clinal pattern. Northeastern specimens are
intermediate in shape between north-central and southern African specimens for
all three views. Further work is needed to determine the functional significance of

the geographic variation in shape.

Copyright by
Teresa Lynn McElhinny

2009

To Rosie

ACKNOWLEDGEMENTS

It is a pleasure to acknowledge the mentors, colleagues, and friends who
have contributed to this dissertation. First, I offer thanks to my advisor, Barbara
Lundrigan. Barb’s encouragement, friendship, and above all, patience have been
valuable. I thank her for the opportunity presented to me. I appreciate the
guidance of my dissertation committee, Drs. David Foran, Michael Gottfried, and
Kay Holekamp.

Funding and logistical support for this dissertation were provided by the
College of Natural Sciences, the Graduate School, the Department of Zoology,
and the program in Ecology, Evolutionary Biology, and Behavior. Additional
funding was provided by the American Society of Mammalogists. The
administrative staff of the Department of Zoology is without peer— friendly,
patient, helpful, supportive.

I am indebted to the following museum curators and collections managers
for access to collections under their care: Laura Abraczinskas, Mike Carleton,
Simon Chaplin, Martyn Cooke, Jacques Cuisin, Linda Gordon, Daphne Hills,
Paula Jenkins, Eileen Lacey, Georges Lenglet, Matt Lowe, Barb Lundrigan, Mark
Mandica, Sue McLaren, Ogeto Mwebi, Bill Stanley, LouiSe Thomsett, Michel
Tranier, Win Van Neer, Geraline Veron, and Wm Wendelin.

My dissertation research took me to fabulous cities across the globe, to
spend hours locked away in dusty rooms with old bones. I would not have been
able to afford to stay and work in many of these localities if not for the kindness

and hospitality of family, friends, and friends of friends. Puja Batra, Xavier

vi

Darche, Monique Fowler-Paul Kerman, Peg and Gene McElhinny, the Outram
family, Russ Romeo, Ginger Scarborough, and Corine Vriesendorp offered me
homes away from home during my travels.

Dr. Miriam Zelditch and Dr. Ian Dworkin offered statistical support. Dr. Bilal
Butt was very helpful in procuring ecological data for Africa. Jessica Moy and her
staff at Remote Sensing and Geographic Information Science Research and
Outreach Services georeferenced the ungulate diversity data for me.

I’ve made many good friends at my time here at MSU. Most have moved
away, but cheered me on from afar; Ed Siuda, Kalynn Schulz, Megan Mahoney,
Heather Richardson, and Heather Watts have been great sources of moral
support. Micaela Szykman Gunther and Anne Engh remain two of my closest
friends. I fondly remember our adventures in Kenya and Madagascar, and look
fonivard to new adventures, now with our children in tow. Stephen Thomas is not
only a wise, supportive friend and colleague, but always seemed to have just the
electronic equipment of which I was in dire need.

My friend and colleague Jaime Tanner has been instrumental in my
completion of this degree. We worked side by side through our coursework,
teaching, and comprehensive exams. When she moved on to greener pastures,
she remained one of my most stalwart supporters. I cannot thank her enough for
her advice, friendship, and much-needed kicks in the ass.

My husband, Pat Bills, has acted as my photography assistant, database
guru, statistics sounding board, and artistic consultant. More importantly, he has

helped make this busy time for me an enjoyable one for our daughter, and has

vii

supported me personally in countless ways. This work would not have been
possible without him.

Finally, I thank my daughter, Rosie Bills, for showing wisdom, patience,
and grace far beyond her years as she waited for Mama’s book to be done so

that we would have more time to play. Here I come, baby.

viii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................. x
LIST OF FIGURES ............................................................................................... xii
GENERAL INTRODUCTION ................................................................................. 1
Overview of Chapters ................................................................................. 3
CHAPTER 1
ANOMALIES AND ATTRITION IN THE DENTITION OF SPOTTED HYENAS,
CROCUTA CROCUTA .......................................................................................... 6
Introduction ................................................................................................. 6
Methods .................................................................................................... 10
Results ...................................................................................................... 13
Discussion ................................................................................................ 21
CHAPTER 2

SEXUAL DIMORPHISM IN THE SPOTTED HYENA (CROCUTA CROCUTA), A
REASSESSMENT: ARE FEMALE SPOTTED HYENAS TRULY LARGER THAN

MALES? .............................................................................................................. 30
Introduction ............................................................................................... 30
Methods .................................................................................................... 34
Results ...................................................................................................... 43
Discussion ................................................................................................ 47

CHAPTER 3

SIZE VARIATION IN A WIDE-RANGING CARNIVORE: DO SPOTTED HYENAS

(CROCUTA CROCUTA) CONFORM TO BERGMANN’S RULE? ...................... 49
Introduction ............................................................................................... 49
Methods .................................................................................................... 53
Results ...................................................................................................... 60
Discussion ................................................................................................ 73

CHAPTER 4

ALLOMETRIC AND GEOGRAPHICAL VARIATION IN SKULL SHAPE IN THE

SPOTTED HYENA (CROCUTA CROCUTA) ...................................................... 77
Introduction ............................................................................................... 77
Methods .................................................................................................... 81
Results ...................................................................................................... 87
Discussion ................................................................................................ 96

APPENDIX ........................................................................................................ 105

LITERATURE CITED ........................................................................................ 120

LIST OF TABLES

Table 1.1. Upper tooth fracture in Crocuta crocuta. l=incisors, C=canines,
P=premolars. ....................................................................................................... 14

Table 1.2. Pathological oligodonty in upper teeth of Crocuta crocuta. l=incisors,
C=canines, P=premolars. .................................................................................... 14

Table 1.3. Order of relative frequency of tooth fracture and/or loss in Crocuta,
based on the number of broken teeth/number observed. l=incisors, C=canines,
P=premolars other than P4. ................................................................................. 22

Table 1.4. Rates of complete tooth loss in the upper teeth of spotted hyenas and
wolves (number lost/number observed, multiplied by 100). ................................ 24

Table 2.1. Published accounts of sexual dimorphism, or lack thereof, in Crocuta
crocuta. BL= body length, CBL= skull condylobasal length, HBL= head—body
length, SH= shoulder height, ZB= zygomatic breadth. van Jaarsveld (1988) n=30
total, but numbers of males and females were not reported. Matthews (1939b)
did not perform statistical analyses, but did demonstrate that the median size of
females was larger than that of males. ............................................................... 32

Table 2.2. Sex differences in size and shape of spotted hyena skulls from two
East African populations, Masai Mara National Reserve (MMNR) and
Ngorongoro Conservation Area (NCA), based on a geometric morphometric
analysis. Landmarks for these analyses are as in Figure 2.1. Skull size: mean
centroid size (7 t SE ) for males and females, and Student’s-t results on centroid
size. Skull shape: resampling-based Goodall‘s F-test results on Procrustes
superimposition. .................................................................................................. 43

Table 2.3. Body mass and linear measures of living animals. Raw mean values
for males and females (7 t SE), mass in kg, all other measures in cm. Student’s

t, with Holm sequential Bonferroni adjustment for multiple tests, on natural log-
transformed variables. ......................................................................................... 44

Table 3.1. Geographical body size variation in spotted hyenas (Crocuta crocuta)
(mean 1: standard error, where available). .......................................................... 52

Table 3.2. Pearson’s correlation coefficients describing the relationship between
head-body length and various skull and tooth measures (d.f.=13). All measures
were observed from free-living spotted hyenas in Kenya. M1=first lower molar,
CBL=condylobasal length, CS= centroid size. .................................................... 63

Table 3.3. Pearson’s correlation coefficients describing the relationships between
first lower molar (M1) length and ventral centroid (CS) and latitude, the absolute
value of latitude ([Iatitude] i.e., distance from the equator), longitude, and
minimum temperature. ........................................................................................ 63

Table 3.4. Multiple regression results of ventral centroid size against
precipitation, minimum temperature, ungulate diversity, and the interaction of
minimum temperature and precipitation. ............................................................. 66

Table 4.1. Former proposed subspecies of Crocuta crocuta (Allen, 1939). ....... 79

Table 4.2. Results of regression of shape variables (partial warp +uniform
component scores) on natural log centroid size. ................................................. 88

Table 4.3. Summary of results for PLS analysis of shape with geographical
variables. ............................................................................................................. 89

Table 4.4. Results of regression of standardized shape variables on the first
singular axis for geography of the 2-block partial least squares analyses. ......... 93

Table A1 Catalogue numbers of specimens examined in Chapter One.

*included in cranium analysis, "(included in mandible analysis. ........................ 106
Table A2. Museums visited and abbreviations. ................................................ 110
Table A3. Presence of supernumerary P1 in Crocuta crocuta .......................... 110

Table A4 Specimens used in the study of sexual dimorphism of the spotted
hyena. * excluded from mandible analysis. 1' excluded from ventral analysis... 111

Table A5. Descriptions of landmark locations. ................................................. 112

Table A6. Catalogue numbers of specimens examined in Chapter Three.
*included in ventral centroid size analyses, Tincluded in first lower molar
analyses, :tused for evaluating centroid size, first lower molar length, and
condylobasal length as proxies of body size. .................................................... 114

Table A7. Catalogue numbers of specimens examined in Chapter Four.

*included in ventral analysis, Tincluded in mandible analysis, 1: included in lateral
analysis. ............................................................................................................ 1 17

xi

LIST OF FIGURES

Figure 1.1. Normal dentition of Crocuta crocuta. ................................................... 8

Figure 1.2. Alveolar overgrowth. A. Left first incisor missing. B. Normal
dentition. .............................................................................................................. 11

Figure 1.3. Scatterplot of the number of broken and missing teeth against age in
months of spotted hyenas. Age was calculated as calculated by a mixed-sex
model developed for this species by Van Horn et al. (2003). The line re resents a

least-square fitted linear regression (r2=0.267, F1,405=148.8, p<_2.2x10'1 ). ....... 15
Figure 1.4. Anomalous P1(NMK-OM 2706). ...................................................... 17
Figure 1.5. Supernumerary molar (BM 65.537). Remnant alveolus of M1 is

indicated by the arrow. ........................................................................................ 17
Figure 1.6. Supernumerary P1 (Cambridge 4065). .............................................. 19

Figure 1.7. Supernumerary ‘P3’. A. Supernumerary tooth in left mandible. B.
Normal dentition. Note the difference in post-canine diastema length (MSU
36364). ................................................................................................................ 19

Figure 1.8. Supernumerary M2 (USNM 161909). ................................................ 20

Figure 1.9. The presence of P1 and M2 in recent ancestors of Crocuta. Phylogeny
after Werdelin and Solounias (1991). ................................................................. 27

Figure 2.1. Landmarks (closed circles) and semi-landmarks (open circles) for A)
ventral cranium, B) lateral cranium, and C) lateral mandible views. Numbers on
landmarks correspond to descriptions in Table A5. ........................................... 37

Figure 2.2. Power curves for A) mass, B) girth, and C) head-body length,
generated by sampling t-test P-values for successively larger sub-sample sizes
of measurement data. Simulated power is calculated as the proportion of
samples with t-test p-values less than 0.05. The crossbars indicate the sample
size needed for each sex to attain acceptable power (0.8). ................................ 46

Figure 3.1. Landmarks (closed circles) and semi-landmarks (open circles) for
ventral cranium A), lateral cranium B), and lateral mandible views C). Numbers
on landmarks correspond to descriptions in Table A5 ...................................... 53

Figure 3.2. Map of the distribution of collection localities, overlaid on the scale for

ungulate diversity. Dark areas with high numbers have the highest numbers of
ungulate species. ................................................................................................ 60

xii

Figure 3.3. Geographic distribution of ventral centroid size. Small animals are
indicated by light colored, small dots, and larger animals indicated by dark
colored large dots. The Equator and Greenwich Meridian are indicated by
dashed horizontal and vertical lines, respectively. .............................................. 63

Figure 3.4. Geographic distribution of lower first molar length. Small teeth are
indicated by light colored, small clots, and larger teeth indicated by dark colored
large dots. The Equator and Greenwich Meridian are indicated by dashed
horizontal and vertical lines, respectively. ........................................................... 64

Figure 3.5. Scatterplots of A) centroid size and the absolute value of decimal
latitude and B) centroid size and decimal latitude. ............................................. 66

Figure 3.6. Scatterplots of A) centroid size and decimal longitude and B) centroid
size and mean minimum temperature. ................................................................ 67

Figure 3.7. Scatterplots of A) centroid size and the absolute value of decimal

latitude and B) lower M1 length and decimal latitude. ......................................... 68
Figure 3.8. Scatterplots of A) lower M1 length and decimal longitude and B) lower
M1 length and mean minimum temperature. ...................................................... 69
Figure 3.9. Scatterplot of centroid size and mean precipitation. ......................... 70
Figure 3.10. Scatterplot of centroid size and ungulate diversity. ......................... 71

Figure 4.1. Landmarks (closed circles) and semi-landmarks (open circles) for
ventral cranium A), lateral cranium B), and lateral mandible views C). Numbers
on landmarks correspond to descriptions in Table A5 ...................................... 81

Figure 4.2. Deformation grid showing allometric changes from a linear regression
of shape on log(centroid size) in the ventral view. The landmarks have been
back-reflected, and the deformation exaggerated 2.5 times for ease of
interpretation. Vectors on landmarks in the deformation grid show the direction
and magnitude of change from the smallest to the largest specimens. .............. 87

Figure 4.3. Deformation grid showing allometric changes from a linear regression
of shape on log(centroid size) in the lateral view. The landmarks have been back-
reflected, and the deformation exaggerated 2.5 times for ease of interpretation.
Vectors on landmarks in the deformation grid show the direction and magnitude
of change from the smallest to the largest specimens. ....................................... 88

xiii

Figure 4.4. Deformation grid showing allometric changes from a linear regression
of shape on log(centroid size) in the mandible view. The landmarks have been
back—reflected, and the deformation exaggerated 2.5 times for ease of
interpretation. Vectors on landmarks in the deformation grid show the direction
and magnitude of change from the smallest to the largest specimens. .............. 89

Figure 4.5. Dendrogram produced by a UPGMA cluster analysis based on a
painivise Euclidean distance matrix of the first singular axis for shape from the
ZB-PLS analysis of the ventral view. The symbols correspond to those used for
groups 1 - 3 in Figures 4.9 — 4.10 ....................................................................... 91

Figure 4.6. Plot of the three groups realized by the UPGMA clustering, for the
ventral cranial view. The Equator and the Tropic of Capricorn are indicated by the
upper and lower lines, respectively. ................................................................... 92

Figure 4.7. Canonical variate analysis results for each view on groups 1
(traingles), 2 (circles), and 3 (squares) identified using UPGMA clustering; A)
ventral cranium, B) lateral cranium, and C) lateral mandible. ............................. 94

Figure 4.8. Dendrogram produced by a UPGMA cluster analysis based on a
painivise Euclidean distance matrix of the first singular axis for shape from the
2B-PLS analysis of the ventral view. The symbols correspond to those used for
groups 1 —3 in Figures 4.9—4.10.............. ......................................................... 95

Figure 4.9. Pair-wise deformations of shape change between all groups in the
lateral cranial view; A) 2>1, B) 2>3, C) 1>3, D) landmark map. The deformations
are exaggerated 5 times for ease of interpretation. ............................................ 96

Figure 4.10. Pair-wise deformations of shape change between all groups in the
lateral mandible view; A) 2>1, B) 2>3, C) 1>3, D) landmark map. The
deformations are exaggerated 5 times for A and C, 3 times for B, for ease of
interpretation. ...................................................................................................... 97

xiv

GENERAL INTRODUCTION

Spotted hyenas (Crocuta crocuta), the most abundant large carnivores in
Africa, are of interest not only for their influence on ecosystems and populations
of prey species, but also for their morphological and behavioral plasticity. The
species has a wide geographical range that encompasses much of sub-Saharan
Africa and occupies a variety of different habitat types. Throughout their range,
spotted hyenas exhibit considerable variation in behavior and morphology, with
groups ranging in size from 8-80 animals, and adult body mass ranging from 40-
80 kg. Exploring the underlying evolutionary processes that produce such
variation is central to understanding evolution by speciation (Gould and
Johnston, 1972; Endler, 1977).

There is considerable variation in body size and pelage characteristics in
the spotted hyena, and many subspecies have been described historically
(Meester et al., 1986). Although Matthews (1939b) found significant variation in
pelage color and skull size, he was able to discern no clear geographical pattern
to the variation. He concluded that the previously described subspecies were
based merely on individual variation, and his monospecific description of the
spotted hyena is the systematic convention followed today.

While Matthews noted no geographical regularity to the variation in
spotted hyena size, other authors have described a geographical cline in body
size, with the smallest forms at the equator, and size increasing to the north and
the south (Kurten, 1957; Turner, 1984; Klein, 1986). These studies had a narrow

geographical scope, and the authors focused almost entirely on surrogates for

body size, ignoring variation in skull shape. The observed size variation was
attributed simply to variation in ambient temperature, disregarding other
environmental variables as well and the influence of evolutionary history.

Spotted hyenas are durophagous carnivores; literally, durus (L) hard,
phago (Gr) to eat. They will consume a kill in its entirety, save for the rumen
contents and the boney bases of ungulate horns in larger species (Estes, 1991).
This capability is facilitated by a robust skull and impressive dental battery, both
of which are built to withstand large amounts of pressure from the action of the
adductor muscles used to close the jaws. Despite being a streamlined machine
built for processing bone, there is considerable variation in skull morphology.

In this dissertation, I focus primarily on morphological variation in the
spotted hyena skull, using museum specimens. I present data on tooth loss and
on supernumerary teeth. I examine three issues regarding size variation within
this species, that between sexes, across the continent, and the effect that skull
size has on shape. I assess shape variation across the continent in the context of
past subspecies designations, and the Pleistocene refugium hypothesis, which
has been proposed to explain mitochondrial DNA variation in this species.

I traveled to 13 museums in the United States, United Kingdom, Europe,
and Kenya to examine over 600 Crocuta crocufa skulls. The privilege of studying
such a large sample of any one species allows for direct observation of
morphological variation, but evaluating patterns within such a large group of
specimens presents a challenge. Geometric morphometrics, a group of methods

for the multivariate statistical analysis of form based on landmark coordinates,

allows for easier interpretation of variation within large samples. These methods
facilitate the analysis of patterns in size and shape variation that would be difficult
to extract using traditional linear measurements. The thin plate spline
interpolation method is used to create deformation grids that permit the

visualization of shape change between groups or along gradients.

OVERVIEW OF CHAPTERS

The following chapters have been written as independent manuscripts,
and are presented as such.

In Chapter One, I report on anomalies and levels of attrition that I
observed in my studies of museum specimens. I found supernumerary teeth,
some anomalous, or of developmental disturbance origin, and others that l
determined to be atavistic. Being durophagous carnivores, spotted hyenas often
break their teeth, but are able to survive the loss of one or more teeth. Tooth
breakage and loss occur at higher rates in spotted hyenas compared to many
other carnivores. Canines are the teeth that are broken most often, but they are
rarely completely lost. The assertion that the teeth of spotted hyenas possess
deficient safety factors, that the bone-crushing break frequently compared to
other teeth and to other taxa is re-evaluated. As expected, dental attrition is
significantly correlated with increasing age in this species. The predicted
correlation between high level of competition and dental attrition is not realized,
in fact the opposite relationship is demonstrated, but this may have been due to

small sample size, and a difference in average age of the specimens.

The first step in any analysis of skull variation is to address potential
sexual dimorphism. In Chapter Two, I address this question in spotted hyenas,
using both skull and linear body measures. Perhaps influenced by the unique
female-dominated social system, Crocuta was long assumed to be sexually
dimorphic. In fact, spotted hyenas were described as a robust example of
female-biased sexual size dimorphism in the definitive review paper on the
subject (Rails, 1976); however, reports of field measurements failed to reach a
consensus. Some authors reported that females were larger than males, others
failed to find a significant difference.

I investigate size differences in body measurements taken from wild
animals, and shape and size differences uSing geometric morphometric analyses
of the skull, between male and female spotted hyenas. Females are indeed
larger than males in certain body dimensions, but skulls are sexually isometric in
size and shape. The differences in body measurements are maintained in captive
populations, and therefore are not the result of higher food quality afforded to the
dominant females. The discord in the literature regarding sexual dimorphism in
this species is likely due to the small sample sizes used in the reports. I use a
resampling—based method to illustrate that the slight dimorphism in head-body
length and chest girth require very large sample sizes to show a significant result,
but that only a handful of animals is required for mass.

I examine another long-held hypothesis about spotted hyenas in Chapter
Three- that they conform to Bergmann’s rule. Bergmann’s rule predicts that,

within a species, larger bodied individuals are found at greater distances from the

equator, and thus at lower temperatures. Spotted hyenas were reported to ‘obey’
Bergmann’s rule based on a correlation between latitude and the length of the
first lower molar (Klein, 1986). It was suggested that the relationship between
body size and temperature in this species was so tight, that Crocuta might be
used as an indicator species for predicting paleotemperature in Eurasia (Klein
and Scott, 1989).

Using ventral centroid size as a proxy for body size, I demonstrate that the
pattern in body size variation is not a strictly linear relationship with latitude.
Rather, the smallest individuals are found concentrated in eastern Africa, and
larger animals span the equator outside of this area. This geographic pattern in
body size is not related to temperature, but to the diversity of the prey available.
Additionally, although the length of the first lower molar is significantly correlated
with average minimum temperature, this relationship explains less than 20% of
the variance in first lower molar length, not an ideal level for using the measure to
predict paleotemperature.

Lastly, in Chapter Four I examine allometric shape variation in the skull,
and shape variation across the continent. Shape variation that is the result of
increasing size is related to muscle size and activity. There is also a cline of
shape change from north-central Africa, through northeastern Africa, to the
southern part of the continent that harkens back to the descriptions of subspecies
past. This cline further supports the notion that spotted hyenas were affected by

Pleistocene climate change, and movements were limited by inhospitable habitat.

CHAPTER ONE

ANOMALIES AND ATI'RITION IN THE DENTITION OF SPOTTED HYENAS,
CROCUTA CROCUTA

INTRODUCTION

Each family within the Class Mammalia can be characterized by its dental
formula, the numbers of incisors, canines, premolars, and molars present in the
permanent adult dentition. However, variations in dental formulae can and
regularly do occur (Hall, 1940). Oligodonty, the absence of one or more teeth
within an individual, may be due to congenital, pathological, or traumatic causes.
V\fith congenital oligodonty, the tooth in question never develops. Alternatively, a
normally developed tooth can be lost as the result of disease or injury. Polydonty
is the development of supernumerary teeth, teeth in excess of the number
specified in the species’ dental formula. These departures from the norm are of
interest because they can provide insight into systematics, tooth development,
and genetic heritability of dental traits (D'Souza and Klein, 2007; Hall, 1984;
Manville, 1963).

Many species within the Order Carnivora possess dentitions that are
reduced in number from the ancestral Eutherian condition, and/or possess highly
specialized morphology for processing animal foods. Deviations from the dental
formula, in the form of lost or supernumerary teeth, within species that exhibit
highly modified dental arcades are of particular interest because there is
seemingly little room for error. Breakage, deletions, or duplications of teeth in a

streamlined, highly specialized dental arcade might result in impaired function.

Reports of dental anomalies within the Order Carnivora are well
represented in the literature. Variations in dental formulae have been described
for most of the caniform families, including Canidae (Andersone and Ozolins,
2000; Buchalczyk et al., 1981; Gisburne and Feldhamer, 2005; Nentvichova and
Andera, 2008; Paradiso, 1966; Szuma, 1999; Vila et al., 1993, Mustelidae
(Wolsan, 1989; Wolsan, 1984; Hauer, 2002), Procyonidae (Hall, 1940), and
Ursidae (Hall, 1940). However, fewer studies have investigated dental anomalies
in the feliform carnivores, including true cats, hyenas, mongooses, and civets
(but see (Manville, 1963). Within this group, the bone-cracking hyaenids, which
includes three extant species (brown, striped, and spotted hyenas), are
particularly intriguing due to their robust and highly specialized dentition designed
to split open large bones.

