.ammvsmru; 2.”:
.. ma.
.

 

I:
I

in?

 

      

 

2,! ,
353%:

it
i

 

l ,

‘ (,5 1‘ "I
x , /'
6%5957?”

This is to certify that the
dissertation entitled

SOCIOENDOCRINOLOGY OF SPOTTED HYENAS:
PATTERNS OF ANDROGEN AND GLUCOCORTICOID
EXCRETION WITHIN A UNIQUE SOCIAL SYSTEM

presented by
Stephanie Marie Dloniak

has been accepted towards fulfillment
of the requirements for the

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

 

 

  
 

% Major rofessor’s S' ure

Date

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

v v -- V ~.—— v-v T—

 

 

 

MW]

MIChIQan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6’01 c-JCIRC/DatoDuopss-sz

PATTERNS

SOCIOENDOCRINOLOGY OF SPOTTED HYENAS:
PATTERNS OF ANDROGEN AND GLUCOCORTICOID EXCRETION WITHIN
A UNIQUE SOCIAL SYSTEM
By

Stephanie Marie Dloniak

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Zoology
Program in Ecology, Evolutionary Biology and Behavior

2004

PATTERNS

 

Sodoe
animal's socia
Here, l exam'II
Crocuta CfOCL
thGSpecies.\I
0f Spotted hyI
9|U00C0rticoic
hyenas.lthe
hYena socioe
interseanl in
aggIBSSil/e Ct
be“ Predicto
interaction wi
vafiauonin aI
inlelSeanl in

FerTlal

ABSTRACT
SOCIOENDOCRINOLOGY OF SPOTTED HYENAS:
PATTERNS OF ANDROGEN AND GLUCOCORTICOID EXCRETION WITHIN A
UNIQUE SOCIAL SYSTEM
By

Stephanie Marie Dloniak

Socioendocrinology is specifically concerned with the effects of an
animal’s social environment on the interactions between hormones and behavior.
Here, I examine how unique aspects of the social system of the spotted hyena,
Crocuta crocuta, are mediated by hormone-behavior interactions in individuals of
the species. Working with data collected over 10 years from a free-ranging clan
of spotted hyenas in Kenya, l validate methods to quantify androgen and
glucocorticoid concentrations in fecal samples collected from known individual
hyenas. I then use these techniques to address three issues central to spotted
hyena socioendocrinology. I first investigate the relationships among intra- and
intersexual interactions on male androgen levels. In other polygynous species,
aggressive competition causes elevated androgen levels in males. I find that the
best predictor of increased androgens in male spotted hyenas is the degree of
interaction with dominant females. This suggests that attempts to explain
variation in androgen-behavior relationships in male vertebrates should consider
intersexual interactions as well as male-male competition.

Female spotted hyenas are both morphologically and behaviorally

masculinized. I test whether maternal effects of androgens are involved in

behavioral In
have greater
ranking moth
rates of aggn
nongenetio,
organization
deposition of
Finalli
show that tir
tor. Social 5
hyenas of e
glucocortict
i"190mm 5
”Mid 3m

SOCIBI Stair

behavioral masculinization in this species. I find that mothers of high social rank
have greater fecal androgen concentrations during pregnancy than do low-
ranking mothers. This variation in maternal androgens is related to variation in
rates of aggression and sexual play in offspring. These results indicate a possible
non-genetic, androgen-mediated effect of a mother’s social environment on the
organization of behavioral patterns in hyenas, similar to the effect of maternal
deposition of androgens in bird eggs on chick development and behavior.
Finally, I examine predictors of glucocorticoid concentrations in hyenas. I
Show that time of day of sample is an important covariate that must be controlled
for. Social status is not a predictor of glucocorticoid concentrations in adult
hyenas of either sex. In female hyenas, reproductive status influences
glucocorticoid concentrations, whereas in males, prey availability has an
important effect. Therefore, different physiological and ecological variables
predict stress hormone concentrations in males and females of this species, and

social status per se does not appear to be an important variable.

Copyright by
STEPHANIE MARIE DLONIAK
2004

Perh
of the wonde
made possih
the US and r
Holekamp, f
the Masai Mi
and has alw;
Dmlecl has I
Worked with
Committee r
times when
comIOIIable
himself, In ,
literally clirr
Sisk and L;
thank them

My r
QladUate C;
Packages, .
My dad‘ Fr;

When I ”Ge

ACKNOWLEDGMENTS

Perhaps the best part of writing this dissertation is being able to thank all
of the wonderful individuals who have helped me complete it. This research was
made possible by help and encouragement from many different people both in
the US and in Kenya. First and foremost, I would like to thank my advisor, Kay
Holekamp, for giving me the opportunity to work in her lab and to live and work in
the Masai Mara for almost two years. Kay has been extremely patient with me
and has always challenged me to live up to my potential. Being a part of her
project has made me a better person, and I consider myself very lucky to have
worked with her. I am also particularty grateful to Jeff French, one of my
committee members. Jeff and his family welcomed me into their home numerous
times when I traveled to Omaha to do lab work, which made those visits
comfortable and fun. Jeff even made a trip to Kenya to do a bit of field work
himself, in addition to a climb up Mt. Kenya. Not many graduate students get to
literally climb a mountain with a committee member! Committee members Cheryl
Sisk and Laura Smale were always available to discuss data and ideas, and I
thank them as well.

My family and friends have given me unconditional support throughout my
graduate career. My mom, Sue Matz, has sent me about a hundred care
packages, as well as countless positive thoughts, in both Michigan and in Kenya.
My dad, Frank Dloniak, has always sent cheesecake and has never let me down

when I needed him. Over the years, my brother, Brian, provided many breaks

 

horn work a
working with
water, but ti
work hard. t
can say. Th'
though we
of tent-time,
Moose alwa
years in Ke
Man
Michigan 3
Bagel men
were all m;
OIfiCemate
With my “q
My labmai
Other rant:
Bills, Erin
Croix’ S at
Hom' Pag
their llien,

illVOlved iI

from work and outlets from the academic world. My family may not quite get why
working with hyena poop is fun or how I can be happy living without running
water, but they have always encouraged me to live the life of my choice and to
work hard. My very best friend, Howard Saunders, has done more than words
can say. The last few years have been the best of all because of him, even
though we have had to spend so much time apart. At least we have many years
of tent-time, birding, and playing in the African bush ahead of us. My buddy
Moose always welcomed me home, both after long days at school and after two
years in Kenya.

Many folks in Michigan provided great friendship. Upon my arrival at
Michigan State, Micaela Szykman and Corinne Vriesendorp were fast friends.
Bagel mornings, coffee breaks, and a few dips in Lake Michigan in December
were all made so much more fun because of them. Russ Van Horn was a great
officemate, colleague, and friend. I apologize for taking up so much of his time
with my “quick” questions and random outbursts of profanity during his last year.
My labmates were a special group of people who endured my statistical and
other rants, and who always had interesting suggestions and help. I thank Pat
Bills, Erin Boydston, Anne Engh, Keron Greene, Joe Kolowski, Suzanne La
Croix, Sarah Lansing, Micaela Syzkman, Jaime Tanner, Kevin Theis, Russ Van
Horn, Page Van Meter, Eva Maria Mueke, Sofi Wahaj, and Heather Watts for
their friendship. Matt Heintz assisted with some of the more tedious work

involved in my dissertation.

vi

 

l was'.
conducting
hormone as
Place, Stev-
challenges I
Berkeley Fie

crucial for m

 

mmwki
Dave Kerse:
Smilhsoniar
DIOVided va!

Num
Pelis Keshe
lime. Josep
in particular
me to the h
latht m e l
the few 8hr
and Sofi Vt-
Up IOr bUsl
Andrew he

bUSheS Wi‘

Simp'y Dro

l was fortunate to collaborate with some excellent scientists while
conducting my research. Jeff French taught me how to do fecal extractions and
hormone assays in his laboratory, and assisted with most of my lab work. Ned
Place, Steve Glickman, and Mary Weldele graciously performed hormone
challenges on the hyenas in the captive colony at the University of California
Berkeley Field Station. Mary collected many hyena fecal samples that were
crucial for my research, and Ned and Steve also provided valuable feedback on
my results. I thank Peggy Ostrom for allowing me to use the Iyophilizer in her lab.
Dave Kersey and Janine Brown, of the Conservation and Research Center of the
Smithsonian Institution, completed the HPLC analyses in Chapter 2, and Janine
provided valuable insight in the interpretation of HPLC results.

Numerous people gave me assistance and friendship in Kenya. John and
Peris Keshe and Moses Sairowa made fisi camp feel like luxury-living most of the
time. Joseph Silantoi and Pilot Nairuori were great night watchmen, and Joseph
in particular always looked out for me. Anne Engh and Keith Nelson introduced
me to the hyenas and life in fisi camp. Anne and Micaela Syzkman both patiently
taught me how to ID hyenas, record hyena behavior, and “check” the cars during
the few short months I spent with them in the field. Jaime Tanner, Joe Kolowski,
and Sofi Wahaj were all great co-workers in the field. Andrew Peart was always
up for bush golf, frisbee, slingshot practice, or explaining the game of cricket.
Andrew helped me many times, by scrambling around the muddy riverbank and
bushes with me looking for hyena dens, by pulling cars out of the mud, and by

simply providing access to the Fig Tree swimming pool and phone. Milton and

vii

Elly Kirkma
visits to fist
heard them
always gaw
Camp. Nas
routine, an
Walpole w.-
was an ext
always an
Watkin we
thank him
was my “I
logistical
Others he
NL
Rebecca
lsla Gral»
N6lson, I
HOm, S C
Gollecuns
'ESearch
F i

Clearanc

Elly Kirkman always provided a hot shower and a break from fisi camp, and their
visits to fisi camp on the pikipiki were always greeted with excitement when we
heard them coming from over a kilometer away. Lars, Tina, Jonas, and Kristofer
always gave me non-vegetarian breakfasts and great conversation at Dream
Camp. Nasha Silantoi was a wonderful and inspirational break from my daily
routine, and our conversations Often reminded me of what’s important in life. Matt
Walpole was always fun to discuss conservation and life issues with. Ian McRae
was an extremely helpful mechanic and friend. Anthony and Fiona Cheffings
always entertained me and allowed me to vent frustrations in Nairobi. John
Watkin was a good friend, a great listener, and a wonderfully bad influence. I also
thank him for sending me random car parts on flights to the Mara. Lastly, Simo
was my “nurse" when I had malaria, and was also the absolute best source of
logistical advice and vehicle help a girl in the bush could ever want. Countless
others helped me out at various times, and I thank them all.

Numerous folks collected data used in this dissertation project, including
Rebecca Bankson, Erin Boydston, Martin Durham, Anne Engh, Paula Garrett,
lsla Graham, Keron Greene, Tyson Harty, Kay Holekamp, Joe Kolowski, Keith
Nelson, Gabe Ording, Laura Smale, Micaela Szykman, Jaime Tanner, Russ Van
Horn, Sofi Wahaj, Heather Watts, and Kim Weibel. I thank them all for diligently
collecting hyena fecal samples and behavioral Observations over the years. This
research project would not have been possible without their efforts.

Finally, I would like to thank all of the organizations that provided

clearance and funding for this research. The Office of the President of Kenya, the

viii

 

 

Kenya Wilc
Masai Mara
in the Mara-
Hyena Pro;
IBN 01131 1
DepartmenI

Foundation

research.

 

Kenya Wildlife Service, the Narok County Council, and the Senior Warden of the
Masai Mara National Reserve gave me permission to live and conduct research
in the Mara. During the period of time encompassing my research, the Mara
Hyena Project was supported by NSF grants IBN 9630667, IBN 9906445, and
IBN 0113170. A Michigan State University Distinguished Fellowship, the
Department of Zoology, and a graduate fellowship from the International
Foundation for Ethical Research provided funding for my graduate studies and

research.

ix

 

 

LIST OF Ti
LIST OF Ft?

CHAPTER
GENERAL

CHAPTER

NON-INVA '

SPOTTED '
Intro
Met
Res
Disc

CHAPTER
FECAL AN
INTERACT
IntroI
Meth
ResI
DISC!

CHAPTER
ANDROGE
BEHAVIOF
Intro
Metl
Res

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. xi

LIST OF FIGURES ............................................................................... xii

CHAPTER 1:

GENERAL INTRODUCTION .................................................................... 1

CHAPTER 2:

NON-INVASIVE MONITORING OF FECAL ANDROGEN EXCRETION IN

SPOTTED HYENA ................................................................................ 15
Introduction ................................................................................ 1 5
Methods .................................................................................... 19
Results ...................................................................................... 28
Discussion ................................................................................. 34

CHAPTER 3:

FECAL ANDROGEN LEVELS IN ADULT MALE SPOTTED HYENAS REFLECT
INTERACTIONS WITH SOCIALLY DOMINANT FEMALES............................42

Introduction ................................................................................ 42

Methods .................................................................................... 47

Results ..................................................................................... 55

Discussion ................................................................................. 62
CHAPTER 4:

ANDROGENS DURING PREGNANCY IN WILD SPOTTED HYENAS: THE
INFLUENCE OF SOCIAL RANK AND MATERNAL EFFECTS ON CUB

BEHAVIOR ......................................................................................... 70
Introduction ................................................................................ 70
Methods .................................................................................... 75
Results ..................................................................................... 81
Discussion ................................................................................. 89

CHAPTER 5:

CORRELATES OF FECAL GLUCOCRTICOID EXCRETION IN WILD ADULT

SPOTTED HYENAS ........................................................................... 102
Introduction .............................................................................. 1 02
Methods .................................................................................. 108
Results .................................................................................... 116
Discussion ............................................................................... 129

CHAPTER 6:

CONCLUSIONS ................................................................................. 138

LITERATURE CITED .......................................................................... 148

 

Table 2.1.
captive sp .
expressed

Table 4.1.
relation to I'
hyenas. V

Table 5.1. I
captive sp
are expres
challenge,

Table 5.2. l
male Talek

Table 5.3.
htenas. W

LIST OF TABLES

Table 2.1. Individual and mean response patterns to LHRH challenge in five
captive Spotted hyenas. Androgen values indicate mean i S.e.m. and are
expressed as ng androgen per g Iyophilized feces.

Table 4.1. Mean t SE fecal androgen and glucocorticoid concentrations in
relation to litter composition during the second half of gestation in spotted
hyenas. Values are given as ng/g feces.

Table 5.1. Individual and mean response patterns to ACTH challenge in five
captive spotted hyenas. Fecal glucocorticoid values indicate mean i- SEM and
are expressed as ng glucocorticoid per g Iyophilized feces. NR = no response to
challenge.

Table 5.2. Best model of fecal glucocorticoid concentrations in adult immigrant
male Talek hyenas. Whole model R2 = 0.30, F5103 = 7.37, p < 0.001.

Table 5.3. Best model of fecal glucocorticoid concentrations in adult female Talek
hyenas. Whole model R2 = 0.11, F5334 = 6.66, p < 0.001.

xi

 

 

Figure 2.1
male (A) a\
testosteror
as counts-
EIA expre. _‘

 
 
 
 
   

Figure 2.2.
in captive

Figure 2.3.
male (A) a

FlSure 3.1.
adult male
males and
adult male .

Figure 3.2.
males and
at the time
Iaeca| Sam

Figure 3.3.
immigiant I

LIST OF FIGURES

Figure 2.1. HPLC profiles of immunoreactive androgen metabolites in separate
male (A) and pregnant female (B) fecal extract pools. Elution of 3H-Iabeled
testosteronelT], androstenedione[A4], and dihydrotestosterone[DHT] expressed
as counts-per-minute per fraction, and immunoreactive androgens measured by
EIA expressed as pg irT per well.

Figure 2.2. Changes in fecal androgen excretion in response to LHRH challenge
in captive spotted hyenas. LHRH was administered on day 0.

Figure 2.3. Relationship between plasma and fecal hormone concentrations in
male (A) and female (B) wild hyenas. ANM = Adult Natal Males; IMM =
Immigrant Adult Males; Lact Fern = Lactating, Nonpregnant Adult Females; Preg
Fem = Pregnant Adult Females.

Figure 3.1. Paired rates of reproductive aggression and courtship Observed in
adult male Talek hyenas from May 2000 through May 2002. N = 6 adult natal
males and 12 adult immigrant males. Rates of behavior were compared within
adult male category.

Figure 3.2. Relationships between fecal androgen concentrations in immigrant
males and their tenure in the Talek clan (A) and their standardized social rank (B)
at the time of sampling. The highest social rank is indicated as 1.0. N = 159
faecal samples from 26 immigrant male hyenas.

Figure 3.3. Differences in fecal androgen concentrations between samples when
immigrant males did and did not emit particular behaviors. (A) N = 13, non-
reproductive aggression (Agg) and no non-reproductive aggression (No Agg), (B)
N = 6, reproductive aggression (Reprod Agg) and no reproductive aggression
(No Reprod Agg), (C) N = 16, courtship (Court) and no courtship (No Court).
Males are paired between bars within each histogram. Note that 5 of 6 males
showing reproductive aggression (B) were also Observed courting a female. No
males showing courtship (C) also showed reproductive aggression. Asterisks
indicate significant differences. '

Figure 3.4. Fecal androgen concentrations in immigrant males are positively
correlated with their association index (AI) with attractive females (A), but not
with lactating females (B). N = 69 fecal samples from 22 immigrant male hyenas.

Figure 4.1. Relationships among maternal social rank, day of gestation, and fecal
androgen concentrations in pregnant female hyenas. By convention, the highest
social rank possible is 1.

Figure 4.2. Relationships among age of mother, day of gestation, and fecal
glucocorticoid concentrations in pregnant female Spotted hyenas.

xii

 

”El

Figure 4.3
of gestatio
gestation.

figure 4.4
day of ges
gestation.

Figure 4.5.
male (A) a
were 2-6

Figure 4.6.
in male (A)
were 243 11

Figure 4.7.
directed by
Period whe

FiQure 5.1
Challenge
indicated I

Figure 5.2
QIUCOCOIIII

Flsure 5,3
QIUCOCOTITI

Figure 5.4
coneentra
availability
prey avail;
. Sathpl
"ldlcate III

”9”“? 5.5
concentra
and Unk :
whereas F
bars Indie,

Figure 5-6
lactatin (‘

Figure 4.3. Relationships between fecal androgens and social rank (A) and day
of gestation (B) in pregnant female spotted hyenas during the second half of
gestation.

Figure 4.4. Relationships between fecal glucocorticoids and social rank (A) and
day of gestation (B) in pregnant female spotted hyenas during the second half of
gestation.

Figure 4.5. Relationships between maternal androgens and mounting rates of
male (A) and female (B) cubs. Rates were calculated for the period when cubs
were 2-6 months old. Note different scales of y axes.

Figure 4.6. Relationships between maternal androgens and cub aggression rates
in male (A) and female (B) cubs. Rates were calculated for the period when cubs
were 2-6 months old.

Figure 4.7. Relationship between maternal androgens and rates of aggression
directed by dominant cubs towards their subordinate littermates, during the
period when cubs were 2-6 months of age.

Figure 5.1. Changes in fecal glucocorticoid excretion in response to ACTH
challenge in five captive spotted hyenas. ACTH was administered on day 0, as
indicated by the arrow.

Figure 5.2. Paired comparison of mean morning (AM) and afternoon (PM) fecal
glucocorticoid concentrations in 10 adult immigrant male hyenas.

Figure 5.3. Paired comparison of mean morning (AM) and afternoon (PM) fecal
glucocorticoid concentrations in 21 adult female hyenas.

Figure 5.4. The influence of prey availability on fecal glucocorticoid
concentrations in adult male spotted hyenas. LOW = lowest quartile of prey
availability. AVE = middle 2 quartiles of prey availability. HI = upper quartile of
prey availability. AM samples were collected between 0530 and 0900h, whereas
PM samples were collected between 1600 and 2000h. Numbers above error bars
indicate numbers of fecal samples in categories.

Figure 5.5. The influence of reproductive state on fecal glucocorticoid
concentrations in adult female spotted hyenas. Preg = pregnant, Lact = lactating,
and Unk = unknown. AM samples were collected between 0530 and 0900h,
whereas PM samples were collected between 1600 and 2000h. Numbers above
bars indicate numbers of fecal samples in categories.

Figure 5.6. Relationships between age and fecal glucocorticoid concentrations in
lactating (A) and pregnant (B) female spotted hyenas.

xiii

 

Be
behavior 0
and psych :
influence t
is influenc
social grou
external st
SDGCITICBII)
interaction
dissertatic
hyenas ((

QIOUDS.

W

Th

SOClal en
the repro

Ziegler, 1
behai’lor

SlsnmCar

CHAPTER 1

GENERAL INTRODUCTION

Behavioral endocrinology is the study of how hormones influence the
behavior of animals. It joins together the fields of endocrinology, neuroscience,
and psychology to ask questions about how changes in the endocrine system
influence the brain and behavior (Becker and Breedlove, 2002). Hormone action
is influenced by both internal and external stimuli, and for individuals living in
social groups, interactions with other group members can represent significant
external stimuli. Socioendocrinology is the branch of behavioral endocrinology
specifically concerned with the effects of an animal's social environment on the
interactions between hormones and behavior (Bercovitch and Ziegler, 1990). My
dissertation investigates various facets of the socioendocrinology of wild spotted
hyenas (Crocuta crocuta), a large carnivore that lives in large, stable social

groups.

Socioendocrinology

The goal of socioendocrinology is to understand the links among the
social environment, hormones, and behavior in order to know how they modulate
the reproductive success of individuals living in social groups (Bercovitch and
Ziegler, 1990). Studies that monitor variation in both endocrine physiology and
behavior are able to link the mechanisms mediating behavior with the functional
significance of behavior. Such studies therefore have the potential to increase

our understanding of the evolution of social organization, mating systems, and

'_

 

life history
a framew
physiolog
hwbda
Wh
cooperate
specific m
in which ll
determine
1997). Sc
beefing
monOpon
311d Fren
n'lediatec
0f Which
releasing
hipothai
hflehfizh
gonadot
Gauges ‘
inchdir1S
and Not

gonadot

life history strategies (Ketterson and Nolan, 1992). Socioendocrinology provides
a framework in which to connect evolutionary biology with reproductive
physiology, and thus bridges the traditional gap between proximate and ultimate
levels of analysis.

Where natural selection has favored individuals who live in groups and
cooperate with group-mates, elaborate social systems have evolved along with
specific mechanisms for maintaining order within social groups. In stable groups
in which individuals interact repeatedly over time, a dominance hierarchy often
determines priority of access to resources within each group (Pusey and Packer,
1997). Social dominance has a dramatic effect on reproduction in cooperatively
breeding species in that dominant members of such societies almost completely
monopolize reproduction (birds: Stacey and Koenig, 1990; mammals: Solomon
and French, 1986). The influence of dominance on reproduction may be
mediated by the hypothalamic—pituitary-gonadal (HPG) axis, the normal function
of which is necessary for reproduction. Within the HPG axis, gonadotropin
releasing hormone (GnRH) is secreted from neurosecretory cells in the
hypothalamus, and acts on the pituitary to regulate release of the gonadotropins,
luteinizing hormone (LH) and follicle stimulating hormone (FSH). The
gonadotropins then act on the gonad, where FSH affects gametogenesis and LH
causes the release of sex steroids. In males, the testes produce androgens
including testosterone (T), whereas in females the ovaries produce estrogens
and progestins (Norman and Litwack, 1997). Concentrations of the

gonadotropins and sex steroids may be directly affected by dominance status,

though relationships between dominance and hormone concentrations vary
across species, indicating that factors other than social dominance may interact
to influence hormones and reproduction. (Bercovitch and Clarke, 1995; Zielinski
and Vandenbergh, 1993).