The normal dental formula for the spotted hyena (Crocuta crocuta) is
incisors 3/3 canines 1/1 premolars 4/3 molars 0-1/1 =32-34 (Figure 1.1); this
formula is considerably reduced along the cheek tooth row from the ancestral
carnivoran formula, I 3/3 C 1/1 P 4/4 M 3/3=44. Spotted hyenas do not deliver a
“killing bite’ to suffocate their prey. Instead, a kill is made by biting the prey with
the incisors and canines, and tearing off flesh, until the prey animal is no longer
able to flee. The carnassial teeth (P4/M1) are then used to cut open the hide and
sever connective tissue (Van Valkenburgh, 1996). Spotted hyenas are
durophagous carnivores, able to consume the long bones of even very large
preerhe robust upper and lower third premolars are used to break open large-

diameter bones to obtain the nutritious marrow inside (Kruuk, 1972).

 

Figure 1.1. Normal dentition of Crocuta crocuta.

The aim of this chapter is to document the occurrences of tooth breakage,
oligodonty, and polydonty in Crocuta crocuta. Spotted hyenas are known to
exhibit high levels of tooth breakage relative to other carnivores, which Van
Valkenburgh (1988, 2009) suggested is the result of high stresses placed on the
teeth due to the durophageous diet. Vila et al. (1993) proposed that one could
infer the importance of a tooth to an animal's survival by the rate of complete loss
of the tooth and reossification (resorption) of the alveolus. That is, teeth that are
less important are lost at higher rates than those that are essential to survival.
Vila et al. (1993) found high levels of attrition in wolves at both upper and lower
premolars 1-3, and concluded that while loss at these teeth may increase
processing time of food, prey capture was likely not affected. In order to compare

the findings of Vila et al. (1993) on oligodonty due to trauma and/or disease in

wolves (Canis lupus) with that in spotted hyenas, tooth fracture and complete
loss separately were examined separately.

Complete tooth loss (oligodonty) due to pathological and/or traumatic
causes can be separated from congenital oligodonty, in which the tooth never
forms. As the total number of teeth is already reduced in the spotted hyena from
the ancestral number, and the teeth are highly specialized for the species’
durophageous diet (Van Valkenburgh, 1996), I expected congenital oligodonty to
be low, as has been shown for felids (Hall, 1940).

Fenton et al. (1998) suggested that incidence of tooth damage may be
associated with longer life spans in mammals. Indeed, Van Valkenburgh (1988;
2009) found a trend involving higher rates ’of damaged or missing teeth in older
individuals. I estimated individual ages from tooth wear to examine the
relationship between dental attrition and age in Crocuta. Van Valkenburgh (2009)
hypothesized that the higher level of tooth fracture in Pleistocene carnivores than
in recent carnivores was due to higher levels of intraspecific competition during
the Pleistocene. I tested this hypothesis in Crocuta using specimens from
eastern and southern Africa, which experience high and low intensities of feeding
competition, respectively.

I classified instances of polydonty according to what is currently known

about the etiology of supernumerary teeth.

METHODS

I examined 472 complete skulls (i.e. crania with mandibles), and an
additional 11 mandibles (without crania) and 23 crania (without mandibles), from
adult and subadult spotted hyenas (Table A1) housed in 13 museums in the
United States, Europe, and Kenya (Table A2). These skulls were collected from
across sub-Saharan Africa, although specimens from western Africa were few.
Sex is excluded from descriptions and analyses, except where noted, because
there is no evidence of sexual dimorphism in the dentition (Van Horn et al.,
2003), and sex assignments of museum specimens are not consistently reliable
due to monomorphism of the external genitalia. All statistical tests were
performed in R (R Development Core Team, 2009).

Broken teeth and instances of oligodonty were identified from digital
photographs of the ventral crania. Photographs of the occlusal view of the
mandibles were not available, so fracture and oligodonty data are from crania
only. Oligodonty was classified as being due to trauma and/or disease (hereafter
referred to as pathological oligodonty), or due to congenital absence.
Pathological loss of a tooth during the animal’s lifetime can be identified by the
resultant gap between the adjacent teeth, and the presence of porous, bony
overgrowth of the alveolous (Figure 1.2); both of these traits are absent when
oligodonty is due to congenital absence (Vila et al., 1993). Loss of a tooth post-
mortem is indicated by an open alveolous. In order to utilize skulls with teeth that

were lost post-mortem without skewing results, individual teeth were counted,

10

and the totals of fractured and lost teeth were compared to the total number of

teeth counted (eg. number |1 broken/total number I1 observed).

 

Figure 1.2. Alveolar overgrowth. A. Left first incisor missing. B. Normal dentition.

To examine the relationship between age and dental attrition, | aged 405
adult skulls (see Table A1) using Van Horn et al.’s (2003) mixed-sex age
estimation models (equations 2 and 3) for Crocuta, and regressed the number of
fractured and lost teeth onto age for each specimen (least squares regression).
As spotted hyenas are physiologically competent to breed at 24 months of age
(Glickman et al., 1992), l Included animals whose estimated ages were between
24 and 281 months.

In an attempt to assess the contribution of feeding competition to dental

attrition, I compared skulls (n=108) from eastern Africa, where clan size is very

high (n=47 — 54) (Frank, 1986a; Hofer and East, 1993; Kruuk, 1972), to those in
southern Africa (n=17), where clan size is much smaller (n=4 - 14) and feeding
competition at kills is far less intense (Henschel, 1986; Mills, 1984; Tilson and
Henschel, 1986; Trinkel et al., 2004; Whateley and Brooks, 1978) I performed
ANCOVA on the number of teeth affected by fracture or loss (total of fracture +
loss/individual) by geographic region, with age as a covariate in R version 2.8.1
(R Development Core Team, 2009).

Instances of polydonty, the presence of teeth in excess of those in the
normal dental formula, were recorded when the skulls were photographed.

The peg-like M1 is variable in the dental formula of Crocuta, and has been
previously described as frequently absent'(Lumsden, 1981; Miles and Grigson,
1990), but this has not been systematically studied. M1 is often difficult to see in
photographs of the ventral cranium as the tooth can be obscured by P4.
Therefore, presence or absence of M1 was investigated with the skulls in hand. I
used 74 specimens collected from a single population in what is now the
Ngorongoro Conservation Area, Tanzania (see Table A. 1 ). I considered the
presence of an open alveolus with no bony overgrowth as evidence that the tooth
was lost from the skull postmortem. As these specimens were dissected upon
collection, sex assignments were reliable, and data on sex were included in this
analysis. Age of the hyenas from which the skulls were collected was estimated
using mixed-sex estimation models developed for this species (Van Horn et al.,
2003), and age was tested as a factor in loss of M1 using a Welch t-test for

unequal variances In R.

12

Abnormal tooth rotation has been examined in other carnivores (Gisburne
and Feldhamer, 2005; Szuma, 1999), but is not considered here because
rotation of the cheek teeth is highly variable in this species. Former subspecies
descriptions used rotation as a character (e.g. Crocuta crocuta fortis, Allen,
1924). This practice has been abandoned and these subspecies of Crocuta are

now collapsed (Matthews, 1939b).

RESULTS

Breakage and Oligodonty

Tooth fracture was detected in 29.1 % (144/495) of individuals, with 3.28%
(250/7634) of all teeth broken. Tooth fractUre was most frequent at the canine,
and least frequent at P4 (Table 1.1). I found no instances of congenital tooth loss,
but 11.5% (57/495) of individuals showed evidence of pathological tooth loss
(Table 1.2). All missing teeth were recognized by alveolar resorption and a gap in
the tooth row where the missing tooth had been. Of the 104 pathologically
missing teeth, nearly half were P1. The least pathological loss was seen at P4.

The number of broken or missing teeth was positively correlated with age
(Figure 1.3) (r2=0.267, F1,405=148.8, p<2.2x10'16). There was a significant
difference in dental attrition between eastern and southern Africa. East African
skulls had a mean of 1.79 affected teeth, compared to 2.39 in southern Africa
(ANCOVA: provenance: F=7.22, p=0.045, age F=30.727, p=0.0001). Calculated
mean ages also differed between regions, 124.1 months in eastern Africa, and

144.3 in southern Africa.

13

Table 1.1. Upper tooth fracture in Crocuta crocuta

P=premolars.

. l=incisors, C=canines,

 

P1

P2

P3

Total

 

Number of
fractured
teeth

12

25

61

71

29

17

29

250

 

Total
number
teeth
examined

847

893

877

874

846

987

989

967

7280

 

Percentage
of fractured
teeth

 

 

1 .42

 

2.8

 

6.96

 

8.12

 

3.33

 

1.72

2.93

 

 

0.62

 

3.43

 

Table 1.2. Pathological oligodonty in upper teeth of Crocuta crocuta. l=incisors,
C=canines, P=premolars.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'1 '2 I3 C P1 P2 P3 P4 Total
Number of
teeth lost 12 5 9 51 14 5 3 104
Total
PeiThbe' 847 893 877 874 846 987 989 967 7280
examined
Percentage
ofteeth lost 1.42 0.56 1.03 0.57 6.03 1.42 0.51 0.31 1.43

 

14

 

 

 

14
I
o

N.‘

,_

£0-

0)?—

a:

q—a

'0

.03.

o

3:

C0

0‘—

Eto— 0000 o
g 00 on
:5,

Zv— co. co

 

 

 

 

Age (months)

Figure 1.3. Scatterplot of the number of broken and missing teeth against age in
months of spotted hyenas. Age was calculated as calculated by a mixed-sex
model developed for this species by Van Horn et al. (2003). The line represents a
least-square fitted linear regression (r2=0.267, F1,405=148.8, p<2.2x10‘1 ).

15

Polydonty

Supernumerary teeth were present in the cheek tooth rows of 11 of 472
(2.33%) complete skulls, plus one additional cranium without a mandible. I
observed no polydonty of the upper or lower canines or incisors, and no

instances of polydonty on both the upper and lower jaws of any single individual.

Anomalous P1

Anomalous first upper premolars were present in 4 specimens: unilaterally
in three specimens from Kenya (NMK—OM 2706 and 7761, and USNM 182082),
and one on each side bilaterally in a skull from the Democratic Republic of
Congo (AMH 52068) (Figure 1.4). In all cases, P1 appeared to be abnormally
large. On close inspection, a normal P1 was seen to be fused to a supernumerary
tooth sitting directly anterior, resulting in a bifid crown. The anterior tooth was
smaller than P1, and the extent of separation of the roots of the teeth was
unclear. Without x-ray examination, it was impossible to determine whether the

teeth shared a pulp chamber.

Supernumerary upper molar

A supernumerary tooth is present at the end of the tooth row in the left
maxilla of a specimen from Chilongozi Game Reserve, Zambia (BM 65.537;
Figure 1.5). A remnant alveolus of M1 is present at the posterior edge of the
palate. The supernumerary tooth appears to have three roots, and is situated

lingual to P4, which was lost postmortem and is represented by its alveolus. The

16

 

Figure 1.5. Supernumerary molar (BM 65.537). Remnant alveolus of M1 is
indicated by the arrow.

presence of the supernumerary tooth does not affect the position of P“. This
anomalous tooth has the appearance of an anterio-posteriorly contracted P4, with
a parastyle, paracone, metastyle, and protocone (Figure 1.5). The paracone is
worn, and although there is no mandible deposited with this specimen, it is
assumed that M would have slid between this supernumerary molar and P4 as it

occluded with the paracone and metastyle of P4.

Supernumerary P1
A supernumerary first lower premolar, or its alveolus, is present in 6
specimens (Table A3). When present, the tooth was single-rooted, small, robust,

and column-like, situated just anterior and'lingual to P2 (Figure 1.6).

Supernumerary ‘P3’

A supernumerary premolar is present in the left mandible of a specimen
from the Masai Mara National Reserve, Kenya (MSU 36364), lingual to P3
(Figure 1.7). Due to its position with reference to P3, I refer to this tooth as
supernumerary ‘P3’. The tooth has 2 roots, and is intermediate in size between P2
and P3. Like P3, the tooth possesses a paracone and a metacone, and unlike P2,
it lacks a parastyle. The tooth is in the correct orientation with reference to the
tooth row; that is, it is not a mirror image of P3. P3 is displaced Iabially and
ventrally (Figure 1.7). The post-canine diastema is 0.3 mm, which is shorter than
diastemata of other skulls collected in this geographic region (0.861003, n=32).

There is evidence that this anomalous premolar was present bilaterally during

18

 

Figure 1.6. Supernumerary P1 (Cambridge 4065).

 

Figure 1.7. Supernumerary ‘P3‘. A. Supernumerary tooth in left mandible. B.
Normal dentition. Note the difference in post-canine diastema length (MSU
36364)

19

life, as there is alveolar overgrowth in the same position of the right mandible. P3

displacement and a shortened diastema are also evident in the right mandible.

Supernumerary M2

A left M2 and a right M2 alveolus are present posterior to M1 in a skull from
Kitanga, Kenya (USNM 161909), and a left M2 is present in the mandible of a
skull from the Balbal plains of Tanzania (BM 39.378). The teeth are very small
and each appears to be singly-rooted, with a single, central, pointed cusp (Figure

1.8).

 

Figure 1.8. Supernumerary M2 (USNM 161909).

20

M’- the variant tooth

I examined the presence of the variable upper molar in a sample of
reliably sexed Crocuta skulls from a single collection site in northeastern
Tanzania. M1 is present bilaterally in 59% (44/74), and on at least one side in
85% (63/74) specimens. There is no difference in mean age between hyenas
that possessed at least one M1, and those that did not (Welch t-test for unequal
variances, t=-0.6808, dh16.548, p=0.5054). Skulls from 86% (37/43) of males,
and 83% (26/31) of females possess M1 on at least one side, indicating that there

is no sex bias (Fisher’s Exact test for small sample sizes, p=0.7592).

DISCUSSION
The perils of durophagy ’

Spotted hyenas regularly process and consume large pieces of bone from
their prey. Van Valkenburgh (1996) demonstrated that spotted hyenas utilize all
tooth types (incisors, canines, premolars, and molars) when processing bone,
and invoked this behavior to explain her finding that the rate of tooth breakage
per individual in spotted hyenas is higher than that of other carnivores (Van
Valkenburgh, 1988). In her most recent study Van Valkenburgh (2009) reported
that the specific teeth found fractured in spotted hyenas, in order of decreasing
relative frequency, were canines, premolars other than F“, incisors, and the
carnassials (Table 1.3). In the current study, the results for breakage of upper
teeth were similar, except that I found relatively more broken incisors than

premolars (Tables 1.1 and 1.3). Here, teeth that were completely lost, with

21

Table 1.3. Order of relative frequency of tooth fracture and/or loss in Crocuta,
based on the number of broken teeth/number observed. l=incisors, C=canines,

P=premolars other than F“.

 

Source of affected teeth

Order of relative frequency of affected
teeth

 

 

 

 

 

Broken (Van Valkenburgh, 2009) C=P>I>P4IM1
Broken (this study) C>I>P>P4
Lost (this study) P>|>C>P4
Broken + lost (this study) C>P>I>P4

 

 

 

resorbing or completely overgrown alveoli, in order of decreasing frequency,
were premolars other than P4, incisors, the canine, and P4 (Table 1.3). The most
striking difference between the fracture data (Van Valkenburgh, 2009; this study)
and the resorption data is that, although canines are broken with the highest
relative frequency in spotted hyenas (Table 1.3), the canines show very little
alveolar overgrowth (Table 1.2). That is, canines are most frequently broken but
rarely ever completely lost. This is probably because once the tip of a canine is
broken the tooth is under less stress during prey capture, i.e. if the tooth is still in
use, the bending stress of the broken tooth would be much lower, thus the tooth
would no longer be under the intense pressure that would lead to further
breakdown and eventual loss. Vila, et al. (1993) proposed that inferences could
be made about the importance of a tooth to the survival of the animal based on
the rate of alveolar resporption; teeth that are lost frequently could be considered
to be less important. While there is little doubt that the canine is more important

to prey capture than is P‘, the fact that the former is rarely completely lost, while

22

the latter is frequently so, probably has more to do with the size and placement of

the tooth than the effect of the loss of these teeth on survival.

Comparative analyses of breakage and oligodonty

Most studies of tooth loss in carnivores pool congenital and pathological
oligodonty. I found no congenital oligodonty in spotted hyenas. Keeping in mind
that the present study only considers teeth in the cranium, the 11.5% oligodonty
in Crocuta is still higher than that seen in Vulpes vulpes (2.8%, Szuma, 1999;
8.6%, Gisburne and Feldhamer, 2005), or Canis lupus (3.4%, Buchalczyk et al.,
1981), but lower than that seen in Urocyon cinemargenteus (19.8%, Gisburne
and Feldhamer, 2005). A study that separated pathological from congenital
oligodonty in gray wolves reported 12.4% pathological loss (Vila et al., 1993).

Van Valkenburgh (1988) found that wolves and spotted hyenas are similar
in terms of the overall incidence of tooth fracture, which is elevated compared to
other large carnivores. Comparison of complete tooth loss in these two species,
both of which include bone in their diets, also reveals many similarities (Table
1.4). As in wolves (Vila et al., 1993) tooth loss in spotted hyenas is highest at the
first premolar, and tends to decrease toward the carnassials. Both spotted
hyenas and wolves show low rates of loss at the canine. Another parallel is in the
loss pattern of the incisors. The relative frequency of loss is the same in both
species: the highest rate of loss is seen in the first incisor, followed by the third

and second incisors (I‘>l3>l2).

23

Table 1.4. Rates of complete tooth loss in the upper teeth of spotted hyenas and
wolves (number lost/number observed, multiplied by 100).

 

l1 l2 l3 c P1 P2 P3 P4 M1M2

 

Spotted
hyenas 1.42 0.56 1.03 0.57 6.03 1.42 0.51 0.31 NA NA
thisstudy)
Wolves

(Vilaetal., 1.31 0.49 0.82 0 2.60 1.31 1.47 0.33 0.65 0.98
1993)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Van Valkenburgh (1988) suggested that the high fracture rate in spotted
hyena premolars indicates inferior safety factors within these teeth relative to
those in other predatory carnivores. A safety factor is the ratio of the breaking
stress of a tooth to the estimated maximum stress expected in ordinary use. She
concluded that the durophagous diet of Crocuta regularly places loads on the
premolars that exceed the ability of these teeth to withstand breakage. However,
because Van Valkenburgh (1988) grouped P1'3 together, her premolar fracture
count was undoubtedly strongly influenced by the high rate of breakage at P1.
Carnivore premolars are quite variable in size and shape (e.g., see Figure 1.1).
Rather than lumping premolars into a single group, and comparing spotted hyena
premolars to premolars in other carnivore species, a more appropriate
comparison would be to compare the teeth responsible for processing bone
among species. In the upper dentition of spotted hyenas, the teeth that perform
the majority of the bone-cracking are P3, and to a lesser extent, the protocone of
P4 (Ewer, 1954). Wolves do not crack bones, but rather crush them using the
post-carnassial molars (Werdelin, 1989). Wolves show higher rates of loss in the

bone-processing teeth (Mm) than do spotted hyenas (P34; see Table 1.4). That

24

is, although spotted hyenas do fracture and sometimes lose their bone-cracking
premolars, this loss occurs at a lower rate than that among wolf teeth with the
same function, indicating that the bone-cracking premolar teeth in spotted

hyenas are indeed relatively resistant to attrition.

Effects of age and competition on dental attrition

Tooth fracture should correlate with the age of the animal (Fenton et al.,
1998; Van Valkenburgh, 1988; Van Valkenburgh, 2009). A trend for increased
tooth fracture with increased tooth wear has been described in several large
carnivore species, indicating an association between age and tooth breakage
(Van Valkenburgh, 2009). Here, I demonstrated a significant positive correlation
of dental attrition with age (Figure 1.3).

Rates of tooth fracture may be associated with the level of feeding
competition to which an animal is exposed. Higher rates of tooth fracture are
seen in Pleistocene carnivores, where feeding competition was relatively high,
than in their Recent counterparts where feeding competition is lower, even where
tooth wear indicates that mean ages do not differ (Van Valkenburgh and Hertel,
1993; Van Valkenburgh, 2009). Here I found a relationship between attrition and
competition level, but attrition was higher in southern Africa where feeding
competition is lower, than in eastern Africa, opposite of expectations. Age may
have affected this result, as the skulls from southern Africa were from older

animals, on average. Although the sample size from southern Africa was too

25

small to test the hypothesis, it is possible that spotted hyenas live longer in areas

of lower competition.

Historical significance of polydonty

Although Wozencraft (1989) cited the loss of P1 as a synapomorphy
uniting Felidae and Hyaenidae, the absences of P1 and of M2 are derived
characters within the advanced bone-cracking hyaenids (Werdelin and
Solounias, 1991) (Figure 1.9). During the evolution of the Hyaenidae, P1 is lost
independently in many lineages at the transition from the late Miocene, when
hyaenids evolved from small insectivore/omnivores to carnivorous cursors
(Werdelin and Solounias, 1990; Werdelin and Solounias, 1991 ; Werdelin and
Solounias, 1996)). The loss of M2 likely occurred more than once, and was often
coupled with the loss of P1 (Werdelin and Solounias, 1991).

Where P1 occurs in extant Crocuta (Figure 1.6), it has similar dimensions
and is in a similar position as seen in its sister taxon (Adcrocuta exima), and in
other early bone-cracking hyaenid species (Figure 1.9). In contrast, M2 as seen
in modern Crocuta does not resemble the larger quadrituberculate M2 seen in the
transitional bone-cracking species of hyaenid (lkelohyaena abronia, (Hendey,
1974); Palinhyaena reperta, (Werdelin, 1988); Lyceeaena lycyaenoides, (Qiu,
1987); Belbus beaumonti, (Solounias and de Beaumont, 1981)), the most recent
hyaenids to possess the tooth (Figure 1.9). Therefore, M2 in extant spotted

hyenas offers a glimpse at a transitional form of M2 during the evolutionary loss

26

330.6 $3690

6256 Snoocobcs
mEmEEoE mSooLoAcoml
.26qu 6309605
82:23 mcomxcmcml
memos: mtomsf
$305on mtomxoooq
ecoEamon mzfiom

mEoEm 333083:

mtoqoc mcomxccx‘ml

mace? mciomcoocom I

mace? 9309938 38:63.... I

 

 

 

..ofiwoocm _8=o£oa>I me a...

 

Figure 1.9. The presence of P1 and M2 in recent ancestors of Crocuta. Phylogeny

after Werdelin and Solounias (1991).

27

of this tooth within the bone-cracking hyaenids, suggesting that a reduction in
size over evolutionary time preceded complete loss (Figure 1.8).

Wolsan (1989) warned against classifying supernumerary teeth as
atavisms or anomalies when they may just be polymorphisms within a range of
morphological variation. Atavism is indicated by the apparent reappearance of a
character that was absent in the animal’s recent ancestors, but was present in all
members of an ancestral population or taxon (Hall, 1984). As evolutionary
reduction in tooth number in mammals begins at the level of the first premolar
and the last molar, and moves inward toward the carnassials (Ziegler, 1971),
supernumerary teeth at the anterior and posterior extremes of the cheek tooth
row are more likely to be atavistic than supernumerary teeth within the tooth row.
The argument that a supernumerary tooth be considered atavistic is further
strengthened by the identification of a homologue in related fossil taxa, as seen
for both P1 and M2.

A supernumerary tooth that cannot be classified as a retained deciduous
tooth, or as an atavistic tooth, may have resulted from developmental
disturbance to the tooth bud. Partial splitting of a single bud, or gemination, can
result in a tooth with a single root and pulp chamber, but two crowns. The tooth
looks abnormally large, and the crowns may be completely split, or separated by
a groove or notch (Langlais and Miller, 2002; Tannenbaum and Alling, 1963).
Complete splitting of a single bud, or twinning, results in a supernumerary tooth
that is the mirror image of the normal tooth (Tannenbaum and Alling, 1963). The

four anomalous P1 teeth described here appear to be the result of gemination

28

(Figure 1.4). Although it is unknown whether the two crowns share the same root .
and pulp chamber, they are separated by a groove in each case, resulting in
what appears to be a single tooth that is larger than normal, but that probably
does not seriously influence function.