In this dissertation, I investigate whether social status and other variables
influence the functioning of the HPG axis in a social carnivore exhibiting female
dominance over males. 1 inquire whether social status, male-male aggressive
competition, or intersexual interactions influence androgen concentrations in
male spotted hyenas. I also inquire whether social status influences androgen
levels during pregnancy in female hyenas, and whether variation in prenatal
androgen exposure corresponds to variation in sexual play and aggression in
cubs. Lastly, I test the relative importance of social dominance in the regulation
of stress hormone concentrations in this species.

One fundamental tenet of socioendocrinology is that individuals exhibit an
adaptive flexibility enabling them to adjust both mating and parental effort on the
basis of their current social surroundings. The occurrence of this flexibility can
result in alternative mating strategies for maximizing individual reproductive
success (Bercovitch and Ziegler, 2002). One excellent example of the effects of
dominance on reproductive activity is found in the naked mole-rat
(Heterocephalus glaber). Naked mole-rats are eusocial subterranean rodents
with extreme division of labor in reproduction. Each group typically has one
queen who mates with one to three breeding males (Jarvis, 1991 ). The rest of

the colony members of both sexes are reproductively suppressed, but not sterile

 

 

(faulkes
females
lower pIa .
Both non
reproduct.
MdNMUai
physiolog
MMSum
Ar
“END;
MMWdua
strategy
macaQUl
Success
thSlOlc
Numbe,
r‘3I3r0c1u
higher ‘
DIXSon

aSSOCia

(Faulkes et al., 1990; 1991; Faulkes and Abbott, 1991; 1997). Non-breeding
females are anovulatory (Faulkes et al., 1990), and non-breeding males have
lower plasma LH and T levels than do breeding males (Faulkes et al., 1991).
Both non-breeding females and non-breeding males in a colony rapidly become
reproductively active when socially suppressive cues from the dominant
individuals are removed (Faulkes and Abbott, 1991). Thus, socially regulated
physiological suppression of the HPG axis has evolved for maintenance of order
in this unusual social system.

Among primate males living in groups, variation in reproductive success is
likely to arise from the adoption of alternative reproductive strategies by
individual group members (Bercovitch and Goy, 1990), and each alternative
strategy may be associated with a different endocrine profile. In rhesus
macaques (Macaca mulatta), socially dominant males have greater reproductive
success than subordinate males because of a suite of social, sexual, and
physiological traits that foster mating with multiple females (Bercovitch and
Numberg, 1996). Alpha male mandrills (Mandn‘llus Sphinx) have greater
reproductive success, more pronounced secondary sex characteristics, and
higher T concentrations than subordinate males (Wickings and Dixson, 1992;
Dixson et al., 1993). In these mandrills, abrupt Changes in social status are also
associated with dramatic changes in physical appearance, circulating T, and
reproductive activity (Setchell and Dixson, 2001). Lastly, orangutan males
(Pongo pygmaeus) show perhaps the most startling physical manifestation of a

socioendocrine mechanism, with alternative paths of development leading to two

 

 

different '
larger 00
Graham
with the I‘-
in this spe
Re

male and

 

female ch
SUggestec
their differ
This pane
hypothest
Socioendc
hYPOlheSi
aggressio
relc’ltionsh
QIVen Spe
male Com
mating 3y
seasOnatt
SUCCESS t1

Alti

among llc

different adult morphs, one with and one without secondary sexual traits such as
larger body size, cheek flanges, and enlarged laryngeal sacs (Kingsley, 1982;
Graham and Nadler, 1990). Social suppression of endocrine levels is associated
with the lack of development of these characteristics, but not with reduced fertility
in this species (Maggioncalda et al., 1999; 2000).

Recently, Sannen et al. (2003) showed that urinary androgen levels in
male and female bonobos (Pan paniscus) are similar, whereas those in male and
female chimpanzees (Pan troglodytes) differ substantially. These authors
suggested that this species difference in male-female androgen ratio is related to
their different social systems and differing levels of aggression among males.
This pattern is predicted by the “Challenge Hypothesis.” The challenge
hypothesis is one of the best known, and most tested, hypotheses in the current
socioendocrine literature. Wingfield and colleagues (1990; 2000) proposed this
hypothesis to account for variation in the male androgen response to male-male
aggression among avian species. In brief, this hypothesis suggests that the
relationship between T and aggression during the breeding season in males of a
given species can be predicted by two variables: the level of aggressive male-
male competition and the extent of paternal care characteristic of that species’
mating system. This hypothesis has mainly been tested in high-latitude,
seasonally breeding birds, but it has more recently been applied with some
success to mammalian species.

Although much current socioendocrine research focuses on relationships

among hormones, reproduction, and intrasexual competition, another key area of

research involves study of effects on hormones exerted by members of the
opposite sex. lntersexual social interactions Can strongly influence hormone
levels in many species of birds and mammals, causing androgen levels in males
to increase following their pairing with females (e.g. birds: Moore, 1983; Pinxten
et al., 2003; mammals: Katongole et al., 1971; Rose et al., 1972; Jainudeen et
al., 1972; Bernstein et al., 1977). One of the clearest examples of intersexual
effects on reproductive endocrinology in mammals is reflex ovulation, whereby
copulation results in a surge of circulating LH in females, and this LH then acts
on follicular tissue in the ovary to induce ovulation (Marie and Anouassi, 1986). A
more subtle intersexual effect occurs in red deer (Cervus elephas), where
ovulation may be accelerated as a result of roaring among males (McComb,
1987). Thus socioendocrine mechanisms can serve to coordinate the sexes for
successful reproduction.

Another broad area of socioendocrine research has focused on the ways
in which individual differences in the stress response can reflect variation in
social dominance status among gregarious animals. In vertebrates, stressors
(aversive stimuli) are known to activate the hypothalamic—pituitary-adrenal (HPA)
axis, resulting in the release of glucocorticoids (GCS, stress hormones) from the
adrenal cortex (Sapolsky, 2002). Individual differences in the physiological stress
response often relate to an individual’s dominance status within its social group
(eg. Bronson and Eleftheriou, 1964; Christian and Davis, 1964; Coe et al., 1979;
Creel et al., 1997). However, the relationship between social rank and patterns of

the stress response is not consistent among mammals, indicating that there is no

simple relationship that holds across species between social status and any
aspect of stress physiology (Clarke and Boinski, 1995; Sapolsky, 1993; Abbott et
al., 1997; Abbott et al., 2003).

Influential early work on this issue was conducted with winner-loser
studies in captive rodents and monkeys (Louch and Higginbotham, 1967;
Bronson and Eleftheriou, 1964; Manogue, 1975). In these studies, both winners
and losers of staged encounters Showed a strong stress response, but the
response was larger among losers. In stable social groups, subordinate olive
baboons (Papio anubis) and naked mole-rats have higher levels of GCs than
dominant individuals (Virgin and Sapolsky, 1997; Faulkes and Abbott, 1997).
However, some recent studies with female mammals have shown that dominant
individuals have higher 60 levels than subordinates (eg. Saltzman et al., 1994;
Creel et al., 1996; 1997; Sands and Creel, 2004). In cooperatively breeding
carnivores, the highest ranking individuals with the vast majority of reproductive
success have higher stress hormone levels than do subordinate individuals in the
group (reviewed in Creel, 2001; Creel and Sands, 2003). The diversity of stress
horrnone-dominance relationships in mammals appears to reflect, among other
things, the extraordinary variety of sOcial systems observed in this taxonomic
group.

One of the most exciting new areas of socioendocrine research is that
involving the study of maternal effects. Socioendocrine factors can affect
ontogenetic trajectories, beginning in utero and lasting until death. For example,

neuroendocrine feedback mechanisms play a prominent role in mediating

reproductive processes in post-pubescent animals, but variation in fetal and
neonatal development influences adult neuroendocrine profiles. Prenatally
stressed rhesus macaques develop different neuroendocrine feedback
sensitivities than do control subjects not exposed to prenatal stress (Clarke et al.,
1994). Blockage of elevated neonatal T in male common marmosets (Callithn'x
jacchus) results in attenuation of the pubertal rise in T (Lunn et al., 1994). Social
interactions can influence a pregnant female's glucocorticoid or sex steroid
levels, which may in turn have permanent effects on offspring development.
These organizational effects can have profound influences on adult behavior and

reproduction, and thus create natural variation on which selection may act.

Sggtted hyenas

Spotted hyenas are the second largest, and most abundant, large
carnivore in sub-Saharan Africa (Feldhammer et al., 1999). The social system of
the spotted hyena is unique, and the hyena is therefore a particularly interesting
species in which to investigate relationships between hormones and behavior.
Hyenas live in dynamic fission-fusion groups of up to 80 individuals, called clans,
that are structured by hierarchical rank relationships (Kruuk, 1972; Tilson and
Hamilton, 1984; Frank, 1986; Holekamp and Smale, 1990). Each clan defends
an exclusive group territory containing one or two active communal denning
Sites, where females rear their offspring (Kruuk, 1972; East et al., 1989). Female
hyenas are philopatric, but almost all males disperse from their natal clans

between the ages of 2 and 5 years (Frank, 1986; Henschel and Skinner, 1987;

Smale et al., 1997; Van Horn et al., 2003). All adult females are socially dominant
to all adult males not born in the clan (Kmuk, 1972; Smale et al., 1993), and adult
females are more aggressive than adult immigrant males (Kruuk, 1972; Frank,
1986; Mills, 1990; Szykman et al., 2003).

Social rank is important in this species because it determines an
individual’s priority of access to food (Kruuk, 1972; Tilson and Hamilton, 1984;
Frank, 1986; Mills, 1990). A individual’s social rank is not determined by fighting
ability or size, but rather it is “inherited” from it’s mother (Holekamp and Smale,
1991; 1993; Smale et al., 1993; Frank, 1986). There is a strong effect of social
rank on female reproductive success, with higher ranking individuals doing much
better than lower ranking individuals (Frank et al., 1995; Holekamp et al., 1996;
Holekamp and Smale, 2000). Social rank of individuals within the adult immigrant
male hierarchy is acquired by queuing, rather than by fighting (East and Hofer,
2001; Engh et al. 2002). Males are seldom observed to engage in aggressive
interactions, and what little aggression does occur is typically very mild.

In addition to this unique social system, female hyenas are about 10%
heavier than males (Hamilton et al., 1986), and their genitalia are virilized to
resemble the genitals of males. The clitoris is enlarged and forms an erectile
pseudopenis, the labia are fused to form a pseudo-scrotum, and the urogenital
tract passes through the pseuduopenis (Matthews, 1939; Neaves et al., 1980;
Frank et al., 1990; Cunha et al., 2003). Female hyenas urinate, copulate, and
give birth through the elongated clitoris. Due to female dominance and

morphological masculinization of the external genitalia, females appear to have

unusually tight control over reproduction (Kruuk, 1972; East et al., 2003).
Molecular genetic data have revealed that high-ranking males do not monopolize
reproduction, that both males and females mate with multiple partners, and that
at least 20% of twin litters are of mixed paternity (Engh et al., 2002; East et al.,
2003). Males show no paternal care (Kruuk, 1972).

The unique suite of social, behavioral, and morphological characteristics
exhibited by spotted hyenas may be shaped and modulated by various endocrine
mechanisms. The chapters of this dissertation examine some of these potential
mechanisms. Exploring how endocrine mechanisms operate in this unusual, but
extremely successful large mammal will shed light on important hormone-
behavior relationships in this species in particular, and on some general

principles that may be widely applicable in other mammals as well.

_(_)_v_e_rview of chapte_rs

All of the data presented in this dissertation were collected from a Single
clan of wild spotted hyenas, the Talek clan, which defends a territory on the
northeastern edge of the Masai Mara National Reserve in Kenya. The Talek Clan
has been monitored extensively since 1988, and the research presented here
was only made possible due to the efforts of all of the people involved with the
Mara Hyena Project over the years. I collected a relatively small proportion of
these data, as I spent only 2 years in Kenya doing field work. Much of the data

were collected by Kay Holekamp, Laura Smale, numerous field assistants, and

10

other graduate students. I therefore use “we” in each data Chapter to indicate that
this research was a collaborative effort.

Acquiring blood samples for traditional endocrine analyses is somewhat
difficult with large free-ranging carnivores. Spotted hyenas must be immobilized
in order to collect blood, and this process disrupts normal behavior. However,
recent advances in non-invasive endocrine methods have made possible the
quantification of steroid hormones in feces. Luckily, researchers on the Mara
Hyena Project had the foresight to begin collecting and freezing fecal samples
from known individual hyenas in 1993. In order to investigate the
socioendocrinology of spotted hyenas using this long-term data set, it was first
necessary to validate methods for non-invasive quantification of steroid hormone
concentrations in fecal samples collected and frozen over the last ten years. In
Chapter 2, l collaborate with Dr. Jeffrey French of the University of Nebraska at
Omaha, and with scientists studying captive hyenas at the University of California
Berkeley, in order to validate an enzymeimmunoassay with which to measure
fecal androgen concentrations in spotted hyenas. We employ HPLC analysis, as
well as challenges with GnRH in captive hyenas. We test the effects of potential
procedural covariates in fecal androgen concentrations to determine whether it is
necessary to control for variables like time of day of sample deposition, bone in
the diet, or time elapsed between deposition and freezing of the sample. Lastly,
we confirm that fecal androgen patterns mirror plasma testosterone patterns,

indicating that our assay measures biologically relevant variation in androgens in

11

wild hyenas. The work presented in Chapter 2 has recently appeared in General
and Comparative Endocrinology (Dloniak et al., 2004).

Next, the noninvasive methods we developed in Chapter 2 are used to
address relationships among fecal androgens, social status, reproductive state,
and behavior in free-ranging spotted hyenas. In Chapter 3, my goals were to
determine the best predictors of elevated androgen levels in adult males of this
species, and to elucidate whether sexually motivated aggression against other
males or interaction with females had a stronger relationship with fecal androgen
concentrations. In many polygynous mammals and birds, elevated androgens
are associated with increased male-male aggression during the breeding season
(eg. Wingfield et al., 1990; 2000). However, male spotted hyenas Show very low
levels of aggressive male-male competition over females, whereas they expend
considerable effort actively courting females. Here, I Show that androgen
concentrations in male spotted hyenas are related to courtship interactions and
degree of association with attractive adult females. These results suggest that
the biggest social challenge a male hyena may face is interacting with attractive,
but socially dominant females. The work presented in Chapter 3 is currently in
review at Animal Behaviour.

The prevailing hypothesis suggested to account for the unusual array of
‘masculine’ morphological and behavioral traits in the female Spotted hyena is
that elevated androgens mediate the development of large, aggressive females
that are able to dominate males in competitive feeding Situations, with a

byproduct of genital masculinization (Gould, 1981; Gould and Vrba, 1982;

12

reviewed in Frank, 1997). In Chapter 2, I confirm that pregnant female hyenas
have higher levels of androgens in both blood and feces than do lactating
females. In Chapter 4, I investigate the effect of social rank on maternal
androgens during pregnancy, and the effect of variation in prenatal maternal
androgens on offspring behavior, thus testing whether maternal effects of
androgens influence behavioral development in this species. I Show that high-
ranking females have higher levels of fecal androgens during gestation than do
low-ranking females. To my knowledge, this is the first documentation of an
effect of social status on androgen levels during gestation in a wild mammal. In
addition, I show that prenatal androgen concentrations are positively correlated
with rates of aggression and sexual play in cubs. These results may demonstrate
a non-genetic form of inheritance related to female social status in mammals that
may be comparable to the phenomenon observed in some birds in which
variation inthe amount of T deposited by mothers in eggs generates variation in
offspring behavior and performance (eg. Schwabl, 1996; Eising et al., 2001 ).
Finally, in Chapter 5, I attempt to determine the best predictors of stress
hormone concentrations in spotted hyenas. In order to investigate what variables
influence stress levels in spotted hyenas, it was first necessary for me to validate
an assay for fecal glucocorticoids in hyenas, with methods similar to those
described for the androgen assay in Chapter 2. Using this new methodology, I
then develop models testing the power of variables such as social status, local
prey availability, and female reproductive status, to predict fecal glucocorticoid

concentrations in adult hyenas. I find that important procedural covariates,

13

namely time of day of fecal sample deposition and female age at time of
sampling, must be controlled for when investigating other variables of potential
interest. In addition, factors other than social status are important as predictors of
fecal glucocorticoid excretion in hyenas. Reproductive state and prey availability
have Significant effects in female and male hyenas, respectively, indicating that
physiological and ecological factors are more important influences on stress

hormones than social status in this species.

14

CHAPTER 2
NON-INVASIVE MONITORING OF FECAL ANDROGENS IN
SPOTTED HYENAS

INTRODUCTION

During fetal development, female spotted hyenas (Crocuta crocuta) are
exposed to unusually high levels of androgens, and females are heavily
masculinized in various aspects of their morphology and behavior (Lindeque and
Skinner, 1982; Glickman et al., 1987; Glickman et al., 1992; Licht et al., 1992).
The external genitalia of females closely resemble those of males in that the
clitoris is elongated to form a fully erectile pseudopenis, and the vaginal labia are
fused to form a pseudoscrotum (Matthews, 1939; Neaves et al., 1980; Frank et
al., 1990). Although adult females may weigh more than adult males, the sexes
are monomorphic with respect to other adult body Size measurements
(Matthews, 1939; Kruuk, 1972; Hamilton et al., 1986, Van Horn et al., 2003).
Females are socially dominant to adult breeding males (Kruuk, 1972; Smale et
al., 1993; Smale et al., 1997), and females are more aggressive than adult males
(Frank, 1986; Hamilton et al., 1986; Monaghan and Glickman, 1992; Szykman et
aL,2003)

Spotted hyenas live in social groups called clans. Adult male and female
clan members have separate stable linear dominance hierarchies (Kruuk, 1972;
Tilson and Hamilton, 1984; Frank, 1986; Holekamp and Smale, 1990), and both
male and female offspring “inherit" their mother’s social rank (Holekamp and

Smale, 1991; Smale et al., 1993). Most males eventually disperse from their

15

natal clans between the ages of 24 and 76 months (Frank, 1986; Henschel and
Skinner, 1987; Smale et al., 1997, East and Hofer, 2001, Van Horn et al., 2003)
and attempt to immigrate into new clans in search Of mating opportunities
(Holekamp and Smale, 1998; Engh et al., 2002). In males of many animal
species, androgens vary according to social rank, aggressive behavior or
dispersal status (e.g., rhesus monkeys: Rose et al., 1971, 1972; red deer: Lincoln
et al., 1972; elephants: Jainudeen et al., 1972; primates: Bernstein et al., 1974;
ground squirrels: Holekamp et al., 1984; vervet monkeys: Steklis et al., 1985;
rabbits: Farabollini, 1987; birds: Wingfield et al., 1987; olive baboons: Alberts et
al., 1992; macaques: Zumpe and Michael, 1996; cattle: Lunstra et al., 1998;
spotted hyenas: Holekamp and Smale, 1998). Availability of a method for
assessing relationships among androgens, behavior, and social rank in free-
living spotted hyenas would therefore allow us to address a wide array of
intriguing questions.

Matthews (1939) first proposed that the behavioral and morphological
masculinization of the female spotted hyena might be associated with androgens,
but this question was not addressed systematically until 40 years later. Although
considerable headway has been made in hormone studies utilizing plasma
samples from captive hyenas (see Glickman et al., 1992; Licht et al., 1992;
Yalcinkaya et al., 1993; Licht et al., 1998; Drea et al., 1998; Glickman et al.,
1998; Place et al., 2002; Drea et al., 2002), the study of Circulating androgen
levels in free-ranging spotted hyenas has yielded conflicting results (Goymann et

al., 20013). For example, Racey and Skinner (1979) found that mean plasma

16

testosterone (T) levels did not differ between the sexes, whereas Frank et al.
(1985) concluded that adult males had significantly higher levels of serum T than
adult females, and suggested that variation in T within each sex is related to
social status. The most recent study addressing plasma androgen levels in wild
spotted hyenas indicated that adult females had lower plasma T and
dihydrotestosterone (DHT) levels than adult males, but androstenedione (A4)
levels did not differ between the sexes (Goymann et al., 2001a).

A review of the existing literature on sex differences in circulating
androgens in spotted hyenas (Goymann et al., 2001a) identified some of the
variables that may contribute to differences among results from previous studies,
including different blood collection procedures, various extraction and assay
methods, and lack of knowledge about the reproductive or social status of
individuals. Furthermore, the problem of small sample size has plagued all
studies involving measures of circulating androgens in free-ranging spotted
hyenas. Drawing blood from large carnivores in the wild involves immobilizing
each sampled animal, and this invasive procedure represents a large effort that
yields limited sample sizes. Multiple samples from the same individual are
difficult to achieve by darting, as it can be stressful to the hyena (van Jaarsveld
and Skinner, 1992) and conditions are often unsuitable for obtaining repeated
measures from Specific individuals. Fecal steroid hormone analysis offers an
appealing alternative to blood sampling in order to answer questions about
hormone-behavior relationships in the spotted hyena. This technique has

recently been used to measure fecal androgens in an array of wild mammals (for

17

example; African wild dogs: Creel et al., 1997; sifakas: Brockmann et al., 1998;
muriquis: Strier et al., 1999; ring-tailed lemurs: Cavigelli and Pereira, 2000;
mongoose lemurs: Curtis et al., 2000; hairy-nosed wombats: Hamilton et al.,
2000; meerkats: MOSS et al., 2001; Japanese macaques: Barrett et al., 2002).
Multiple fecal samples can be collected from an individual non-invasively, without
disrupting normal behavior. Within a well-studied population, each sample can be
placed in the context of the individual’s known age, reproductive status, social
rank, and observed behavior. In contrast to plasma samples, large numbers of
fecal samples can often be collected from animals assigned to particular
reproductive and social rank categories, yielding sample sizes appropriate for
analyses of the relationships between these variables and androgens. Although
Goymann et al. (2001b, 2003) have recently examined the relationship between
ecological and social variables and glucocorticoid hormones excreted in hyena
feces, no methods have been described to measure excreted androgens in this
species.

A Our aim here was to validate a fecal androgen enzymeimmunoassay (EIA)
for use in spotted hyenas. After demonstrating assay parallelism and
confirmation of immunoreactive androgen metabolites by HPLC, we assessed
the biological validity of our androgen EIA. We first injected captive hyenas with
LHRH, expecting to see a post-injection increase in fecal androgens. We also
determined whether variation in diet might affect fecal androgen levels by
systematically manipulating the proportion of the diet derived from bone in

captive hyenas, and then measuring excreted androgens in samples collected

18

after feeding. Using archived frozen fecal samples collected from wild hyenas
over the last decade, we inquired whether time of day of sample deposition, time
elapsed before freezing the sample, or time elapsed between freezing and assay
had any systematic effects on fecal androgen concentrations. Finally, we
compared patterns of excreted and circulating androgens collected from
particular groups of wild hyenas (immigrant vs. natal adult males, pregnant vs.
lactating females) among which we expected to observe differences in plasma

androgens.

METHODS

 

Captive study site, subject animals, and sample collection

All captive hyenas were housed at the University of California Berkeley
Field Station for Behavioral Research. These individuals were of known age and
reproductive status. Some hyenas were housed individually and others were
housed in small groups (for more details see Berger et al., 1992). To identify
feces produced by group-housed hyenas, their food was treated with food
coloring. All fecal samples were collected between 0800 and 1200. These were
immediately mixed thoroughly and stored in individual containers at -80° C until
extraction and assay.

LHRH challenges were conducted on 5 gonadally—intact adult hyenas (3
males and 2 females). Fecal samples were collected daily for 7 days prior to
treatment to establish baseline levels of fecal androgen excretion. On the day of

LHRH challenge, animals were immobilized with ketamine and xylazine

19

administered by blow dart, and anesthetized with isoflorane inhalant. Each hyena
then received a single iv. injection of gonadotropin releasing hormone (1 mg/kg
LHRH, L-7134, Sigma Chemical 00., St. Louis, MO). Each hyena was allowed to
recover from anesthesia and released back to its home enclosure. Fecal samples
were collected on the day of challenge as well as for 8 days after the challenge.
Although wild hyenas ingest highly variable amounts of meat and bone as
they feed on ungulate prey, captive hyenas in this colony are normally fed a
standard zoo carnivore mix (Nebraska Brand Feline Food, Central Nebraska
Packing, Inc., North Platte, NE) and small amounts of bone every day (Berger et
al., 1992). In order to assess whether variation in the amount of bone in the diet
influences measurement of fecal androgens, we varied the bone content in the
diet of 5 adult females. Two of these females were ovariectomized for another
experiment. At the start of this experiment some individuals were fed only feline
diet while others were fed a small amount of feline diet plus 3 or 4 sheep neck
bones. After 1 to 3 days on their respective diets, diets were reversed in all
subjects, and reversed again another 1 to 3 days later. Fecal samples were
collected each day from all subjects. For statistical analysis, a given fecal sample

represented the previous day’s diet.

Field study site, subject animals, and sample collection

Our field study site was the Talek area of the Masai Mara National
Reserve in southwest Kenya. The subject population was one large, stable

Crocuta clan inhabiting a home range of approximately 65 km2 (Boydston et al.,

20

2001). The Talek hyenas have been monitored intensively since June 1988, and
all hyenas in the clan were identifiable based on each individual’s unique spot
pattern and other distinguishing marks. Sex was determined by the dimorphic
glans morphology of the erect phallus (Frank et al., 1990). Ages of individuals
born in the Talek clan were estimated to within 7 days based on pelage, size,
appearance, and behavior of cubs when first observed. Adult natal males were
Talek-bom males between 24 and 60 months of age that had not yet dispersed
from the Talek home range. Immigrant males were adult males that had
dispersed from natal clans elsewhere and had been present in the Talek clan for
at least 6 months. Ages of immigrant males were estimated based on toothwear
(Van Horn et al., 2003). Date of first appearance in the Talek clan was recorded
for each immigrant male, and tenure was calculated as time elapsed since joining
the clan (Holekamp and Smale, 1998). Female reproductive state was
determined by behavioral observations or by assessment during immobilization.
A female was pregnant if she gave birth to cubs within 110 days after sampling
(Matthews, 1939), or if fetuses were observed in her uterine horns during
immobilization, using a Hitachi portable ultrasound machine. A female was
lactating if she was Observed to nurse cubs around the time of sampling and/or
milk could be expressed from teats when she was immobilized. Critical incident
sampling (Altmann, 1974) of all observed aggressive and appeasement
behaviors was used to determine social ranks of individuals. Social ranks were
assigned based on a matrix of outcomes of dyadic agonistic interactions (Martin

and Bateson, 1988), as described previously (Smale et al., 1993).

21

Fecal samples were collected either during early morning (0530 — 0900h)
or evening (1700 - 2000h) observation periods. Samples were collected
whenever a hyena defecated, upon direct and unambiguous observation.
Samples were first collected into plastic bags at the site of defecation, and later
approximately 6 ml of mixed sample were transferred to multiple 2 ml cryovials
for freezing in liquid nitrogen. Ninety-four percent of the samples were frozen
within 12 h of collection, and all samples were frozen within 48 h of collection.
Samples were stored in liquid nitrogen until shipped on dry ice to the United
States, where they were stored at -20° C or colder until extraction and assay.

Between 1990 and 2002, 33 adult immigrant males, 13 adult natal males,
16 pregnant females, and 33 lactating females in the Talek clan were
anesthetized with Telazol (2.5 mg/kg body mass) administered in a lightweight
plastic dart fired from a 002 rifle. All immobilizations took place between 0630
and 0900 hours, when hyenas were found resting. Within 10-17 minutes of
anesthetic injection, we drew a blood sample from each hyena's jugular vein into
a heparinized vacutainer tube, and then proceeded to take an array of body and
dental measurements, as described elsewhere (Van Horn et al., 2003). Hyenas
typically recovered from anesthesia within 60 min. All immobilizations were
performed in accordance with Kenyan law and with NIH animal treatment
guidelines. Blood was centrifuged at 10009 for 5 min, then plasma was drawn off
and stored in liquid nitrogen until it was Shipped on dry ice to the United States,

where it was stored at -80° C until radioimmunoassay.

22

Radioimmunoassay of plasma testosterone in wild hvenas

Duplicate aliquots of plasma from each sample from each wild hyena were
assayed for total testosterone (T) using coated tube l125 kits from Diagnostic
Products Corporation (Los Angeles, CA), as described by Holekamp and Smale
(1998). The T kit was previously validated for use with this species by
demonstrating parallelism between serial dilutions of plain and T-spiked plasma
and the standard curve generated using kit calibrators. The detection limit of the
assay was 0.04 ng/ml. Cross-reactivity of the T anti-serum with 5-alpha-
dihydrotestosterone was 3.3%, and was less than 0.1% with any other androgen.
The mean coefficient of variation between T assays (N = 11) was 7.1%. The
mean intra-assay coefficient of variation for high and low T control tubes run with

each T assay was 6.8%.

High performance liguid chromatography (HPLC) Of fecal samplg

Fecal samples from 6 adult male and 6 pregnant adult female wild spotted
hyenas were selected for two pools (male and female, respectively) to be
subjected to HPLC for determination of androgen metabolites. Samples were
collected between 1995 and 2001, and stored frozen until extraction. Each
animal contributed approximately 0.5 grams of feces to its sex-specific pool, and
steroids were extracted with ethanol (see below). Samples were spiked with
~7000 cpm of 3H-testosterone, 3H-androstenedione, and 3H-dihydrotestosterone
and air-dried. Samples were reconstituted in 500 pl of phosphate buffered saline

(PBS; pH 5.0), sonicated, and vortexed to remove any particulates sticking to the

23

surface of the glass tube. Fecal metabolites were first recovered by: 1) priming 3
Spice “1 C18 sample preparation cartridge (AnalteCh, lnc., Newark, DE) with 3ml
of methanol followed by 3 ml of dHZO; 2) loading the 500 pl PBS sample onto the
cartridge; 3) pushing 5 ml of dHZO through the cartridge; and 4) pushing 5 ml of
methanol through the cartridge, and collecting this portion into a 12 x 75 mm
glass tube. The methanol portion was dried under air and the residue was
resuspended in 300pl of methanol. HPLC was conducted by injecting 50 pl of the
reconstituted sample onto the column (Reverse Phase MicrosorbTM MV 100 Ct 8,
Sum diameter particle size, Varian Analytical Instruments, Woburn, MA). A
mobile phase of 45% acetonitrile in water over 80 min at room temperature (RT)
was used at a rate of 1 ml/min. For determination of retention times of 3H
reference standards, aliquots (100 pl) from each fraction were counted for
radioactivity. The remainder of the fraction was dried down and resuspended in
250 pl PBS and analyzed by enzyme immunoassay (EIA) to evaluate androgen

immunoreactivity (see below).

Extraction and Asgy of Mens from Fecal Samples

Approximately 0.5 g wet weight of each fecal sample were placed frozen
in an open whirI-pak bag and lyophilized overnight (Labconco Freeze-Dry
System 10-269). After lyophilization, each sample was ground to a fine powder
with a mortar and pestle. In a 16 x 125 mm culture tube, 0.2 g of powdered feces
were mixed with 5 ml absolute ethanol. Tubes were capped and vortexed, and

then placed on an orbital shaker for 12-14 h. The samples were again vortexed

24

to resuspend particulate matter from the side of the tube, placed in a block
heater, and gently boiled for 20 min. The remaining suspension was centrifuged
for 15 min at 10009 to pellet the solid fecal material. The ethanol supernatant
was poured into a Clean 12 x 75 mm culture tube and evaporated to dryness
under compressed air in a warm-water bath. Samples were reconstituted in 1.0
ml PBS and stored frozen at -20° C until assay.

Fecal androgens were assayed using a modified version of a previously
described testosterone EIA (Nunes et al., 2000). The assay utilized a
testosterone antibody (R156/7) and a testosterone conjugate (horseradish
peroxidase; HRP) kindly provided by Dr. Bill Lasley and Coralie Munro of the
University of California, Davis. When used with plasma, this antibody cross-
reacts 57.4% with dihydrotestosterone, < 0.3% with androstenedione, and <
0.1% with androsterone, dihydroepiandrosterone, estradiol, and progesterone.
Stock antibody was diluted 1:15.000 in bicarbonate coating buffer and T-HRP
was diluted 1:15,000 in PBS. Testosterone standards (Sigma) ranged from 7.8 —
1,000 pg per well in halving dilutions. Wells of microtiter plates (Nunc Maxisorp)
were coated with 50 pl antibody, and diluted fecal sample extract (1 :50 in PBS)
and 50 pl T-HRP were incubated in duplicate wells for 2h. Unbound hormone
was removed by washing the plates, and 100 pl of substrate (ABTS-H202) were
then added to each well. Plates were read on a Dynex plate reader when optical
density in Bo wells reached 1.0. Serial dilutions of pooled extracts from multiple
males and females produced displacement curves that were parallel to the

displacement curve produced by T standards. Quality control in the assay was

25

monitored by measuring T concentrations in two sets of pools on each plate.
One pool was diluted to yield a high concentration of T in 50 pl (~ 30% binding)
and the other to yield a low concentration of T (~ 70% binding). The inter-assay
coefficients of variation were 7.8% (high pool) and 4.5% (low pool). The intra-
assay coefficient of variation, based on these same pools, averaged 9.58 +/- 3.9

% (n = 35 assays).

Archived sample collection, processing at storage conditions

Conditions and time in storage can affect assessments of steroid hormone
concentrations in fecal samples (Khan et al., 2002, Terio et al., 2002), and
circadian variation has also been reported for some excreted steroids (Sousa
and Ziegler, 1998). We therefore inquired whether time of collection or variation
in processing and storage conditions systematically affected fecal androgen
levels in archived samples from wild hyenas. We extracted and assayed 539
samples from 39 adult females and 302 samples from 32 adult immigrant males,
collected between 1993 and 2001 in the Talek clan, for which time of day of
sample and time elapsed between defecation and freezing had been recorded.
We tested for effects of these variables, as well as time elapsed between
freezing and assay, on fecal androgen levels measured in male and female

hyenas separately.

26

Comparison of fecal and plasma androgens

A strong biological validation of a fecal hormone assay includes
demonstration of differences in fecal hormone concentrations among groups that
also vary in circulating hormone concentrations. In order to do this in spotted
hyenas, we analyzed archived frozen fecal samples collected from the Talek clan
between January 1993 and June 2001 that could be compared with plasma
samples acquired during immobilizations. Fecal samples from 16 adult natal
males, 25 immigrant males, 25 pregnant females, and 22 lactating females were
accessed in order to compare patterns in fecal androgens with patterns in
plasma testosterone among similar groups of immobilized hyenas. A single fecal
sample for each hyena was selected for assay. Plasma T may vary with
immigrant male tenure (Holekamp and Smale, 1998; Goymann, 2000), so we
assigned each immigrant a tenure for the day on which his sample was collected.
Tenure of males whose samples were used in the immigrant male group showed
a random distribution between 6 and 36 months, so samples were not biased
toward males with long or short tenure. Age estimates for immigrant males at

time of fecal sampling ranged between 32 and 98 months.

Statistical analysis

All statistical treatment of data followed Zar (1996). Prior to statistical
analysis, all data sets were tested for departures from normality and

homoscedasticity. When such departures were detected, we employed the

27

nonparametric equivalents of the appropriate parametric statistical tests.

Analyses were considered significant when P < 0.05.

RESULTS
HPLC analysis

HPLC fractions assayed with the testosterone EIA revealed the presence
of 6 immunoreactive fecal metabolites in males and 7 metabolites in females
(Figure 1). Three of the major peaks in both sexes were clearly associated with
the elution of T, A4, and dihydrotestosterone (DHT), respectively. Highly polar
fecal metabolites were noted in fractions 10 — 25, with 2 distinct peaks within
these fractions for both males and females. One apolar peak was detected in
males (fractions 72-83) and 2 apolar peaks in females (fractions 65-71 and 75-
79), but these accounted for only a small proportion of the total immunoreactivity.
Overall, patterns of immunoreactive androgen metabolites were strikingly similar
in the fecal pools from males and females, although A4 and DHT were present at
higher levels in females, relative to the T peak, than they were in males. Thus,
our EIA has a broad ability to detect multiple androgen metabolites excreted in

the feces of both male and female hyenas.

LHRH challenges
All five captive hyenas treated with LHRH responded physiologically to the

challenge with an increase in fecal androgens from average baseline levels

(Figure 2, Table 1). The latency to peak androgen varied from one to three days

28

post-LHRH, and averaged 2.4 days. Peak fecal androgen levels after Challenge
were Significantly higher than average levels prior to challenge, but levels Six to
eight days after challenge were no longer different from initial baseline levels
(Repeated measures ANOVA followed by Tukey test for multiple comparisons, F

= 7.32, p = 0.016).

Effect of pone in di_e_t

The presence or absence of bone in the diet did not affect mean androgen
concentrations in feces collected the next day .from adult female hyenas in
captivity. No differences in fecal androgens were found when hyenas switched
from a meat plus bone diet to one consisting of only meat (meat plus bone =
625.74 +/- 381.02 ng/g, meat = 614.41 +/- 191.10; Wilcoxon paired sample test:
Z = 0.944, p = 0.35) or the reverse (meat = 839.96 +/- 186.12 ng/g, meat plus
bone = 1218.10 +/- 529.58 ng/g; Wilcoxon paired sample test: 2 = 0.404, p =
0.68).

Archived sample collection. processing and storage conditions

Variation in sample collection time, processing methods, and storage
conditions had no systematic influence on measured concentrations of fecal
androgens. Although there was considerable variation in hormone
concentrations, there were no pronounced effects of time of day of deposition on

androgen levels. Mean levels in morning samples did not differ from evening

29

150

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A: Adult Male CPM
. + IR Testosterone
2000 T . 120
O
8 1 A4
2 + . DHT 3
“- l-
\ h
1000 . j .—
2 60
8 , 3'
500 “ I I A r 30
" I 1H .1.
0 ~ - “‘1 ' t: '4 - .- r 0
0 20 40 60 80
Fraction
2000 150
8: Pregnant Female T A4 + CPM
+ IR Testosterone
DHT ' 120
1500 . ‘
I: l
.2 90 3
g P
:3 B
II. 1000 - - ~
2 1 ( I 60 5
2 I T m
3 ‘ A a
500 J '
. \ . 30
l 1‘ I
o -1 ' . T ‘_- 1"" - i . ~ 0
0 20 4O 60 80
Fraction

Figure 2.1. HPLC profiles of immunoreactive androgen metabolites in separate
male (A) and pregnant female (B) fecal extract pools. Elution of 3H-Iabeled
testosteronel'l'], androstenedione[A4], and dihydrotestosterone[DHT] expressed
as counts-per-minute per fraction, and immunoreactive androgens measured by

EIA expressed as pg irT per well.

30

8000

 

 

 

 

—.__ Male 13
/V\ ........ O ........ Female 35

_. _ _—+—— M01045
./ ‘\ _..—V-—--- Female“

6000 ‘ Y 3 _ —. — Male 51

 

4000 -

2000 -

Fecal Androgens (nglg feces)

 

 

 

 

 

-8 -6 -4 -2 0 2 4 6 8 10
Day Relative to LHRH Challenge

Figure 2.2. Changes in fecal androgen excretion in response to LHRH challenge
in captive spotted hyenas. LHRH was administered on day 0.

31

Table 2.1. Individual and mean response patterns to LHRH challenge in five
captive spotted hyenas. Androgen values indicate mean i S.e.m. and are

expressed as ng androgen per g Iyophilized feces.

 

 

 

 

”ye” '° 55115351? raggf/ Chm ,, MDym‘m
Male 13 298.46 i 48.84 5167.0 1733.89 3
Female 35 325.71 3: 37.45 629.35 193.32 2

Male 45 431.61 i 84.45 1331.67 308.82 3
Female 49 2371.88 1: 414.03 7281.50 307.09 3

Male 51 673.01 :I: 209.24 3671.11 545.47 1

All 820.14 :1: 393.52 3616.13 1: 1224.39 440.92 2.4 :t 0.89

 

32

samples in adult immigrant male hyenas, whether levels were analyzed within (t
tests; t’s < 2.30, p’s > 0.06), or across individuals (am = 652.64 +/- 193.88 ng/g,
pm = 509.62 +/- 165.38 ng/g; paired sample t-test; t = 0.552, p = 0.592).
Although two adult females had significantly different morning and evening mean
values, they differed in opposing directions, and there was no overall difference
in morning and evening levels within individual females (all other 1 tests; t’s <
3.64, p's > 0.06) or across all females (am = 1325.336 +/- 136.311 ng/g, pm =
1125.739 +/- 125.392 ng/g; paired sample t test; t = 1.456, p = 0.161 ). Likewise,
the latency to process and freeze samples was not associated with significant
variation in excreted androgen. The number of minutes elapsed between sample
collection and freezing was not correlated with fecal androgen levels in either
females (Rp = - 0.057; F = 1.742; p = 0.187) or males (Rp = -0.082; F = 1.998; p =
0.158). Finally, long-term storage of samples while frozen was not associated
with systematic changes in fecal androgen. The number of days frozen until
extraction and assay (ranging from 46 to 2673 days) was not correlated with
fecal androgen levels in either males (Rp = 0.024; F = 0.169; p = 0.681) or

females (Rp = 0.092; F = 0.824; p = 0.115).

Qomgrisons of patterns in fecal androgens wit_h_patterns in plasma_T
Differences in fecal androgens mirrored those in plasma T in wild males

and females. Mean plasma T and fecal androgens both varied with dispersal

status in adult males (Figure 3A). Adult immigrant males had higher plasma T

levels than adult natal males (Mann Whitney U test: U = 97.5, p = 0.0043), as

33

well as higher fecal androgen concentrations (Mann Whitney U test: U = 125, p =
0.045). Mean fecal androgens and plasma T also varied with reproductive status
in adult females (Figure 3B). Pregnant females had higher fecal androgen
concentrations (Mann Whitney U test: U = 104, p = 0), and higher plasma T
values than did lactating females (Mann Whitney U test: U = 58.5, p = 0). In
addition, we tested whether either category of adult males differed from pregnant
females in either plasma T or fecal androgen concentrations. Immigrant males
had significantly higher concentrations of both plasma T and fecal androgens
than both adult natal males and pregnant females, which did not differ from each
other (ANOVAs followed by Tukey tests: plasma T, F = 12.54, p = 0; fecal

androgens, F = 10.99, p = 0).

DISCUSSION

The goal of this study was to determine the extent to which we can reliably
measure androgens in the feces of spotted hyenas and make available a non-
invasive measurement tool to investigate relationships between androgens and
behavior in this species. Our results show that multiple fecal androgen
metabolites can be reliably measured by use of our EIA in both male and female
spotted hyenas. The assay measures variation in fecal androgen concentrations
that is biologically significant, and we have also shown similar relationships in
plasma and feces for both males and females. In addition, frozen archived

samples can be analyzed without concern regarding variation in the amount of

34

2°“ 1000

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:E? MALES l A
31 - _ A
5 1.5 « PlasmaT 23;... 1:, 800 E:
3 “3 Fecal Androgens z
9 ~ 600 a.
3 a:
3 1.0 " 1' e
8 " ........ ' 400 2
._ ....... a
G 1;"
0.5 - - o
5 - 200 a
E
:L
0.0 " 1 ' o
ANM IMM
2°° 1000
35‘ FEMALES - PIasmaT B
2’ 1.5 - [:l Fecal Androgensi 800 g
T; 5
c C
9 - 600 3,
a 1.0 " e
2 s
8 IT " 400 2
a
0.5 ‘ 5.3;. U
'2 T 53".; 1.51
0.0 -__i 1 13"“- " 1 -.-.';.'-‘..';-'-.-_'- 0
Lact Fem Preg Fem

Figure 2.3. Relationship between plasma and fecal hormone concentrations in
male (A) and female (B) wild hyenas. ANM = Adult Natal Males; IMM =
Immigrant Adult Males; Lact Fem = Lactating, Nonpregnant Adult Females; Preg
Fem = Pregnant Adult Females.

35

bone in the diet, the time of defecation, or the length of time the sample has been
frozen. We subjected male and female fecal sample pools to HPLC analysis to
determine which androgen metabolites show immunoreactivity in our T assay
system. The HPLC analysis revealed that both male and female spotted hyenas
excrete multiple immunoreactive androgen metabolites, with markedly similar
patterns for the two sexes. Three of the large peaks of immunoreactivity
corresponded to the elution times for testosterone, androstenedione, and
dihydrotestosterone. A broad peak Of immunoreactivity was detected in early
fractions, where more polar steroids are eluted. Similar broad, early peaks in
metabolite excretion have been noted in studies of androgen excretion in
marmosets (MOhle et al., 2002) and maned wolves (Velloso et al., 1998), and in
estrogen and progestin excretion in felids (Brown et al., 1994). These peaks
likely represent conjugated androgens (glucuronide and/or sulfate conjugates)
that nonetheless cross-react with the T antibody (MOhle et al., 2002; Brown et al.,
1996; J.L. Brown, pers. comm.). Our assay thus permits estimation of multiple
androgen metabolites in spotted hyena feces.

Steroid hormone conversion and metabolism, excretion dynamics in the
gut, variation in Uptake of hormones by target tissues, binding proteins, and even
defecation rate all undoubtedly influence the makeup of the pool of androgen
metabolites in fecal samples. Although the exact makeup of the metabolite pool
is highly variable, the other carnivore species sampled to date excrete multiple
androgen metabolites, some of which remain unidentified (domestic cats: Brown

et al., 1996; wild dogs: Monfort et al., 1997; Cheetahs: Brown et al., 1998; maned

36

wolves; Velloso et al., 1998; red wolves: Walker et al., 2002). In these Species,
the number of androgen metabolites that Show immunoreactivity in a particular T
or androgen assay is also quite variable. In the case of the spotted hyena, we
have Shown that all of the detected fecal androgen metabolites Show
immunoreactivity in our T assay, in both males and females.