The supernumerary ‘P3’ (Figure 1.7) is also the result of developmental
disturbance. The tooth has no homologue in the recent evolutionary history of the
bone-cracking hyaenids, and is of similar size and shape as P3. The
supernumerary tooth is not a mirror image of the normal tooth, but this is not a
requirement for gemination according to most authors (see Verstraete, 1985).

Taken together, the data presented in this chapter support the assertion
that a durophagous diet is damaging to an animal’s dentition, but these data also
demonstrate that the teeth of hyenas are up to the challenge. Although the teeth
of spotted hyenas are regularly subject to high forces, the teeth responsible for
the majority of the bone-cracking action, P34, experience less attrition than the
bone processing teeth in wolves. These data would be strengthened by work that
models the forces exerted on the bone-cracking teeth, and the ability of the teeth

to withstand those forces.

As there is little polydonty or oligodonty at the carnassials of carnivores, it
has been suggested that these teeth are evolutionarily conserved e.g. (Gisburne
and Feldhamer, 2005). The data presented here indicate that the bone-cracking
P3 in spotted hyenas, with low levels of loss during life and zero observed

polydonty, is also highly conserved.

29

CHAPTER TWO
SEXUAL DIMORPHISM IN THE SPOTTED HYENA (CROCUTA CROCUTA), A
REASSESSMENT: ARE FEMALE SPOTTED HYENAS TRULY LARGER THAN
MALES?
INTRODUCTION
Sexual dimorphism is defined as any difference between males and
females of the same species, manifested as a physiological, behavioral, or
morphological difference (Glucksmann, 1974). Sexual dimorphism is widespread
among mammals, where it is most commonly expressed as a sex difference in
weapon or body size. In contrast to other vertebrate Classes, sexual dimorphism
in mammals is typically male-biased, such that males are larger and/or have
more robust weapons than conspecific females (Andersson, 1994). The reason
for this male-biased pattern of dimorphism is well established, with sexual
selection shaping and maintaining differences between the sexes (Darwin, 1871;
Trivers, 1972). Specifically, in most mammalian species, males compete directly
or indirectly with other males for mating access to females, and as a result,
males have been favored that possess relatively large bodies or weapons. In
conjunction with their larger size and superior armaments, males are socially

dominant to females in most species of mammals, they have higher priority of

access to food than females, and they are also often more pugnacious than

females (Bouissou, 1983; Trivers, 1972).

There are only a few species of mammals in which females are socially
dominant to males. We define female dominance as the ability of adult females to

win against adult males in contests over resources, and to evoke submissive

30

behavior from males in dyadic contexts. Mammalian species in which females
dominate males include two species of mole-rats (Clarke and Faulkes, 1997),
several species of lemurs (Kappeler, 1993), and spotted hyenas (Crocuta
crocuta) (Kruuk, 1972). In many of these cases, it is not clear whether social
dominance by females is associated with a female-biased sexual dimorphism in
body size (Brett, 1991; Kappeler, 1991). Here we explore this question in spotted
hyenas.

Naturalists have historically experienced substantial difficulty
distinguishing male from female spotted hyenas. Sexual dichromatism is absent,
and the external genitalia of the female are so heavily “masculinized,” that
spotted hyenas were long believed to be hermaphrodites (Glickman, 1995). The
sexual dimorphism in the shape of the glans of the phallus, which allows sex
determination in the field, was not described until 1990 (Frank et al., 1990).
Although Matthews (Matthews, 1939a) unequivocally put the myth of the
hermaphroditic hyena to rest with his detailed description of internal and external
reproductive anatomy in both sexes, distinguishing between living male and

female Crocuta nonetheless remains difficult. The literature contains conflicting

reports regarding whether or not female-biased sexual size dimorphism occurs in
this species, and estimates of the degree of sexual size dimorphism in Crocuta
\Jary considerably for several characteristics, including head and body length,
body mass, girth, skull condylobasal length, and zygomatic arch breadth (Table
2.1).

31

Table 2.1. Published accounts of sexual dimorphism, or lack thereof, in Crocuta
crocuta. BL= body length, CBL= skull condylobasal length, HBL= head-body
length, SH= shoulder height, ZB= zygomatic breadth. van Jaarsveld (1988) n=30
total, but numbers of males and females were not reported. Matthews (1939b)
did not perform statistical analyses, but did demonstrate that the median size of
females was larger than that of males.

 

 

 

 

 

 

 

 

 

 

 

 

 

Females>Males Females=Males # # Locality Reference
males females
HBL CBL, ZB 63 40 Tanzania Matthews
(1939b)
Mass*** 12 8 Tanzania Kruuk
(1972)
BL*, CBL*, Mass, SH 5 5 South Skinner
ZB*** Africa (1976)
Mass, HBL 13 12 South Whateley
Africa . (1980)
Masst 5 5 Kenya Neaves et
aL(1980)
Mass”, girth" HBL 25 18 Kenya Hamilton
et al.
(1986)
Mass*, girth* SH 8 6 South Henschel
Africa @986)
HBL, SH South van
Africa Jaarsveld
et al.
(1988)
Masst, girth* HBL 9 7 South Mills
Africa (1990)
Mass, girth, 5 9 Kenya Sillero—
HBL Zubiri and
Gottelli
(1992b)

 

 

 

 

 

 

"I'P<0.001, ***P<0.01. **P<0-02. *P50—05

32

There are a number of possible explanations for the conflicting data
summarized in Table 2.1. First, whether or not individuals are measured during
the breeding season might contribute to variation in mass and girth in both males
and females. However, Crocuta are aseasonal breeders (Kruuk, 1972; Lindeque
and Skinner, 1982), so we can discount that possibility here. Second, the
discrepancies may be due to small sample sizes and inadequate statistical
power. In those studies reporting female-biased sexual size dimorphism in
Crocuta, the difference between males and females was slight. The PM ratio in
raw body mass ranges from 1.09—1.20 (Hamilton et al., 1986; Kruuk, 1972; Mills,
1990; Neaves et al., 1980; Skinner, 1976; van Jaarsveld et al., 1988) compared
to a mean F/M ratio of 0.62 for the most male-biased mammalian taxa (Weckerly,
1998). As the differences documented between male and female Crocuta have
been slight, and sample sizes in most studies have been small (Table 2.1), larger
sample sizes may be required to demonstrate significant differences between
males and females in some measures. Third, age may play a role if subadult

individuals are measured along with adults. Spotted hyenas exhibit a relatively
slow postnatal growth rate, continuing to develop both cranially (Tanner et al.,
2009) and postcranially (van Jaarsveld et al., 1988) after attainment of sexual
rmaturity. Finally, the disparities may reflect geographic variation in the strength of
sexual size dimorphism in this species. The studies included in Table 2.1' were
undertaken over a range of about 30 degrees in latitude. When sexual
dimorphism is present within other mammalian species, the magnitude of the sex

differences can vary with geography (e.g., raccoons, (Ritke and Kennedy, 1993);

33

bobcats, (Dobson and Wigginton, 1996); Mustela spp, (Rails and Harvey, 1985).
Ralls (1976) suggested that geographic variation in the strength of sexual
dimorphism among Crocuta might be substantial.

The conflicts apparent in earlier published accounts of dimorphism in
Crocuta led Frank (1986a) to conclude “earlier assertions that the female is
larger are not strictly correct” (p. 1524). Our aims here were twofold. The first
was to use large samples of museum specimens to compare male and female
spotted hyenas with respect to cranial size and shape, and use large samples of
living individuals to compare males and females with respect to body mass and
linear body size measures. Our second goal was to evaluate the effect of sample

size on the detection of sexual dimorphism in this species.

METHODS
Skull Samples
Given the difficulties inherent in assigning sex to spotted hyenas, we were

reluctant to trust the sex assignments of specimens housed in museum

collections. Many Crocuta specimens at visited museums were catalogued
\Nithout sex information, and several specimens in the Natural History Museum,
London were labeled “both sex”. Here we used only two large and reliably sexed
groups of specimens. The first, from the Natural History Museum, London,
consisted of specimens collected by L. Harrison Matthews on the OI Balbal plain
i In what is now the Ngorongoro Conservation Area (NCA), northeastern Tanzania

(decimal latitude and longitude, -3.0, 35.467). Matthews (1939a) used dissection

34

to conduct a detailed study of the reproductive tracts of the same animals from
which he collected the skulls measured here, so we are confident that his sex
assignments were correct. For our analyses, we used 75 (44 adult males and 31
adult females) of the 103 specimens that Matthews collected, excluding
specimens that were damaged or not adults (Table A4). To be considered adult,
a specimen had to have complete or nearly complete closure of the lambdoid
and basilar sutures, and tooth wear such that the bilateral mean of the length of
the occlusal surface of the third lower premolar was at least 5.0 mm. Age, as
estimated to +/- 6 months based on tooth wear (Van Horn et al., 2003), did not
differ significantly between the males and females in this sample (t=-0.718,
d.f.=73, P=0.4749).
The second set of specimens was collected by KEH and colleagues in the
Masai Mara National Reserve (MMNR) in southwestern Kenya (-1.417, 34.917);
this set is housed in the collections of the Michigan State University Museum.
This set includes 10 adult males and 22 adult females (Table A4). These
specimens were judged to be adult based on the same criteria morphological
outlined above, and average age of the sexes did not differ significantly in this
sample (l=0.189, d.f.=30, P=0.8511). We kept the NCA and MMNR samples

separate throughout our analyses to avoid the potential confound of geographic

Variation in dimorphism.

35

Image capture and data acquisition

We used two-dimensional geometric morphometrics to evaluate sexual
dimorphism in the spotted hyena skull. The use of geometric morphometrics to
assess sexual dimorphism offers not only powerful methods for looking at size
differences, but also allows evaluation of variation in shape (Hood, 2000; Leigh
and Cheverud, 1991). Photographs of skulls were taken with a Fuji FinePix S1
Pro digital camera fitted with a Nikon AF Nikkor 28-80 mm lens. Images were
saved directly to a laptop computer. A 1 cm scale was included in all
photographs for all views. Three views were captured for each skull: ventral
cranium, lateral cranium, and lateral mandible (Figure 2.1). Views of some NCA
specimens were excluded from analysis because of damage that obscured one
or more landmark locations (see Table A5 for landmark locations, and Table A4
for specimen list).

Landmarks, points presumed homologous across specimens, were
selected so as to provide even coverage, and digitized using tpsDig 2.10 (Rohlf,
2005). Fusion of the sutures associated with the braincase is typically complete
in adult Crocuta (Schweikher, 1930), particularly along the sagittal crest, and

there are few structures that make for suitable homologous landmarks on the
mandible. Therefore, semi-landmarks, points evenly spaced along curves
beginning and ending at homologous landmarks, were used in the lateral
Cranium and mandible views to capture the overall shape of the specimen

( Bookstein, 1997).

36

 

F igure 2.1. Landmarks (closed circles) and semi-landmarks (open circles) for A)
Ventral cranium, B) lateral cranium, and C) lateral mandible views. Numbers on
Iandmarks correspond to descriptions in Table A5

37

In the ventral view (Figure 2.1A), specimens were arranged such that the
palate was parallel to the photographic plane. Landmarks were placed bilaterally
and at the midline, but to avoid inflating degrees of freedom for subsequent
analyses, the coordinates of bilaterally homologous landmarks were reflected
across a baseline extending from landmark 1 to landmark 5 (Figure 2.1A) and
averaged (BigFix, Sheets, 2001), yielding 19 total landmarks. In the lateral
cranium view (Figure 218), specimens were oriented such that the sagittal plane
was parallel to the camera lens. Lateral crania were digitized with 14 landmarks
and 35 semi-landmarks. In the lateral mandible view (Figure 2.1C), specimens
were oriented such that the corpus of the mandible bone was parallel to the
camera lens. Lateral mandibles were digitized with 13 landmarks and 44 semi-

landmarks.

Analyses
Landmark and semi-landmark data were aligned with a Generalized
Least-Squares Procrustes superimposition algorithm using CoordGen (Sheets,
2006) for the ventral cranium view, and SemiLand (Sheets, 2003a) for the lateral
cranium and lateral mandible views. This removes data unrelated to shape, such
as position, scale, and rotation (Rohlf and Slice, 1990; Zelditch et al., 2004).
Procrustes distance was used to quantify the magnitude of sex differences in
Shape, which were tested using a resampling-based Goodall’s F-test with 1600
b()otstraps (Bookstein, 1996; Zelditch et al., 2004). Goodall’s F tests were

Qonducted using TwoGroup (Sheets, 2003c).

38

Our measure of skull size was centroid size, defined as the square root of
the summed distances of the landmark points from the geometric center of the
specimen. Differences between sexes in skull size were evaluated in R (R

Development Core Team, 2009) using Student’s-t tests.

Body measurements

We used three data sets to explore sex differences in body measures of
spotted hyenas. The first consisted of matched head-body length measures from
the 75 (44 male, 31 female) NCA specimens used in the analyses of skull size
and shape. These data were collected from the catalogue and specimen tags at
the Natural History Museum, London. The animals were measured by Matthews
(Matthews, 1939b) post-mortem in the field to the nearest 0.5 in, which we
converted to cm.

The second data set contained body measurements taken from wild
NIMNR animals anaesthetized during a longitudinal behavioral study conducted
by KEH and colleagues. The data included measurements from 182 adults (75

males and 107 females) between the ages of 36 and 130 months, although
sample sizes for some measures are slightly smaller. The mean ages of males
and females did not differ significantly in this sample (t=0.63, d.f.=181,
P=0.5319).
Hyenas were anesthetized with Telazol (6.5 mg/kg; Fort Dodge Animal
Health, Fort Dodge, IA) administered in a dart using a COZ-powered rifle

(Telinject lnc., Agua Dulce, CA). lmmobilizations were carried out according to

39

the guidelines of the American Society of Mammalogists for the use of wild

mammals in research (Gannon et al., 2007), with the approval of the Michigan
State University Institutional Animal Use and Care Committee and the Kenyan
Ministry of Education, Science, and Technology. Total body mass was measured
for each darted hyena with a digital scale to within 0.1 kg. Linear body measures
were taken with a cloth tape measure to within 0.1 cm, along one side of the
body.

Thirty-one morphological measurements were routinely collected from
darted hyenas, but to facilitate comparison of our data with those from previous
studies, we focused on measures that were used most commonly by earlier
workers. In addition to body mass, these included the following linear
measurements: 1) head-body length: distance from the tip of the rhinarium to the
base of the tail, measured along the sagittal midline and vertebral column; 2)
neck circumference: circumference of the neck, measured midway between the
shoulders and the head; 3) girth: circumference of the torso, measured
immediately posterior to the forelegs, with the forelegs perpendicular to the body;
4) shoulder height: distance from the bottom of plantar pad of the forepaw to the
cranial angle of the scapula, measured with foreleg extended perpendicular to
the vertebral column; 5) head circumference: circumference of the head
measured at the widest point of the zygomatic arches; 6) zygomatic to top of

Crest: distance from the widest point of the zygomatic arch dorsally to the sagittal

Crest, measured at a right angle to the sagittal plane; 7) zygomatic to back of

40

crest: distance from the widest point of the zygomatic arch posterio-dorsally to
the external occipital protuberance.

The third data set contained body measurements from captive spotted
hyenas housed at the University of California, Berkeley Field Station for the
Study of Behavior, Ecology, and Reproduction (UCB). These animals were all
born in captivity, bred from animals collected in Narok District, Kenya, near the
MMNR. Measurements were taken from 32 adults (19 males and 13 females)
between the ages of 36 and 130 months. The mean ages of males and females
did not differ significantly in this sample (t= -0.51, d.f.=30, P=0.6124). Each
captive hyena was fed a uniform daily diet of 09-14 kg of Nebraska Brand
standard feline diet (Central Nebraska Packing, Inc., North Platte, Nebraska,
69103). Animals were measured while immobilized with a combination of
ketamine (maximum 10mg/kg; 100mg/ml), xylazine (1mg/kg; 100 mg/ml), and
atropine (0.045 mg/kg; 15 mg/ml) administered via a dart (Telinject Inc., Agua
Dulce, CA) using a blow-pipe (Addison Biological Laboratory, Inc., Fayette, MO).
‘l'otal body mass, head-body length, and shoulder height were measured as
described above. After all measurements were taken, xylazine sedation was
reversed with yohimbine (0.075-0.12 mg/kg intravenously or intramuscularly; 2

mg/ml; Lloyd Incorporated, Shenandoah, IA).

Analyses

All analyses of body measures were performed in R on natural log-

tr a nsformed values. Student’s t-tests were performed on postcranial measures,

41

and P—values were corrected with a Bonferroni sequential step-down adjustment
to control for use of multiple comparisons (Holm, 1979). Postcranial measures
and body mass were analyzed together, separately from cranial measures.
We also conducted power analyses on head-body length, body mass, and
girth of MMNR hyenas to determine the effect of sample size on detection of a
significant sex difference in body size. These variables were selected because
they are commonly presented in published accounts that address sexual size
dimorphism (e.g., Hamilton et al., 1986; Mills, 1990; Sillero-Zubiri and Gottelli,
1992a). For each variable, we simulated power curves by calculating t-test P-
values for successive sample sizes, created using the original measurement
data, from N=5 males and 5 females untilthe simulated power was asymptotic.
For each successive sample size (N), we created 500 independent, random sub-
samples (with replacement) of the original data. For each value of N, power was
calculated as the proportion of those 500 sub-samples with a significant
difference between males and females at (1:005. We a priori selected 80%
power as the minimum a test would need to consistently detect a significant
difference. Our simulations estimate the minimum sample size needed to achieve
80% power. We calculated the mean and confidence intervals of the minimum
sample size from 50 replications of the simulation described above, using the

adjusted bootstrap percentile (BCa) method (Efron, 1987) from the “boot” library

Of the R program (Canty, 1998).

42

RESULTS
Skull Morphology
Males and females from our two populations of East African spotted
hyenas (MMNR and NCA) did not differ significantly with respect to skull centroid
size in the ventral, lateral cranium, or lateral mandible views, although females
tended to be larger in both populations (Table 2.2). Furthermore, we found no
significant sexual dimorphism in skull shape in ventral, lateral cranium, or lateral

mandible views (Table 2.2).

Table 2.2. Sex differences in size and shape of spotted hyena skulls from two
East African populations, Masai Mara National Reserve (MMNR) and
Ngorongoro Conservation Area (NCA), based on a geometric morphometric
analysis. Landmarks for these analyses are as in Figure 2.1. Skull size: mean

centroid size (3? :1: SE ) for males and females, and Student’s-t results on centroid

size. Skull shape: resampling-based Goodall’s F-test results on Procrustes
superimposition.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Skull size Skull shape

View N Sex cenil/rlcfiznsize t‘“ P F P

Lateral :2 “If 2:33:33 0.9330 0.36 0.72 0.70

MMNR Ventral :3 “If $33233 0.8330 0.41 1.04 0.37

L Mandible £2 E 233:1: 1.1930 0.24 0.68 0.65
Lateral :1 “F" 33:23:? 1.2273 0.23 0.89 0.51

NCA Ventral 3? if 3:32:13 1.3671 0.18 0.59 0.86

Mandible :3 “'3 (6:13:33 0.0770 0.94 1.16 0.28

 

 

 

 

 

 

43

Table 2.3. Body mass and linear measures of living animals. Raw mean values
for males and females (3? 1 SE), mass in kg, all other measures in cm. Student’s

t, with Holm sequential Bonferroni adjustment for multiple tests, on natural log-
transformed variables.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Males Females
N X 18E N X 1 SE M/F* tdf. P

Body mass

MMNR 65 51716.4 87 58.4163 0.96 6.51150 4.35x10'9

Berkeley 19 49011.3 13 55612.0 0.96 2.8930 0.0071
Head-body length

NCA 44 118.5109 31 123.1110 0.96 3.3373 0.0014

MMNR 75 12471144 107 12731123 0.98 3.0130 0.0068

Berkeley 19 1272111 13 131.4112 0.97 2.2430 0.0193
Girth

MMNR 74 80519.4 106 83.7181 0.96 4.10178 0.0003
Neck circumference

MMNR 74 48615.7 104 50.4150 0.96 3.81176 0.0005
Shoulder height

MMNR 74 77.9191 106 78.2176 1.00 0.66173 0.5382

Berkeley 19 81.7104 13 82.010.54 1.00 0.5330 0.5996
MMNR Cranial
circufizf'ence 75 50.9159 104 52.4151 0.97 4.47177 4.16x10'5
ZY9°matI°t° 72 12811.5 107 13211.3 0.96 3.26177 0.0027

top crest
Eygmaficm 71 16.9120 107 17311.7 0.98 2.34175 0.0202

back crest

 

 

 

 

 

 

 

 

 

*The ratio of the raw mean male value/ raw mean female value, which is
indicative of the strength of sexual dimorphism for a particular measure. Note
that for mass, to facilitate comparison with other measures, the ratio presented is
that of the cube roots of the means (Ralls, 1976).

44

Linear measures

Females were larger than males for all seven of the body measurements
taken in the field (Table 2.3). In both captive (UCB) and wild populations
(MMNR), female spotted hyenas displayed significantly greater body mass. In
captive animals and in wild populations from Kenya and Tanzania, the head-body
length of females was significantly longer than that in males. Female spotted
hyenas from MMNR had significantly greater neck circumference and chest girth
than males (these measurements were not collected from either the Tanzanian
specimens or captive animals). Neither captive animals nor those from MMNR
showed any significant sex difference in height at shoulder (this measure was not
collected from Tanzanian specimens). MMNR females were also significantly
larger than males in head circumference, and the distances from the widest point
of the zygomatic arch to both the top and to the back of the sagittal crest (Table
2.3). F/M ratio of dimorphism for significant linear measures ranged from 1.02 to
1.04. The only measurement for which we did not observe a significant sex

difference was shoulder height.

Power analyses of body measures

The results of our power analyses indicate that the sample sizes required
to demonstrate female-biased sexual size dimorphism ranges widely, depending
upon the variable in question (Figure 2.2 A—C). Detection of a dimorphism in
body mass requires a mean sample size of only 14.4 individuals for each sex

(95% Cl 14.06-14.66). Girth requires a larger sample size, with a mean of 38.9

45

Simulated Power

Simulated Power

Simulated Power

1.0

(13 (14 (15 (16 (17 (18 (19

LG

 

‘prb

 

 

Oooon

°<3

 

A

 

 

T I I I

40 60 80 100

l

120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I T

50 100
Sample size for each sex

150

46

Figure 2.2. Power curves for A)
mass, B) girth, and C) head-body
length, generated by sampling t-
test P-values for successively
larger sub-sample sizes of
measurement data. Simulated
power is calculated as the
proportion of samples with t-test
p-values less than 0.05. The
crossbars indicate the sample
size needed for each sex to attain
acceptable power (0.8).

individuals of each sex (95% CI 38.22-39.58). Head-body length, which was the
characteristic showing the lowest level of dimorphism (Table 2.3), requires a

mean sample containing 70.9 members of each sex (95% Cl 69.88—71.92).

DISCUSSION

Although our geometric morphometric measurements revealed no sex
differences in skull size or shape among adult spotted hyenas, female-biased
sexual size dimorphism was nonetheless clearly apparent for body
measurements among the animals in our sample. Using data from large numbers
of adults in East Africa, we found that female spotted hyenas are approximately
11% heavier, 4% stouter, and 2-4% longer than males. Although females are not
taller than males at the shoulder, overall females are huskier than their male
counterparts (Table 2.3).

As female spotted hyenas are socially dominant to males, and can
displace males from kills in nature, we might reason that males do not reach
female size for the simple reason that they have inferior access to food.
However, we demonstrated that these female-biased differences in body mass
(12%) and head-body length (3%) are evident even among captive-bred animals
fed a uniform diet, indicating that sex differences seen in free-living animals are
not due to a higher food volume ingested or a higher quality diet afforded to the
socially dominant females (see also Glickman et al., 1992).

That we found no sexual dimorphism in skull size or height at the shoulder

was initially surprising, given the significant differences in other measurements.