LHRH is a potent stimulator of the hypothalamic-pituitary—gonadal (HPG)
axis, therefore the first step of the biological validation of our assay was to
administer LHRH challenges to 5 captive adult spotted hyenas. As predicted,
LHRH treatment led to substantial increases in excreted fecal androgen in the
three males and two females tested in this study. There was an average 5-fold
increase in excreted androgens across individuals, ranging from 2 to 17-fold.
The timing of peak androgen excretion was somewhat variable in that three
hyenas responded maximally after three days, one responded after two days,
and one responded the day after injection with LHRH. Goymann et al. (1999)
found Similar variation in peak glucocorticoid excretion in feces of captive spotted
hyenas after ACTH injection. While the average latency to peak excretion of fecal
glucocorticoids after injection was 24 h, some hyenas exhibited peak excretion
rates more than 48 h after injection. Variation in the efficiency of iv. infusion,
variable thresholds in sensitivity to LHRH, variation in dietary intake or fecal
output during the course of the experiment, and individual differences in steroid
hormone metabolism and excretion might all factor into the variation in the peak
response to LHRH challenge. Regardless, the predicted response to LHRH

challenge was seen in all five hyenas, showing that our EIA is capable of

37

measuring physiologically-induced elevations in androgen activity within
individuals. In addition, we now know that changes in Circulating levels of
androgen are typically expressed in fecal samples 1 to 3 days later, which is
important for future studies investigating relationships between behavior and
fecal androgen levels in this species.

Dietary variation may influence concentrations of excreted steroid
hormones. For example, an increase in dietary fiber has a negative effect on
progestogen excretion in adult female baboons (Wasser et al., 1993). While fiber
is not necessarily of concern in carnivores, bone constitutes a variable proportion
of the Spotted hyena diet. Social rank affects nutritional status in female hyenas
by means of priority of access to food (Holekamp et al., 1996), and higher
ranking hyenas may be ingesting less bone than lower ranking hyenas. Finding
variation in fecal androgens with the presence or absence of bone could force us
to treat high and low ranking hyenas differently. Therefore, we tested whether the
presence of bone had an effect on fecal androgen concentrations in captive
females. Our results indicate no obvious effect of bone, therefore allowing us to
treat high and low ranking hyenas similarly in future studies.

Slight variations in collection and storage conditions could also affect the
ability to extract and detect androgen in hyena fecal samples. Hot weather and
humidity could potentiate bacterial degradation of the hormone product. In the
field, reluctance to disturb our study animals, behavioral sampling protocols, the
presence of other large mammals such as lions, and the distance back to

basecamp can make it difficult to quickly collect, process, and freeze samples. In

38

addition, for this project, samples were collected opportunistically in both
mornings and evenings. Other species exhibit diurnal variation in excreted
hormone levels (e.g., marmosets: Sousa and Ziegler, 1998). However, we found
no effect of variation in collection and storage conditions, and our data revealed
no significant differences in fecal androgen concentrations in morning and
evening samples, suggesting little circadian variation in the excretion of this class
of steroids.

Steroid concentrations in fecal samples may change with long-term
storage. For example, concentrations of fecal glucocorticoids in baboon feces
stored in ethanol at either room temperature or -20° C for 90-120 days differed
from concentrations in the same samples measured after 180 days of storage
(Khan et al., 2002). However, our data suggest that long-term frozen storage
without the use of ethanol does not systematically affect androgen
concentrations in spotted hyena feces. These findings are significant, because
they suggest that minor variations in sample collection and processing, even
delays of up to 48 hours between deposition and freezing, do not affect fecal
androgen concentrations systematically in this species. Furthermore, the ability
to measure meaningful levels of androgenic steroids in samples that have been
frozen for long periods of time allows us to retrospectively examine hormone-
behavior-Iife-history relationships by accessing samples collected during the
long-term study of the Talek hyenas that has endured for more than a decade.

In most mammals, males have significantly higher circulating T than

females. Thus, a good test of the validity of a particular androgen assay is often

39

the ability to differentiate males and females on the basis of excreted androgens.
However, as described earlier, the demonstration of consistent sex differences in
plasma androgens in free-ranging spotted hyenas has been problematic. In the
LHRH challenges presented here, one female (#49) had higher baseline
androgen than the three males (and had the highest post-LHRH levels), but
LHRH-induced elevations in excreted androgens reduced the differences
between this female and the three males. Inspection of the plasma T and A4
levels of this female, which were measured as part of another study (Place et al.,
2002), showed that she had circulating T levels higher than the other female, but
well below the levels of the males. However, this female had very high circulating
A4 levels at the beginning of the experiment, and her plasma A4 levels 120
minutes after the LHRH challenge were five times higher than levels in the
males. Thus it appears that the relative concentrations of each of the androgen
metabolites shown in figure 1 may vary among the endocrine responses of
individual hyenas challenged with LHRH.

The results from our comparisons of plasma T and fecal androgens in wild
hyenas provide valuable insight towards answering questions concerning sex
differences in androgens in this species. Previous studies have shown that
plasma T increases in captive females during pregnancy (Licht et al., 1992;
Glickman et al., 1992). In one study of wild hyenas, no differences in plasma T
were noted between pregnant and lactating females (Goymann et al. 20013), but
that sample included only four pregnant females. Here, using large numbers of

both fecal and plasma samples from adult females in various reproductive states,

40

we found that wild pregnant females do indeed Show higher levels of both
circulating T and excreted androgens than do lactating females. Also of
significance is the fact that pregnant females had Similar levels of androgens to
those of adult natal males, but not to those of immigrant males. Lactating females
had the lowest androgen levels among all groups of adults. This convincingly
shows that it is important to account for female reproductive state and male
dispersal status when comparing androgenic hormones in male and female
hyenas.

In conclusion, we are confident that fecal androgen analysis will be a
useful tool in future research with wild spotted hyenas. Our results indicate that
we can measure biologically meaningful variation in fecal androgen
concentrations among groups of hyenas known to vary with respect to in
concentrations of plasma androgens. Future studies can now focus on
investigating the unusual relationships between sex, social status, reproductive
state, aggression, sexual behavior and levels of excreted androgens in this

species.

41

CHAPTER 3
FECAL ANDROGEN LEVELS IN ADULT MALE SPOTTED HYENAS REFLECT
INTERACTIONS WITH SOCIALLY DOMINANT FEMALES
INTRODUCTION
Linking physiology with reproductive behavior is critical to our
understanding of the proximate mechanisms involved in the regulation of
reproductive success, and socioendocrine studies can shed light on the sources
of individual variation in reproductive success (Bercovitch, 1999; Bercovitch and
Ziegler, 2002). In male vertebrates, androgens, particularly testosterone (T),
influence a variety of morphological and behavioral processes related to
reproduction and social behavior. Testosterone is essential for the formation of
secondary sexual characteristics (Wickings and Dixson, 1992; Dixson, 1998;
Maggioncalda et al., 1999; Gonzalez et al., 2001), plays a crucial role in the
initiation and maintenance of sperm production and sexual behavior (Wickings et
al., 1986; Baum, 2002), and has been linked to the expression of aggressive
behavior and social status (Harding, 1981; Bouissou, 1983; Albert et al., 1990;
Whitten, 2000). Androgen levels can be quite variable among individuals in a
population, and this variation may be associated with individual differences in
behavior, physiology, and reproductive success. The relationship between
androgens and behavior is often modulated by social factors. For example, levels
of T in primates often do not correlate with dominance status or aggressive
behaviour under stable social conditions (Sapolsky, 1983; Cavigelli and Pereira,

2000), whereas a positive correlation is found in situations characterized by

42

unstable hierarchies or other social challenges (Sapolsky, 1983; Steklis et al.,
1986; Bernstein et al., 1992; Cavigelli and Periera, 2000; Brockman et al., 2001).
Although originally developed for use in seasonally-breeding avian
species, the ‘Challenge Hypothesis’ (Wingfield et al., 1990; Wingfield et al., 2000)
has recently been employed as a theoretical framework to help predict patterns
in the relationship between T and aggression associated with mating in various
species of mammals (dwarf mongoose, Helogale parvula: Creel et al., 1993; wild
dog, Lycaon pictus: Creel et al., 1997; ring-tailed lemur, Lemur catta: Cavigelli
and Periera, 2000; Verreaux’s sifaka, Propithecus verreauxi: Brockman et al.,
2001; tufted capuchin monkey, Cebus ape/la nigritus: Lynch et al., 2002;
redfronted lemur, Eulemur fulvus rufus: Ostner et al., 2002; spotted hyena,
Crocuta crocuta: Goymann et al., 2003; arctic ground squirrel, Spennophilus
pan'yii: Buck and Barnes, 2003; chimpanzee, Pan troglodytes: Muller and
Wrangham, 2004). As originally formulated, the main premise of the challenge
hypothesis is that the relationship between T and aggression during the breeding
season in males of a given species can be predicted by the level of aggressive
male-male competition and extent of paternal care characteristic of that species.
In species where males are intensively involved in parental care and engage in
relatively low rates of male-male aggressive competition (often monogamous
Species), the hypothesis predicts that males should maintain a breeding baseline
level of T and respond to occasional aggressive challenge with a relatively large
increase in T to facilitate aggressive competition. By contrast, in species showing

no paternal care and higher rates of male-male aggression (often polygynous

43

species), the hypothesis predicts that males should maintain relatively high levels
of T during the entire breeding season to facilitate a heightened state of
aggression. Therefore, the physiological response of a male to challenge in this
situation Should be smaller than in the previous category of males, because T is
already at or near the effective maximum (Wingfield et al., 1990; Creel et al.,
1993; Wingfield et al., 2000).

Although the vast majority of studies testing the challenge hypothesis
have focused on the relationship between androgens and reproductive
aggression among same-sexed individuals, the original hypothesis also suggests
that physiologically influential social challenges may occur during interactions
with potential mates as well as those occurring during interactions with rival
males (Wingfield et al., 1990, Wingfield et al., 2000). Male white-crowned
sparrows (Zonotrichia Ieucophyrs) that are caged with estradioI-implanted
females have higher levels of T than do males housed with non-receptive
females (Moore, 1983). Pinxten et al. (2003) demonstrated that testosterone
levels in European starlings (Stumus vulgan's) were positively influenced by
female presentation. Most recently, Moore et al. (2004) found that the challenge
hypothesis was not supported in an equatorial population of rufous-collared
sparrows (Zonotn‘chia capensis), and these authors suggested that male-female
interactions may be responsible for patterns of T observed in this species.
Plasma T levels also rise in many male mammals after mating or exposure to a
novel receptive female (Katongole et al., 1971; Rose et al. 1972; Jainudeen et

al., 1972; Purvis and Haynes, 1974; Sanford et al., 1974; Macrides et al., 1975).

44

These studies suggest that, when investigating the sources of variation in
androgen concentrations in males, it may be important to consider the influence
of intersexual interactions in addition to the influence Of aggressive competition
with rival males.

Here, we set out to evaluate the influence of inter- and intrasexual
interactions on male androgen excretion in a gregarious carnivore, the spotted
hyena. These hyenas live in dynamic fission-fusion groups of up to 80
individuals, called clans. Each clan contains one to several matrilines of adult
females and their offspring, as well a number of immigrant males. Clans are
structured by hierarchical rank relationships (Kruuk ,1972; Tilson and Hamilton,
1984; Frank,1986; Holekamp and Smale, 1990), and all adult females are
socially dominant to all adult males not born in the clan (Kruuk ,1972; Smale et
al., 1993). Adult males are less aggressive than adult females (Kruuk, 1972;
Frank, 1986; Mills, 1990; Szykman et al., 2003), and the social rank of individuals
within the male hierarchy is acquired by queuing, rather than by aggression (East
and Hofer, 2001; Engh et al., 2002). Males are seldom observed to engage in
aggressive interactions, and what little aggression does occur is typically very
mild. Due to female dominance and morphological masculinization of the external
genitalia, females appear to have unusually tight control over reproduction
(Kruuk, 1972; East et al., 2003), and the awkward courtship behaviour of the
male hyena reflects his extreme motivational conflict with regard to approaching
females (Holekamp and Smale, 1998; Szykman, 2001). Molecular genetic data

have revealed that high-ranking males do not monopolize reproduction, that both

45

males and females mate with multiple partners, and that at least 20% of twin
litters are of mixed paternity (Engh et al., 2002; East et al., 2003). Males Show no
paternal care (Kruuk, 1972), although they can recognize their own offspring
(Van Horn et al., in press).

Female hyenas are philopatric, but males disperse sometime after
reaching reproductive maturity at approximately 2 years of age. Males may
remain in their natal clans for up to 52 months after puberty (Smale et 31.1997;
East and Hofer, 2001; Van Horn et al., 2003), so there are two classes of adult
males in every clan: natal males born in the clan and immigrant males born
elsewhere. Although natal males are socially dominant to all immigrant males,
immigrants engage in higher rates of sexual and courtship behavior, have higher
average levels of plasma testosterone (Holekamp and Smale, 1998; Holekamp
and Sisk, 2003; Dloniak et al., 2004), and sire 30 times as many offspring as do
natal males (Engh et al., 2002).

Goymann et al. (2003) recently described a relationship in spotted hyenas
between elevated plasma T and the occurrence of male defense of one or more
females. However, low intensity male-male reproductive aggression in the form
of female defense is not very common among male hyenas, and we therefore
suggest that the relationship between mate defense and androgens may be
confounded by the presence of a female and/or by courtship interactions
between the sexes. We therefore set out to test the hypothesis that intersexual
interactions constitute a Significant social challenge for male hyenas, and that

these interactions correlate with androgen concentrations. We recently

46

developed an assay to quantify fecal androgen (fA) concentrations in this species
non-invasively (Dloniak et al. 2004), and we used that assay here with samples
collected from adult male spotted hyenas to inquire about the influence of the
following variables on fA concentrations: social status, tenure in the clan,
reproductive aggression among males, male-male aggression not associated
with the defense Of a female, courtship, and degree of association with females
in particular reproductive states. If male-male competition represents a significant
challenge for male hyenas, fA concentrations should vary with the occurrence of
male-male aggression. On the other hand, if interactions with behaviorally
dominant yet attractive females constitute significant challenges for males, we
expected fA concentrations in males to vary with social measures related to
interactions with females. We also predicted that relationships between
androgens and behavior would differ between adult natal and immigrant males,
based on differences in behavior and physiology between these two groups of

males that were documented in earlier studies.

METHODS
Study Site and Animals

The study was conducted in the Talek area of the Masai Mara National
Reserve, an area of open rolling grassland in southwest Kenya. The subject
population was one large, stable spotted hyena clan, inhabiting a home range of
about 65km2 (Boydston et al., 2001). The Talek clan has been intensively studied

since 1988, and the boundaries of the clan’s home range have been stable at

47

least since 1979 (Frank, 1986). All hyenas in the clan were known individually by
their unique spot patterns, and sex was determined by the dimorphic glans
morphology of the erect phallus (Frank et al., 1990). Data used in this study were
gathered between July 1995 and July 2002, during which time observers
monitored the Talek hyenas 20-31 days per month for approximately 6 hours per
day. During this study period, the number of adult immigrant males, adult natal
males, and adult females in the clan ranged from 11-21, 1-9, and 21-31,
respectively.

Adult natal males were those that had been born in the Talek clan, and
were older than 24 months (the approximate age of puberty) but had not yet
dispersed. Ages of adult natal males were known +/- 7 days, using methods
described by Holekamp and Smale (1998). Ages of the adult natal males
monitored in this study ranged from 26 to 68 months. Adult immigrant males
were those that had been born outside Talek and then immigrated into the Talek
clan at dispersal. Because immigrant males originated in clans other than our
study population, exact ages of immigrants were unknown. However, we could
reliably estimate (to +/- 6 months) the ages of immigrant males based on tooth-
wear measures Obtained during routine immobilizations (Van Horn et al., 2003).
Estimated ages of immigrant males in this study ranged from 30 to 150 months.
In addition, we recorded the date of first appearance in the Talek home range of
each immigrant male, thus allowing us to calculate his tenure in the clan at the
time of each sampling, based on time elapsed Since his initial appearance in the

clan.

48

We monitored the extent to which adult males associated with adult
females that were either attractive (close to conception) or lactating (relatively
unattractive). Spotted hyenas in East Africa breed year-round (Holekamp et al.
1999), thus at any given time there may be pregnant, lactating, and cycling
females present in a clan. Lactating females were within the third or fourth month
of lactation, based on direct observation of nursing bouts and known birth dates
of cubs. Attractive females were within one month prior to a known conception.
We calculated the conception date for each litter born to each female by
subtracting 110 days, the length of the gestation period in Crocuta, from the birth
date of that litter (Schneider, 1926). Birthdates could be reliably estimated to +/- 7

days (Smale et al., 1993).

Behavioral Observations

Behavioral data presented here were collected throughout the 7-year
study period, except for courtship data, which were collected only from May 1997
through August 1999, and May 2000 through May 2002. We conducted daily
behavioral observations from vehicles between 0530 and 0930 hours and
between 1630 and 2000 hours. We located hyenas while driving daily circuits
around the Talek home range, visiting high points in the area and scanning with
binoculars, and by radiotracking collared individuals. We initiated an observation
session when we first drove up to one or more hyenas separated from others by
at least 200m, and the session ended when all hyenas moved out of sight, or

when we drove to a new location. Data documenting agonistic behavior during

49

observation sessions were collected using critical incident sampling (Altmann
1974) of all aggression and appeasement behaviors observed in all social
contexts. All courtship behaviors were also recorded as critical incidents. Social
ranks were assigned based on a matrix of outcomes of dyadic agonistic
interactions (Martin and Bateson, 1988), as described previously (Smale et al.,
1993). Adult females (together with their offspring) and adult immigrant males
were ranked in separate dominance hierarchies. Since the number of immigrant
males present during the study varied over time, we calculated standardized
social rank for each immigrant at each time of sampling by assigning the highest-
ranking male a rank of +1 and the lowest-ranking male the rank -1, with all other
males spread evenly between these two individuals (East and Hofer, 2001;

Goymann et al., 2003).

Agonistic Behavior

Appeasement behaviors used to determine social ranks included head-
bob, carpal crawl, giggle, squeal, back-off, and submissive-posture (tail between
the legs, head down, body lowered and bent) (Kruuk 1972; Holekamp and
Smale, 1998). Aggression included lunge, snap, bite, chase, displace, push,
stand over, and intention movement to bite (Kruuk, 1972; Holekamp and Smale,
1998). We used social context to distinguish between reproductive and non-
reproductive aggression among males. In order for a male to show reproductive
aggression, he had to be within 10m of an adult female and show aggression

(including at least a low-level threat, such as displace) toward another adult male

50

who approached to within 10m of that female. If the context of the aggression
was not related to a female (such as during feeding with other males), we
considered it to be non-reproductive aggression. We discarded any aggressive

interactions for which context was ambiguous.

Courtship Behavior

Courtship behavior included extended following, approach/avoid display,
present, bowing display, paw the ground, and mount (see Holekamp and Smale,
1998; Szykman, 2001 for full definitions of behaviors). These behaviors were
exhibited by adult male hyenas only toward females and indicated male sexual
interest (Holekamp and Smale, 1998; Szykman, 2001). Although these behaviors
were often accompanied by appeasement gestures, courtship behaviors alone
could not be used to determine social rank because males never directed them
at other males. We did not include either male aggression towards females
(baiting behavior: Kruuk, 1972; Szykman et al., 2003) or copulations in our

analyses due to small sample sizes for these rarely observed events.

Association Index

We wanted to assess whether the degree of association between a male
and a particular female influenced androgen concentrations in the male’s feces.
Due to the highly dynamic, fission-fusion nature of spotted hyena society, the
best measure of association between any two individuals is an association index

(Al). We therefore calculated Als around a male’s fecal sample date, between

51

that male and all members of each of the 2 classes of adult female, attractive and
lactating. We used the twice-weight index of association (Cairns & Schwager
1987x

(A+B)together / [(AwithoutB) + (BwithoutA) "' (A "' B)togetner]
where (Ammo... 3) represents the number of observation sessions in which male A
was observed but female B was not present, (Bwimou. A) represents the number of
observation sessions in which female B was present but male A was not, and (A
+ BMW represents the number of sessions in which both male A and female B
were present (as in Syzkman et al., 2001). Specifically, for 69 fecal samples from
22 immigrant males we separately calculated male Als with both lactating and
attractive females present in the clan for the time period encompassing three

weeks before, and one week after, the date of fecal sample collection.

Fecal Sample Collection, Processing, and Androgen Assav

During the study period, fecal samples were collected from 16 adult natal
males and 26 adult immigrant males, yielding totals of 25 and 159 samples,
respectively. A fecal sample was collected whenever a hyena was directly
observed to defecate during an observation session. Samples were first collected
into plastic bags at the site of defecation, then mixed and transferred to 3ml
cryovials. Samples were stored in liquid nitrogen until transported on dry ice to
the United States, where they were stored at —20°C or colder until processing.
Samples were then extracted and assayed for androgens as described by

Dloniak et al. (2004). In brief, steroids in 0.29 Iyophilized feces were extracted

52

with 100% ethanol. Fecal extracts were reconstituted and diluted 1:50 in PBS
and measured for immunoreactive androgens by enzymeimmunoassay (EIA).
The assay utilized a testosterone antibody (R156/7) and a testosterone
conjugate (horseradish peroxidase) kindly provided by Dr. Bill Lasley and Coralie
Monroe of the University of California at Davis. The mean coefficient of variation
between assays (N = 10) was 8.5%, based on high and low pools in each assay.
The intra-assay coefficient of variation for high and low pools run within each
assay was 7.3%. Dloniak et al. (2004) demonstrated that neither time of day at
which the sample was collected, nor the length of time the sample was kept in
frozen storage before extraction and assay systematically influence fA
concentrations in spotted hyenas, so we did not include these variables in our

analyses.

Statistical Analyses

Adult natal and immigrant male hyenas behave very differently, have
different mean plasma T and fA concentrations, respond differently to challenge
with gonadotropin releasing hormone (Holekamp and Smale, 1998; Dloniak et
al., 2004; Holekamp and Sisk, 2003), and we collected many more faecal
samples from immigrants than from natal males. Therefore, we analyzed the
relationships between fA and behavior in each of these two groups of males
separately. We first inquired whether rates of courtship were higher than rates of
reproductive aggression in 6 adult natal males and 12 adult immigrant males

present in the Talek clan for the entire period between May 2000 and May 2002,

53

using Wilcoxon matched pairs analyses. In adult natal males we then
investigated relationships among fA, age, and social rank at the time of sample
using multiple regression. Due to non-independence of some data points, we
included individual hyena identity (ID) as a random effect variable in a
generalized linear mixed model with backward selection. We next inquired
whether natal males showing non-reproductive aggression, reproductive
aggression, or courtship behavior during the three days prior to, and on the day
of, a fecal sample had greater fA concentrations than natal males not showing
these behaviors at the time of sampling. We chose this time-frame based on the
fact that circulating androgens are represented in hyena feces one to four days
later (Dloniak et al., 2004). Three separate analyses were done using Mann-
Whitney U tests, and since the same samples were used in multiple analyses we
corrected for the experimental error rate by applying a Bonferroni adjustment in
significance testing (Rice, 1989). Due to the small number of samples obtained
from natal males, we did no further analysis of this dataset.

ln immigrant males, we first used multiple regression to investigate
relationships among fA, estimated age, standardized social rank, and tenure in
the clan. The majority of adult immigrant males were sampled repeatedly
throughout the study, and the number of samples per individual varied greatly (1
— 17). Due to this unbalanced data set, we included hyena ID as a random effect
variable in the initial regression. lD had no significant effect, so we repeated the
analysis without ID included. We next used subsets of the available samples to

analyze the relationships between fA and behavior in immigrants. Here we used

54

Wilcoxon matched pairs to determine whether fA concentrations in males
showing non-reproductive aggression, reproductive aggression, or courtship
behavior were different than concentrations in the same males when they were
not showing these behaviors. Here, no sample was used in more than one
analysis, and a male was represented by only one paired sample within an
analysis. Lastly, we again used multiple regression to investigate the relationship
between immigrant male fA and Als with lactating and attractive females. All
statistical analyses followed Zar (1996) and were done using Statistica 8.0.
Statistical significance was set at P s 0.05 and nonparametric tests were

employed when necessary. Results are presented as means +l- standard errors.