47

However, it may be that males and females are skeletally isometric, but sex
differences in head-body length and girth are explained by differences in muscle
or fat mass. Despite finding no difference in skull size, we did observe a
significant female bias in head measures from live animals, highlighting the mass
of the temporalis and masseter muscles in living hyenas; this too supports the
notion that female bias is caused by larger muscle and/or fat mass.

Here we were able to avoid all of the confounding factors that may have
been responsible for conflicting results in earlier studies comparing male and
female hyenas with respect to body size, including small sample sizes,
geographic variation, or inclusion of subadult animals in samples. We could rule
out these possibilities here by using large, age-matched samples from eastern
Africa.

Our power analyses demonstrated that large sample sizes are required to
achieve statistical significance in female-biased sexual size dimorphism in the
spotted hyena. The mean total sample size required to elucidate sexual size
dimorphism varied substantially among three morphological measures obtained
from living hyenas (Figure 2.2 A—C). Of the previous studies of sexual size
dimorphism in this species (Table 2.1), only one reached the required sample
size for the variable measured (Hamilton et al., 1986), and demonstrated female-
biased sexual size dimorphism for body mass. The results of the power analyses
highlight the importance of adequate sample sizes required in to detect a

dimorphism in taxa where differences are slight.

48

CHAPTER THREE

SIZE VARIATION IN A WIDE-RANGING CARNIVORE: DO SPOTTED HYENAS
(CROCUTA CROCUTA) CONFORM TO BERGMANN’S RULE?

INTRODUCTION

Ecogeographical rules seek to describe biogeographic patterns of
morphological variation within and among species. Among endotherms, rules
have been proposed to explain geographic variation in body size (Bergmann,
1847; Foster, 1964), coloration (Gloger, 1833), and relative appendage size
(Allen, 1877). The explanatory power of these rules has been debated in the past
(e.g. (Geist, 1987)), but a recent resurgence of interest has been driven by the
potential usefulness of these rules in predicting animal responses to global
climate change (e.g. Meiri et al., 2009; Mi‘llien et al., 2006; Teplitsky et al., 2008).
Of particular interest is Bergmann’s rule, which predicts that, within genera,
species with larger body sizes are found at higher latitudes, and at therefore
cooler temperatures (Bergmann, 1847). Mayr (1954) and James (1970) revised
the rule to address intraspecific body size clines that vary with temperature.

Early studies inquired whether or not certain species show a latitudinal
cline in body size (eg. Barnett, 1977; Brown and Lee, 1969). While some authors
found Bergmann’s rule a suitable construct to describe patterns of variation in
body size, arguments arose regarding the physiological basis of body size
increase in response to decreasing ambient temperature. Scholander (1955)
pointed out that insulation from fur and vascular modulation were more effective
means to control heat balance than changes in body size, and that the increases

in body size observed were negligible in terms of physiological change. Geist

49

(1987) argued that in order for ambient temperature to be a valid driving force for
evolution in body size, changes in body size would have to be nearly exponential
to compensate for temperature changes.

Today, Bergmann’s rule is regarded as an empirical generalization, and
species that show a latitudinal size cline are thought to be responding to some
environmental variable or set of variables for which latitude acts as a proxy. What
exactly that variable or set of variables may be is open to interpretation. Many
environmental variables are intercorrelated, and resolving their contributions and
interactions can be challenging (Vtfigginton and Dobson, 1999). Most treatments
of Bergmann’s rule address two broad relationships: 1) a negative correlation of
body size with temperature, measured as- wet bulb temperature (James, 1970),
temperature maxima (Yom-Tov and Geffen, 2006), or mimima (Castro et al.,
1992), and 2) a positive correlation with habitat productivity, measured as actual
evapotranspiration (Rosenzweig, 1968), seasonality (Boyce, 1978), or duration of
the annual productivity pulse (Geist, 1987).

Klein (1986) proposed that carnivores are more likely to ‘obey’
Bergmann’s rule (defined as response to temperature) than herbivores, as
carnivores are removed from direct reliance on herbaceous food supplies, which
are affected by temperature variation. McNab (1971) postulated that prey
availability may drive latitudinal body size clines among carnivores, and that this
may involve not only distribution of prey, but competition with other sympatric
carnivorous species. Meiri et al. (2007) found that variation in brown bear (Ursus

arctos) body size correlates with salmon availability, indicating that spatial

50

distribution of prey affects body size in this species. Raia and Meiri (Raia and
Meiri, 2006) also suggested that prey size might affect variation in carnivore body
size.

Species with broad geographical ranges can be used to test the
hypotheses that body size varies with latitude or specific ecological variables that
might covary to some extent with latitude (e.g., Fuentes and Jaksic, 1979; Gay
and Best, 1996; Kennedy and Lindsay, 1984; Sikes and Kennedy, 1992). The
spotted hyena, Crocuta crocuta, is an excellent candidate for such analyses.
Although now fragmented and contracted, the historical range of spotted hyenas
extended across most of sub-Saharan Africa to the southern tip of the continent,
a range of about 55 degrees in latitude (Kingdon, 1977). Reports of mean body
size measures indicate that southern African specimens are larger than those
from eastern Africa (Table 3.1). Kruuk (1972) also observed variation in body
size, supporting the eastern Africa-southern Africa size dichotomy, and further
noted that skull size was larger in specimens from Uganda than from Tanzania.
Klein (1986) reported that extant Crocuta conform to Bergmann’s rule south of
the equator, and attributed the observed size cline to temperature-related
vafiafion.

Geographic variation in body size in Crocuta is also reported in the
paleontological literature. Kurten (1956) compared size in extant (South Africa,
Tanzania, Uganda, and Somalia) and fossil (Europe and Syria- late Pleistocene)
spotted hyenas. He described a cline in size that increased with distance from

the equator, both to the north and to the south, which he also attributed to

51

Bergmann’s rule. Similarly, Baryshnikov (1999) invoked Bergmann’s rule to

explain a northerly increase in size among Pleistocene Crocuta of Russia.

Table 3.1. Geographical body size variation in spotted hyenas (Crocuta crocuta)
(mean :1: standard error, where available).

 

 

 

 

 

 

Locality Mass (kg) Head-32:1); length Source

figzgflegark 50212.97 125 7 (N=14) Sillero-Zubiri and

Kenya (N=14) Gottelli (1992b)

3:323:33 Hamilton et al.
51.6 (N=43) 126,114.70 (N=43) (1986);

Resewe' Frank (1986b)

Kenya

Hluhluwe and

Umfolozi Game

Reserves, 68.4 (N=23) 133.3 (N=25) Whateley (1980)

Natal, South -

Africa

Transvaal, _ _ .

South Africa 58,412.18 (N-9) 130810.09 (N-9) Skinner (1976)

 

 

 

 

 

Several authors have also found body size variation in fossil specimens
from the same latitude, but deposited during different geologic time periods.
Crocuta from glacial periods of Pleistocene Europe are significantly larger than
animals found in interglacial sediments from similar latitudes (Klein and Scott,
1989; Kurten and Poulianos, 1977). A similar relationship between temperature
and Crocuta body size is found in Holocene specimens from South Africa (Klein
and Scott, 1989), and Pleistocene specimens from Russia (Baryshnikov, 1999).
The variation in body size through geologic time is attributed to climate variation,

such that larger body size in Crocuta is associated with colder ambient

52

temperatures (Baryshnikov, 1999; Klein and Scott, 1989; Kurten and Poulianos,
1977)

Observed body size variation in extant Crocuta, coupled with
paleontological evidence, indicates that this species may indeed conform to
Bergmann’s rule. The paleontological data indicate that a latitudinal cline in body
size within the species may correlate with a cline in ambient temperature. In fact,
temperature and body size seem to be so tightly correlated, that Klein and Scott
(1989) suggested that the size of fossil spotted hyenas could be used to predict
paleotemperature. The purpose of this chapter is to document and describe the
pattern of size variation in extant Crocuta on both sides of the equator, and
examine the relationship between size variation and selected environmental
variables. First I determine the relevance of centroid size, condylobasal length,
and the length of the first lower molar as proxies for body size. I then evaluate
the relationship of body size variation with that of latitude, longitude, minimum
temperature, precipitation, and ungulate diversity. Finally, I re-evaluate the notion

that spotted hyenas conform to Bergmann’s rule.

METHODS
Body size data are often not recorded with large mammal specimens
deposited in museums, so it is necessary to use a proxy measure to examine
body size variation. Traditionally, the condylobasal length of the skull (CBL), the
distance from the anterior edge of the premaxillae to the posterior facets of the

occipital condyles, has been used to estimate body size. When skulls are

53

incomplete, as is often the case for fossils, the length of the first lower molar (M1
length) is used. With the advent of geometric morphometric approaches to the
study of skull morphology, authors have used centroid size (CS) as a proxy for
body size (e.g., Cardini et al., 2007; Frost et al., 2003). Centroid size, the size
metric used in geometric morphometrics, is the square root of the summed
distances of the landmark points from the geometric center of the specimen.

To determine which metric would be the best proxy for body size in
spotted hyenas, I measured CBL, M1 length, and ventral cranial, lateral cranial,
and lateral mandible CS for 338 adult spotted hyena skulls from museums in the
US, Europe, and Kenya (sample sizes for some measures are lower due to
incomplete or broken specimens, see Table A6). Photographs were taken with a
Fuji F inePix S1 Pro digital camera fitted with a Nikon AF Nikkor 28—80 mm lens.
Images were saved directly to a laptop computer. A 1 cm scale was included in
all photographs for all views. M1 length and CBL were measured from digital
photographs of the left lateral mandible and the ventral cranium, respectively,
using TMorthen (Sheets, 2002), which calculates traditional linear
measurements with reference to the ruler in the photograph.

Three views were captured for each skull: ventral cranium, lateral cranium,
and lateral mandible (Figure 3.1 ). CS was calculated for all three views.
Landmarks were selected so as to provide even coverage (Table A5), and
digitized using tpsDig 2.10 (Rohlf, 2005). Fusion of the sutures associated with
the braincase is complete in adult Crocuta (Schweikher, 1930), especially along

the sagittal crest, and there are few structures that make for suitable homologous

54

 

Figure 3.1. Landmarks (closed circles) and semi-landmarks (open circles) for
ventral cranium (A), lateral cranium (B), and lateral mandible views (C). Numbers
on landmarks correspond to descriptions in Table A5

55

landmarks on the mandible. Therefore, semi-landmarks, points evenly spaced
along curves beginning and ending at homologous landmarks, were used in the
lateral cranium and mandible views to capture the overall shape of the specimen
(Bookstein, 1997).

In cranium ventral view photographs, specimens were arranged so that
the palate was parallel to the photographic plane. Landmarks were placed
bilaterally and at the midline, but to avoid inflating degrees of freedom for
subsequent analyses, the coordinates of bilaterally homologous landmarks were
reflected and averaged across a baseline extending from landmark 1 to landmark
5 (Figure 3.1A), yielding 19 total landmarks (BigFix, Sheets, 2001).

In cranium lateral view photographs, the left side of the specimens was
oriented such that the mid-sagittal plane was parallel to the photographic plane. If
the left side of the specimen was not usable due to damage, the right side was
photographed, and the photograph was digitally reflected. Lateral crania were
digitized with 14 landmarks and 10 semi-landmarks (Figure 3.1 B).

In mandible photographs, specimens were arranged so the left horizontal
ramus was parallel to the photographic plane. As for the lateral cranium, if the left
mandible was not usable due to damage, the right side was photographed, and
the photograph was digitally reflected. Lateral mandibles were digitized with 13
landmarks and 16 semi-landmarks (Figure 3.1C).

Landmark and semi-landmark data were aligned with a Generalized

Least-Squares Procrustes superimposition algorithm using CoordGen (Sheets,

56

2006) for the ventral cranium view, and SemiLand (Sheets, 2003a) for the lateral
cranium and lateral mandible views. Both of these programs calculate CS.

To determine the appropriateness of M1 length, CBL, and CS as proxy
measures for body size in Crocuta, I examined pairwise correlations between
these measures and head-body length in animals for which I had both the skull
and body measurements that were taken while the animals were alive. Body
measurements were taken during routine immobilization as part of a long-term
behavioral study for 14 adult spotted hyenas (Table A6) from the Masai Mara
National Reserve in Kenya, and their skulls were deposited at the Michigan State
University Museum post-mortem. Head-body length was used as the body size
estimate because it is a more accurate metric of body size in this species than is
body mass, which is highly variable due to variation in recent meal size (Kruuk,
1972)

Female-biased sexual size dimorphism is present in this species, but it is
not evident in either the skull (Chapter 2 of this dissertation) or the dentition (Van
Horn et al., 2003). Because of this, and the fact that sex assignments for many
museum specimens of Crocuta are dubious due to the male-like genitalia of the

females, sex is excluded from these analyses.

Geographical and environmental variables
Conventional latitude and longitude coordinates recorded for specimen
collection sites were converted to decimal latitude and longitude using the

Federal Communications Commission’s online converter

57

(http://www.fcc.gov/mb/audio/bickel/DDDMMSS-decimal.html). For specimens
with place-name or landmark collection information, I determined the decimal
latitude and longitude from the National Geospatial-lntelligence Agency’s
GEOnet Names Server (GNS, http://earth-info.nga.mil/gns/html/index.html). For
specimens that had headings recorded (e.g. “80 mi N of Maun”), I took the
decimal latitude and longitude from GNS, and calculated the coordinates given
the heading with the Mammal Networked Information System (MANIS)
Georeferencing Calculator (http://manisnet.org/gci2.html). Both decimal latitude
and the absolute value of decimal latitude were used in analyses. Since the
range of the spotted hyena straddles the equator, the absolute value of latitude,
or distance from the equator, is a more appropriate measure than latitude in
evaluating Bergmann’s rule.

Georeferenced climate data for Africa were based on the Climatic
Research Unit (CRU) TS 2.1 global climatic dataset, produced by the Climatic
Research Unit of University of East Anglia. The CRU T8 2.1 Global Climate
Dataset is comprised of 1224 monthly time-series of climate variables, for the
period 1901-2002, covering the global land surface, excluding Antarctica, at 0.5
degrees resolution. I used a version of these data available from the Consortium
for Spatial Information of the Consultative Group on International Agricultural
Research (CGIAR-CSI), which has been reformatted for use in ArcGIS Grid
format in decimal degrees using the World Geodetic System 1984 datum
(WG884) (Mitchell and Jones, 2005). This is the same system as was used for

determining coordinates for the specimens. Annual means were calculated from

58

each monthly variable from 1901—1960, and a grand mean was calculated over
the 60 years. Data from 1961 and later were not used. Climate values were
assigned to specimens based on the 0.5 degree grid in which the specimen was
located, by overlaying specimen locations on to the CGIAR-CSI grid file using
ArcMap.

As the clinal relationship of body size with temperature predicted by
Bergmann’s rule has been related to extremes of temperature, especially
extreme cold, I used mean minimum temperature as the ambient temperature
variable. Exploratory analyses of other ambient temperature variables available
through the Global Climate Dataset, daily mean temperature, and maximum
temperature revealed high correlations with mean minimum temperature (mean
temperature: r=0.97, t=49.74, df=154, p<2.2x10'16; maximum temperature:
r=0.87, t=21.52, df=154, p<2.2x10'16). '

In addition to temperature, studies of Bergmann’s rule also traditionally
examine water availability. Later studies have incorporated this variable into a
measure of primary productivity, usually a combination of temperature and water
availability (e.g. Le Houerou, 1984). However, since the effect of primary
productivity would most likely be realized indirectly via the prey ingested by
spotted hyenas, here I use a simpler measure of water availability, that of mean
precipitation.

I chose the number of ungulate species present in an area (ungulate
diversity) as a proxy for food availability. I used a distributional map of ungulate

diversity for 89 species of ungulates ranging in body size from Kirk’s dik-dik

59

(Modoqua kirkir) to the giraffe (Giraffa camelopardalis) (after Turpie and Crowe,
1994). The ungulate diversity cline map was digitized and georeferenced for use
in ArcMAP based on the WGS84 datum (Figure 3.2). Ungulate diversity values
were assigned to specimens according to the cline in which the specimen’s
collection locality was spatially contained, and were extracted using ArcMap's
Spatial Analyst extension. Prey diversity varied from 0 to 35 in five-unit
increments (Figure 3.2).

As latitude and longitude are not meaningful ecological variables, I did not
use them in the multiple regression model; rather, I present correlations of both
M1 length and ventral CS with latitude and longitude. As the spotted hyena is
suggested as an indicator species for predicting temperature (Klein and Scott,
1989), I also present correlations of both M1 length and ventral CS with minimum
temperature. I calculated Pearson’s correlations in R (R Development Core
Team, 2009). I used multiple regression in R to evaluate the effect of minimum
temperature, precipitation, ungulate diversity, and the interaction between

minimum temperature and precipitation on ventral CS.

RESULTS
CBL, ventral CS, and lateral CS all showed significant positive correlations
with head-body length in 14 specimens from Kenya; M1 length and mandible CS
were positively correlated with head-body length, but not significantly so (Table
3.2). Ventral CS explained the most variation in head-body length, and was thus

used in analyses to investigate the pattern of body size. Although CBL has long

60

 

Figure 3.2. Map of the distribution of collection localities, overlaid on the scale for
ungulate diversity. Dark areas with high numbers have the highest numbers of
ungulate species.

61

been the traditional measure for analyses of body size in carnivores, this
measure was highly correlated with ventral CS across all specimens (r=0.983,
t=91.882, d.f.=294, p<2.2x10'16), and therefore did not warrant further
investigation. M1 length was not significantly correlated with head-body length in
this subset of 14 animals, but M1 length was significantly correlated with ventral
CS in the full data set (r=0.438, t=8.204, d.f.=284, p=8.216x10‘15). As M1 length
has been used in several studies that invoke Bergmann’s rule for fossil and
extant Crocuta, the geographical pattern of this measure is also described below.
Ventral CS exhibits a clear geographical pattern that is easily visualized,
with the smallest animals concentrated in eastern Africa (Figure 3.3). The pattern
of M1 length is similar, but this measure shows a greater tendency to increase
with distance from the equator (Table 3.3, Figure 3.4). Both M1 length and ventral
CS exhibit a significant positive relationships with the absolute value of latitude
and mean minimum temperature, and a significant negative relationship with both
latitude and longitude (Table 3.3, Figures 3.5A—B, 3.6A-B, 3.7A—B, 3.8A—B).
Multiple regression of ventral CS on minimum temperature, precipitation,
ungulate diversity, and the interaction between minimum temperature and
precipitation was also highly significant, with all environmental variables tested
making significant contributions to the model (Figures 3.8A—B, 3.9, 3.10).
Ungulate diversity and the interaction between minimum temperature and
precipitation made the most significant contributions to the model (R2=0.41,

F4,319=54.23, p<2.2x10"6, Table 3.4).

62

Table 3.2. Pearson’s correlation coefficients describing the relationship between
head-body length and various skull and tooth measures (d.f.=13). All measures
were observed from free-living spotted hyenas in Kenya. M1=first lower molar,
CBL=condylobasal length, CS= centroid size.

 

 

 

 

 

 

Measure r t p

M length 0.26 0.96 0.36
CBL 0.66 2.04 0.01
Ventral CS 0.68 3.32 0.01
Lateral CS 0.57 2.50 0.03
Mandible CS 0.49 3.16 0.06

 

 

 

 

 

Table 3.3. Pearson's correlation coefficients describing the relationships between
first lower molar (M1) length and ventral centroid (CS) and latitude, the absolute
value of latitude ([Iatitude] i.e., distance from the equator), longitude, and
minimum temperature.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r tor. P
M length
[Latitude] 0.461 9.215314 <2.2x10'16
Latitude -0223 4051314 6.431x10‘5
Longitude -0234 4265314 2.646x10‘5
@3322ng 0.168 3.020314 0.003
Ventral CS
[Latitude] (1452 9.083322 <2.2x10‘16
Latitude -0225 4149322 4.286x10’5
Longitude -0543 -1 1 .602322 <2.2x10'16
twat/122:: 0.349 6.687322 1.010x10'10

 

63

 

 

       

 

Ventral Centroid Size I

Smaller :

O
O Larger

 

 

 

Figure 3.3. Geographic distribution of ventral centroid size. Small animals are
indicated by light colored, small dots, and larger animals indicated by dark
colored large dots. The Equator and Greenwich Meridian are indicated by
dashed horizontal and vertical lines, respectively.

64

 

   

M1 Length

Smaller

 

o ..
O Larger

 

Figure 3.4. Geographic distribution of lower first molar length. Small teeth are
indicated by light colored, small dots, and larger teeth indicated by dark colored
large dots. The Equator and Greenwich Meridian are indicated by dashed
horizontal and vertical lines, respectively.

65

Table 3.4. Multiple regression results of ventral centroid size against
precipitation, minimum temperature, ungulate diversity, and the interaction of
minimum temperature and precipitation.

 

 

 

 

 

 

 

 

 

Flt p
Model 54234, 319 < 2.2x10'16
Precipitation -2.67 0.01
Minimum _2 32 0 02
temperature ' '
Ungulate ~16
diversity -9.84 < 2x10
Minimum
temperature: 3 49 0 001
precipitation ' ‘
interaction

 

 

66

 

Centroid size

 

 

 

 

I l l l l f T
0 5 1 0 1 5 20 25 30

[Decimal latitude]

 

320

Centroid size

300

280

 

 

 

 

-30 —20 -—1 0 0 10

Decimal latitude

Figure 3.5. Scatterplots of A) centroid size and the absolute value of decimal
latitude and B) centroid size and decimal latitude.

67

 

>
340
I

320
l

Centroid size

280
I

 

 

 

 

l l l l I
0 1 0 20 30 40

Decimal longitude

 

DJ
Centroid size
320 340

300

280

 

 

 

 

l T l
10 15 20

Mean minimum temperature (C)

F iQUre 3.6. Scatterplots of A) centroid size and decimal longitude and B) centroid
size and mean minimum temperature.

68

 

Lower M1 length

 

22

 

 

 

I I I I I I I
0 5 1O 1 5 20 25 30

[Decimal latitude]

 

Lower M1 length

 

22
l

 

 

 

l l l I l
-30 —20 -10 0 1O

Decimal latitude

Figure 3.7. Scatterplots of A) centroid size and the absolute value of decimal
latitude and B) lower M1 length and decimal latitude.

69

 

 

 

 

 

 

 

 

 

 

A.
.C
‘61
C
.9
E
'5
3
O
_J
O o
N _
N
o
I I r T l I
-10 0 10 20 30 40
Decimal longitude
B.
.C
‘61
C
2
E
6
3
O
..J
o o
N _
N
o
F l l
10 15 20

Mean minimum temperature (C)

Figure 3.8. Scatterplots of A) lower M1 length and decimal longitude and B) lower
M1 length and mean minimum temperature.

70

 

340
J

 

 

 

 

 

o
N—
on

an

.21
(D
:9

9

E8-

0:0

0
o
m—
or

O

 

 

 

I I l r | I I
20 40 6O 80 1 00 1 20 1 40

Mean precipitation (mm)

Figure 3.9. Scatterplot of centroid size and mean precipitation.

71

Centroid size

Figure 3.10. Scatterplot of centroid size and ungulate diversity.

340

320

300

280

 

 
 

 

CXDO

   

OCIDOGD CD ("I

O O OO OOCEJO 00M. W00) 0

 

O
O
O 3
. O
l l l l I l l
0 5 10 15 20 25 35

Ungulate diversity (# of species)

72

 

DISCUSSION

The traditional interpretation of Bergmann’s rule is that, within a species, a
decrease in ambient temperature drives an increase in body size with increasing
latitude. That is, animals get larger in the cold. Based on the work of Klein (1986,
Klein and Scott, 1989), spotted hyenas have long been considered an example
of a carnivore species that ‘obeys’ Bergmann’s rule. As with Klein’s work with M1
length, l have shown here that there is indeed a correlation between ventral CS
and latitude; spotted hyenas tend to get larger with increasing latitude (Figure
3.5A—B). Thus, at first blush, it would seem that spotted hyenas do obey
Bergmann’s rule, using latitude as a proxy for temperature. There are, however,
two caveats to this conclusion.

The first is that the correlation between size and latitude is not based on a
cline of increasing size from the equator, as predicted by Bergmann’s rule. The
significance of this relationship is driven by the cluster of the smallest animals
east of 33°E longitude between 5°S and 10°N latitude, and this belies the true
geographical pattern of body size (Figure 3.2). In fact, longitude actually explains
more body size variation than latitude in this species (r=-0.54 vs. r=0.45 for
[latitude], Table 3.3 Figure 3.6A), but we would certainly not conclude that there
is a longitudinal cline in body size. Simply stating that spotted hyenas conform to
Bergmann’s rule based on the correlation of size with latitude is a
misrepresentation of the actual geographical pattern.

The second caveat is that while the relationship with latitude is relatively

strong (r=0.45 for [latitude]), the relationship with temperature is less so (r=0.35),

73

and it is in the opposite direction from that predicted by Bergmann’s rule. That is,
in spotted hyenas, animals tend to get smaller in the cold (Table 3.3 and Figure
3.68). Thus, the complex geographical pattern in body size variation in the
spotted hyena, based on the proxy of ventral CS, is not best described by
invoking Bergmann’s rule.

CBL and M1 length are often used as proxy measures for body size in
mammals. M1 length is used when complete specimens are unavailable, and is
particularly useful for fossil studies. Here, I have shown that ventral CS is a more
appropriate proxy for head-body length for spotted hyenas than either CBL or M
length (Table 3.2), and CS should therefore be used when available.

Based on extant and fossil data, it was suggested that the variation in
spotted hyena M1 length was driven by temperature (Klein, 1986; Klein and
Scott, 1989). In the current study, M1 length shows similar patterns to those seen
in ventral CS (Figures 3.7A—B, 3.8A-B) and is indeed significantly correlated with
average temperature, but the relationship is weakly positive (Figure 3.88). Again,
this runs counter to the predicted relationship with temperature based on
Bergmann’s rule; some of the largest spotted hyenas hail from the warmest
climates. Based on these data from extant spotted hyenas, it would be imprudent
to use Crocuta as an indicator of paleotemperature. Dayan, et al (1991) urged
caution when using dental data to infer paleoclimate, suggesting that competition
may be a stronger forcethan climate on evolution of the dental arcade. That

indeed may be the case here.

74

Meiri et al. (2007) proposed that some aspect of prey base drives
conformation to Bergmann’s rule in carnivores. In spotted hyenas, ungulate
diversity was highly significant among environmental variables examined, such
that smaller animals were found in areas of higher prey diversity (Figure 3.10).
Why are larger spotted hyenas found in hot areas with low prey diversity? The
most ecological information is available for populations in eastern and southern
Africa. On average, animals in the southern part of the continent have much
larger home ranges and lower prey density than their conspecifics in eastern
Africa. Larger home ranges and lower prey density would mean travelling farther
to defend the territory and to find food. McNab (1963) found a positive
association between territory size and body mass in mammals. Thus, it would
seem that body size variation in this species has more to do with socioecological
factors.

Over 15 years after Geist (1987) deemed Bergmann’s rule invalid, citing
spurious correlations between body size and temperature, Meiri, et al. (2003)
pronounced the rule sound based on meta-analysis of previously published tests
of the rule, including the data from spotted hyenas (Klein, 1986). However, if
meta-analyses such as these are based on specious correlations, then the
results of these analyses may themselves be misleading. Bergmann’s rule was a
useful construct for describing patterns of variation early in the 20th century, but
today, ecological data beyond latitude are available. A more direct path is to

describe patterns of variation within species, and test hypotheses posed to

75

explain those patterns, rather than try to fit the data to an outdated empirical

generalization.

76

CHAPTER FOUR

ALLOMETRIC AND GEOGRAPHICAL VARIATION IN SKULL SHAPE IN THE
SPOTTED HYENA (CROCUTA CROCUTA) '

INTRODUCTION

For more than a century, studies of geographical variation within and
between taxa have sought to advance our understanding of the mechanisms of
evolution. As Gould and Johnston (1972) stated in their review article on the
subject, "the foundation of most evolutionary theory rests upon inferences drawn
from geographic variation or upon the verification of predictions made about it"
(p. 457). Exploring the underlying evolutionary processes that produce
morphological and behavioral variation within species is central to understanding
evolution by speciation (Endler, 1977; Gould and Johnston, 1972).

Geographical variation in size is relatively easy to study, but because it is
multivariate, shape presents more of a challenge. Traditional studies of shape
using linear morphological measures are difficult to interpret because the
description of shape is up to the investigator. Also, evaluating shape
independently of size is difficult for classical morphologists, as linear shape
measurements are confounded by size. Newer approaches, such as geometric
morphometrics, permit assessment of shape independent of size and shape, use
powerful multivariate statistical methods to test for shape variation, and illustrate
shape change as a deformation grid. These methods have proven particularly
useful for quantifying geographical variation in skull shape, especially with
respect to phylogeography (e.g., Cardini and Elton, 2009; D'Anatro and Lessa,

2006; Frost et al., 2003).

77

For a number of reasons, the spotted hyena (Crocuta crocuta) is an
excellent subject for a study that addresses phenotypic variation with geography.
The species has a wide geographical range that encompasses much of sub-
Saharan Africa, and includes a variety of different habitat types (Kingdon, 1977).
The behavioral ecology of the species is well-studied (e.g., Hofer and East, 1995;
Holekamp and Smale, 1991; Kruuk, 1972), although most workers have focused
on animals inhabiting the eastern and southern parts of the species’ range.
Spotted hyenas are remarkably successful predators, and throughout their range
they exhibit considerable variation in behavior, with groups ranging in size from
5—90 animals (Kruuk, 1972; Mills, 1990).

Because of the wide range of variation in body size (40-86 kg, (Kingdon,
1977) and pelage coloration, many forms of extant Crocuta were described and
given scientific names, resulting in a long list of subspecies containing many
likely synonyms. Between the years 1777 and 1924, no fewer than 21 spotted
hyena subspecies were described in the literature (Meester et al., 1986). Most of
these taxa seemed to have been established on the basis of pelage
characteristics (eg. Heller, 1910), although some researchers also considered
differences in body size, and/or skull morphology. Among the skull characteristics
used to distinguish taxa were size, shape of the posterior edge of the palate,
degree of inflation of the auditory bullae, shape of the silhouette of the braincase,
and breadth of the palate and of the skull overall (Allen, 1924; Cabrera, 1911;

Matschie, 1900). In a re-evaluation of the many subspecific descriptions, Allen

78

Table 4.1. Former proposed subspecies of Crocuta crocuta (Allen, 1924).

 

Subspecies Locality Synonymous taxa

Hyaena maculata Thunberg
Hyaena rufa Desmarest
Crocuta crocuta . Hyaena capensis Desmarest
crocuta Erxleben SOUth Africa Hyaena encn'ta Smith
Hyaena wissmanni Matschie
Hyaena gan'epensis Matschie

 

 

C. c. fisi Heller Northern Kenya

 

. Democratic Republic
C. c. fortrs Allen of the Congo

 

Crocotta kibonotensis Lonnberg
C. c. genninans Tanzania, southern Crocotta panganensis Lonnberg
Matschie Kenya, Malawi Crocuta nzoyae Cabrera
Crocuta nyasae Cabrera

 

 

 

 

 

C. c. . . . Hyaena (Crocuta) Ieontiewi Satunin
habessynica 52:3: Somalia, Croctua rufopicta Cabrera
Blainville g Crocuta thomasi Cabrera

C. c. theirryi Hyaena togoensis Matschie
Matschie Togo, Cameroon Hyaena noltei Matschie

 

 

(1924). revised the taxonomy of the species to contain six subspecies based on
morphological variation and geographical locality (Table 4.1), but he noted that
additional study was needed of the species across its range. Allen (1924)
suggested restricting the subspecies of Crocuta to Crocuta crocuta crocuta in
South Africa, C.c. fortis in the Democratic Republic of the Congo, C.c. theinyi in
western Africa, C.c. gerrninans in Malawi, Tanzania, and southern Kenya, C.c.
fisi in northern Kenya, and Co. habessynica in Uganda, Ethiopia and Somalia.
Matthews (1939b) later set out to systematically quantify morphological variation
within the species, by documenting variation in pelage color, body size, and skull
Characters in the approximately 170 specimens of spotted hyenas then at the

Natural History Museum, London. He observed considerable variation in this

79

sample, but was not able to discern a clear geographical pattern, and concluded
that the previously described subspecies were invalid. His monospecific
designation of extant Crocuta crocuta is the systematic convention still followed
today (e.g., Wozencraft, 2005).

Recent molecular studies have attempted to clarify the evolutionary history
of spotted hyenas and other African carnivores. A 2005 study of the spotted
hyena (Rohland et al., 2005) revealed significant geographical structuring of
cytochrome b sequences, with southern and northern clades that overlap at the
equator. A southern/eastern distribution, with varying levels of admixture, is also
seen among African wild dogs (Girman et al., 2001), lions (Dubach et al., 2005),
and cheetahs (Freeman et al., 2001). The extent to which morphological variation
corresponds to these phylogeographical patterns is unknown.

In this chapter, I assess the contribution of skull size to variation in skull
shape in the spotted hyena, and examine the shape change associated with size.
I determine the extent to which geographical variables explain the pattern of
shape variation. l re-evaluate earlier taxonomic work on the species in light of the
pattern of shape variation revealed with geometric morhometric methods. Finally,
I evaluate the Pleistocene refugia hypothesis as a means to describe the current

pattern in shape variation.

80

METHODS
Skull samples

I examined 385 skulls of adult Crocuta crocuta with associated locality
data housed in 13 museums in the United States, Europe, and Kenya (Tables
A2 and A7). To be considered as an adult in this study, a specimen had to have
complete or nearly complete closure of the lambdoid and basilar sutures, and
tooth wear such that the bilateral mean of the length of the occlusal surface of
the third lower premolar (P3) was at least 5.0 mm. P3 is a large tooth, dominated
by a robust, central, cone-shaped cusp. The length of the occlusal surface of P3
is commonly used as a variable in equations for estimating the age of individual
adults of this species (Van Horn et al., 2003). Using equation 3 from Van Horn et
al. (2003) a P3 occlusal surface length of 5.0 mm results in an estimated age of
38.37 months, which is over one year beyond the age of reproductive maturity
(Glickman et al., 1992). In a recent study, Tanner et al. (2009) demonstrated that
skull maturation is complete by 35 months of age, suggesting that the above
guidelines for classifying a skull as fully adult are conservative. As we found no
significant sexual dimorphism in skull centroid size or in skull shape in Crocuta
(Chapter 2 of this dissertation), sexes were pooled for these analyses.

Where available, latitude and longitude were recorded from the collection
information at the museum; in each case, I converted the conventional
coordinates to decimal latitude and longitude using the Federal Communications
Commission’s online converter (http://www.fcc.gov/mb/audio/bickellDDDMMSS-

decimal.html). For specimens with place-name or landmark collection

81

information, I determined the decimal latitude and longitude from the National
Geospatial-lntelligence Agency’s GEOnet Names Server (GNS, httpzllearth-
info.nga.mil/gnslhtmlflndex.html). For specimens that had headings recorded

‘ (eg. “80 mi N of Maun”), I took the decimal latitude and longitude from GNS, and
calculated the coordinates given the heading with the Mammal Networked
Information System (MANIS) Georeferencing Calculator
(httpzllmanisnet.org/gci2.html). Some specimens were excluded from the

analysis of one or more views due to damage (see Table A7).

Data collection

Photographs were taken with a Fuji FinePix 81 Pro digital camera fitted
with a Nikon AF Nikkor 28-80 mm lens. Images were saved directly to a laptop
computer. A 1 cm scale was included in all photographs for all views. Three
views were captured: ventral cranium, lateral cranium, and lateral mandible
(Figure 4.1).

Landmarks were digitized at suture intersections and other presumed
homologous points using tpsDig 2.10 (Rohlf, 2006). Fusion of the sutures
associated with the braincase is mostly complete in adult Crocuta (Schweikher,
1930), especially along the sagittal crest, and there are few points that make
suitable homologous landmarks on the mandible. Therefore, semi-landmarks,
points evenly spaced along curves beginning and ending at homologous

landmarks, were used in the lateral cranium and mandible views to capture the

82

overall shape of the specimen (see Table A5 for a list and description of
landmarks).

In cranium ventral view photographs, specimens were arranged so that
the palate was parallel to the photographic plane. Landmarks were placed
bilaterally and at the midline, but to avoid inflating degrees of freedom for
subsequent analyses, the coordinates of bilaterally homologous landmarks were
reflected and averaged, yielding 19 total landmarks (BigFix, Sheets, 2001, Figure
4.1A).

ln cranium lateral view photographs, the left side of each specimen was
oriented such that the mid-sagittal plane was parallel to the photographic plane. If
the left side of the specimen was not usable due to damage, the right side was
photographed, and the photograph was digitally reflected. Lateral crania were
digitized with 14 landmarks and 10 semi-landmarks (Figure 4.1B).

In mandible photographs, specimens were arranged so the left horizontal
ramus was parallel to the photographic plane. As for the lateral cranium, if the left
mandible was not usable due to damage, the right side was photographed, and
the photograph was digitally reflected. Lateral mandibles were digitized with 13
landmarks and 16 semi-landmarks (Figure 4.10).

Landmark and semi-landmark data were aligned using a Generalized
Least-Squares Procrustes superimposition algorithm (GPA). This removes data
unrelated to shape, such as position, scale, and rotation (Rohlf and Slice, 1990;
Zelditch et al., 2004). Superimposition was performed with CoordGen (Sheets,

2006) for the ventral view, and SemiLand (Sheets, 2003a) for the lateral and

83

 

Figure 4.1. Landmarks (closed circles) and semi-landmarks (open circles) for
ventral cranium A), lateral cranium B), and lateral mandible views C). Numbers
on landmarks correspond to descriptions in Table A5.

84

mandible views. Procrustes distances were used as shape variables in the
statistical tests described below, and the differences were visualized with

deformation grids generated using the thin-plate spline algorithm.

Data analysis

Allometry

While GPA separates shape from geometric scale, shape variation that is
correlated with size (i.e. allometric shape variation) remains. Skull size varies
geographically in Crocuta, with the smallest skulls found in eastern Africa
(Chapter Three of this dissertation). To examine the influence of size on shape,
for each view, I performed a linear regression of shape on log centroid size to
illustrate the changes in shape from small to large skulls (Regress6n, Sheets,
2008).

I then standardized the data for size, using Standard6 (Sheets, 2003b),
which regresses shape on ln(centroid size) and calculates residuals. These
standardized data, which allow for analysis on shape independent of allometric
variation, were used for all subsequent analyses. In addition, in an effort to clean
up the geographical signal by removing localized individual variation, I took
means of shape data for specimens having identical locality data, resulting in 157

total localities.

85

Geographical shape variation

l explored the contribution of drift, or isolation by distance, to non-
allometric shape variation by performing Mantel tests on Procrustes and
geographical distance matrices (Mantel, 1967). The Mantel test is a permutation
test of the correlation between two matrices. Pairwise Procrustes distance
matrices were generated for each view in CoordGen6h (Sheets, 2006), and a
distance matrix of latitude and longitude was generated in R. Mantel tests for
each view were performed in R 2.8.1 (R Development Core Team, 2009), using
the mantelrtest platform with 1000 permutations.

To further examine the relationship between shape and geography, I
tested for covariation between shape and geographical variables with two-block
partial least squares (PLS) analysis (Rohlf and Corti, 2000). PLS models the
covariation between two sets of variables. PLS uses singular value
decomposition (SVD) to create a pair of singular axes, one for each block of
data, which maximizes the covariance between the blocks. The first block of data
contained the geometric shape variables. The second block of data contained
normalized latitude and longitude. Permutation tests were used to determine
whether the covariance between the blocks was greater than that expected by
chance (PLSMaker, Sheets, 2004; n=1000 permutations). I then regressed the
standardized shape data for each skull view onto the first singular axis of the PLS
analyses to determine how much variation the covariation between shape and
geography explains, and regressed shape onto the dominant geographical

variable to visualize the shape change with deformation grids.

86

Using the individual singular axis scores from the first shape vector from
the PLS analysis, I calculated a paiwvise Euclidean distance matrix. I used this
distance matrix to perform an unweighted pair-group method with arithmetic
mean (UPGMA) cluster analysis in R using the h.clust platform, and inspected
the output for groups. I evaluated identified groups with canonical variate
analysis (CVA), using CVAGen6o (Sheets, 2007) and illustrated shape between
each pair of groups using TwoGroup6h (Sheets, 2003c). The robustness of the
groups described by UPGMA was evaluated with a jack-knife reassignment test
that randomly removed 10% of the sample and recalculated the canonical
function, reassigning specimens to groups (Sheets, 2007). The rate of correctly

classified specimens over 500 trials is reported.

RESULTS

Allometric shape van'ation

Centroid size of the spotted hyena skull is a significant predictor of skull
shape in all views examined (ventral cranium: F34,11325= 8.70, p<0.01; lateral
cranium: F44,14372= 18.61, p<0.01; lateral mandible: F34,13224= 12.85, p<0.01), but
size explains only a small proportion of the variation in shape (Table 4.2). The
shape changes that accompany increasing size are depicted by vectors on
landmarks, and associated deformation grids (Figures 4.2—4.4). Shape change
from small to large skulls in the ventral cranial view is dominated by a relative
lengthening of the anterior basicranial region, and a broadening of the palate.

The streamlining seen through the zygomatic arches is a function of changes

87

primarily along the jugal bone. In smaller skulls, the zygomatic arch bows
outward at the jugal, whereas in the larger skulls, the lateral edge of this bone is
flattened. The postglenoid process also lengthens medially (Figure 4.2). In the
lateral view, shape change is dominated by a relative heightening of the cranial
vault. This change is most noticeable just posterior to the orbit and along the
anterior braincase. The zygomatic arch becomes broader dorso-ventrally, as
shown by the ventral displacement of the lower zygomatic landmark (landmark
9). The posterior sagittal and upper nuchal crests are ventrally displaced, and the
incisor is rotated downward (Figure 4.3). In the mandible, changes in shape are
primarily observed in the coronoid process, the angular process, and the
horizontal ramus. In the coronoid process, there is an increase in relative
breadth, and a posterior displacement of the apex. The angular process is
decreased in length and is ventrally displaced. Additionally, there is a dorso-
ventral expansion of the horizontal ramus, particularly ventral to P4 and M. A
relative increase in length of the anterior horizontal ramus is indicated by an
increase in post-canine diastema size, and posterior displacement of the incisors

and the end of the mandibular symphysis (Figure 4.4).

Table 4.2. Results of regression of shape variables (partial warp +uniform
component scores) on natural log centroid size.

 

 

 

 

View Glgodall’s % variation p-value
an, df2. explalned

Ventral 8.7034, 11523 2.5 <0.01

Lateral 186144, 14372 5.22 <0.01

Mandible 12.8534, 13224 3.48 <0.01

 

 

 

 

 

 

88

Geographical shape variation

Mantel tests on geographic distance and size-adjusted Procrustes
distance data were significant for all views (ventral: r=0.183, p=0.001; lateral:
r=0.183, p=0.003; mandible: r=0.107, p=0.018), indicating that drift does have an
effect on skull shape. That is, skulls collected geographically close to one
another are more similar in shape to one another than to skulls collected farther
away.

Additionally, a PLS analysis of shape variables revealed significantly
greater covariance with latitude and longitude than would be expected by chance
for all three skull views (Table 4.3). The first singular axis was significant for all
three views, with the loadings for this axis almost completely dominated by
latitude (Table 4.3). Taken together, these results indicate a structuring of shape
variance along a latitudinal gradient. Regression of shape on the first singular
axis for geography indicates that this covariation accounts for a modest portion of
the overall variation in shape (Table 4.4).

Table 4.3. Summary of results for PLS analysis of shape with geographical
variables.

 

 

 

 

View Axis 6X53: e d p-value Correlation [[33:96 L%2%::ge
Ventral SA1 98.52 0.00 0.503 0.996 0.094
Lateral SA1 98.36 0.02 0.495 0.991 0.134
Mandible SA1 91.68 0.03 0.318 0.992 0.132

 

 

 

 

 

 

 

 

 

89

 

Figure 4.2. Deformation grid showing allometric changes from a linear regression
of shape on log(centroid size) in the ventral view. The landmarks have been
back-reflected, and the deformation exaggerated 2.5 times for ease of
interpretation. Vectors on landmarks in the deformation grid show the direction
and magnitude of change from the smallest to the largest specimens.

90

     
   
   
    

     

    

   

     

 

 

Figure 4.3. Deformation grid showing allometric changes from a linear regression
of shape on log(centroid size) in the lateral view. The landmarks have been back-
reflected, and the deformation exaggerated 2.5 times for ease of interpretation.
Vectors on landmarks in the deformation grid show the direction and magnitude

of change from the smallest to the largest specimens.

91

 

 

 

 

 

Figure 4.4. Deformation grid showing allometric changes from a linear regression
of shape on log(centroid size) in the mandible view. The landmarks have been
back-reflected, and the deformation exaggerated 2.5 times for ease of
interpretation. Vectors on landmarks in the deformation grid show the direction
and magnitude of change from the smallest to the largest specimens.

92

Table 4.4. Results of regression of standardized shape variables on the first
singular axis for geography of the 2-block partial least squares analyses.

 

View

 

 

 

 

Goodall’s % variation
. p value
Fdfi' dfz explained
Ventral 2574344420 16.53 <0.01
Lateral 14,554,5632 10,27 <0.01
Mandible 3545485576 2056 <0.01

 

 

 

 

 

Three major groups were identified from the UPGMA cluster analysis on
the first singular axis for shape for all three skull views. The dendrogram for the
ventral view is depicted in Figure 4.5, and the patterns for the other two views
were similar. Plotted on a map of Africa, the three groups overlap at the equator
(Figure 4.6). For each view of the skull, the area north of the equator is
dominated by two groups, group 1 to the east, and group 3 to the west, whereas
the area south of the equator is dominated by group 2 (coded specimens for the
ventral view are mapped in Figure 4.6). For each skull view, CVA analysis found
one significant canonical variate separating the three groups (ventral cranium: A:
0.11, x2=247.97, df=68, p<2.22x10'16; lateral cranium: x: 0.05, x2=295.05,
df=88, p<2.22x10’16; lateral mandible: x: 0.05 x2=295.12, df=108, p<2.22x10"6).
When a jack-knifing procedure was applied to the canonical function, 87.5% of
ventral, 83.8% of lateral, and 79.3% of mandible individual samples were
correctly classified. For all three skull views, groups 1-3 represent a cline of
shape change such that group 1 was situated between groups 2 and 3 when

canonical variates 1 and 2 were plotted (Figure 4.7A—C).

93

 

 

 

 

 

 

 

 

 

 

C
N
Q—
0
In
\—
(U 0
fl
.9
D
c
N
a, C
.12 5-
‘5’ d
LIJ
In
C
Q...
C
o
o——l
o.
O
T v ‘ f w
1A 20 3'”
$252313?!

Figure 4.5. Dendrogram produced by a UPGMA cluster analysis based on a
painlvise Euclidean distance matrix of the first singular axis for shape from the
2B-PLS analysis of the ventral view. The symbols correspond to those used for
groups 1 — 3 in Figures 4.9 — 4.10.

94

 

 

 

A 0 #
Ventral UPGMA Grouping

\\ A / A 1
‘1,va “NJ/ 2 :

 

 

 

 

 

 

Figure 4.6. Plot of the three groups realized by the UPGMA clustering, for the
ventral cranial view. The Equator and the Tropic of Capricorn are indicated by the
upper and lower lines, respectively.