RESULTS
fites of Reproductye Aggression and COlLSth

A matched-pairs comparison of rates of courtship and reproductive
aggression in 6 adult natal males, all of which were present in the Talek clan
throughout a tvvo-year period, showed that these males directed courtship
behaviour toward females at significantly higher rates than those at which they
directed reproductive aggression toward other males (Wilcoxon matched pairs, Z
= 2.201, P = 0.041; Figure 1). The corresponding analysis in 12 immigrant males
present in the Talek clan during the same time interval as the 6 natal males
showed that immigrants also courted females at much higher rates than they

showed reproductive aggression (Wilcoxon matched pairs, 2 = 3.059, P = 0.002;

55

Figure 1). Reproductive aggression in adult males generally consisted of low-

level threats, and these were always directed toward subordinate males.

Adult Natal Males

We found no relationship among social rank, age, and fA concentrations in
adult natal males, controlling for hyena ID as a random effect in a mixed model
multiple regression (whole model R = 0.413. F321 = 1.4389, P = 0.26). Because
we found no effect of social rank or age in natal males, we did not control for
these variables in our subsequent analyses of behavior. Concentrations of fA
were similar between males showing non-reproductive aggression and males
failing to show any non-reproductive aggression (Mann-Whitney U test: 2 = -0.28,
P = 0.77, aggression N = 9, >'< i SE = 386.32 i115.48 nglg, no aggression N =
16, >7 i SE = 315.27 1 88.70 ng/g). We found no significant difference between
fA concentrations in natal males showing courtship behavior prior to a faecal
sample and those in males showing no courtship behavior (Mann Whitney U test:
2 = 0.11, P = 0.93, courtship N = 10, >‘< 1- SE = 308.12 x 87.62 ng/g, no courtship
N = 15, >7 :1: SE = 362.66 $101.63 ng/g). Lastly, although there was a slight
tendency for natal males showing reproductive aggression to have greater fA
concentrations than natal males failing to show reproductive aggression, this
difference was not statistically significant when the Bonferroni adjustment was
applied, possibly because so few natal males were observed showing
reproductive aggression (Mann-Whitney U test: 2 = -2.00, P = 0.045,

reproductive aggression N = 2, >'< 1: SE = 851.66 1' 52.71 nglg, no reproductive

56

 

   

 

El Reproductive aggression

§ lCourtship

x 60 4

8

2

t5

'3 4° 1

g l

3 l

1: zoj

o

I I

o ‘ _
Natal lmmlgrant
Adult Male Category

Figure 3.1. Paired rates of reproductive aggression and courtship observed in
adult male Talek hyenas from May 2000 through May 2002. N = 6 adult natal
males and 12 adult immigrant males. Rates of behavior were compared within
adult male category.

57

aggression N = 23, >‘< t SE = 296.43 :l: 67.49 ng/g). Interestingly, the fecal
samples of the two males who showed reproductive aggression had the highest
concentrations of fA among all samples collected from adult natal males, and
both of these males were also involved in courting females prior to the sample

collection.

Immigrant Males

We first used multiple regression to investigate whether there was an
influence of male social status, age, or tenure on fA concentrations among
immigrants. The overall model here was significant (whole model R = 0.25, F4,134
= 4.78, P = 0.01 ), with tenure (P = 0.03, Figure 3.2A) and standardized social
rank (P = 0.03, Figure 3.28) retained, but hyena ID and estimated age were
removed from the model by backwards selection. However, social rank and
tenure were closely and positively correlated (R = 0.90), neither variable
explained significant variation in fA (social rank partial R = 0.03, tenure partial R
= -0.13), and the partial correlations with fA concentration were in opposing
directions. These results show that there is no simple, systematic relationship
between social status or tenure and fA concentrations in immigrant male hyenas.

Samples from immigrant males were sufficiently abundant to permit use of
paired analyses of the relationships between androgens and specific behaviors.
Although a male's fA concentrations did not vary significantly based on whether
or not he exhibited any non-reproductive aggression (Wilcoxon matched-pairs: Z

= 1.29, N = 13, P = 0.20, Figure 3.3A), fA concentrations were significantly

58

>

4000.
E i

a l

l=3000: 0

V e

2 l .

3: . e

22°°°T . e

'2 e e

< 0
Ti 0
E e

120

 

""3 flesh

Tenure in months

 

B

3 ;

a .

5 3000 i 0

o 'e

c

0

g 2000 6 . .

.3 O

c C :

E 0 ' .

§1ooo‘i.. J. .e‘: :3. e

u. o .0“ ‘ .’ I;
-1.0 -0.5 0.0 0.5 1.0

Standardized Rank

Figure 3.2. Relationships between fecal androgen concentrations in immigrant
males and their tenure in the Talek clan (A) and their standardized social rank (B)
at the time of sampling. The highest social rank is indicated as 1.0. N = 159
faecal samples from 26 immigrant male hyenas.

59

elevated in samples collected from immigrant males showing reproductive
aggression compared to samples collected from the same males when they were
not showing reproductive aggression (Wilcoxon matched-pairs: Z = 1.99, N = 6,
P = 0.046; Figure 3.38). In addition, fA concentrations were greater in samples
from immigrant males showing courtship behaviour than in samples from the
same males when they did not show courtship (Wilcoxon matched-pairs: Z =
3.31, N = 16, P = 0.0009, Figure 3.30). Very few males ever showed
reproductive aggression that was associated with a fecal sample (6 out of 26
males). In addition, five of the six males engaged in reproductive aggression
(3.3B) that was associated with a fecal sample were also seen courting a female
during the days prior to the sample. In the courtship analysis (3.30) we only
included samples collected from males showing courtship behavior without
reproductive or non-reproductive aggression, therefore our results indicate that
there is an effect of courtship that is independent of reproductive aggression, and
that there is not a significant additional effect of reproductive aggression on fA
concentrations after courtship is considered.

Thus far, there is a potential confound in our behavioral analyses in that
female presence is required for courtship and reproductive aggression, but not
for non-reproductive aggression. Therefore we next tested whether or not there
was an influence on fA concentrations of the extent to which immigrant males
associated with females. The multiple regression model of Al was significant
(whole model R = 0.565, F255 = 15.506, P < 0.00001). The Al between males

and lactating females was not correlated with the Al between these same males

60

/\ E3
.. 120° 7 1200
a: ‘ A
a l 9 *
5 900 ‘ 2’
g ‘ v 900
3 ‘
600
'u g 600
c 1:
5 300 F
4'
LI. 0
0 l v ,3
A99 No Aug 0 ‘
Reprod Agg No Reprod A99

1200 r,

«a
o
c
3(-

Fecal Androgens (nglg)
u as
8 8

. .4 .l

Court No Court

Figure 3.3. Differences in faecal androgen concentrations between samples
when immigrant males did and did not emit particular behaviours. (A) N = 13,
non-reproductive aggression (Agg) and no non-reproductive aggression (No
Agg), (B) N = 6, reproductive aggression (Reprod Agg) and no reproductive
aggression (No Reprod Agg), (C) N = 16, courtship (Court) and no courtship (No
Court). Males are paired between bars within each histogram. Note that 5 of 6
males showing reproductive aggression (B) were also observed courting a
female. No males showing courtship (C) also showed reproductive aggression.
Asterisks indicate significant differences.

61

and attractive females (R = 0.035). fA concentrations were positively correlated
with AI between males and attractive females (partial R = 0.558, P < 0.00001;
Figure 3.4A), but not with Al between males and lactating females (partial R =
0.127, P = 0.213; Figure 3.48). This suggests that the degree of association
between a male and females near their times of conception influences his

androgen levels.

DISCUSSION

Whereas behavior was not a useful predictor of fA concentrations in adult
natal males, reproductive aggression and courtship were good predictors of
elevated fA concentrations in adult immigrant males. Furthermore, degree of
association with females near their times of conception was positively related to
fA levels in immigrant males, whereas association with lactating females showed
no such relationship. These data suggest that associating closely with cycling
females influences reproductive physiology in the male spotted hyaena.

Our results indicate that age does not systematically influence fA
concentrations in adult male spotted hyaenas, regardless of dispersal status.
Furthermore, neither adult natal nor immigrant males showed systematic status- '
dependent variation in fA concentrations within their social hierarchies. Although
a lack of a simple relationship between plasma T and social status has been
shown previously in immigrant male hyenas (Goymann et al., 2003; Holekamp
and Sisk, 2003), we demonstrate here for the first time that fA concentrations

show the same pattern as plasma T, and that the lack of relationship between

62

>

 

 

{a 2500 -~

0 I

8 I e

- 2000 t .

9 e

2’ I e

v 1500 _. '

g e

3: e

9 1000 4 . . ' . .

“U

5 ' ' " ' a.

_ 500 . .

§ r..

IL 0 _ Y , 7 ,AA #2 j
0 0.1 0.2 0.3 0.4

Average AI with Attractive Females

  

‘a 2500 1

O

0

e 2000 4 o

9 e

2’ e

I," 1500 I '

c e

8:

2 1000 .

'2

< 500 ' . .

a .0 ..

8 000 .0

u. 0 _ . 2.---
0 0.1 0.2 0.3 0.4

Average Al with Lactating Females

Figure 3.4. Fecal androgen concentrations in immigrant males are positively
correlated with their association index (Al) with attractive females (A), but not
with lactating females (B). N = 69 fecal samples from 22 immigrant male hyenas.

63

androgens and social rank also holds true for adult natal males. Social ranks in
natal males are based on the social ranks of their mothers (Holekamp and
Smale, 1993; Smale et al., 1993; Engh et al., 2000). Furthermore, immigrant
males only gain status in their new clan by advancing in a queue as higher-
ranking males die or emigrate (East and Hofer, 2001; Engh et al., 2002). Thus,
neither natal nor immigrant males gain social status through fighting ability or
size. These data therefore resemble those from previous studies in other species
showing that androgen levels are not closely related to male social status within
a stable hierarchy (Eaton and Resko, 1974; Gordon et al., 1976; Bernstein et al.,
1983; Steklis et al., 1986; Alberts et al., 1992; Sapolsky .1993; Goymann et al..
2003).

The challenge hypothesis is often interpreted to predict high testosterone
levels in males during bouts of reproductive aggression, regardless of other
important social factors that might also affect relationships between androgens
and behavior in social mammals. Goymann and colleagues (2003) recently
showed that male spotted hyenas defending a female had higher levels of
plasma T than did non-defending males in the Serengeti. We report here a
similar result in Talek hyenas, in that immigrant males had higher concentrations
of fA when showing reproductive aggression and courting than when they were
not showing reproductive aggression. However, our data are not consistent with
Goymann et al.'s (2003) conclusion that defending males have higher levels of T
due to a higher interaction rate with other males. If a higher interaction rate with

other males per se were enough to raise androgen levels, we should have

64

detected significantly higher concentrations of fecal androgens in males showing
non-reproductive aggression than in males showing no aggressive behavior.
Instead, our results indicate that the context in which aggression occurs is quite
important. Also, our results suggest that male-male interactions over a female do
not constitute a particularly challenging situation for male hyenas. Males rarely
fight over females, aggression intensities are low and consist mostly of low-level
threats among males, and the outcome of any male-male interaction is nearly
perfectly predictable due to the stable social hierarchy and queuing system
among immigrants.

In their efforts to apply the challenge hypothesis to spotted hyaenas,
Goymann et al. (2003) acknowledged that there was a possibility that T levels
were elevated in male hyenas defending prospective mates due to being close to
females. However, male-female interactions were not investigated in that study.
The present work directly addressed this possibility by investigating whether the
main social challenge for a male spotted hyena is male-male competition or
interacting with dominant females. Although we did find a significant relationship
between reproductive aggression and androgens, most males showing
reproductive aggression were also concurrently engaged in active courtship.
Males showing courtship alone also had elevated androgen levels. These results
demonstrate that interacting with females is associated with elevated androgen
levels independent of whether or not a male is also engaging in defensive
aggressive interactions with other males over access to the females. This

supports our hypothesis that interacting with females represents a substantial

65

challenge to adult immigrant males. It is quite obvious upon observation that
spotted hyena males are conflicted between tendencies to approach and flee
during interactions with dominant females. Males sometimes even engage in
brief running mounts and awkward dismounts when interested in a female, as
well as masturbation next to a female. Males may need to reach some threshold
of sexual stimulation in order to overcome this approach-avoidance conflict, to
actively court and eventually mate with a female.

The challenge hypothesis has proven to be a useful starting point for
formulating predictions concerning androgen-behavior relationships in male
vertebrates. However, the results presented here, as well as a small number of
recent studies (Creel et al., 1993; Peters et al., 2001; Lynch et al., 2002; Moore
et al., 2004), suggest that the hypothesis may be limited in its direct application to
social mammals and some birds. The challenge hypothesis appears to lack
predictive power for patterns in androgen responsiveness to challenge in species
with low ratios of paternal care to male-male aggression (Fig. 4.7 in Wingfield et
al., 2000). This suggests that factors influencing male fitness other than direct
male-male aggressive competition with other males modulate hormone-behavior
relationships in these species. Instead of attempting to modify the challenge
hypothesis for application to complex social systems, we suggest that future
research should focus on testing the influence of additional variables that may
affect horrnone-behavior relationships in free-living animals, including alternative

reproductive tactics and male-female interactions.

66

We found that degree of association with attractive females positively
influenced fA concentrations in immigrant male hyenas. The reproductive state of
the females with which a male associated was clearly important, as males
associating closely with females in the month prior to a conception had greater fA
concentrations than males associating with lactating females. An endocrine
response of a male to a female may play a role in synchronizing reproductive
behavior and physiology between the sexes when mating opportunities are
somewhat unpredictable. Our findings suggest that, in spotted hyenas, androgen
levels are increased by interaction with stimulating females, and that this
increased androgen might then function to facilitate courtship and sexual
behavior, as well as occasional aggressive defense of females (Moore, 1984;
Wingfield et al., 1987, 1990; Beletsky et al., 1995; Balthazart et al., 1996).

We were unable to investigate the relationship between fA concentrations
in adult natal males and the intersexual association patterns involving these
males because we had too few fecal samples from adult natal males in the
months before known conceptions. However, we found no relationship between
courtship or reproductive aggression and fA concentrations in adult natal males
whereas we did find such relationships in immigrant males. The results
presented here, as well as those by Holekamp and Smale (1998) and Holekamp
and Sisk (2003), strongly support the idea that males who are adults but have not
yet dispersed differ both behaviorally and physiologically from males who have
dispersed and immigrated into a new clan, and should be considered as such.

East and Hofer (2002) and Goymann et al. (2003) maintain that some natal male

67

hyenas in the Serengeti never disperse. Instead, these natal males are reported
to drop in social rank to the top of the immigrant male hierarchy in their natal
clans, and to begin competing for females at that time. These non-dispersing
natal males submit to all adult females, including those of lower maternal rank, in
all contexts (M.L. East pers. comm.). Although some adult natal males in the
Talek clan do show reproductive aggression and court females, they all do so at
lower rates than immigrant males (Holekamp and Smale, 1998; this study). The 6
adult natal males for which we calculated rates of reproductive aggression and
courtship were aggressive towards, but were never seen to submit to, lower-
ranking females in contexts other than courting, such as feeding. Thus adult natal
males in the Talek clan do not ever fall in social rank, as they are sometimes
reported to do in the Serengeti. This discrepancy appears to represent a very
interesting difference between the Talek and Serengeti populations, because it
suggests an alternative male strategy occasionally adopted by males in one
population but not the other.

Szykman et al. (2001) recently showed that immigrant male spotted
hyaenas associate most closely with females that are most likely to be fertile, and
now we have shown that male androgen levels increase with this association.
However, Szykman (2001) also examined the strength of associations between
females and the males known to sire litters born to these females, and found that
known sires varied greatly in how tightly they associated with their future mates
during the month preceding conception. This suggests that immigrant male

hyenas may adopt alternative reproductive tactics. If elevated androgen levels

68

are costly to male hyenas as they are reported to be in other vertebrate species
(Folstad and Karter, 1992; Wedekind and Folstad, 1994), then only males who
can afford the cost of associating closely with females are likely to do so. Another
potential cost to male hyenas of associating closely with breeding females is the
energetic cost of having these dominant females monopolize food resources. In
response to these potential costs, alternative reproductive tactics in the male

spotted hyena may be mediated by variation in androgenic and stress hormones.

69

CHAPTER 4
ANDROGENS DURING PREGNANCY IN WILD SPOTTED HYENAS:
THE INFLUENCE OF SOCIAL RANK AND MATERNAL EFFECTS ON CUB
BEHAVIOR
INTRODUCTION
Absolute female dominance is uncommon among mammals, but occurs in
some lemurs (Jolly, 1984; Kappeler, 1990; Young et al., 1990) and in spotted
hyenas (Crocuta crocuta). Hyenas live in dynamic fission-fusion groups of up to
80 individuals, called clans, that are structured by hierarchical rank relationships
(Kruuk, 1972; Tilson and Hamilton, 1984; Frank, 1986; Holekamp and Smale,
1990). All adult females are socially dominant to all adult males not born in the
clan (Kruuk, 1972; Smale et al., 1993), and adult females are more aggressive
than adult males (Kruuk, 1972; Frank, 1986; Mills, 1990; Szykman et al., 2003).
In addition, females are about 10% heavier than males (Hamilton et al., 1986)
and their genitalia are virilized and resemble the genitals of the males. The
clitoris is enlarged and forms an erectile pseudopenis, the labia are fused to form
a pseudo-scrotum, and the urogenital tract passes through the pseduopenis
(Matthews, 1939; Neaves et al., 1980; Frank et al., 1990; Cunha et al., 2003).
Female hyenas thus urinate, copulate, and give birth through the elongated
clitoris.
The prevailing hypothesis suggested to account for this unusual array of
masculine morphological and behavioral traits in the female spotted hyena has
been that elevated androgens produce large, aggressive females that are able to

dominate males in competitive feeding situations, with a byproduct of genital

7O

masculinization (Gould, 1981; Gould and Vrba, 1982; reviewed in Frank, 1997).
In mammals, including humans, masculinization and defeminization of the
reproductive system and behavior are induced by testosterone (T) or its
metabolites (androgens) during certain critical periods of perinatal development,
and hyperandrogenism during fetal life results in the physical virilization of female
fetuses (reviewed in Breedlove and Hampson, 2002). The current evidence
suggests that, in general, adult female hyenas do not have unusually elevated
levels of T when compared to male hyenas (reviewed in Goymann et al., 2002,
Dloniak et al., 2004), indicating that females are not simply more aggressive than
males due to higher levels of circulating T. However, in captivity, pregnant female
hyenas experience a great increase in various androgens during the second half
of gestation, and these concentrations cross the placenta and reach the
developing fetus (Yalcinkaya et al., 1993; Licht et al., 1998). In wild hyenas, it
has recently been shown that females have higher plasma T and fecal androgen
metabolite concentrations when pregnant than when lactating (Dloniak et al.,
2004). These data indicate that spotted hyenas are exposed to high
concentrations of androgens during fetal life, which could account for both the
external genital and behavioral masculinization of female offspring (Glickman et
al., 1987; Yalcinkaya et al., 1993; Glickman et al., 1998; Licht et al., 1998).
Recent research on captive hyenas has cast some doubt on this
hypothesis, however. Drea et al. (1998) showed that pregnant female hyenas
given anti-androgens during pregnancy still produce male and female cubs with

masculinized external genitalia. Although there were definite de-masculinizing

71

effects of prenatal anti-androgens on both male and female external genitalia,
this experiment indicated that the formation of the external genitalia of both male
and female spotted hyenas involves more than early androgen exposure, and
therefore presently can not be a mere side-effect of selection for a hormonally
mediated trait (Drea et al., 2002). Based on this experiment and the fact that
adult female hyenas do not always have unusually elevated circulating T levels,
East and Hofer (2002) concluded that the assumed link between androgens and
aggression does not exist in this species. However, although research has begun
to investigate the links between prenatal androgens and external genital
masculinization, the relationship between prenatal exposure to androgens and
the organization of behavior patterns has not yet been examined in spotted
hyenas. The organizational effects of androgens have been able to account for
the development of virtually all behavioral sex differences among juvenile
mammals studied to date (Goy, 1997; Stockman et al, 1986; Pellis et al., 1992;
Olioff and Stewart, 1978; Orgeur, 1995). In particular, sexual play and aggression
are known to be affected by organizational androgen effects of in a wide array of
mammals (reviewed in Becker et al., 2002). In addition, the influence of
androgens on behavioral and genital masculinization of a fetus can be uncoupled
(Goy et al., 1988; Goy, 1996). Therefore, the current evidence is certainly not
sufficient to rule out a maternal effect of prenatal androgens on behavior in this
species.

Maternal effects can occur whenever variation among maternal

phenotypes provides an additional, usually non-genetic, source of variation

72

among offspring phenotypes (Mousseau and Dingle, 1991). These maternal
effects represent a mechanism by which current environmental conditions
experienced by the mother can influence offspring phenotype, potentially
enhancing offspring fitness (Bernardo, 1996; Mousseau and Fox, 1998). In birds
and other oviparous species, the egg contents are an important route through
which maternal effects can be transmitted. Mothers in good condition often
produce relatively large eggs, and egg size is positively correlated with hatchling
body size, growth rate, and survival prospects (reviewed in Williams, 1994;
Christians, 2002). Variation in maternal investment also occurs by differential
transfer of hormones to eggs during their production. Recent studies on birds
have shown that nestlings hatching from eggs with relatively high maternal
androgen concentrations exhibit increased growth rates (Schwabl, 1996; Eising
et al., 2001), accelerated embryonic development (Eising et al., 2001), enhanced
development of the hatching muscle (Lipar and Ketterson, 2000), and enhanced
social rank (Schwabl, 1993). Importantly, the social environment of a laying bird
can influence the amount of androgen transferred to the eggs. Breeding density
(Schwabl, 1997; Reed and Vleck, 2001), female social rank (Miiller et al., 2002),
and aggressive interactions (Whittingham and Schwabl, 2002) have all been
shown to influence the amount of T allocated to eggs.

Among wild female spotted hyenas living in a clan, social rank determines
priority of access to food (Kruuk, 1972; Tilson and Hamilton, 1984; Frank, 1986;
Mills, 1990), and thus profoundly influences energy intake. A female’s social rank

is not determined by fighting ability or size, but rather is “inherited" from her

73

 

mother (Holekamp and Smale, 1991; 1993; Smale et al., 1993; Frank, 1986).
There is a strong effect of social rank on female reproductive success, with
higher ranking individuals displaying earlier age at first parturition, shorter
interbirth intervals, earlier weaning, and enhanced growth of cubs (Frank et al.,
1995; Holekamp et al., 1996; Holekamp and Smale, 2000). Within a clan, female
spotted hyenas therefore vary considerably with respect to condition depending
on their social rank. Interestingly, the highest level of circulating T thus far
reported in a free-ranging female spotted hyena was in an alpha female that
happened to be in the last trimester of gestation, suggesting a possible effect of
rank on circulating T levels in pregnant hyenas (Frank et al., 1985). Thus,
maternal transfer of androgens to hyena cubs may represent a mechanism for
adaptive maternal control of offspring phenotype in hyenas, as is also suggested
to occur in birds. In this study, we tested the hypothesis that maternal effects of
androgens during pregnancy influence offspring behavior in wild hyenas. This
hypothesis predicts that variation in fecal androgen concentrations in pregnant
female spotted hyenas should be related to female social status, and this
variation in turn should be related to rates of sexual play and aggression
observed in offspring.

ln rodents, stress experienced by a mother during pregnancy has also
been shown to influence offspring development and behavior via increased
glucocorticoids. Prenatally stressed males show reduced male-typical sexual
behavior (reviewed in Ward and Reed, 1985) and increased aggression

(Marchlewska-Koj et al., 2003). Prenatally stressed females are partially

74

masculinized (Kinsley and Bridges, 1988) and show decreased fecundity and
fertility (Herrenkohl, 1979). Because social rank may influence stress hormones '
levels in female hyenas (Goymann et al., 2001), we also investigated the role of
female glucocorticoid concentrations as a covariate in our analyses of maternal

androgen levels and offspring behavior.