95

Deformations showing painlvise differences between all three groups
reinforce the notion that the groups exist in a cline. In the ventral view (Figure
4.8), shape change along this cline is dominated by a relative shortening and
narrowing of the palate, and the anterior displacement of the jugal-squamosal
suture on the zygomatic arch. The anterior displacement of landmark 9 relative to
landmark 10 highlights a shift from a U-shaped palatal margin to a V-shaped
margin. Shape change along the cline in the lateral cranium view (Figure 4.9) is
mainly in the anterior displacement of the ventral jugal-squamosal suture on the
zygomatic arch. The angle of this suture is not as oblique in the northern skulls.
This displacement seems to be coupled with a dorsad flare of the squamosal
bone that is not captured by the deformation due to a lack of landmarks in this
area. In the lateral mandible view (Figure 4.10), there is a decrease in relative
breadth of the coronoid process, and the apex is displaced posteriorly, such that
the entire process seems to slope backward. Additionally, the angular process is

shortened, and the posterior portion of the mandibular ramus loses depth.

DISCUSSION
Allometric shape variation
Skull size makes a small but significant contribution to shape variation in
the spotted hyena. Thus, the allometry in adult skulls is not merely a continuation
of the shape trajectory seen with growth in early ontogeny (Tanner et al., 2009),
where the skulls experience global changes to allow for an increase in jaw

adductor mass, as well as an increase in surface area for muscle attachment.

96

 

 

 

 

 

 

Figure 4.7. Canonical variate analysis results for each view on groups 1
(traingles), 2 (circles), and 3 (squares) identified using UPGMA clustering; A)
ventral cranium, B) lateral cranium, and C) lateral mandible.

97

 

Figure 4.8. Pair-wise deformations of shape change between all groups in the
ventral cranial view; A) 2>1, B) 2>3, C) 1>3, D) landmark map. The deformations
are exaggerated 3 times for ease of interpretation.

98

 

Figure 4.9. Pair-wise deformations of shape change between all groups in the
lateral cranial view; A) 2>1, B) 2>3, C) 1>3, D) landmark map. The deformations
are exaggerated 5 times for ease of interpretation.

99

 

 

 

 

 

 

 

 

 

Figure 4.10. Pair-wise deformations of shape change between all groups in the
lateral mandible view; A) 2>1, B) 2>3, C) 1>3, D) landmark map. The
deformations are exaggerated 5 times for A and C, 3 times for B, for ease of
interpretation.

100

Rather, among the adult spotted hyenas examined here, shape change from
smaller to larger skulls is highlighted by changes in structures associated with the
origin or insertion of one or more muscles of the head and neck. As for early
ontogeny, allometric shape change in adults is associated with structures that are
acted upon by muscles, but unlike early ontogeny, the changes are more
localized. Instead of the anterior-posterior expansion of the zygomatic arches
seen in subadults, the zygomae in adults expand dorso-ventrally. Among adult
spotted hyena skulls, larger specimens show evidence of epigenetic restructuring
in response to the action of increased adductor muscle mass (e.g. Herring,
1993). The increase in the height of the vault, however, is not likely an increase
in surface area for muscle attachment (Joeckel, 1998), but may be related to an
allometric increase in the size of the fronto—parietal sinus, which has been shown
to dissipate stress during bone-cracking while feeding (Tanner et al., 2008). The
dramatic changes in shape with increasing skull size most likely contributed to
the difficulty early researchers experienced in attempting to document a

geographic pattern in skull shape.

Shape variation and the taxonomy of Crocuta crocuta

Many wide-ranging mammalian species containing multiple subspecies
described in the late 18‘", 19‘“, and early 20‘“ centuries have since been
collapsed such that they now contain far fewer subspecies. Wider sampling has
allowed for more rigorous descriptions of species and subspecies based on both

morphological and molecular studies. In the spotted hyena, variation in body

101

size, coloration, and skull morphology were the bases for many subspecific
assignments. It was the individual nature of this variation that led Matthews
(1939b) to conclude that subspecific patterns could not be identified. Wlthin a
group of 103 specimens collected from one geographical locality in Tanzania,
Matthews (1939b) found specimens that he claimed could have been placed in
any of four subspecies (C.c. gemlinans, C.c. fisi, C.c. leontewi, C.c. fortis), based
on their original descriptions. Matthews” Tanzania sample undoubtedly contained
individuals from more than one clan, but there is evidence that variation exists at
even lower levels. Wlthin a single clan in the Masai Mara National Reserve,
Tanner (2007) found that skulls from adult females low in the dominance
hierarchy are larger than those from adult females of higher rank.

Large-scale sampling of skulls across the African continent has revealed a
pattern of morphological variation formerly obscured by a high level of individual
variation. The three groups indicated by UPGMA analysis represent a cline of
variation running from north-central Africa, through northeastern Africa, and
southward. While the variation in skull characteristics does not fall into distinct
groups that permit delineation of clear subspecies, the pattern of shape variation
does correspond to previously described subspecies. Group 2, the northwestern
group, allies with Crocuta crocuta fortis (Allen, 1924). This subspecies, originally
described from the northeastern extreme of the democratic republic of the
Congo, was invoked on the basis of its large size, inflated auditory bullae,
straight-sided braincase, and palate with a deeply incised, V-shaped posterior

border. While the auditory bullae and the shape of the braincase were not

102

evaluated here, the V-shaped palatal border of Group 2 dominates the shape
deformations between this group and Groups 1 and 3 (Figure 4.8).

The historic subspecies from the areas dominated by Groups 2 and 3 are
C. c. germinans to the northeast, and Co. crocuta to the south. The distinction
made between these two groups was based on size and coloration (Matschie,
1900), and no skull characteristics have been described. There is no evidence for
the additional taxa described for northern Kenya (C. c, fisr) or

Ethiopia/Somalia/Uganda (C. c. habessynica).

Pleistocene refugia hypothesis

The shape data described here parallel results of Crouta cytochrome b
analysis by Rohland et al. (2005), who found two clades, northeastern and
southern, with overlap through Kenya and Tanzania. With shape, I found a cline
stretching from the north-central part of the contient, through the northeast, to the
south. A phylogeographical pattern with an eastern/southern Africa split has been
demonstrated for many African mammals (e.g., Arctander et al., 1999), often with
an additional western lineage (Flagstad et al., 2000; Muwanika et al., 2003). The
driving force behind this disjunct pattern in geographical range and/or genetic.
variation is suggested to be range restriction at the end of the Pleistocene.
Global warming likely restricted savannah habitat (Flagstad et al., 2000),
resulting in 2—3 (depending upon the species in question) Pleistocene refugia.
The inhospitable habitat separating eastern and southern Africa in the

Pleistocene may have left a footprint on the ranges and in the DNA of modern

103

taxa. However Werdelin (2008) suggested that large carnivores are not likely to
show such distinct patterns, owing primarily to their dispersal abilities and
catholic feeding habits.

Similar to spotted hyenas, African wild dogs have an eastern/southern
clade pattern to mtDNA and microsattelite data, but with a larger zone of
admixture (Girman et al., 2001). African wild dogs have larger home range sizes
than spotted hyenas (Creel and Creel, 2002), and this larger zone of haplotype
overlap may reflect a higher level of mobility in African wild dogs. Geographical
skull size variation among African wild dogs is similar to that among spotted
hyenas in that the specimens collected in east Africa are smaller than those from
central and southern Africa (Girman et al., 1993), but to date, there have been no
morphological analyses examining associated variation in shape.

The cline in shape described here for Crocuta strengthens the Pleistocene
refugia hypothesis suggested by Rohland et al. (2005). More genetic analyses,
with both wider sampling, and more genes sequenced, are needed to elucidate
the phylogeographic pattern between northern and southern spotted hyenas.
Matched genetic and skull samples will allow for further development of the
thesis that skull morphology follows a cline in shape that is a remnant of

geographic isolation in Plesitocene refugia.

104

APPENDIX

105

Table A1. Catalogue numbers of specimens examined in Chapter One.

*included in cranium analysis, Tincluded in mandible analysis.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AMH 114225*1 BM 39.414*1 MSU 35077*1
AMH 114227*1 BM 39.415*1 MSU 35078*1
AMH 114256*1 BM 39.415*1 MSU 35079*1
AMH 147880*T BM 39.417*1 Msu 35080*1
AMH 165118*1' BM 39.419*1 Msu 36081*1'
AMH 155119*1 BM 39.420*1 MSU 36082*1
AMH 173511*1 BM 39.421*1 MSU 36083*1'
AMH 187759*1 BM 39.422*1 MSU 36084*1'
AMH 1377701 BM 39.423*1 MSU 35094*1
AMH 137771*1 BM 39.424*1 MSU 36156*1
AMH 137772*1 BM 39.425*1 MSU 36150*1
AMH 187773*1 BM 39.426*1 MSU 35151*1
AMH 187774*1' BM 39.427*1 MSU 35153*1
AMH 187776*1 BM 39.428*1 MSU 35155*1
AMH 187777*1 BM 39.428a*1 MSU 36168*1'
AMH 187778*1 BM 39.429*1 MSU 35550*1
AMH 187779*1 BM 39.430*1 MSU 36551*1
AMH 187780*1 BM 39.431*1 Msu 35552*1
AMH 187781*1 BM 39.432*1 MSU 35553*1
AMH 187782*1' BM 39.433*1 MSU 36558*1'
AMH 205150*1 BM 39.435*1 MSU 36568*1'
AMH 20809*1' BM 39.436*1 MSU 36569*1'
AMH 20810*1 BM 39.437*1 MSU 35570*1
AMH 215355*1 BM 39.438*1’ MSU 35571*1
AMH 27755*1 BM 46.8.3.31 MSU 35581*1
AMH 27757*1 BM 5.4.3.4*1 MSU 8048*1'
AMH 35389*1 BM 58.208*1 MSU F987*1'
AMH 363901 BM 59.272*1 Msu Bl=T*1
AMH 35391*1 BM 52.705*1 MSU Eco*1
AMH 52059*1 BM 62.707*1 Msu vcs*1
AMH 52060*1 BM 65.537* MSU NHM114*1'
AMH 52052*1 BM 66.790*1 MSU NHM115*1*
AMH 52053*1 BM 56.791*1 MVZ 165159*1
AMH 52064*1 BM 56.7921 Mvz 155160*1

 

AMH 52065*1‘

BM 69.2.2.13*L

MVZ 165162*1'

 

AMH 52067*‘|'

BM 70.706*T

MVZ 165163*1‘

 

AMH 52068*T

BM 73.1955*1’

MVZ 155155*1

 

AMH 52069*1’

BM 8.7.24.13*1'

MVZ 165166*1’

 

AMH 52097*1’

BM 8.7.24.14*‘|’

MVZ 165167*1'

 

 

 

 

 

 

 

AMH 54243*1' BM 9.6.1.14*‘|’ MVZ 165169*1’
AMH 54244*1' BM 92.8.1 .4*1’ MVZ 165170*1’
AMH 54312*T BM 92.8.1.5*1' MVZ 165171*‘|'
AMH 55467*1' Cambridge 4064*1' MVZ 165173*‘|'

106

 

Table A1. Chapter One specimens, continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AMH 59447*1 Cambridge 4052*1 Mvz 155174*1
AMH 80103*1 Cambridge 4055*1 MVZ 155175*1
AMH 805211 Cambridge 4055*1 Mvz 155175*1
AMH 81833*1' Cambridge 4057*1 MVZ 155177*1
AMH 83591*1 CM 2073*1 MVZ 155178*1
AMH 83592*1 CM 20871*1 MVZ 155179*1
AMH 83593*1 CM 5852*1 MVZ 165180*1'
BM 0.10.3.1*1 CM 5855*1 Mvz 155181*1
BM 0.3.18.22*1 CM 5873*1 MVZ 155182*1
BM 0862* CM 53108*1 MVZ 173733*1
BM 1.8.9.27*1 CM 5454*1 MVZ 173734*1
BM 11.4.4.1*1 CM 5827*1 MVZ 173735*

BM 11.8.2.10*1 FMNH 104021*1 Mvz173737*1
BM 11.8.2.9*1 FMNH 104022*1 Mvz 173738*1
BM 12331*1 FMNH 104981*1 Mvz173740*1
BM 15.3.5.90*1 FMNH 127825*1 MVZ173741*1'
BM 19.5.1.3*1 FMNH 127825*1 MVZ173743*1'
BM 2.2.8.1*1 FMNH 127829*1' MVZ 173744*1
BM 2.8.5.4*1 FMNH 188551 MVZ 173745*1
BM 21 .29.10.30*1 FMNH 270071 MVZ 173745*1
BM 23.1.1.811 FMNH 32933*1 MVZ 173747*1
BM 23.3.4.111 FMNH 32935* MVZ 173748*1

 

BM 23.3.4.14*1’

FMNH 34582*1‘

MVZ 173751 *T

 

BM 23.3.4.15*1‘

FMNH 34583*1

MVZ 173758"

 

BM 23.3.4.16*

F MNH 73034*1'

MVZ 173759*

 

BM 23.3.4.19*1'

FMNH 73035*1’

Mvz 173752*1

 

BM 25.12.4.233*'|'

FMNH 93866*1’

MVZ 173768*T

 

BM 27.2.9.10*1’

F MNH 98739*1'

MVZ 173770*1’

 

 

 

BM 27.2.9.9*‘|' FMNH 98952*1’ MVZ 173771*1'
BM 27.7.3.8*1’ IRSNB 102501' MVZ 173773*1’
BM 27.7.3.8A*1‘ IRSNB 10336*1’ MVZ 175801*1'

 

BM 28.11.6.3*1'

IRSNB 11799*1'

MVZ 184088*1'

 

BM 28.9.11.133*1’

IRSNB 1 1801*1’

MVZ 184089*1‘

 

BM 29.11.3.8*1‘

IRSNB 118021

MVZ 4823*1

 

 

 

 

 

 

 

 

 

 

 

 

 

BM 30.12.182*1' IRSNB 21278*1 NMK-OM 2703*1
BM 30.12.2.4*1 IRSNB 21302*1 NMK-OM 2705*1
BM 30.12.2.5*1 IRSNB 21435*1 NMK-OM 2705*1
BM 30.3.5.13*1 IRSNB 4512*1 NMK-OM 27131

BM 30.3.5.4*1 IRSNB 7705*1 NMK—OM 3444*1
BM 31.1.2.11*1 IRSNB 8532*1 NMK—OM 3445*1
BM 31.4.1.11* IRSNB 8533*1 NMK-OM 3575*1
BM 31 4.1121 IRSNB 8534*1 NMK-OM 4750*1
BM 31.4.1 .13*1 IRSNB 8535*1 NMK-OM 5085*1
BM 34.4.1.135*1 IRSNB 9480*1 NMK—OM 51941

107

 

Table A1 Chapter One specimens, continued.

 

BM 34.4.1.137*1’

IRSNB 9967*1'

NMK—OM 5314*1’

 

BM 34.4.1.138*1'

MNHN-AC 1894-54*1

NMK—OM 7189*1’

 

BM 34.4.1 .139*1'

MNHN-AC 1896-450*

NMK-OM 7465*1’

 

BM 34.4.1.140*1'

MNHN-AC 1901 -662*1'

NMK-OM 7754*1'

 

BM 38.10.18.471

MNHN-AC 1910-162*1‘

NMK-OM 7755*1'

 

BM 38.5.10.1*1’

MNH N-AC 1927-175*1'

NMK-OM 7756*1’

 

BM 38.5.10.2*1‘

MNHN-OM 1962-1 533*1‘

NMK-OM 7757*1’

 

BM 38.5.10.3*1‘

MNHN-OM 1962-15351’

NMK-OM 7759*1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BM 39.3371' MNHN-OM 1962-15362*1’ NMK-OM 7760*1'
BM 39.338*:L MNHN—OM 1962-1537*1’ NMK-OM 7761*1‘
BM 39.339*1‘ MNHN-OM 1972.399*T NMK-OM 7762*1'
BM 39.340*1’ MNHN-OM 1972.400*1’ NMK-OM 7850*1’
BM 39.341*1' MNHN-OM 1973.125*T NMK-OM 7893*1‘
BM 39.342*'|' MNHN-OM 1985-1858*1' NMK-OM u*‘|'

BM 39.343*1' MNHN-OM 1986-1090*'|’ RCSOM137.41*1’
BM 39.344*1‘ MNHN-OM 1996-2514*1’ RCSOM137.42*1‘
BM 39.345*1' MNHN-OM 1997-415* RCSOM137.421*1’
BM 39.346*1' MRAC 11376*1’ RCSOM137.43*1‘
BM 39.347*'|' MRAC 11602*1‘ RCSOM137.60*T
BM 39.348*1‘ MRAC 11701*1' RCSOM137.61*1'
BM 39.349*1‘ MRAC 1182*1' RCSOM137.62*1‘
BM 39.350*1‘ MRAC 1183-m*1' RCSOM137.63*j‘
BM 39.351*1' MRAC 12096*1' RCSOM16.5*‘|‘
BM 39.352*1‘ MRAC 12442*1' USNM 015202*‘|‘
BM 39.353*1' MRAC 128141' USNM 020874*1'
BM 39355“ MRAC 13843*1' USNM 122544*1’
BM 39.356*1’ MRAC 14367*1’ USNM 161909*1‘
BM 39.357* MRAC 14369*1' USNM 162920*1’
BM 39.358*'|' MRAC 14813* USNM 162921*'|'
BM 39.359*1' MRAC 15644*1' USNM 162923*‘|‘
BM 39.3599*1' MRAC 15928” USNM 162924*1’
BM 39.360*1' MRAC 16719*1‘ USNM 163099*1‘
BM 39.361*1‘ MRAC 16785*1' USNM 163100*1'
BM 39.362*‘|‘ MRAC 16786*1‘ USNM 163101*'|'
BM 39.363*1' MRAC 16787*1‘ USNM 163102*1'
BM 39.364*1‘ MRAC 17619*T USNM 163103*1’
BM 39.365*1' MRAC 17701*1' USNM 163104*‘|'
BM 39.366*1‘ MRAC 177401’ USNM 163299*1'
BM 39.367*1' MRAC 18000*'|' USNM 163344*1'
BM 39.368*1’ MRAC 18001*‘|’ USNM 164502*1'
BM 39.369*1’ MRAC 18495*1’ USNM 164506*1‘
BM 39.370*1' MRAC 18627*1' USNM 164549*‘|'
BM 39.371*1' MRAC 1897*1' USNM 164834*1'
BM 39.372*'|' MRAC 19272*1’ USNM 172924*1'

108

 

Table A1. Chapter One specimens, continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BM 39.373*1 MRAC 19273*1 USNM 1730031

BM 39.374*1 MRAC 19274*1 USNM 173004*1
BM 39.375*1 MRAC 20325*1 USNM 181515*1
BM 39.375*1 MRAC 2152*1 USNM 181517*1
BM 39.377*1 MRAC 2172*1 USNM 181518*1
BM 39.378*1 MRAC 22802*1 USNM 181519*1
BM 39.379*1 MRAC 2907*1 USNM 181520*1'
BM 39.380*1 MRAC 35328*1 USNM 181521*1
BM 39.381*1 MRAC 35543*1 USNM 181522*1
BM 39.382*1 MRAC 35545*1 USNM 181524*1
BM 39.383*1 MRAC 3728*1 USNM 181525*1
BM 39.384*1 MRAC 3788*1 USNM 181525*1
BM 39.385*1 MRAC 3794*1 USNM 181527*1
BM 39.385*1 MRAC 39401 USNM 181529*1
BM 39.3873 MRAC 3870*1 USNM 181530*1
BM 39.388*1 MRAC 5934*1 USNM 181533*1'
BM 39.389*1' MRAC 5154*1 USNM 181534*1
BM 39.390*1 MRAC 5330*1 USNM 182032*1
BM 39.391*1 MRAC 79541 USNM 182078*1
BM 39.392*1 MRAC 8005*1 USNM 1820821

BM 39.393*1 MRAC 9292*1 USNM 182084*1
BM 39.394*1 MRAC 95791 USNM 182085*1
BM 39.395*1 MRAC 95891 USNM 182091*1
BM 39.395*1 MSU 12391*1 USNM 182095*1
BM 39.397*1 MSU 22401*1 USNM 182101*1
BM 39.398*1 MSU 24292*1 USNM 182103*1'
BM 39.399*1 Msu 25055*1 USNM 182105*1
BM 39.400*1 Msu 2714*1 USNM 182110*1
BM 39.401*1 MSU 35852*1 USNM 182113*1
BM 39.402*1 MSU 35853*1 USNM 1821 14*1
BM 39.403*1 MSU 35854*1 USNM 182117*1'
BM 39.404*1 MSU 35855*1 USNM 182210*1
BM 39.405*1 MSU 35855*1 USNM 201010*1
BM 39.407*1 MSU 35857*1 USNM 239151*1
BM 39.408*1 MSU 35858*1' USNM 2527741

BM 39.409*1 MSU 35859*1 USNM 357384*1
BM 39.410*1 Msu 35008*1 USNM 357385*1
BM 39.411*1 MSU 35009*1 USNM 358502*1
BM 39.412*1 MSU 35011*1 USNM 4291751

BM 39.413*1 Msu 35074*1

 

 

 

109

 

Table A2. Museums visited and abbreviations

Field Museum, Chicago (FMNH)

Museum of Vertebrate Zoology, Berkeley (MVZ)

Royal Museum for Central Africa, Tervuren (MRAC)

Royal Belgian Institute of Natural Sciences, Brussels (IRSNB)
National Museums of Kenya, Nairobi (NMK—OM)

Natural History Museum, London (BM)

National Museum of Natural History, Paris (MNHN)

American Museum of Natural History, New York (AMH)

The Smithsonian Institution National Museum of Natural History, Washington,
DC. (USNM)

Carnegie Museum of Natural History, Pittsburgh (CM)

Michigan State University Museum, East Lansing (MSU)

University Museum of Zoology, Cambridge (Cambridge)

Royal College of Surgeons Odontological Collections, London (RCSOM)

Table A3. Presence of supernumerary P1 in Crocuta crocuta.

 

 

 

 

 

 

 

 

Specimen Tooth details Collection locality
Cambridge 4065 Left P1 present Hargeisa, Somalia
RCSOM 16.5 Right P1 present Samburu, Kenya
USNM 020874 Left P1 alveolus present Victoria Falls
USNM 367385 Broken left P1 present, Chioco, Mozambique

right P1 alveolus present
BM 39.420 Both P1 present Balbal, Tanzania
CMNH 15020 Both P1 present Pittsburgh Zoo

 

 

 

 

Note that the zoo specimen was not included in the total anomaly count.

110

Table A4. Specimens used in the study of sexual dimorphism of the spotted
hyena. * excluded from mandible analysis. 1' excluded from ventral analysis.

 

The Natural History Museum, London

Michigan State
University Museum

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Catalogue Sex Catalogue Sex Catalogue Sex
number number number
BM39 3371' M BM39 391 F 35852 M
BM39 339 M BM39 394 F 35853 M
BM39 340 M BM39 395 F 35854 M
BM39 342 M BM39 396 F 35856 F
BM39 343 M BM39 397 F 36008 F
BM39 344 M BM39 399 F 36011 F
BM39 345 M BM39 400 F 36074 F
BM39 346 M BM39 401 F 36077 F
BM39 348 M BM39 402 F 36078 M
BM39 349 M BM39 403 F 36079 M
BM39 351 M BM39 404 F 36080 F
BM39 353 M BM39 407 F 36083 F
BM39 355* M BM39 408 F 36084 M
BM39 356 M BM39 409 F 36094 F
BM39 358 M BM39 410 F 36160 F
BM39 359 M BM39 411 F 36163 M
BM39 360 M BM39 412 F 36165 F
BM39 361 M BM39 413 F 36168 M
BM39 362 M BM39 414 F 36550 F
BM39 363 M BM39 416 F 36551 F
BM39 364 M BM39 417 F 36552 F
BM39 366 M BM39 419 M 36553 F
BM39 368 M BM39 420 M 36558 F
BM39 369 M BM39 421 M 36567 F
BM39 370 M BM39 422 M 36568 F
BM39 373 M BM39 423 M 36569 F
BM39 375 M BM39 424 F 36570 F
BM39 376 M BM39 425 M 36571 F
BM39 378 M BM39 427 M 36581 F
BM39 381 M BM39 428 M 486ECO F
BM39 382* M BM39 429 M 897BFT M
BM39 383 M BM39 430 F 225VGS M
BM39 3861' M BM39 431 F
BM39 387' M BM39 432 F
BM39 388 M BM39 433 F
BM39 389 F BM39 434* F
BM39 390 F BM39 435 F

F

 

 

BM39 437

 

 

 

111

 

Table A5. Descriptions of landmark locations.

Ventral Landmarks

 

1 Premaxilla-premaxilla suture at the posterior edge of the first upper incisor
alveoli

Premaxilla-maxilla suture at the lingual edge of the canine*

Posterior-most point of the incisive foramen*

Premaxilla-maxilla suture at the midline

Posterior edge of the occipital bone at the midline of the foramen magnum

Metacone of P2*

Posterior palatine foramen*

Maxilla-palatine suture at the midline

Palatine-palatine suture at the posterior edge of the palate

10 MaxilIa-palatine suture at the posterior edge of the palate*

11 Medial-most edge of the protocone of P4*

12 Medial-most edge of the maxilla-jugal suture*

13 Lateral-most edge of the jugal-squamosal suture*

14 Anterio-lateral corner of the glenoid fossa*

15 Medial-most extension of the postglenoid process*

16 Medial aspect of the foramen ovale*

17 Palatine-pterygoid suture at the presphenoid*

18 Posterior-most edge of the jugular/hypoglossal foramen, medial aspect*

19 Anterior edge of the external auditory meatus*
*bilateral landmark

(OQNODO'I-bOJN

Lateral landmarks

 

Anterior-most point of the l3 alveolus

Anterior-most edge of the canine at the alveolus

Posterior-most edge of the canine at the alveolus

Anterior-most edge of the nasal-premaxilla suture

Dorsal edge of the infraorbital foramen

Dorsal edge of the lacrimal foramen

Tip of the postorbital process of the frontal bone

Dorsal-most edge of the jugal-squamosal suture

Ventral-most edge of the jugal-squamosal suture

10 Ventral-most edge of the maxilla-jugal suture

11 Dorsal aspect of the junction of pterygoid hamulus with the body of the
pterygoid

12 Anterio-dorsal edge of the external auditory meatus

13 Anterior-most dorsal edge of the occipital condyle

14 Posterior-most extreme of the curvature of the sagittal crest

(OGDVODU‘I-thA

112

Table A5. Descriptions of landmark locations, continued.

Mandibular landmarks

 

Anteriodorsal-most point of the mandiblular symphysis

Posterior-most edge of the canine at the alveolus

Posterior-most edge of M1 at the alveolus

Anterior-most edge of P2 at the alveolus

Posterior-most extreme of the curvature of the coronoid process

Posterior edge of the articular facet of the mandibular condyle

Posterior-most point of the angular process

Anterior edge of the articular facet of the mandibular condyle

Ventral apex of the curve of the dentary

10 Posterior-most point of the mandibular symphysis

11 Dorsal apex of the alveolus between the two roots of P4

12 Dorsal-most aspect of the curve between the angular process and the
mandibular condyle

13 Dorsal-most projection of the angular process

(DWNODU‘IhODNé

113

 

Table A6. Catalogue numbers of specimens examined in Chapter Three.
*included in ventral centroid size analyses, 1'included in first lower molar
analyses, :lzused for evaluating centroid size, first lower molar length, and
condylobasal length as proxies of body size.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AMH 1142251 BM 39.409*1 MSU 36008*1't
AMH 114227*1 BM 39.410*1 Msu 35011*11
AMH 114255*1 BM 39.411*1 MSU 35074*1
AMH 165118*1' BM 39.412*1 MSU 35077*11
AMH 155119*1 BM 39.413*1 MSU 35078*1:l:
AMH 187769*1' BM 39.414*1 Msu 35079*1
AMH 187771*1' BM 39.415*1 MSU 35080*1
AMH 187772* BM 39.417*1 MSU 36083*1':1:
AMH 187776*1' BM 39.419*1 MSU 36084*1'
AMH 187777*1' BM 39.420*1 MSU 35094*1
AMH 187779* BM 39.421*1 Msu 35150*1
AMH 187782* BM 39.422*1 MSU 35151*1
AMH 20809*1' BM 39.423*1 MSU 35153*11:
AMH 20810*1' BM 39.424*1 MSU 35155*11
AMH 215355*1 BM 39.425*1 MSU 36168*1':l:
AMH 27755*1 BM 39.427*1 MSU 35550111
AMH 2775711 BM 39.428*1 MSU 35551*11
AMH 52059*1 BM 39.429*1 MSU 35552*1
AMH 52050*1 BM 39.430*1 MSU 35553111 ‘
AMH 52053*1 BM 39.431* MSU 36558*1'
AMH 52054*1 BM 39.432*1 MSU 35557*1:1
AMH 52055*1 BM 39.433* MSU 36568*1'
AMH 52068*1' BM 39.435*1 MSU 35559*11
AMH 52059*1 BM 39.437*1 MSU 35570*1
AMH 52097*1 BM 58.208*1' MSU 35571*11:
AMH 54243*1 BM 59.2721 Msu 36581*1'
AMH 54244*1 BM 52.705*1 MSU 8048*1
AMH 55457*1 BM 52.707*1 MSU BFT*1':1:
AMH 81833*1' BM 55.537* Msu ECO*1':1:
AMH 83591*1' BM 55.790*1 MSU VGS*1'
AMH 83592*1 BM 55.7921 MVZ 155159*
AMH 83593*1 BM 70.705* MVZ 155150*1
BM 0.10.3.1*1 BM 9.5.1.14*1 Mvz 155152*1
BM 0.3.18.22*1‘ BM 92.8.1.4*1 MVZ 155153*1
BM 0862* Cambridge 4052*1 Mvz 155155*
BM 1.8.9.27*1' Cambridge 4055*1 MVZ 155155*1
BM 19.5.1 .3*1 Cambridge 4057*1 MVZ 155157*1
BM 2.2.8.1*1 CM 20871*1' Mvz 155159*1
BM 2.8.5.4*1' CM 5862*1' MVZ 155170*1
BM 21 29.10301 CM 53108*1 Mvz 155175*1

 

 

 

 

114

 

Table A6. Chapter Three specimens, continued.

 

BM 23.3.4.111

CM 68271"

MVZ 165176*1'

 

BM 23.3.4.14*1‘

FMNH 104021*1

MVZ 165179*1'

 

BM 23.3.4.15*1

FMNH 104981 *1'

MVZ 165180*1'

 

 

 

 

 

 

 

 

BM 23.3.4.16* FMNH 127825*1' MVZ 165181*1'
BM 23.3.4.19*1‘ FMNH 127826*1’ MVZ 165182*j
BM 24.8.3.41‘ FMNH 127829*1’ MVZ 173733*‘|'
BM 24.8.3.741‘ FMNH 135072* MVZ 173734*1'
BM 25.12.4.233*‘[’ FMNH 32933*‘|’ MVZ 173737*1'
BM 27.2.9.9*‘|‘ FMNH 34582*1‘ ‘ MVZ 1737411'

BM 27.7.3.8*1' FMNH 34583*1' MVZ 173743*1'
BM 27.7.3.8A*1' FMNH 73034*1‘ MVZ 173744*1'

 

BM 28.11.6.3*‘|’

FMNH 73035*1‘

MVZ 1737451"

 

BM 29.11.3.8*1'

F MNH 938661‘

MVZ 173745*1

 

BM 30.12.182*1‘

FMNH 98739*1'

MVZ 173747*1‘

 

BM 30.12.25“

FMNH 98952*1'

MVZ 173748*1’

 

BM 30.3.6.13*1'

IRSNB 10250*1‘

MVZ 173751 *1“

 

BM 31.1.2.11*1

IRSNB 10336*1'

MVZ 173754*

 

BM 31.4.1.1'3*1'

IRSNB 11799*1

MVZ 173758*

 

BM 34.4.1.1341‘

IRSNB 1 1801*1’

MVZ 173759*

 

BM 34.4.1.136*1’

IRSNB 118041

MVZ 173768*1'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BM 34.4.1.137*1' IRSNB 21278* MVZ 173770*
BM 34.4.1.138*1’ IRSNB 21302*1‘ MVZ 173771*1'
BM 34.4.1.139*T IRSNB 21436*1' MVZ 175801*1'
BM 34.4.1.140*1‘ IRSNB 4612*1' MVZ 184089*T
BM 38.10.18.47*1‘ IRSNB 77051“ NMK-OM 27031'
BM 38.5.10.2*1‘ IRSNB 8632*1’ NMK—OM 2705*1’
BM 38.5.10.3*1’ IRSNB 8633*1’ NMK-OM 3580*
BM 39.3371' IRSNB 8634*1' NMK—OM 7189*1'
BM 39.339*‘|’ IRSNB 9480*1' NMK—OM 7755*1’
BM 39.340*1‘ IRSNB 99671' NMK-OM 7757*‘1’
BM 39.342*1' MNHN-AC 1894-54*1’ NMK-OM 7761*1‘
BM 39.343*1' MNHN—AC 1896-450* NMK-OM 7762*1‘
BM 39.344*1' MNHN-OM 1962-1537*1' NMK-OM 7850*1'
BM 39.345*1’ MNHN-OM 1972.400*‘|' RCSOM137.41*1'
BM 39.346*1' MNHN-OM 1996-2514*‘|' RCSOM137.42*1‘
BM 39.348*‘|' MNHN-OM 1997-415* RCSOM137.43*1’
BM 39.349*1' MRAC 11376*'|' RCSOM16.5*T
BM 39.351*1' MRAC 11602*1' USNM 020874*1'
BM 39.353*‘|‘ MRAC 11701*1' USNM 1225441'
BM 39.355* MRAC 12096*1' USNM 162920*'[
BM 39.356*1‘ MRAC 12442*1' USNM 162924*
BM 39.359*1' MRAC 14367* USNM 163099*1‘
BM 39.360*1‘ MRAC 14369*1’ USNM 163100*1‘
BM 39.361* MRAC 14813* USNM 163101*1‘

 

 

 

115

 

 

Table A6. Chapter Three specimens, continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BM 39.352*1 MRAC 15719* USNM 153102*1
BM 39.353* MRAC 17519*1 USNM 153103*1
BM 39.354*1 MRAC 180001 USNM 154502*1
BM 39.355*1 MRAC 18495*1 USNM 154505*1
BM 39.368*1' MRAC 18627*1' USNM 154549*1
BM 39.359*1 MRAC 18971 USNM 181516*1'
BM 39.370*1 MRAC 192721 USNM 181518*1
BM 39.373*1 MRAC 192731 USNM 181519*1'
BM 39.375* MRAC 20325*1 USNM 181520*1'
BM 39.375*1 MRAC 2152*1 USNM 181521*1
BM 39.378*1' MRAC 22802*1 USNM 181524*1'
BM 39.381*1' MRAC 2907*1 USNM 181525*1
BM 39.382* MRAC 36328*1' USNM 181526*1'
BM 39.383*1' MRAC 35543*1 USNM 181527*1'
BM 39.3851 MRAC 35545*1 USNM 181530*1'
BM 39.3851 MRAC 3728*1' USNM 181533*1'
BM 39.387*1' MRAC 3788*1' USNM 181534*1
BM 39.388*1' MRAC 3841 USNM 182032*1'
BM 39.389*1 MRAC 3870*1' USNM 182085*1'
BM 39.390*1 MRAC 5934*1 USNM 182091*1
BM 39.391*1 MRAC 9292*1 USNM 182095*1'
BM 39.394*1 MRAC 95791 USNM 182103*1
BM 39.395*1 Msu 12391*1 USNM 1821051
BM 39.395*1 MSU 22401*1 USNM 182113*1
BM 39.397*1 Msu 242921 USNM 1821171
BM 39.399*1 MSU 25055* USNM 182210*1'
BM 39.400*1 MSU 2714*1 USNM 201010*1
BM 39.401*1 MSU 35852*1' USNM 239151*1
BM 39.402*1 MSU 35853*1' USNM 367384*1'
BM 39.403* MSU 35854*1 USNM 357385*1
BM 39.404*1 MSU 35856* USNM 358502*1
BM 39.407*1 Msu 35857*1' USNM 429175*1
BM 39.408* MSU 35858*1'

 

 

116

 

Table A]. Catalogue numbers of specimens examined in Chapter Four.

*included in ventral analysis, 1'included in mandible analysis, :1: included in lateral

 

 

 

 

analysis.
AMH 1142261 BM 39.408*1'1 MSU 36008*'|':1:
AMH 114227*1'1 BM 39.409*1':t MSU 3601 1*1'1
AMH 114256*1':1: BM 39.410*1:1: MSU 36074*1'1:
AMH 165118*1' BM 39.41 1*1'21: MSU 36077*1':1:

 

AMH 155119*11

BM 39.412*1':1:

MSU 36078*1':1:

 

AMH 187769*1‘:I:

BM 39.41 3*1'1:

MSU 36079*T:1:

 

AMH 187771 *1:

BM 39.414*T:|:

MSU 36080*1':1:

 

AMH 187772*1':1:

BM 39.415*11

MSU 36083*1':1:

 

AMH 187776*1':1:

BM 39.41 7*1'1:

MSU 36084*1':I:

 

AMH 187777*1:|:

BM 39.419*11:

MSU 35094*1¢

 

AMH 187779*:t

BM 39.420*1':1:

MSU 36160*1':I:

 

AMH 1877801

BM 39.421 *1'21:

MSU 36161*1':1:

 

AMH 187782*‘|‘

BM 39.422*11

MSU 35153*1:1

 

AMH 20809*1':1:

BM 39.423*‘|".1:

MSU 36165*‘|'1:

 

AMH 20810*1'1:

BM 39.424*'|':l:

Msu 36168*1':1:

 

AMH 21 6355*1'11:

BM 39.425*T:|:

MSU 36550*1‘:l:

 

AMH 27755*11

BM 39.427*11

MSU 35551*1.1

 

AMH 27767*T¢

BM 39.428*T:|:

MSU 36552*‘|':1:

 

AMH 52059*11

BM 39.429*11

MSU 35553*1':1:

 

AMH 52060*1’:I:

BM 39.430*1'i

MSU 36558*'|’:I:

 

AMH 52063*1':1:

BM 39.431*1‘:1:

MSU 36567*1':1:

 

AMH 52054*11

BM 39.432*1':1:

MSU 36568*1':1:

 

AMH 52065*1‘:I:

BM 39.433*1':1:

MSU 36569*T:I:

 

AMH 52068*1':l:

BM 39.435*1’:t

MSU 35570*11

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

AMH 52059*1¢ BM 39.437*1: MSU 35571*1.1
AMH 52097*11 BM 58.208*11 MSU 36581*1':l:
AMH 54243*11 BM 59.2721 MSU 8048*1'1:
AMH 54244*11 BM 52.705*11 MSU F9871:
AMH 5545711 BM 52.707*11 MSU BFT*1':1:
AMH 81833*11 BM 55.7921 MSU ECO*1':1:
AMH 83591’j'; BM 9.5.1 .14*1:1 MSU vos*11
AMH 83592*'|'1: BM 92.8.1.4*1:l: MSU NHM1151'11:
AMH 83593*1':1: Cambridge 4052*1: Mvz 155159*:
BM 0.10.3.1*1:1 Cambridge 4055*11 MVZ 155150*1:
BM 0.3.18.22*1':l: Cambridge 4057*11. Mvz 155152*1:
BM 0.8.6.2*:1: CM 20871*1':l: Mvz 155153*11:
BM 1.8.9.27*1':1: CM 5862*1'1: MVZ 155155*11
BM 15.35.9011. CM 63108*1':1; Mvz 155155*:
BM 19.5.1.3*1: CM 58271 Mvz 155157*11

 

BM 2.8.5.4*1’

FMNH 104021 *1'21:

MVZ 165169*1':1:

 

BM 21 .29.10.30*1‘

FMNH 104981*1'1:

MVZ 165170*1’:1:

 

 

BM 23.3.4.111‘

 

FMNH 127825*1':I:

 

MVZ 155175*11

 

117

 

Table A7 Chapter Four specimens, continued.

 

 

 

BM 23.3.4.14*1’:1: FMNH 127826*‘|':1: MVZ 165176*1':1:
BM 23.3.4.15*1':1: FMNH 127829*1’:|: MVZ 165179*‘|’i
BM 23.3.4.16*: FMNH 135072* MVZ 165180*1".1:

 

BM 23.3.4.19*1‘:I:

FMNH 32933*1’¢

MVZ 1651 81*1'3:

 

 

 

 

 

 

BM 24.8.3.41' FMNH 34582*11 Mvz 155182*11
BM 24.8.3.741 FMNH 34583*11 MVZ 173733*11
BM 25.12.4.233*1 FMNH 73034*11 MVZ 173734*11
BM 27.2.9.9*11 FMNH 7303531 MVZ 173737*11
BM 27.7.3.8*11 FMNH 938551 MVZ 17374111

BM 27.7.3.8A*11 FMNH 98739*11 MVZ 173743*11

 

BM 28.11.6.3*T:1:

FMNH 98952*1':|:

MVZ 1737451'

 

BM 29.11.3.8*11

IRSNB 10250*11

MVZ 173746*1’$

 

BM 30.12.182*1’:1:

IRSNB 10336*1’:l:

MVZ 173747*:I:

 

 

 

BM 31.1.2.11*1':1: IRSNB 11799*‘|':1: MVZ 173748":
BM 31.4.1.13*T:|: IRSNB 11801*1'; MVZ 173751*'1':1:
BM 34.4.1.1341‘ IRSNB 118041' MVZ 1737541:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BM 34.4.1.135*11 IRSNB 21278*11 MVZ 173758*1
BM 34.4.1.137*11 IRSNB 21302*11 MVZ 173759*1
BM 34.4.1.138*11 IRSNB 21435*11 Mvz 173758*11
BM 34.4.1.139*11 IRSNB 4512*11 Mvz 173770*1
BM 34.4.1.140*11 IRSNB 77051 MVZ 173771*11
BM 38.10.18.47*11 IRSNB 8532*11 MVZ 175801*11
BM 38.5.10.2*11 IRSNB 8533*11 MVZ 184088*11
BM 38.5.10.3*11 IRSNB 8534*11 MVZ 18408911
BM 39.33711 IRSNB 9480*11 NMK-OM 270311
BM 39.339*11 IRSNB 99571 NMK-OM 2705*11
BM 39.340*11 MNHN-AC 1894-54*1':1: NMK-OM 3580*1
BM 39.342*11 MNHN-AC 1896-450*:1: NMK-OM 7189*1
BM 39.34311 MNHN-OM 1952-1537*11 NMK—OM 7755*1
BM 39.344*11 MNHN-OM 1972.400*11 NMK-OM 7757*1
BM 39.345*11 MNHN-OM 1995—2514*11 NMK-OM 7751*1

 

BM 39.346*1': MNHN-OM 1997-415*:I: NMK-OM 7762*1’

 

BM 39.348*1':1:

MRAC 1 1376*1’1:

NMK-OM 7850*1’1:

 

 

 

 

BM 39.349*‘|'1: MRAC 1 1602*1'21: RCSOM137.42*1':|:
BM 39.351 *1: MRAC 11701 *1' RCSOM137.41*1‘$
BM 39.353*1’:t MRAC 12096*1’:1: RCSOM16.5*1':|:

BM 39.355*:1: MRAC 12442*1':1: RCSOM137.43*1‘:1:

 

BM 39.355*11

MRAC 14367*1'11:

USNM 020874*1':1:

 

BM 39.358*‘1’1:

MRAC 14369*1':|:

USNM 1225441‘

 

BM 39.359*1‘:I:

MRAC 14813*1

USNM 163099*1’:I:

 

BM 39.360*1':l:

MRAC 16719*1’:I:

USNM 163100*1':1:

 

 

 

 

 

 

BM 39.361*1’:|: MRAC 167861: USNM 163101*‘[:|:
BM 39.362*T:|: MRAC 17619*'1':1: USNM 163102*1’:I:
BM 39.363*1'1: MRAC 177401‘ USNM 163103*1’:1:

 

118

 

Table A7 Chapter Four specimens, continued.

 

BM 39.364*11

MRAC 1800011

USNM 164502*11

 

BM 39.366*11

MRAC 18495*11

USNM 154505*11

 

BM 39.358*11

MRAC 1852731

USNM 154549*11

 

BM 39.369*11

MRAC 1897*11

USNM 181516*11

 

 

 

 

 

 

 

 

 

 

BM 39.370*11 MRAC 192721 USNM 181518*11
BM 39.373*11 MRAC 192731 USNM 181519*11
BM 39.375*1'1 MRAC 2162*11 USNM 181520*11
BM 39.376*11 MRAC 22802*11 USNM 181521*11
BM 39.378*11 MRAC 2907*11 USNM 181524*11
BM 39.381*11 MRAC 36328*11 USNM 181525*11
BM 39.382*1 MRAC 36543*11 USNM 181526*11
BM 39.383*11 MRAC 36545*1 USNM 181527*11
BM 39.38511 MRAC 3728*11 USNM 181530*11
BM 39.38611 MRAC 3788*11 USNM 181533*11

 

BM 39.387*11

MRAC 3841

USNM 181534*11

 

BM 39.388*11

MRAC 3870*11

USNM 182032*11

 

BM 39.389*11

MRAC 5934*11

USNM 182085*11

 

BM 39.390*11

MRAC 9292*11

USNM 182091 *11

 

BM 39.391*11

MRAC 95791

USNM 182095*11

 

BM 39.394*11

MSU 12391 *11

USNM 182103*11

 

BM 39.395*11

MSU 22401*11

USNM 18210511

 

BM 39.396*11

MSU 2429211

USNM 182113*11

 

BM 39.397*11

MSU 26055*11

USNM 18211711

 

BM 39.39911

MSU 2714*11

USNM 182210*11

 

BM 39.400*11

MSU 35852*11

USNM 201010*11

 

BM 39.401*11

MSU 35853*11

USNM 239161*11

 

BM 39.402*11

MSU 35854*11

USNM 357384*11

 

BM 39.403*11

MSU 35856*11

USNM 367385*11

 

BM 39.404*11

MSU 35857*11

USNM 358502*11

 

 

BM 39.407*11

 

MSU 35858*11

 

USNM 429175*11

 

119

 

LITERATURE CITED

120

Allen, J.A. (1924). Carnivora collected by the American Museum Congo
Expedition. Bulletin of the American Museum of Natural History 47: 73—
281.

Allen, J.A. (1877). The influence of physical conditions in the genesis of species.
Radical Review 1: 108—140.

Andersone, Z., and Ozolins, J. (2000). Craniometrical characteristics and dental
anomalies in wolves Canis lupus from Latvia. Acta Theriologica 45: 549—
558.

Andersson, M. (1994). Sexual Selection. (Princeton, New Jersey: Princeton
University Press).

Arctander, P., Johansen, C., and Coutelle-Vreto, MA. (1999). Phylogeography of
three closely related African bovids (Tribe Alcelaphini). Molecular Biology
and Evolution 16: 1724—1739.

Barnett, R.J. (1977). Bergmann's rule and variation in structures related to
feeding the gray squirrel. Evolution 31: 538—545.

Baryshnikov, G. (1999). Chronological and geographical variability of Crocuta
spelaea (Carnivora, Hyaenidae) from the Pleistocene of Russia. Deinsea
6: 155—174.

Bergmann, C. (1847). Uber die Verhéltnisse der Warmedkonomie der Thiere zu
ihrer Grdsse. Gottinger Studien 3: 595—708.

Bookstein, FL (1996). Combining the tools of geometric morphometrics. In
Advances in morphometrics, Marcus, L.F., Corti, M., Loy, A., Naylor,
G.J.P., and Slice, D.E., eds (New York: Plenum Press), pp. 131-151.

Bookstein, FL. (1997). Landmark methods for forms without landmarks:
morphometrics of group differences in outline shape. Medical Image
Analysis 1: 225—243.

Bouissou, MP. (1983). Androgens, aggressive behaviour and social relationships
in higher mammals. Hormone Research 18: 43—61.

Boyce, MS. (1978). Climatic variability and body size variation in the muskrats
(Ondatra zibethicus) of North America. Oecologia 1—19.

Brett, RA. (1991). The population structure of naked mole-rat colonies. In The
Biology of the Naked Mole-Rat, Sherman, P.W., Jarvis, J.U.M., and
Alexander, R.D., eds (Princeton, New Jersey: Princeton University Press),
pp.97-136.