METHODS

Studv Population

i“-.

The study was conducted in the Talek area of the Masai Mara National
Reserve, an area of open rolling grassland in southwest Kenya. The study
animals were members of one large, stable spotted hyena clan, inhabiting a
home range of about 65km2 (Boydston et al., 2001). Life history data are
available for all hyenas in the Talek clan from 1988 to the present. Adult hyenas
were known individually by their unique spot patterns, and cubs were identified
on the basis of distinctive scars, relative sizes, and patterns of molt from the
black pelage of young cubs to the spotted patterns of older ones. Sex was
determined by the dimorphic glans morphology of the erect phallus (Frank et al.,
1990). Data presented here were collected between July 1993 and July 2002,
during which time observers monitored the Talek hyenas 20-31 days per month
for approximately 6 hours per day. Data documenting agonistic behavior among
adult females during observation sessions were collected using critical incident
sampling (Altmann, 1974) of all aggressive and appeasement behaviors

observed in all social contexts. Social ranks were assigned based on a matrix of

75

outcomes of dyadic agonistic interactions (Martin and Bateson, 1988), as

described previously (Smale et al., 1993).

Fecal Sample Processing and Assays

We collected 53 fecal samples from a total of 43 pregnancies of 27 adult
female hyenas over 10 years. We could calculate a female’s day of gestation for
a given fecal sample by back-calculating from the birthdate of cubs resulting from
each pregnancy. When cubs were first seen, day of birth +/- 7 days was
estimated (Smale et al., 1993). The number of days between the sample and the
birthdate was subtracted from 110 days, the length of the gestation period in
spotted hyenas (Schneider, 1926), to give the day of gestation corresponding to
each sample. Age and current social rank of each female at each time of
sampling were also known from long-term records.

A fecal sample was collected when a hyena was directly observed to
defecate during an observation session. Each sample was first collected into a
labeled plastic bag at the site of defecation, then mixed and transferred to 3ml
cryovials. Samples were stored in liquid nitrogen until transported on dry ice to
the United States, where they were stored at —20°C or colder until processing.

Samples were extracted and assayed for fecal androgen metabolites (fA)
as described by Dloniak et al. (2004). In brief, steroids in 0.29 Iyophilized feces
were extracted with 100% ethanol. Fecal extracts were reconstituted and diluted
1:50 in phosphate buffered saline (PBS) and measured in duplicate for

immunoreactive androgens by enzymeimmunoassay (EIA). The androgen assay

76

utilized a testosterone antibody (R156/7) and a testosterone conjugate
(horseradish peroxidase) kindly provided by Dr. Bill Lasley and Coralie Monroe of
the University of California Davis. In hyena feces, this assay measures both
conjugated and unconjugated androgens, including testosterone,
androstenedione, and dyhydrotestosterone. Samples were assayed as part of a
much larger study (Dloniak et al., 2004). The mean duplicate coefficient of
variation for only the samples used in this study was 9.4%.

Fecal extracts were also assayed for fecal glucocorticoid metabolites
(fGC) with a modified version of an assay previously validated for use with
spotted hyena feces (Goymann et al., 1999). Reconstituted fecal extracts were
diluted 1:20 in steroid diluent and measured in duplicate for immunoreactive
glucocorticoids by radioimmunoassay kit (RIA). This assay kit utilizes a
corticosterone antibody (ICN Biomedicals, Costa Mesa, CA) that crossreacts with
the major glucocorticoid metabolites present in the feces of spotted hyenas, as
well as many other mammals (Goymann et al., 1999; Wasser et al., 2000). The
samples used in this study were assayed as part of a larger study. The
interassay coefficient of variation for the entire study was 11.11% (N = 9 assays,
a total of 788 samples). The intra-assay coefficient of variation was 5.3 +/- 0.4%.
The mean duplicate coefficient of variation for only the samples used in this study

was 4.5%.

77

Offspring Behavior

Subject animals for cub behavioral analyses included all cubs born to
females for which we had fecal samples from the second half of gestation, that
survived to at least 6 months of age. We chose to use only samples from the
second half of gestation in order to limit the effects of day of gestation of sample,
and to have comparable measures of fA in females at the time concentrations
are the most variable. This included 35 cubs in 8 mixed-sex twin litters, 2 male-
male litters and 4 female-female litters, as well as 4 singleton males and 3
singleton females. Spotted hyenas are born in an isolated natal den where they
live either alone or with a littennate sibling (East et al., 1989). At 2 to 4 weeks of
age they are brought to the clan’s communal den where multiple litters of various
ages live together in a shared burrow system. Cubs reside at the communal den
until 8-12 months of age. The communal den serves as the social center of the
clan and is periodically visited by all clan members, including adult females, adult
males, and older juveniles that no longer reside at the den. Den-dwelling cubs
spend much of the day underground, typically entering the den 2-3 hours after
sunrise and emerging in the late afternoon, approximately 2-3 hours before
sunset. Behavioral data from cubs were therefore collected between 0600 and
0900h, and between 1600 and 2000h.

Behavioral data were collected using critical incident sampling (Altmann,
1974) of all cubs between the ages of 2-6 months, and cubs were observed for
an average of 36.1 +/- 3.8 hours during this period. We recorded all mounts and

aggressions of cubs. Aggression included the following behaviors: lunge, snap,

78

bite, chase, displace, push, stand over, and intention movement to bite (Kruuk,
1972; Smale et al., 1993). Aggressive behavior could easily be distinguished
from rough and tumble play because it was always accompanied by distinctive
elements of a threatening posture with ears forward and the head and tail raised
(Kruuk, 1972). We considered aggression against a Iittermate separately from all
other aggressions, since a different set of variables may affect intra-Iitter
aggression than those affecting aggression among non-littennates. Play-
mounting involved one cub approaching another from behind, raising up on the
hind legs, and placing the forepaws on the other cub’s back with a posture like
that exhibited by adult males during copulation. The phallus was often, but not
always, erect during play mounting. For each cub, hourly rates of mounting and
aggression were calculated by dividing the numbers of observed behaviors of
each type by the number of hours each animal was observed during the four-
month interval when cubs were 2-6 months of age. At this age, cubs have not yet
attained their maternal social ranks (Holekamp and Smale, 1993; Smale et al.,
1993), so patterns of aggressive behavior among cubs are not yet shaped by the
rank relationships existing among their mothers. We focused on this age interval
because play-mounting generally disappears from the behavioral repetoire of

cubs by 6 months of age (Holekamp and Smale, 1998).

Statistica_l Analvses

We first investigated the relationship between social rank and fecal steroid

concentrations among pregnant females. We employed mixed model multiple

79

regressions (one for fA and one for fGC) that included the continuous predictor
variables of social rank, day of gestation, and age. Due to an unbalanced design
with some repeated measures, we also included individual hyena (ID) as a
random effect variable. This method effectively doubled our sample size, but may
still be considered pseudoreplication due to the fact that we had more than one
sample from some females. Therefore we also re-ran these two analyses on a
data set consisting of only a single sample from each female (N=27, one sample
from each female chosen at random) to determine the robustness of the
significant effects found in the larger data set.

Based on the multiple regression results, we next ran additional analyses
on the subset of fecal samples that would be used to investigate the relationship
between maternal hormones and cub behavior. We knew the sex composition of
each of these litters, therefore we could also include litter composition (all male,
all female, or mixed sex) as a variable. We investigated the relationship between
maternal hormones and cub behavior with separate ANCOVAs for aggression
and play mounting. We included sex of cub, maternal fA, residuals of maternal
fGC from the previous analysis, and maternal rank as predictor variables. The
residuals of the fGC values were used in these analyses in order to control for
the significant effect of day of gestation on fGC that was present in this data
subset (see below). Lastly, we investigated whether maternal fA, residuals of
fGC, or maternal social rank were related to dominant cub aggression towards
it’s subordinate littennate with a multiple regression. We only considered

dominant cub aggression because aggression in the other direction happens too

80

rarely to consider here. Significance was set at p = 0.05. All statistical analyses

were performed using Statistica 8.0 and followed Zar (1995).

RESULTS
M_atemal hormones durinq Gestation

We first investigated patterns in fA and fGC concentrations throughout
gestation in pregnant female hyenas. The multiple regression for fA
concentrations was significant (whole model R = 0.61, F5,“ = 9.62, p = 0.004;
Figure 4.1 ). Female lD did not have a significant effect and was therefore
removed from the model by backwards selection. fA concentrations increased as
gestation progressed (day of gestation partial R = 0.45, p < 0.001 ), and higher-
ranking females had higher fA concentrations than did lower-ranking females
during pregnancy (rank partial R = -0.46, p < 0.001). Age of the mother did not
have a significant effect on fA concentrations among pregnant females (age F =
0.71, p = 0.41 ). In addition, when we re-ran the multiple regression on just one
sample from each female (N=27), we still had a significant model (whole model R
= 0.62, F234 = 7.39, p = 0.003) with female social rank (partial R = -0.35, p =
0.049) and day of gestation (partial R = 0.41, p = 0.02) significantly influencing fA
concentrations.

The multiple regression model for fGC concentrations was also significant
(whole model R = 0.62, F547 = 10.43, p < 0.001; Figure 4.2). Female ID was not
significant and was removed from the model by backwards selection. Day of

gestation of sample explained a significant proportion of the variance in fGC

81

1500

1000

Focal androgens (“9’9 W)

 

Figure 4.1. Relationships among maternal social rank, day of gestation, and fecal

androgen concentrations in pregnant female hyenas. By convention, the highest
social rank possible is 1.

82

i

400

 

 

Fecal Glucocorticolds (nele M”)

 

 

r ....
0,, 5°

' 2000 990‘

Figure 4.2 Relationships among age of mother, day of gestation, and fecal
glucocorticoid concentrations in pregnant female spotted hyenas.

83

concentrations, with fGCs increasing in the second half of gestation (day of
gestation partial R = 0.57, p < 0.001). Social rank had no effect on fGC
concentrations in pregnant females (rank F = 0.63, p = 0.43), but older females
had greater fGC concentrations than younger females during gestation

(age partial R = 0.38, p = 0.006). When we re-ran the multiple regression on just
one sample from each female (N = 27), we still had a significant model (whole
model R = 0.70, F224 = 11.29, p < 0.001), including significant effects of maternal
age (partial R = 0.47, p = 0.016) and day of gestation (partial R = 0.62, p <
0.001).

We next determined whether these relationships between maternal rank
and day of gestation and fecal hormone concentrations still held when we only
looked at the samples collected during the second half of gestation that would be
used for our analyses of cub behavior. We used a mixed model multiple
regression and tested for an effect of litter composition on maternal hormones in
these samples by including it as a covariate. Within the subset of data there was
still a significant multiple regression model of fA concentrations (whole model R =
0.67, F446 = 4.54, p = 0.016). A significant proportion of the variance was
explained by maternal rank, with females of high rank having greater fA
concentrations than females of low rank (rank partial R = -0.46, p = 0.02, Figure
4.3A). However, among the samples collected exclusively during the second half
of gestation, day of gestation was not significant (day of gestation F = 3.16, p =
0.10, Figure 4.38). Furthermore, there was no effect of litter composition on

maternal fA concentrations (litter composition F = 0.12, p = 0.73; Table 4.1). The

84

multiple regression model for fGC concentrations was significant (whole model R
= 0.83, F415 = 8.82, p < 0.001), with day of gestation influencing fGC
concentrations (partial R = 0.75, p < 0.001, Figure 4.4A), but with age of mother
not significant (F = 0.512, p = 0.48). There was no effect of social rank (F = 0.20,
p = 0.66, Figure 4.4B) or litter composition (F = 0.24, p = 0.63, Table 4.1) on

maternal fGC concentrations during the second half of gestation.

Maternal Hormones and Cub Behavior

We found that rates of aggression and mounting in cubs were positively
related to maternal fA concentrations during the second half of gestation. Play
mounting was almost always directed toward other den-dwelling cubs, rather
than towards older juveniles or adults. The ANCOVA for cub mounting rate was
significant (Whole model R = 0.681, F430 = 6.48, p < 0.001), with cub sex (F =
8.13, p= 0.008) and maternal androgens (F = 7.91, p = 0.009) significant, as well
as a significant interaction between the two variables (sex x fA interaction F =
17.97, p < 0.001). In general, male cubs showed higher rates of play mounting
than did female cubs (Figure 4.5A and B). Maternal fA concentrations were
significantly and positively related to mounting rates in both sexes, however there
was a more robust relationship in males than in females, as indicated by different
slopes in the two sexes. Maternal fGC and maternal rank did not significantly
influence mounting rates of cubs (fGC residuals F = 0.86, p = 0.36; rank F = 0.16,
p = 0.69).

85

 

 

 

 

..
3 °°°I '
3 e
5 600 .
8
3‘ . "
“ e
" I . .
g 200 I .
IL
0
o J__... ##0##, __
0 5 1O 15 20 25
SocialRank
B

1000— . .
g 000- 0
g e
8
S: .' '

400 Q
2 I’ ' -
- .0.
g 2004'. .
“ I

,, l _ We

50 10 90 110
DayofGestatlon

Figure 4.3. Relationships between fecal androgens and social rank (A) and day
of gestation (B) in pregnant female spotted hyenas during the second half of
gestation.

86

 

 

1000 .
3 ml
0 = O
a I 0
5 I
,8 5001
‘5 I
:‘3’ I
§ 4001 0
3 j .
E9 200' . O .0 .
§ I“ ' 0. . e
u.
0 .' . :2”. _____
0 5 10 15 20 25
SocialRank
10004
g 3004
a e
a I e
5
.3 sooI
3
~43 I
g 400. 0
3 e
‘3 200 ’ . e
8 :0 e 0
ll.
0 ' e.
50 70 90 110
DayofGestatlon

Figure 4.4. Relationships between fecal glucocorticoids and social rank (A) and
day of gestation (B) in pregnant female spotted hyenas during the second half of
gestation.

87

Table 4.1. Mean iSE fecal androgen and glucocorticoid concentrations in

relation to litter composition during the second half of gestation in spotted
hyenas. Values are given as ng/g feces.

 

 

Litter Composition fA fGC N
Female 447.28 4 80.97 263.66 5 95.98 7 (4 twin, 3 single)“
Mixed 429.38 5117.99 113.09 3: 25.60 8
Male 436.53 4 34.42 369.28 5126.53 6 (2 twin, 4 single)

 

88

Cubs directed aggressive behavior towards other den-dwelling cubs and older
juveniles, and occasionally towards adult immigrant males and adult

females. Aggressive behavior occurred in the context of competition over access
to solid food as well as in other contexts. The ANCOVA applied to the aggression
data was significant (whole model R = 0.53, F430 = 2.92, p = 0.03; Figure 4.6 A,
B). Rates of aggression did not vary with cub sex (F = 0.0001, p = 0.99), and
maternal rank did not directly influence aggression in cubs (F = 0.10, p = 0.75).
However, there was a significant positive influence of maternal androgens on cub
aggression rates (F = 7.37, p = 0.011), and a significant cub sex by maternal
androgen interaction (F = 8.0813, p = 0.001). In male cubs, the relationship
between maternal androgens and aggression was weak, whereas in female cubs
the relationship was dramatic. Interestingly, the multiple regression model
investigating the relationships between maternal fA, fGC residuals, maternal
social rank, and rates of aggression directed only towards the subordinate sibling
by the dominant cub within a twin litter was not significant (Whole model R =
0.52, F = 1.30, p = 0.33). This suggests that a different relationship exists
between prenatal androgens and intralitter aggression than that existing between
prenatal androgens and aggression towards members of the clan other than

Iittermates (Figure 4.7).

DISCUSSION
We have shown here that fecal androgens increase during pregnancy in

wild spotted hyenas, corroborating prior results obtained from plasma of captive

89

§ 60 O
5 I
i I 0
~ 40
8 e 0
§ . o
5 20 ~
I o o
0 WWW m2-_9__.
0 200 400 600 800 1000

Maternal androgens (nglg feces)

12]

10

4—. e e

Mounts I hour it 100
O)

24:
. C

+__—-

o W—O—-~
0 200 400 600 800 1000

Maternal androgens (no/g feces)

Figure 4.5. Relationships between maternal androgens and mounting rates of
male (A) and female (B) cubs. Rates were calculated for the period when cubs
were 2-6 months old. Note different scales of y axes.

90

>

100 -
I
75 .

50 3 O

I O O
25IO O
i 00 O O

o
0 02-20222- ~-+ - we

0 200 400 600 800 1000
Maternal androgens (no/g feces)

Aggressions I hour x 100

100 ~,
I
75 1,
. e
50

. C
O
o . 8
_ 294222222 2 ,
200 400

600 800 1000
Maternal androgens (no/g feces)

25

Aggressions I hour it 100

I
C
0

0

Figure 4.6. Relationships between maternal androgens and cub aggression rates

in male (A) and female (B) cubs. Rates were calculated for the period when cubs
were 2-6 months old.

91

100 .. i Ofemale '
I

§ I L Omale
x 75 -' —- . —————-
E I:
E “7
‘3 25Il .. o
.. . .. . O o.
o ;,2,2_2_-2.____2 74—”;

0 200 400 600 800 1000
Maternal androgens (nglg feces)

Figure 4.7. Relationship between maternal androgens and rates of aggression
directed by dominant cubs towards their subordinate littermates, during the
period when cubs were 2-6 months of age.

92

hyenas (Licht et al., 1992; Yalcinkaya et al; 1993; Licht et al., 1998). We have
also shown that wild female hyenas vary in the degree to which fA concentrations
increase during the second half of gestation. Females of higher social rank
excrete greater concentrations of fA than females of lower social rank. In turn,
this variation in maternal fA concentrations appears to influence cub behavioral
development. Cubs born to females that have greater fA concentrations during
the second half of gestation show higher rates of both aggression and sexual
play than do cubs born to females with lower fA concentrations. Fecal
glucocorticoid concentrations, on the other hand, were not influenced by
maternal social rank, and natural variation in fGC concentrations during the
second half of gestation did not appear to influence offspring aggression or

mounting in this species.

GLJcocortjcofls in females

Although we did not detect an effect of maternal social rank on fGC
concentrations in hyenas during pregnancy, we did see a significant effect of
maternal age. Older females had greater fGC concentrations than younger
females throughout pregnancy. F ree-ranging yellow baboons (Papio
cynocephalus) show a general pattern of increased basal plasma cortisol
concentrations and glucocorticoid feedback resistance with increased age
(Sapolsky and Altmann, 1991), and there is a considerable increase in basal
glucocorticoid concentrations in aged rats (Sapolsky, 1992). Our finding that

older female spotted hyenas have elevated fGC profiles during pregnancy

93

therefore probably does not indicate that older females are more “stressed”
during pregnancy, but more likely indicates that fGC concentrations increase with
female age in general. In the only published study investigating correlates of fGC
concentrations in female spotted hyenas, age of female was not addressed
(Goymann et al., 2001). Based on these results and the pattern of glucocorticoid
increase with age in mammals, any future work in this species investigating
correlates of fGC concentrations should certainly consider the influence of age as
a covariate.

Goymann et al. (2001) showed that, in general, lactating female spotted
hyenas had higher fGCs than non-lactating females, but these authors did not
have a sufficient sample size to investigate whether fGCs were elevated in
pregnant females. In addition, Goymann et al. reported a significant influence of
social rank on fGCs in non-lactating females, with lower ranking females having
higher fGC concentrations than higher ranking females. However, the amount of
variance in fGC concentrations that this relationship explained was not reported,
and the correlation was weak at best (Figure 3 in Goymann et al., 2001). In the
same study, no relationship was observed between social rank and fGC
concentrations in lactating females. Here, we did not detect an influence of social
rank on fGC concentrations during pregnancy, but we did find a significant effect
of day of gestation on fGC concentrations. A rise in glucocorticoids, beginning at
mid-gestation and continuing through parturition, also occurs in female primates.
This rise is attributed to the development of the transitional zone of the fetal

adrenal glands, which begin to synthesize cortisol at about mid-gestation (Coulter

94

and Jaffe, 1998; Smith et al., 1999; Umezaki et al., 2001). The pattern of fGC
excretion in female hyenas during pregnancy fits this explanation. Based on
these results, it appears that low ranking female spotted hyenas are not
abnormally stressed during pregnancy, and that gestational stress probably does
not contribute to the lower reproductive success of lower ranking females in this
species. Additional research is now needed to determine the possible effects of
relatively old maternal age and extremely elevated fGCs (due to certain social

situations) on cub development in this species.

Maternal effects of androgens

Our work contributes to an extensive body of research indicating that the
prenatal sex hormone environment is extremely important with respect to both
the morphological and behavioral development of mammalian offspring. Many
studies have now investigated the effects of androgen administration or removal
on mammalian development (reviewed in Ward and Ward, 1985; Baum, 1987;
Hutchinson and Hutchinson, 1990; Goy, 1996). Fewer studies have investigated
natural variation in prenatal exposure to androgens, and virtually all of these
studies have been in rodents. In rats and mice, variation in intra-uterine position
results in variation in the exposure of fetuses to prenatal hormones of opposite-
sexed siblings, and thus to the organizing effects of those steroids on the
developing fetus (Clemens et al., 1978; Drickamer, 1996). Female mice that have
their first litter when they are middle-aged or adolescent have different T and

estradiol profiles than do young adult females during pregnancy. In turn, the male

95

offspring of middle-aged and adolescent mice have lower body, epididymis, and
testes weights than do the male offspring of young adult females (Wang and vom
Saal, 2000). To our knowledge, ours is the first study to demonstrate a
relationship between a free-ranging female mammal’s social environment and
her androgen profile during pregnancy. We do not yet know the mechanism by
which higher ranking females have higher androgen levels during gestation. An
increase in aggressive interactions or a nutritional mechanism seem plausible,
and future research will investigate these possibilities.

In addition to showing that higher ranking female spotted hyenas have
higher fA concentrations than lower ranking females during the second half of
gestation, we also showed that this variance in fA was related to offspring
phenotype. Cubs born to females with higher fA concentrations showed
increased rates of both aggression and play mounting than did cubs born to
females with lower fA concentrations. Thus the naturally-occurring variation in
maternal fA concentrations in pregnant female hyenas appears to influence
offspring behavior in ways similar to the effects of experimental administration or
removal of androgens in lab mammals. Male hyena cubs exposed to low prenatal
androgen concentrations are similar to male rats given flutamide (anti-androgen)
prenatally, which show reduced rates of mounting and intromissions (Casto et al.,
2003). Female spotted hyena cubs exposed to high prenatal androgen
concentrations are similar to female infant rhesus monkeys given prenatal
exogenous androgens, which results in masculinization of rough-and-tumble play

and mounting behavior (Goy et al., 1971; 1988). Females treated prenatally with

96

T propionate showed increased amounts of rough-and-tumble play and mounting
behavior, to levels intermediate between those of normal males and females
(Goy et al., 1971; 1988). Interestingly, prenatal exposure to T propionate had no
effect on the physical or behavioral development of male infant rhesus monkeys,
suggesting that additional administration of maternal androgens does not always
have similar effects in male and female offspring.