121

Brown, J.H., and Lee, AK (1969). Bergmann's rule and climatic adaptation in
woodrats (Neotoma). Evolution 23: 329-338.

Buchalczyk, T., Dynowski, J., and Szteyn, S. (1981). Variations in the number of
teeth and asymmetry of the skull in the wolf. Acta Theriologica 26: 23—30.

Cabrera, A. (1911). On the specimens of Spotted Hyaenas in the British Museum
(Natural History). Proceedings of the Zoological Society, London 1911:
93—99.

Canty, A.J. (1998). An S-Plus Library for Resampling Methods. 30th Symposium
on the Interface: Computing Science and Statistics 236-241.

Cardini, A., and Elton, S. (2009). Geographical and taxonomic influences on
cranial variation in red colobus monkeys (Primates, Colobinae):
introducing a new approach to 'morph‘ monkeys. Global Ecology and
Biogeography 18: 248-263.

Cardini, A., Jansson, A.-U., and Elton, S. (2007). A geometric morphometric
approach to the study of ecogeographical and clinal variation in vervet
monkeys. Journal of Biogeography 34: 1663—1678.

Castro, G., Myers, JP, and Ricklefs, RE. (1992). Ecology and Energetics of
Sandlerlings Migrating to Four Latitudes. Ecology 73: 833—844.

Clarke, FM, and Faulkes, CG. (1997). Dominance and queen succession in
captive colonies of the eusocial naked mole-rat, Heterocephalus glaber.
Proceedings of the Royal Society of London Series B-Biological Sciences
264: 993-1000.

Creel, S., and Creel, NM. (2002). The African wild dog: behavior, ecology, and
conservation.

D'Anatro, A., and Lessa, ER (2006). Geometric morphometric analysis of
geographic variation in the Rio Negro tuco-tuco, Ctenomys n’onegrensis
(Rodentia: Ctenoymidae). Mammalian Biology 71: 288-298.

D'Souza, RN, and Klein, 0D. (2007). Unraveling the molecular mechanisms
that lead to supernumerary teeth in mice and men: current concepts and
novel approaches. Cells Tissues Organs 186: 60—69.

Darwin, C. (1871). The descent of man, and selection in relation to sex. (London:
J. Murray).

Dayan, T., Simberloff, D., Tchernov, E., and Yom-Tov, Y. (1991). Calibrating the
paleothermometer: climate, communities, and the evolution of size.
Paleobiology 17: 189—199.

122

Dobson, PS, and Wigginton, JD. (1996). Environmental influences on the
sexual dimorphism in body size of western bobcats. Oecologia 108: 610—
616.

Dubach, J. et al. (2005). Molecular genetic variation across the southern and
eastern geographic ranges of the African lion, Panthera leo. Conservation
Genetics 6: 15—24.

Efron, B. (1987). Better bootstrap confidence intervals (with Discussion). Journal
of the American Statistical Association 82: 171—200.

Endler, J.A. (1977). Geographic variation, speciation, and clines. (Princeton:
Princeton University Press).

Estes, RD. (1991). The behavior guide to African mammals. (Berkeley:
University of California Press).

Ewer, RF. (1954). Some adapative features in the dentition of hyaenas. Annals
and Magazine of Natural History 7: 188—194.

Fenton, M.B., Waterman, J.M., Roth, J.D., Lopez, E., and Fienberg, SE. (1998).
Tooth breakage and diet: a comparison of bats and carnivorans. Journal
of Zoology, London 246: 83—88.

Flagstad, O., Syversten, P.O., Stenseth, NC, and Jakobsen, KS. (2000).
Environmental change and rates of evolution: the phylogrographic pattern
within the hartebeest complex as related to climatic variation. Proceedings
of the Royal Society of London Series B-Biological Sciences 268: 667-
677.

Foster, JB. (1964). The evolution of mammals on islands. Nature 202: 234-235.

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

Frank, L.G. (1986b). Social organization of the spotted hyaena Crocuta crocuta.
ll. Dominance and reproduction. Animal Behaviour 34: 1510—1527.

Frank, L.G., Glickman, SE, and Powch, I. (1990). Sexual dimorphism in the
spotted hyaena (Crocuta crocuta). Journal of Zoology, London 221: 308—
313.

Freeman, A.R. et al. (2001). Sequence variation in the mitochondrial DNA control
region of wild African cheetahs (Acinonyx jubatus). Heredity 86: 355—362.

123

Frost, S.R., Marcus, L.F., Bookstein, F.L., Preddy, DP, and Delson, E. (2003).
Cranial allometry, phylogeography, and systematics of large-bodied
papionins (Primates: Cercopithecinae) inferred from geometric

morphometric analysis of landmark data. The Anatomical Record Part A
275A: 1048—1072.

Fuentes, ER, and Jaksic, FM. (1979). Latitudinal size variation of Chilean
foxes: tests of alternative hypotheses. Ecology 60: 43—47.

Gannon, W.L., Sikes, RS, and Comm, A.C.U. (2007). Guidelines of the

Amerlcan Society of Mammalogists for the use of wild mammals in
research. Journal of Mammalogy 88: 809—823.

Gay, SW, and Best, TL. (1996). Relationships between abiotic variables and
geographic variation in skulls of pumas (Puma concolor. Mammalia

Felldea) In North and South America. Zoological Journal of the Linnean
Society 117: 259—282.

Geist, V. (1987). Bergmann's rule is invalid. Canadian Journal of Zoology 65
1035-1038.

Glrman, D. J. et al. (2001). Patterns of population subdivision, gene flow and

genetic variability in the African wild dog (Lycaon pictus). Molecular
Ecology 10: 1703—1723.

Glrman, DJ. et al. (1993). Molecular genetic and morphological analyses of the
African wild dog (Lycaon pictus). Journal of Heredity 84: 450—459.

Gisburne, T.J., and Feldhamer, GA. (2005). Dental anomalies in the gray fox

Urocyon cmereoargenteus and the red fox Vulpes vulpes. Acta
Theriologica 50: 515-520.

Glickman, SE. (1995). The spotted hyena from Aristotle to the Lion King
reputation is everything. Social Research 62: 501—537.

Glickman, S.E., Frank, L.G., Pavgi, S., and Licht, P. (1992). Hormonal correlates

of mascullnlzatlon' in female spotted hyaenas (Crocuta crocuta). 1.
Infancy to sexual maturity. Journal of Reproduction and Fertility 95: 451-
462.

Gloger, CL. (1833). Das Abéndern der Vogel durch Einfluss des Klimas
(Breslau: August Schulz).

Glucksmann, A. (1974). Sexual dimorphism in mammals. Biological Reviews 49
423-475.

Gould, S.J., and Johnston, RF. (1972). Geographic variation. Annual Review of
Ecology and Systematics 2: 457-498.

124

Hall, B.K. (1984). Developmental mechnisms underlying the formation of
atavisms. Biological Review 59: 89-124.

Hall, ER. (1940). Supernumerary and missing teeth in wild mammals of the
orders lnsectivora and Carnivora, with some notes on disease. Journal of
Dental Research 19: 103-143.

Hamilton, W.J., Tilson, R.L., and Frank, LG. (1986). Sexual monomorphism in
Spotted Hyenas, Crocuta crocuta. Ethology 71: 63—73.

Hauer, S. (2002). Population analysis of dental anomalies of otters Lutra lutra
from eastern Germany. Acta Theriologica 47: 339-362.

Heller, E. (1910). Four new subspecies of large mammals from equatorial Africa.
Smithsonian Miscellaneous Collections 61: 1—7.

Hendey, QB. (1974). Late Cenozoic Carnivora of the south-western Cape
Province. Annals of the South African Museum 63: 1—369.

Henschel, JR. (1986). The socio-ecology of a spotted hyaena Crocuta crocuta
clan in the Kruger National Park. Zoology D.Sc.: 215.

Herring, SW. (1993). Epigenetic and functional influences on skull growth. In
The skull: development, Hanken, J., and Hall, B.K., eds (Chicago, IL:
University of Chicago Press), pp. 153-206.

Hofer, H., and East, ML. (1993). The commuting system of Serengeti spotted
hyaenas: how a predator copes with migratory prey. l. Social organization.
Animal Behaviour 46: 547-557.

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

Holekamp, KB, and Smale, L. (1991). Dominance acquisition during mammalian
social development: the "inheritance" of maternal rank. American
Zoologist 31: 306—317.

Holm, S. (1979). A simple sequentially rejective multiple test procedure.
Scandinavian Journal of Statistics 6: 65—70.

Hood, CS. (2000). Geometric morphometric approaches to the study of sexual
size dimorphism in mammals. Hystrix 11: 77-90.

James, FC (1970). Geographic size variation in birds and its relationship to
climate. Ecology 51: 365—390.

125

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

Kappeler, PM. (1991). Patterns of sexual dimorphism in body weight among
prosimian primates. Folia Primatologica 57: 132—146.

Kappeler, PM. (1993). Female dominance in primates and other mammals. In
Behavior and Evolution, Bateson, P.P.G., Klopfer, PH, and Thompson,
N.S., eds (New York: Plenum Press), pp. 143-158.

Kennedy, M.L., and Lindsay, S.L. (1984). Morphologic variation in the raccoon,
Procyon lotor, and its relationship to genic and environmental variation.
Journal of Mammalogy 65: 195-205.

Kingdon, J. (1977). Carnivores. (London: Academic Press).

Klein, R.G., and Scott, K. (1989). Glacial/interglacial size variation in fossil
spotted hyenas (Crocuta crocuta) from Britain. Quanternary Research 32:
88—95.

Klein, RC. (1986). Carnivore size and quaternary climatic change in southern
Africa. Quaternary Research 26: 153-170.

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

Kurten, B. (1956). The status and affinities of Hyaena sinensis Owen and
Hyaena ultima Matsumoto. American Museum Novitates 1764: 1-48.

Kurten, B., and Poulianos, AM (1977). New stratigraphic and faunal material
from Petralona cave. Anthropos 47—103.

Langlais, RP, and Miller, CS. (2002). Color atlas of common oral diseases.
Lippincott, Williams, and Wilkins).

Le Houerou, H.N. (1984). Rain use efficiency: a unifying concept in arid-land
ecology. Journal of Arid Environments 7: 213—247.

Leigh, SR, and Cheverud, J.M. (1991). Sexual dimorphism in the baboon facial
skeleton. American Journal of Physical Anthropology 84: 193—208. _

Lindeque, M., and Skinner, JD (1982). Aseasonal breeding in the spotted
hyaena (Crocuta crocuta, Erxleben), in southern Africa. African Journal of
Ecology 20: 271-278.

126

Lumsden, A.G.S. (1981). The comparative anatomy of dentition: the orders of
placental mammals. In Dental anatomy and embryology, Osborn, J.W., ed
(Oxford: Blackwell Scientific), pp. 399—419.

Mantel, N. (1967). The detection of disease clustering and a generalized
regression approach. Cancer Research 27: 209—220.

Manville, RH. (1963). Dental anomalies in North American lynx. Zeitschrift fur
Saugetierkunde 28: 166—169.

Matschie, P. (1900). Geographische formen der Hyaenen. Gesellschaft
naturforschender F reunde 18—58.

Matthews, L.H. (1939a). Reproduction in the spotted hyaena, Crocuta crocuta
(Erxleben). Philosophical Transactions of the Royal Society of London,
Series B 230: 1-78.

Matthews, L.H. (1939b). The subspecies and variation of the spotted hyaena,
Crocuta crocuta Exrl. Proceedings of the Zoological Society of London,
Series B 109: 237-260.

Mayr, E. (1954). Change of genetic environment and evolution. In Evolution as a
Process, Huxley, J., Hardy, AC, and Ford, E.B., eds (London: Allen and
Unwin), pp. 157—180.

McNab, B.K. (1963). Bioenergetics and the determination of home range size.
American Naturalist 97: 133—140.

McNab, B.K. (1971). On the ecological significance of Bergmann's rule. Ecology
52: 845—854.

Meester, J.A.J., Rautenbach, l.L., Dippenaar, NJ, and Baker, CM. (1986).
Classification of southern African mammals. Transvaal Museum
Monograph 5: 122-123.

Meiri, 8., Guy, 0., Dayan, T., and Simberloff, D. (2009). Global change and
carnivore body size: data are stasis. Global Ecology and Biogeography
18: 240—247.

Meiri, S., Yom-Tov, Y., and Geffen, E. (2007). What determines conformity to
Bergmann's Rule? Global Ecology and Biogeography 16: 788-794.

Miles, A.E.W., and Grigson, C. (1990). Colyer's variations and diseases of the
teeth of animals. (Cambridge: Cambridge University Press).

Millien, V., Lyons, S.K., Smith, PA, leon, AB, and Yom-Tov, Y. (2006).
Ecotypic variation in the context of global climate change: revisiting the
rules. Ecology Letters 9: 853-869.

127

Mills, M.G.L. (1984). The comparative behavioural ecology of the brown hyaena
Hyeana brunnea and the spotted hyaena Crocuta crocuta in the southern
Kalahari. Koedoe (Supplement) 237-247.

Mills, M.G.L. (1990). Kalahari Hyaenas. (London: Unwin Hyman).

Mitchell, TD, and Jones, PD. (2005). An improved method of constructing a
database of monthly climate observations and associated high-resolution
grids. International Journal of Climatology 25: 693—712.

Muwanika, V.B., Nyakaana, S., Siegismund, HR, and Arctander, P. (2003).
Phylogeography and population structure of the common warthog
(Phacochoerus afn'canus) inferred from variation in mitochondrial DNA
sequences and microsatellite loci. Heredity 91: 361-372.

Neaves, W.B., Griffin, J.E., and Vlfilson, JD. (1980). Sexual dimorphism of the

phallus in spotted hyaena (Crocuta crocuta). Journal of Reproduction and
Fertility 59: 509—513.

Nentvichova, M., and Andera, M. (2008). Dental anomalies and dental variations

in the red fox Vulpes vulpes in teh Czech Republic. Acta Theriologica 53:
217-228. ,

Paradiso, J. (1966). Notes on the supernumerary and missing teeth in the
coyote. Mammalia 30: 120—128.

Qiu, Z.-X. (1987). Die Hyaeniden aus dem Ruscinium und Villafranchium Chinas.
Munchner Geowissenschaftliche Abhandlungen 9: 1—108.

R Development Core Team (2009). R: A language and environment for statistical

computing. Version 2.8.1. (Vienna, Austria: R Foundation for Statistical
Computing).

Raia, P., and Meiri, S. (2006). The island rule in large mammals: paleontology
meets ecology. Evolution 60: 1731-1742.

Ralls, K. (1976). Mammals in which females are larger than males. Quarterly
Review of Biology 51: 245-276.

Ralls, K., and Harvey, PH. (1985). Geographic variation in size and sexual
dimorphism of North American weasels. Biological Journal of the Linnean
Society 25: 119—167.

Ritke, ME, and Kennedy, ML. (1993). Geographic variation of sexual

dimorphism in the raccoon Procyon lotor. American Midland Naturalist
129: 257—265.

128

Rohland, N. et al. (2005). The population history of extant and extinct hyenas.
Molecular Biology and Evolution 22: 2435—2443.

Rohlf, F.J. (2005). tpsDig version 2.10. (Stonybrook, New York: Department of
Ecology and Evolution, State University of New York at Stonybrook).

Rohlf, F.J., and Slice, D. (1990). Extensions of the Procrustes method for the
optimal superimposition of landmarks. Systematic Zoology 39: 40—59.

Rohlf, F .J., and Corti, M. (2000). Use of two-block partial least squares to study
covariation in shape. Systematic Biology 49: 740—753.

Rosenzweig, ML. (1968). Net primary productivity of terrestrial communities:
prediction from climatological data. American Naturalist 102: 67—74.

Scholander, PF. (1955). The evolution of climatic adaptation in homeotherms.
Evolution 9: 15-26.

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

Sheets, H.D. (2001). BigFix6. (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H.D. (2002). tmorphgen6. (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H. D. (2003a). SemiLand6. (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H. D. (2003b). Standard6. (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H.D. (2003c). TwoGroup6h. (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H.D. (2004). PLSMaker6g (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H.D. (2006). CoordGen6h. (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H.D. (2007). CVAGen6o. (Buffalo, New York: Department of Physics,
Canisius College).

Sheets, H.D. (2008). Regress6n. (Buffalo, New York: Department of Physics,
Canisius College).

129

Sikes, RS, and Kennedy, ML. (1992). Morphologic variation of the bobcat (Felis
rufus) in the eastern United States and its association with selected
environmental variables. American Midland Naturalist 128: 313—324.

Sillero-Zubiri, C., and Gottelli, D. (1992a). Feeding ecology of spotted hyaena
(Mammalia: Crocuta crocuta) in a mountain forest habitat. Journal of
African Ecology 106: 169—176.

Sillero-Zubiri, C., and Gottelli, D. (1992b). Population ecology of spotted hyaena
in an equatorial mountain forest. African Journal of Ecology 30: 292-300.

Skinner, JD. (1976). Ecology of the brown hyaena Hyaena brunnea in the
Transvaal with a distribution map of southern Africa. South African Journal
of Science 72: 262—269.

Solounias, N., and de Beaumont, G. (1981). Brief remarks on certain large-sized
Turolian hyaenids (Mammals, Carnivores) from Pikermi and Samos,
Greece. Archives des Sciences, Geneve 34: 293—3-4.

Szuma, E. (1999). Dental abnormalities in the red fox Vulpes vulpes from Poland.
Acta Theriologica 44: 393—412.

Tannenbaum, K.A., and Alling, EB (1963). Anomalous tooth development, case
reports of gemination and twinning. Oral Surgery, Oral Medicine, and Oral
Pathology 16: 883—887.

Tanner, JB (2007). Behavioral and morphological development in a female-
dominated species, Crocuta crocuta. Zoology PhD: 148.

Tanner, J.B., Dumont, E.R., Sakai, S.T., Lundrigan, BL, and Holekamp, K.E.
(2008). Of arcs and vaults: the biomechanics of bone-cracking in spotted
hyenas (Crocuta crocuta). Biological Journal of the Linnean Society 95:
246—255.

Tanner, J.B., Zelditch, M.L., Lundrigan, BL, and Holekamp, K.E. (2009).
Ontogenetic change in skull morphology and mechanical advantage in the
spotted hyena (Crocuta crocuta). Journal of Morphology. In press.

Teplitsky, C., Mills, J.A., Alho, J.S., Yarral, J.W., and Merila, J. (2008).
Bergmann's rule and climate change revisited: disentangling
environmental and genetic responses in a wild bird population.
Proceedings of the National Academy of Sciences of the United States of
America 105: 13492-13496.

Tilson, R.L., and Henschel, JR. (1986). Spatial arrangement of spotted hyaena
groups in a desert environment, Namibia. African Journal of Ecology 24:
173-180.

130

Trinkel, M., Fleischmann, P.H., Steindorfer, AF, and Kastberger, G. (2004).
Spotted hyenas (Crocuta crocuta) follow migratory prey. Seasonal
expansion of clan territry in Etosha, Namibia. Journal of Zoology 264:
125—133.

Trivers, R.L. (1972). Parental investment and sexual selection. In Sexual
selection and the descent of man, Campbell, 3., ed (Chicago: Aldine),

Turpie, J.K., and Crowe, TM. (1994). Patterns of distribution, diversity and
endemism of larger African mammals. South African Journal of Zoology
29: 19—31.

Van Horn, R.C., McElhinny, TL, and Holekamp, K.E. (2003). Age estimation and
dispersal in the spotted hyena. Journal of Mammalogy 84: 1019—1030.

van Jaarsveld, A.S., Skinner, JD, and Lindeque, M. (1988). Growth,
development and parental investment in the spotted hyaena, Crocuta
crocuta. Journal of Zoology, London 216: 45—53.

Van Valkenburgh, B. (2009). Costs of carnivory: tooth fracture in Pleistocene and
Recent carnivorans. Biological Journal of the Linnean Society 96: 68—81.

Van Valkenburgh, B., and Hertel, F. (1993). Tough times at La Brea: tooth
breakage in large carnivores of the late Pleistocene. Science 261: 456-
459.

Van Valkenburgh, B. (1988). Incidence of tooth breakage among large, predatory
mammals. American Naturalist 131: 291—302.

Van Valkenburgh, B. (1996). Feeding behavior in free-ranging, large African
carnivores. Journal of Mammalogy 77: 240—254.

Verstraete, F.J.M. (1985). Anomalous development of the upper third premolar in
a dog and a cat. Journal of the South African Veterinary Association 56:
131 —1 34.

Vila, C., Urios, V., and Castroviejo, J. (1993). Tooth losses and anomalies in the
wolf (Canis lupus). Canadian Journal of Zoology 71: 968—971.

Weckerly, PW. (1998). Sexual-size dimorphism: influence of mass and mating
systems in the most dimorphic mammals. Journal of Mammalogy 79: 33-
42.

Werdelin, L. (1988). Studies of fossil hyaenas: the genera Thalassictis Gervais
ex Nordmann, Palhyaena Gervais, Hyaenictitherium Kretzoi, Lycyaena
Hensel and Palinhyaena Qiu, Huang and Guo. Zoological Journal of the
Linnean Society 92: 211—265.

131

Werdelin, L., and Solounias, N. (1990). Studies of fossil hyaenids: the genus
Adrocuta Kretzoi and the interrelationships of some hyaenid taxa.
Zoological Journal of the Linnean Society 98: 363-386.

Werdelin, L. (2008). Biogeographic relationships of African carnivoran faunas 7-
1.2 Ma. Comptes Rendus Palevol 7: 645-656.

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

Werdelin, L., and Solounias, N. (1996). The evolutionary history of hyaenas in
Europe and Western Asia during the Miocene. In The evolution of Western
Eurasian Neogene mammal faunas, Bernor, R.L., Fahlbusch, V., and
Mittman, H.-W., eds (New York: Columbia University Press), pp. 291—306.

Werdelin, L., and Solounias, N. (1991). The Hyaenidae: taxonomy, systematics
and evolution. Fossils and Strata 30: 1-104.

Whateley, A. (1980). Comparative measurements of male and female spotted
hyaenas from Natal. The Lammergeyer 28: 40—43.

Whateley, A., and Brooks, PM. (1978). Numbers and movements of spotted
hyaenas in Hluhluwe Game Reserve. The Lammergeyer 26: 44-52.

Wigginton, JD, and Dobson, PS. (1999). Environmental influences on
geographic variation in body size of western bobcats. Canadian Journal of
Zoology 77: 802—813.

Wolsan, M. (1984). Two dental anomalies in the weasel (Mustela nivalis L): a
supernumerary premolar and a reduced upper molar. Zoologiche
Abhandlungen, Staatliches Museum fur Tierkunde, Dresden 40: 67—70.

Wolsan, M. (1989). Dental polymorphism in the genus Martes (Carnivora:
Mustelidae) and its evolutionary significance. Acta Theriologica 34: 545—
593.

Wozencraft, WC. (1989). The phylogeny of the recent Carnivora. In Carnivore
behavior, ecology, and evolution, Gittleman, J.L., ed (Ithaca: Comstock
Publishing Associates).

Wozencraft, WC. (2005). Order Carnivora. In Mammal species of the world: a
taxonomic and geographic reference, Wllson, DE, and Reeder, D.M., eds
(Baltimore: Johns Hopkins University Press), pp. 532—628.

Yom-Tov, Y., and Geffen, E. (2006). Geographic variation in body size: the
effects of ambient temperature and precipitation. Oecologia 148: 213—218.

132

Zelditch, M.L., Swiderski, D.L., Sheets, H.D., and Fink, W.L. (2004). Geometric
morphometrics for biologists: a primer. (San Diego: Elsevier Academic
Press).

Ziegler, AC. (1971). A theory of the evolution of therian dental formulas and
replacement patterns. The Quarterly Review of Biology 46: 226—249.

133

 

IIIIIIIII

I

7
“5
4
o
3
8
o
3
o
3
9
2
1
3

 

"
I
T“
"
||
I
I
TI|||
“-
"ll‘ll
ill"
HI