We found a sex difference in rates of mounting behavior in spotted hyena
cubs that appeared to be independent of maternal androgen exposure. The
same basic sex difference is seen in most juvenile mammals, and in other
species this sex difference during the juvenile pre-pubertal period is promoted by
differential exposure to testosterone earlier in development (Goy, 1997).
Although our results indicate no effect of litter composition on fA concentrations
in pregnant hyenas, it remains possible that males are exposed to slightly
different hormone concentrations in utero. Secretory activity of the prenatal
testes may elevate androgens locally in male fetuses at a stage of perinatal
development during which androgen levels have not yet been measured. Within
the hyena medial preoptic area and adjacent anterior hypothalamus there is a
sexually dimorphic nucleus that is twofold larger in males (Fenstermaker et al.,
1999). Although this is a modest difference compared to that observed in other
mammals, it does support the idea male and female offspring may be exposed to
slightly different prenatal hormonal environments in this species.

In red fronted lemurs (Eulemur fulvus rufus), another species exhibiting

female dominance and masculinized external genitalia, it was also found that fA

97

concentrations were similar in 5 pregnant females regardless of litter
composition. However, fecal estrogen (E) levels were much higher in pregnant
females carrying a male fetus than in females carrying a female fetus (Ostner et
al., 2003). These authors suggested that the relative ratio of E to T to which
fetuses are exposed may be the variable that is important as a masculinization
mechanism. This possibility now needs to be explored in hyenas. Another
possible explanation of sex differences in hyenas is that testosterone during the
early postnatal period, which appears to be elevated in males (Frank et al.,
1991), might promote the development of play mounting. In addition, female
hyena cubs that were exposed to prenantal antiandrogens had lower levels of
circulating androstenedione during the first six months of life than did control
cubs (Drea et al., 1998). Taken together with the results presented here, these
data support the epigenetic hypothesis proposed by Yalcinkaya et al. (1993),
which suggests that this antiandrogen-induced decrease in androgen levels in
females results from modifications in ovarian morphology and function. In order
to test this hypothesis directly, we would need to determine whether cubs
resulting from pregnancies associated with higher androgens have higher
androgens levels themselves than cubs resulting from pregnancies associated
with lower androgens. Hyena cubs rarely defecate during observation sessions at
the communal den and we do not attempt darting young cubs until they have left
the den, so we currently have very little information concerning concentrations of
excreted or circulating androgens in very young cubs and can’t directly test this

hypothesis yet.

98

Adaptive maternal effects?

The idea that naturally occurring variation in fetal exposure to androgens
may be adaptive in mammals has basically been ignored. However, it seems
reasonable that the plasticity of the prenatal organization of this physiological
system would allow environmental factors acting on the mother to “hard-wire" the
differentiated functions of an organ or tissue system to prepare the unborn
animal optimally for the environmental conditions it will soon encounter after birth
(Welberg and Seckl, 2001), analogous to the scenario proposed in some birds.
vorn Saal and Bronson (1980) proposed that, under certain ecological conditions,
female mice with a particular set of characteristics might be more likely to
reproduce than other females. Female house mice (Mus musculus) exposed to
higher levels of androgens in utero, due to intrauterine position between 2 males,
may have a reproductive advantage over other females when population density
is high, because they are highly aggressive toward other females, they fiercely
defend their young when lactating, and they enter puberty sooner when housed
in groups (vom Saal and Bronson, 1980; Zielinski et al., 1992).

Maternal effects can have adaptive value if they improve the performance
of offspring. The maternal transfer of androgens to cubs may be a mechanism for
adaptive maternal control of offspring phenotype in hyenas, as it has often been
suggested to be the case in birds. Female spotted hyena cubs showed a tight
relationship between maternal androgen concentrations and rates of aggression.
Increased aggressiveness in female offspring as they develop can easily be

considered adaptive in this species due to the high level of competition for

99

access to food resources (Kruuk, 1972). Male hyena cubs showed play mounting
behavior at much higher rates than females, and in both sexes the rate of
mounting was positively correlated with maternal fA concentration. An increase in
mounting rate may be considered adaptive for male hyenas, for, if any male
mammal needs to practice mounting as much as possible, it is the male spotted
hyena. Complete female masculinization makes copulation awkward for males,
and in fact it has been shown that males exposed to less T in utero do not
successfully copulate with females in captivity (Drea et al., 2002). It was
suggested that this was an effect of feminized penile morphology. Based on our
results, we suggest that prenatal androgen exposure may be just as important for
behavioral as for morphological masculinization in spotted hyenas. It would be
very interesting to test whether higher ranking sons exposed to higher prenatal
androgen concentrations were more successful at immigrating into neighboring
clans and siring cubs than sons exposed to lower prenatal androgen
concentrations.

The results presented here represent the first definitive link between
prenatal androgens and aggression in female spotted hyenas, and suggest an
adaptive maternal effect of prenatal hormone exposure in a mammal. Our results
do not support the alternative hypothesis put forth by East and colleagues that
higher androgen levels in neonates were selected for increased sibling rivalry
and facultative siblicide (East et al., 1993; East and Hofer, 2002). The lowest
ranking cubs in this study showed relatively higher rates of sibling aggression

than higher-ranking cubs, with no relationship to maternal androgen

100

concentrations. These results do, however, strongly indicate that the epigenetic
hypothesis of selection for increased androgens influencing aggressive behavior

in females in this species should not be thrown out just yet.

101

CHAPTER 5
CORRELATES OF FECAL GLUCOCORTICOID EXCRETION IN WILD ADULT
SPOTTED HYENAS
INTRODUCTION

Animals respond to stress with a series of endocrine responses that
increase the immediate availability of energy, in part by inhibiting physiological
processes that are not required for immediate survival (Munck et al., 1984;
Wingfield, 1994; Sapolsky, 2002). One of the primary responses to stress is an
increase in the activity of the hypothalamic-pituitary-adrenal (HPA) axis, resulting
in an increase in quantities of “stress hormones,” or glucocorticoids (GCs),
released into the bloodstream. The 60s most commonly released from the
mammalian adrenal cortex are cortisol and corticosterone (Sapolsky, 2002).
These steroids stimulate cardiovascular and musculoskeletal activity, and inhibit
a variety of costly anabolic processes such as digestion, energy storage, growth,
and reproduction (reviewed in Sapolsky et al., 2000; Sapolsky, 2002). Short-term
elevations of GCs lead to adaptive behavioral and physiological processes
(Wingfield et al., 1998), but chronic elevation may cause reproductive failure,
disease, and loss of muscle mass (Pottinger, 1999; Sapolsky, 2002).

A general theoretical framework has been proposed to help explain how
animals keep stress and 60s in balance. Allostasis is the process by which the
stability of central physiological “life support" parameters is maintained in the face
of environmental variation or variation among life history stages (McEwen, 1998;

McEwen and Wingfield, 2003). Through their release during times of stress, GCs

102

and other hormones, such as cytokines and catecholamines, are primary
mediators of allostasis (McEwen and Windfield, 2003). The “allostatic load” of an
individual refers to it’s cumulative physiological burden (Seeman et al., 2001) as
it adjusts to variation over time in conditions associated with seasonal changes,
parasite load, reproductive state, social status, injury, infection, and ageing. An
increase in allostatic load is typically accompanied by a rise in GC levels, a
response which may help the individual alter certain behavioral or physiological
processes in order to avoid or resist the potential for chronic stress (Wingfield et
al., 1998; Sapolsky et al., 2000).

A stressor is any stimulus (behavioral, environmental, demographic) that
provokes a physiological stress response, as measured by an increase in GO
concentrations (Creel et al., 2001; Sapolsky, 2002). For animals living in groups,
psychosocial factors such as agonistic interactions may represent stressors
influencing allostatic load (Goymann and Wingfield, 2004). In fact, in social
species with dominance hierarchies, 60 levels may be linked to dominance
status within the group (reviewed in Creel et al., 2001; Abbott et al., 2003; Creel
and Sands, 2003). A number of studies have found elevated cortisol levels in
subordinate individuals based on winner-loser situations in captivity (e.g.
Bronson and Eleftheriou, 1964; Louch and Higginbotham, 1967; Manogue,
1975). However, results are less clear in surveys of stable groups in social
species, where escalated antagonistic encounters are relatively rare. For
example, there is no consistent relationship between social status and 60

measures in primates. Only a few primate studies have demonstrated that

103

subordinates have higher GC levels than dominants (squirrel monkeys: Manogue
1975; olive baboons: Virgin and Sapolsky, 1997). Opposite results, in which
higher GC levels occurred among dominant individuals than among
subordinates, were obtained in marmosets (Saltzman et al., 1998), ring-tailed
lemurs (Cavigelli, 1999; Cavigelli et al., 2003), and Japanese macaques (Barrett
et al., 2002). No reliable rank-related variation in G0 measures was found in
rhesus macaques (Bercovitch and Clarke, 1995), long-tailed macaques (van
Schaik et al., 1991), cynomolgus monkeys (Stavisky et al., 2001), black tufted-
ear marmosets (Smith and French, 1997), or capuchin monkeys (Lynch et al.,
2002). Rather than social status per se, these primate studies suggest that better
predictors of relationships between GC levels and dominance status may be
whether or not subordinates are actually stressed more frequently than
dominants, and whether or not individuals enjoy opportunities for social support
by conspecifics (Abbott et al., 2003).

As in primates, among the few gregarious carnivore species in which 605
and social dominance have been studied there is no clear simple relationship
between these variables. Among wolves, dwarf mongooses, and African wild
dogs, dominant individuals typically have higher basal GC levels than
subordinates, although in dwarf mongooses this only holds for females (Creel et
al., 1992; 1996; 1997; Sands and Creel, 2004; reviewed in Creel et al., 2001;
Creel and Sands, 2003). All of these species are cooperative seasonal breeders,
in which reproductive success within any particular social group is generally

monopolized by the alpha pair. In the one well-studied carnivore species that is a

104

plural breeder, the spotted hyena, no relationship was found between dominance
status and GCs in males (Goymann et al., 2003), and a weak negative
relationship was observed between social status and 60 concentrations in non-
lactating females (Goymann et al., 2001).

In addition to social status or its correlates, allostatic load and GC
concentrations in gregarious mammals may also be strongly influenced by a
number of other variables, including age and reproductive condition of sampled
animals. In female mammals, reproduction often entails large energetic costs.
The energy-intensive processes associated with reproduction activate catabolic
processes that are part of HPA axis activation. Ovulation, pregnancy, and
lactation can all be associated with elevated GC levels (rabbits: Kriesten and
Murawski, 1988; humans: Allolio et al., 1990; Lockwood et al., 1996; marmosets:
Smith and French, 1997; Saltzman et al., 1998; ringtailed lemurs: Cavigelli, 1999;
Cavigelli et al.,2003; spotted hyenas: Goymann et al., 2001). A number of studies
also indicate that ovarian hormones stimulate the HPA axis. For example,
exogenous estrogen treatment increases cortisol levels in many primate species
(Coe et al., 1986; Pepe et al., 1982; Smith and Norman, 1987; Stavisky et al.,
2003)

In addition to social status and reproductive state, another potential
influence on GC levels in gregarious species may be local food availability. Food
deprivation increases circulating GC levels, such that mammals in the wild show
clear negative correlations between measures of GC concentrations and food

availability (mule deer: Saltz and White, 1991; elephants: Foley et al., 2001;

105

chimpanzees; Muller and Wrangham, 2003). Severe drought, during which food
is inevitably scarce, causes a decrease in social interaction rates and a decrease
in male plasma testosterone (T) levels in olive baboons (Sapolsky, 1986).
Although GC concentrations were not measured in that particular study, elevated
GC levels are associated with decreases in T in male baboons from this study
population (Sapolsky, 1982; 1985; 1986), and it was concluded that the
ecological stress of drought caused an increase in the amount of time animals
spent foraging, a decrease in social interaction rates, and a decrease in T in
baboons.

Here we investigated what factors predict GC levels in wild spotted
hyenas, in order to determine the relative influence of social status and other
factors on the HPA axis in this species. In addition to social status per se, we
also examined effects on GCs of reproductive state and local prey availability.
Spotted hyenas live in large, dynamic, fission-fusion social groups called clans.
Hyenas are unique among social carnivores in that they are plural breeders with
moderate reproductive skew (Holekamp and Smale, 2000). Each clan contains
one to several matrilines of adult females and their offspring, as well a number of
immigrant males. Clans are structured by hierarchical rank relationships
(Kruuk,1972; Tilson and Hamilton, 1984; Frank, 1986; Holekamp and Smale,
1990), and all adult females are socially dominant to all adult males not born in
the clan (Kruuk, 1972; Smale et al., 1993). Adult males are less aggressive than
adult females (Kruuk, 1972; Frank, 1986; Mills, 1990; Szykman et al., 2003), and

the social rank of individuals within the immigrant male hierarchy is acquired by

106

queuing, rather than by fighting (East and Hofer, 2001; Engh et al., 2002). Males
are seldom observed to engage in aggressive interactions, and what little
aggression does occur among them is typically very mild.

Spotted hyenas kill 65-95% of their own prey, mainly medium- and large-
bodied ungulates (Kruuk, 1972; Mills, 1990; Holekamp et al., 1997, Cooper et al.,
1999), and variation in local food availability may affect their ability to satisfy their
energetic and nutritional needs. Although Crocuta usually breed throughout the
year, in various populations distributed across the range of this species
reproductive seasonality is known to be associated with seasonal variations in
local prey abundance (Kruuk, 1972; Lindeque and Skinner, 1982; Mills, 1990;
Cooper, 1993; Holekamp et al., 1999). Fluctuations in prey availability may be
very important in affecting allostatic load and therefore GCs in this species.

We had two specific goals in this study. Our first objective was to extend
the biological validation of a fecal glucocorticoid (fGC) assay for use in spotted
hyenas. Goymann et al. (1999) developed a fGC enzymeimmunoassay for use in
spotted hyenas, but certain potentially important procedural covariates have not
been addressed until now. These covariates include time of day at which
samples were deposited, length of time samples were in storage, and age of the
hyena depositing the sample. Here, we employed a radioimmunoassay (RIA) kit
utilizing the same antibody (ICN corticosterone) as in Goymann et al. (1999). We
confirmed the biological validity of the hyena fGC RIA by injecting captive hyenas
with adrenocorticotrophic hormone (ACTH), expecting to see a post-injection

increase in fGC comparable to the results obtained in a similar experiment by

107

Goymann et al. (1999). We determined whether variation in diet affected fGC
concentrations by systematically manipulating the proportion of the diet derived
from bone in captive hyenas, and then measuring excreted glucocorticoids in
samples collected after feeding. We also inquired whether time of day of sample
deposition, time elapsed before freezing the sample, time elapsed between
freezing and assay, or age of the sampled individual had any systematic effects
on fGC concentrations in wild hyenas. After determining whether any of these
procedural covariates significantly influenced fGC concentrations in hyenas, our
second goal was to identify the best predictors of fGC concentrations in adult
male and female hyenas, using samples collected over nine years from one large
clan in the wild. Here we developed models testing the power of each of the
following variables to predict fGC concentrations: social status, female
reproductive state, local prey availability, male tenure in the clan, and interactions

among these variables.

METHODS
Captive study site. spbiect animapspgnd sample collection

All captive hyenas were housed at the University of California Berkeley
Field Station for Behavioral Research. These individuals were of known age and
reproductive status. Some hyenas were housed individually and others were
housed in small groups (for more details see Berger et al., 1992). To identify
feces produced by group-housed hyenas, their food was treated with food

coloring. All fecal samples were collected between 0800 and 1200h, were

108

immediately mixed thoroughly, and were then stored in individual containers at
-80° C until extraction and assay.

Adrenocorticotrophic hormone (ACTH) challenges were conducted on 5
captive adult hyenas (4 males and 1 female). Fecal samples were collected for 7
days prior to treatment to establish baseline levels of fGC excretion. On the day
of ACTH challenge, animals were immobilized with ketamine and xylazine
administered by blow dart, and anesthetized with isofiorane inhalant. Each hyena
then received a single injection of ACTH (100 IU ACTH, Sigma Chemical Co., St.
Louis, MO). Each hyena was allowed to recover from anesthesia and released
back to its home enclosure. Fecal samples were collected on the day of
challenge as well as for 10 days after the challenge.

Although wild hyenas ingest highly variable amounts of meat and bone,
captive hyenas in the Berkeley colony are fed a standard zoo carnivore mix
(Nebraska Brand Feline Food, Central Nebraska Packing, Inc., North Platte, NE)
and small amounts of bone every day (Berger et al., 1992). In order to assess
whether variation in the amount of bone in the diet influences measurement of
fGC concentrations, we varied the bone content in the diet of 5 adult females (as
in Dloniak et al., 2004). At the start of this experiment some individuals were fed
only feline diet while others were fed feline diet plus 3 or 4 sheep neck bones
each day. After 1 to 3 days on their respective diets, diets were reversed in all
subjects, and reversed again another 1 to 3 days later. Fecal samples were
collected each day from all subjects. For statistical analysis, a given fecal sample

represented the previous day’s diet.

109

Field study site, subject animals, and sample collection

Our field study site was the Talek area of the Masai Mara National
Reserve in southwest Kenya. The subject population was one large, stable
Crocuta clan inhabiting a home range of approximately 65 km2 (Boydston et al.,
2001). The Talek hyenas have been monitored intensively since June 1988, and
data used in this study were collected between January 1993 and July 2002. All
hyenas in the clan were identifiable based on each individual’s unique spot
pattern and other distinguishing marks. Sex was determined by the dimorphic
glans morphology of the erect phallus (Frank et al., 1990). Adult females were
those born in the Talek clan that were at least 36 months of age. Ages of females
at any given time were calculated from known birthdates, which were estimated
+/- 7 days when cubs were first seen (Smale et al., 1993). Ages of females in this
study ranged from 36 to 199 months. Immigrant males were adult males that had
dispersed from natal clans elsewhere before joining the Talek clan. Ages of
immigrant males were estimated based on measures of toothwear obtained
during immobilizations (Van Horn et al., 2003). Estimated ages of immigrant
males in this study ranged from 30 to 153 months. Date of first appearance in the
Talek clan was recorded for each immigrant male, and his tenure on a given date
was calculated as time elapsed since joining the clan (Holekamp and Smale,
1998). Critical incident sampling (Altmann, 1974) of all observed aggressive and
appeasement behaviors was used to determine social ranks of individuals. Social

ranks within each sex were assigned based on a matrix of outcomes of dyadic

110

agonistic interactions (Martin and Bateson, 1988), as described previously
(Smale et al., 1993).

Female reproductive state was determined by behavioral observations or
by assessment during immobilization. A female was considered to be pregnant if
she gave birth to cubs within 110 days after sampling (Schneider, 1926), or if
fetuses were observed in her uterine horns during immobilization, visualized with
a Hitachi portable ultrasound machine. A female was lactating if she was
observed to nurse cubs around the time of sampling and/or milk could be
expressed from teats when she was immobilized. Reproductive state was
considered to be “unknown” for females that were neither pregnant nor lactating -
at the time of sampling. These females may have been cycling, but they may
also include individuals for which it was impossible to determine whether or not
their cubs had been weaned at the time of sampling.

The Talek clan defends a territory at the northernmost edge of the
Serengeti ecosystem, an area of open, rolling grasslands grazed year round by
concentrations of several different ungulate species. The majority of the resident
population consists of Thomson’s gazelle (Gaze/Ia thomsonir), topi (Damiscilus
korrigum), and impala (Aepyceros melampus). Each year, the resident antelope
are joined for 3-4 months (usually June-September) by large migratory herds of
wildebeest (Connochaetes taun'nus) and zebra (Equus burchellr). Therefore
although there are resident prey year-round, there is also a period of high
abundance, and there are intermittent periods of relatively low prey availability

(Holekamp et al., 1993; Holekamp et al., 1999). Local availability of food to Talek

111

hyenas was estimated by counting all prey animals found within 100m of 4 km-
Iong transect lines in two different areas of the Talek clan home range (Holekamp
et al., 1999). Both transects were run 2 times each month, between 0800 and
1000h. A fecal sample could thus be related to a prey count that occurred within
14 days of the sample. We divided prey counts into three categories (low,
average, and high) for male and female samples separately. Within each sex, we
calculated the median prey count associated with all fecal samples, and
considered average prey abundance to be represented by the counts falling
within one interquartile interval above and below the median prey count. High
prey abundance then corresponded to counts in the top interquartile interval, and
low prey abundance corresponded to counts in the bottom interquartile interval.
In males, prey count ranges were as follows: low = 20-190; average = 199-452;
high = 457-4094. In females, prey counts ranged as follows: low = 0-172,
average = 173-515, high = 587- 4094.

Fecal samples were collected either during early morning (0530 — 0900h,
AM samples) or evening (1600 - 2000h. PM samples) observation periods.
Samples were collected whenever a hyena defecated, upon direct and
unambiguous observation. Samples were first collected into plastic bags at the
site of defecation, and later approximately 6 ml of mixed sample were transferred
to multiple 2 ml cryovials for freezing in liquid nitrogen. Ninety-four percent of the
samples were frozen within 12h of collection, and all samples were frozen within

48h of collection. Samples were stored in liquid nitrogen until shipped on dry ice

112

to the United States, where they were stored at -20° C or colder until extraction

and assay.

Mn and Assav of Glucocorticoids from Fecal Samples

Fecal samples were extracted in ethanol as described by Dloniak et al.
(2004). Samples were then diluted 1:20 in ICN steroid diluent and fGCs were
assayed with a double-antibody corticosterone radioimmunoassay kit (ICN
corticosterone RIA, Cat. No. 07-120102). This kit utilizes a corticosterone
antibody that demonstrates high cross-reactivities to the major glucocorticoid
metabolites present in the feces of many mammalian and avian species
(Goymann et al., 1999; Wasser et al., 2000). Serial dilutions of pooled extracts
from multiple male and female hyenas, respectively, produced displacement
curves that were parallel to the displacement curve produced by corticosterone
standards. The interassay coefficient of variation for the entire study was 11.11%

(N = 9 assays). The intra-assay coefficient of variation was 5.3 +/- 0.4%.

Procedural covariates

Collection conditions and amount of time spent in frozen storage before
assay varied among fecal samples. Conditions and time in storage can affect
assessments of steroid hormone concentrations in fecal samples in other species
(Khan et al., 2002, Terio et al., 2002), and circadian variation has also been
reported for some excreted steroids (Sousa and Ziegler, 1998; Muller and

Lipson, 2003). We therefore inquired whether time of defecation or variation in

113

i.

processing and storage conditions systematically affected fGC concentrations in
samples from wild male and female hyenas separately. Individuals in this study
were sampled repeatedly over almost 10 years. Since glucocorticoid levels are
known to increase with age in some other mammals (Sapolsky and Altmann,
1991; Sapolsky, 1992), we also investigated whether age of individual hyenas at

the time of sampling influenced their fGC concentrations.

Statistical analysis

All statistical treatment of data followed Zar (1995) and Quinn and Keough
(2002). Prior to statistical analysis, all variables were tested for departures from
normality and homoscedasticity. For the analyses involving wild hyenas, we used
a total of 110 fecal samples from 10 adult males and 341 fecal samples from 21
adult females. Each male and female contributed at least 5 samples to the data
sets, and samples from each individual were separated by at least 1 month. We
first used general linear models to determine whether any of the following
procedural covariates might systematically influence fGC concentrations in male
or female hyenas: time of day of sample (AM or PM), time elapsed before
freezing of sample (minutes), time elapsed before assay (days), and age in
months of individual at the time of sampling. The distributions of two variables,
number of minutes between collection and freezing of sample and fGC
concentrations for both males and females, were obviously skewed when
examined by use of normal probability plots. When natural log transformations

were applied to these variables their distributions were normal, therefore all

114

analyses were done on the transformed variables. Although analyses used log-
transformed fGC concentrations to meet the normality assumption, we present
the results as raw fGC values (and mean +/- sem) for ease of interpretation of
figures.

After we determined whether any of the procedural covariates were
important predictors of variation in fGC concentrations, we sought the models
that explained the greatest amount of variance in fGC excretion in adult hyenas
of each sex. Here, we used general linear models with mixed categorical and
continuous predictor variables. We included any significant covariates from the
procedural covariate analyses as variables within these models. In building the
male model, we investigated effects on fGC concentrations in each fecal sample
of each of the following variables: prey abundance (high, average, or low), social
rank (high, middle, low), tenure in the clan, and time of day of sample. Because
tenure and social rank are closely correlated in immigrant male hyenas (Smale et'
al., 1997), these variables were never included concurrently in a model. In
building the female model, we investigated the predictive power of prey
abundance (high, average, or low), social rank (high, middle, low), reproductive
state (pregnant, lactating, or unknown), and time of day of sample and age.
Because repeated measures were obtained from all individuals, we also tested
hyena identity (ID) as a possible predictor variable within each model to
determine whether substantial differences existed among individuals over time,
and whether ID alone might be important in any models of hyena fGC. We

added and removed variables and interaction terms from the models until we had

115

 

a significant model that explained the greatest amount of variance in fGC

concentrations within each sex (following Quinn and Keough, 2002).

RESULTS
ACTH challenges

Four of five captive hyenas treated with ACTH responded physiologically
to the challenge with an increase in fecal glucocorticoids from average baseline
levels (Figure 5.1, Table 5.1). The latency to peak fGC concentration varied from
one to four days after ACTH injection, and averaged 2.25 days. Among the 4
hyenas that responded to ACTH, peak fGC concentrations after challenge were
significantly higher than average baseline concentrations prior to challenge, but
levels six to eight days after challenge were no longer different from initial
baseline levels (Friedman ANOVA and Kendall coefficient of concordance, N = 4,

df = 2, )8: 8.0, p = 0.018).

Effect of bone i_n diet

The presence or absence of bone in the diet did not affect mean fGC
concentrations in feces collected the next day from adult female hyenas in
captivity. No differences in fGC were found when hyenas switched from a meat
plus bone diet to one consisting of only meat (meat plus bone = 131.03 +/- 32.24
nglg, meat = 196.85 +/- 83.83; Wilcoxon paired sample test: Z = 0.67, p = 0.50)
or the reverse (meat = 198.52 +/- 83.29 ng/g, meat plus bone = 183.12 +/- 51.12

ng/g; Wilcoxon paired sample test: Z = 0.13, p = 0.89).

116

I—o—“HWQEE;

   

12001 ,------Maie9

I ° :—-e—- Male 10

10002 I--0--Mal913
L-O—__ "___ngale 35'

 

 

  

200 '-
° ' /

Fecal glucocortlcoids (nglg feces)
OJ
8

 

o 17 tfi—r T ' v r-r i-i

-7-6-5-4-3-2-1012 3 4 5 6 7 8 910
Day relative to ACTH challenge

Figure 5.1. Changes in fecal glucocorticoid excretion in response to ACTH
challenge in five captive spotted hyenas. ACTH was administered on day 0, as
indicated by the arrow.

117

Table 5.1. Individual and mean response patterns to ACTH challenge in five
captive spotted hyenas. Fecal glucocorticoid values indicate mean 2 SEM and
are expressed as ng glucocorticoid per g Iyophilized feces. NR = no response to

 

 

challenge.
Hyena ID Baseline . Maximum. % Change Days to Maximum
glucocortrcord glucocorticoid
Male 3 325.52 2 60.16 959.67 295 1
Male 9 278.78 i 23.21 266.35 NR NR
Male 10 225.15 :t 38.25 646.50 287 4
Male 13 236.52 1 42.72 979.88 414 2
Female 35 235.11 1 55.45 1128.94 480 2
All responders 255.58 :I: 23.45 928.75 i 101.37 369 2.25

 

118

Procedural covariates

Males

The general linear model investigating effects on fGC concentrations of the
various procedural covariates in males was significant (R2 = 0.27, F4405 = 8.85, p
< 0.001). Within this model, neither the number of minutes elapsed between
sample collection and freezing (F = 0.585; p = 0.446) nor the number of days
frozen until extraction and assay (F = 0.044; p = 0.834) influenced fGC
concentrations in males. In addition, estimated age of male at the time of
sampling was not related to fGC concentrations (F = 0.099, p = 0.754). However,
the model did show a pronounced effect of time of day of sample deposition on
fGC concentrations in males (F = 30.98, p < 0.001). We investigated the
robustness of this effect by comparing paired mean AM and PM fGC
concentrations in males. Each male had at least two AM and two PM samples. In
all ten immigrant males, mean AM fGC values were higher than mean PM values

(paired t-test; t = 4.623, p = 0.001, Figure 5.2).

Females

In females, the general linear model of procedural covariates was also
significant, although it explained little variance in fGC concentrations (R2 = 0.08,
F4336 = 6.710, p < 0.001). As observed in the male model, neither the number of
minutes elapsed between sample collection and freezing (F = 2.94; p = 0.12), nor

the number of days frozen until extraction and assay (F = 1.47; p = 0.23),

119

78'

N
8

x
_...x._.—._..__ __ _

N
8

 

Mean fecal glucocorticoids (nglg feces)
8

 

100 7| \ '
50 % \\
V
0 2 _ t ._
AM PM

Figure 5.2. Paired comparison of mean morning (AM) and afternoon (PM) fecal
glucocorticoid concentrations in 10 adult immigrant male hyenas.

120

significantly influenced fGC concentrations in females. However, we found
significant effects of both age and time of day of sample in females. As in males,
fGC concentrations in females were greater in samples collected in the morning
than in those collected in the afternoon (F = 13.087, p < 0.001). We investigated
the robustness of the time of day effect by paired comparison of mean AM and
PM fGC concentrations in the 21 females. Each female had at least 2 AM and 2
PM samples. Of these females, 16 had greater mean fGC values for AM samples
than for PM samples, two had greater PM than AM means, and three had similar
AM and PM means (Paired t-test, t =2.493, p = 0.021, Figure 5.3). Lastly, as

females got older, fGC concentrations generally increased (F = 7.26, p = 0.008). '

Models of glucocortjcoid excretion in wild hvenas

Males

Due to the fact that time of day of sample had a significant effect on
concentrations of fGCs in male hyenas, we included this variable within our
model of ecological and social factors influencing male fGC levels. The best
model of male fGC concentrations was one that included male tenure, time of
day of sample, prey availability, and an interaction term for time of sample by
prey availability. Of these variables, only time of sample and prey availability
were significant (Table 5.2). Although included in the model, the interaction term
of time of day and prey abundance was not significant, indicating that the
influence of prey availability on fGC concentrations in males was similar in both

AM and PM samples. Immigrant males had higher fGC concentrations when prey

121

 

150 -

 

Mean fecal glucocorticoids (nglg feces)

AM PM

Figure 5.3. Paired comparison of mean morning (AM) and afternoon (PM) fecal
glucocorticoid concentrations in 21 adult female hyenas.

122

Table 5.2. Best model of fecal glucocorticoid concentrations in adult male Talek
hyenas. Whole model R2 = 0.30, F6403 = 7.37, p < 0.001.

 

 

Effect SS df MS F P
Intercept 369.55 1 369.55 438.94 < 0.0001
Tenure 1.52 1 1.52 1.80 0.183
Time“ 23.57 1 23.57 28.00 <0.0001
Prey” 5.79 2 2.90 3.44 0.036
Time*Prey 0.25 2 0.15 0.15 0.863
Error 86.72 103 0.84

 

“ Time represents time of day of sample (AM or PM)
b Prey represents prey abundance at time of sample (high, average, or low)

123

 

availability was low than when prey availability was average or high (Figure 5.4).
Social status did not have a significant effect on fGC concentrations in any
model, and it’s inclusion as a predictor variable actually caused the amount of
variation explained by the final model to decrease. Therefore, social status was

not included in the best model of fGC concentrations in males.

Females

Because we found significant effects of age and time of day of sample on fGC
concentrations in female hyenas in our previous analysis of procedural
covariates, we included these variables when building our model of social and
ecological effects on fGC concentrations in females. The best significant model
only explained a small amount of the variance in fGC concentrations in female
spotted hyenas (Table 5.3). This model included the variables age, time of
sample, reproductive state, and an interaction term for time of sample by
reproductive state. Although included in the model, the interaction term was not
significant, indicating that the influence of reproductive state was similar in both
AM and PM samples among females. Adult females had greater concentrations
of fGCs when pregnant than when lactating or when reproductive status was
unknown (Figure 5.5). Older females generally had greater fGC concentrations
than younger females, and this effect of age was apparent in both pregnant and
lactating females (Figure 5.6). Social status and prey availability did not have
significant effects on fGC concentrations in any model, nor did their inclusion as

predictor variables increase the amount of variance explained by the model.

124

I uAMsarrfles ‘
I .PMsarrples z

N
8

cut

8
_s
A

29 11

_..—24L ...

28

8

 

 

 

 

   

T

 

Fecal glucocorticoids (nglg feces)
8

LOW AVE HI
Prey availability

Figure 5.4. The influence of prey availability on fecal glucocorticoid
concentrations in adult male spotted hyenas. LOW = lowest quartile of prey
availability. AVE = middle 2 quartiles of prey availability. HI = upper quartile of
prey availability. AM samples were collected between 0530 and 0900h, whereas
PM samples were collected between 1600 and 2000h. Numbers above error bars
indicate numbers of fecal samples in categories.

125

Table 5.3. Best model of fecal glucocorticoid concentrations in adult female Talek
hyenas. Whole model R2 = 0.11, F6334 = 6.66, p < 0.001.

 

 

Effect ss df MS F P
Intercept 872.62 1 872.62 1059.39 < 0.0001
Age 3.41 1 3.41 4.14 0.042
Time“ 5.41 1 5.41 6.57 0.011
Repro” 19.60 2 9.80 11.90 <o.0001
Time*Repro 0.63 2 0.32 0.38 0.682

Error 275.12 334 0.82

 

a Time represents time of day of sample (AM or PM)
” Repro represents reproductive status at time of sample (pregnant, lactating, or
unknown)

126

h
8

 

DAM samples I
I PM samples I

 

 

U
C
O

109

 

'8‘

 

Fecal glucocortlcolds (nglg feces)
N
8

 

 

 

     

Preg Lact Unk
Female reproductive state

Figure 5.5. The influence of reproductive state on fecal glucocorticoid
concentrations in adult female spotted hyenas. Preg = pregnant, Lact = lactating,
and Unk = unknown. AM samples were collected between 0530 and 0900h,
whereas PM samples were collected between 1600 and 2000h. Numbers above
bars indicate numbers of fecal samples in categories.

127

8 §

Fecal glucocorticoids (nglg feces)
#
O
0
e
e

 

Age In months

B
1600I
‘3 .
3 I 0
2
g 1200
D
5
i I
'6 800 .
.0 O
a O
3 I . - -
g 400I 0 .
'51 I 0 e
_ O 0
ts ‘ . . . ’ O .
E 012023»: '2 - °

25 50 75 100 125 150 175 200

Age in months

Figure 5.6. Relationships between age and fecal glucocorticoid concentrations in
lactating (A) and pregnant (B) female spotted hyenas.

128

Therefore social status and prey availability were not included in the best model
of fGC concentrations in female hyenas. Lastly, in order to compare our results to
those of another study, we also ran one separate analysis using just the samples
from females in an unknown reproductive state to determine whether social
status was related to GC concentrations within this category. The correlation
between fGC concentrations and social status was not significant in these

females (R2 = 0.001, 1:1,...0 = 0.12, p = 0.73)

DISCUSSION

ACTH is a potent stimulator of the hypothalamic-pituitary-adrenal (HPA)
axis. The first step in the biological validation of our assay was to administer
ACTH challenges to 5 captive adult spotted hyenas. ACTH challenge led to
substantial increases in excreted fGC in 4 of the 5 hyenas challenged in this
study, with an average increase of 369% in excreted glucocorticoids across the
responding individuals, ranging from 287 to 480%. The timing of peak
glucocorticoid excretion was somewhat variable in that one hyena responded
maximally after one day, two hyenas responded after two days, and one
responded four days after injection with ACTH. A previous ACTH challenge
experiment found similar variation in peak glucocorticoid excretion in feces of
captive spotted hyenas after ACTH injection (Goymann et al., 1999). In that
study, while the average latency to peak excretion of fGC after injection was
about one day, some hyenas exhibited peak excretion rates more than two days

after injection. Regardless of the slight variation in response patterns, the

129

predicted response to ACTH challenge was seen in all but one subject, showing
that the ICN corticosterone RIA kit is capable of measuring physiologically-
induced elevations in fGC concentrations within individual spotted hyenas.
Although a response in fGCs to ACTH challenge in spotted hyenas had
been shown prior to this study (Goymann et al., 1999), possible procedural
covariates that might be influencing the measurement of fGC in samples from
wild hyenas had never been addressed. In spotted hyenas, bone constitutes a
variable proportion of the diet, and finding variation in fGC concentrations with
the presence or absence of bone might require correction for bone concentration
in samples. Therefore, we tested whether the presence or absence of bone in the
diet had an effect on fGC concentrations in captive female hyenas. Our results
indicate no obvious effect of bone on fGC concentrations, so we do not need to
adjust for bone content within samples. Slight variations in collection and storage
conditions might also potentially affect the ability to extract and detect fGCs in
hyena fecal samples, for reasons outlined previously (Dloniak et al., 2004). A
study in baboons showed that fGC concentrations in fecal samples stored in
ethanol at either room temperature or -20° C for 90-120 days differed from
concentrations in the same samples measured after 180 days of storage (Khan
et al., 2002). Therefore we explored this possibility, as well as the influence of the
length of time between sample deposition and freezing, with our spotted hyena
samples. Fortunately, our data suggest that variation in length of time elapsed
before freezing and long-term frozen storage do not systematically affect fGC

concentrations in spotted hyena feces.

130

We also investigated potential effects of two other procedural covariates
on fGC concentrations in hyenas: time of day of sample and age of individual at
the time of sampling. Numerous studies have described a circadian rhythm in the
release of glucocorticoids in various mammals (e.g. Bottoms et al., 1972;
Fulkerson et al., 1980; Hoffsis et al., 1970; Hudson et al., 1975; Ingram et al.,
1999; Irvine and Alexander, 1994; Knutsson et al., 1997). Serum cortisol
concentrations in diurnal mammals increase in the early morning and decrease in
the evening, and concentrations in nocturnal animals show the opposite pattern.
It is commonly assumed that using fecal samples integrates such variation by
“pooling” an animal’s daily fluctuations in circulating GC concentrations.
However, the effect of time of day of sample on fecal steroid hormones is rarely
examined. The results presented here show a clear effect of time of day of
sample on fGC concentrations in spotted hyena feces. Both male and female
hyenas had greater concentrations of fGCs in samples deposited in the morning
than in samples deposited in the evening. This effect was less pronounced in
females, but was nonetheless significant. In female marmosets, significantly
higher levels of fecal cortisol were found in afternoon samples than in early
morning samples (Sousa and Ziegler, 1998). Clearly it is important to consider
time of day of sample when evaluating predictors of fGC concentrations. Certain
variables, such as prey availability, may affect the general physiological state of
an individual, and thus influence morning and evening samples in parallel.
However, investigations of predictor variables such as acute social stress may

need to control for circadian variation. For example, if the average time lag

131

between circulating and excreted GCs is between 12 and 24 hours, care must be
taken to only collect fecal samples at one time of day, and corresponding
behavioral observations should be done 12 to 24 hours prior to the sample. In
spotted hyenas, our best estimate of the average time lag between circulating
and excreted 60s is between 24 and 72 hours. A more precise estimate of time
lag in wild hyenas will be necessary prior to any investigation of the influence of
acute stressors on fecal GC concentrations in this species.

In addition to time of day, a female’s age had a systematic effect on her
fGC concentrations. fGC concentrations increased with age in female hyenas,
although this relationship did not appear to be very strong and the model
including this variable did not explain much of the variance in female fGC
concentrations. This relationship seems to be driven by a lot of low fGC values
from relatively young females under 5 years of age (Figure 5.6). GC
concentrations also rise with age in rats (Sapolsky, 1992) and baboons
(Sapolsky and Altmann, 1991). The mechanism proposed to account for this rise
is increased GC feedback resistance with age. Although we do not yet know
whether this mechanism is responsible for the age-related increase in GCs in
female hyenas, our results clearly show that covariates such as age and time of
day of sample simply should be checked before investigating potential behavioral
and ecological predictor variables of steroid hormone excretion in any species.

In our models of fGC excretion in male and female hyenas, social status
did not explain any of the variation in fGC concentrations in either sex, and

therefore was not included in either model. In another study that investigated the.

132

relationship between social status and fGC concentrations in male hyenas in the
Serengeti, there was also no effect of social status on fGCs. We agree with the
interpretation of these results offered by Goymann et al. (2003), that the lack of a
relationship between social status and fGCs is probably a result of the low
frequency and intensity of agonistic interactions among male spotted hyenas
(Frank et al., 1986; Szykman et al., 2003; Dloniak et al., 2004). Low or high
social status within the male social hierarchy may not affect variation in the
allostatic load borne by this species. Our results in male hyenas thus support a
growing body of evidence in the literature that GC levels are not primarily
influenced by social status in species with low frequencies and intensities of
agonistic interactions among males (reviewed in Sapolsky, 2002).

Social status also explained none of the variance in female hyena fGC
concentrations in this study, and this variable was not included in the best female
model. Similarly, no relationship between social status and GCs was found in
studies of female macaques (van Schaik et al., 1991) or female chacma baboons
(Weingrill et al., 2004). However, Goymann et al. (2001) reported a significant
influence of social rank on fGCs in non-lactating female hyenas in the Serengeti, -
with lower ranking females having higher fGC concentrations than higher ranking '
females. However, the amount of variance in fGC concentrations that this
relationship explained was not reported, and the relationship was weak (Figure 3
in Goymann et al., 2001). In the same study, no relationship was found between
social rank and fGC concentrations in lactating females. Although we cannot

directly compare our results with those of Goymann et al. due to methodological

133

differences, we did run a simple correlation between social rank and fGC values
in Talek females assigned to our “unknown” category of reproductive state. In
that analysis we found no relationship between social status and fGC
concentrations.

The degree to which social support is available for subordinates is an
important variable that predicts rank-related variation in G0 levels in primates
(Abbott et al., 2003). Female hyenas live in large stable groups with close kin as
well as unrelated individuals. Female hyenas often do have access to social
support from relatives, and this may explain a lack of a relationship between their
dominance status and fGC concentrations. However, lower ranking females often
have fewer potential allies than higher-ranking females, so we might expect
higher ranking females to have lower fGC values if social support is important. In
order to test this possibility it will be necessary to examine fGC concentrations in
females of similar social rank with and without close relatives present in the clan.
Alternatively, due to the fission-fusion nature of spotted hyena society, females
may simply minimize psychosocial stress by avoiding other individuals as
necessary.

In female spotted hyenas, we found that reproductive state influenced fGC
concentrations. Pregnant females had higher fGC concentrations than did
lactating females or females in which reproductive state was unknown. In
Chapter 4, we showed that fGC concentrations increase at about mid-gestation
in pregnant female spotted hyenas (Figure 4.2). A rise in glucocorticoids,

beginning at mid-gestation and continuing through parturition, also occurs in

134

pregnant female primates, and is attributed to the development of the transitional
zone of the fetal adrenal glands, which begin to synthesize cortisol at about mid- '
gestation (Coulter and Jaffe, 1998; Smith et al., 1999; Umezaki et al., 2001). In a
previous study addressing the influence of female reproductive state on fGC
concentrations in hyenas, Goymann et al. (2001) showed that, in general,
lactating female spotted hyenas had higher fGCs than non-lactating females.
However, these authors did not have a sufficient sample size to investigate
whether fGCs were elevated in pregnant females. There might be an interesting
difference between the two hyena populations, with Serengeti females having a "
“more stressful” time during lactation than Talek females. However, we suspect
that part of the difference between studies may be explained by an influence of
an age bias in the earlier study. Goymann and colleagues considered female
hyenas to be adults at the age of 2, although many females do not start to
reproduce for at least another year or two after this. Therefore, the “non-
pregnant” category in that study may be biased towards young females, which
tend to have lower fGC levels than older females regardless of reproductive
state.

In our hyena study population, previous research has shown that an
annual birth trough occurred approximately one gestation period after the phase
of the annual cycle during which prey animals were least abundant in the home
range of the study clan (Holekamp et al., 1999). Litters were least likely to be
conceived when game was relatively scarce, indicating that prey availability has

important effects on the timing of reproduction in Talek hyenas. Although we did

135

not find evidence of an influence of prey availability on female fGC
concentrations in this study, it remains to be determined whether females who
have higher concentrations of GCs while cycling have less chance of conceiving.
Alternatively, the influence of prey availability on conception in female hyenas
might be mediated by a mechanism other than one involving glucocorticoids.

Although we found no effect of prey availability on fGC concentrations in
adult female hyenas, our results suggest that food deprivation has an important
effect on fGC concentrations in males. All adult male hyenas are subordinate to
all adult females. Therefore, when food becomes scarce, males may suffer the
consequences more than females. The results presented here indicate that
female dominance over males may indeed buffer females from effects of
variation in the food supply.

Within limits, an increase in allostatic load represents an adaptive
response to seasonal and other demands. Overall, spotted hyenas seem
marvelously well-adapted to their environments, and they are extremely
successful. We found no evidence of chronic social stress or periods with
extreme allostatic load in the Talek hyenas. However, if additional stressors such
as human disturbance are superimposed, allostatic load might still increase
dramatically in this species. The results presented here suggest that periods of
low prey availability may represent a time when male spotted hyenas might be
particularly vulnerable to added pressures. By contrast, female spotted hyenas
do not seem to be as strongly influenced by a relative decrease in prey

availability. However, the fact that stress hormones increase during pregnancy

136

 

suggests that this may be a life history stage when female hyenas are more
vulnerable to additional pressures. Future work will investigate whether
individuals in these potential higher-risk categories do indeed suffer more from

added disturbance.

137

 

CHAPTER 6

CONCLUSIONS

The original motivation for studying the socioendocrinology of spotted
hyenas was sparked by their complex social system, and unusual characteristics
such as female dominance and behavioral masculinization. As stated in the
introduction, the goal of socioendocrinology is to understand the links among the
social environment, hormones, and behavior in order to know how they modulate
the reproductive success of individuals living in social groups (Bercovitch and
Ziegler, 1990). Evolution has led to individuals with physiological flexibilities that
enable them to respond to their social surroundings and environment in ways
that maximize reproductive success, and socioendocrine factors can mediate
ontogenetic trajectories that span the entire period between embryonic
development and death. A socioendocrine perspective provides a unique
framework for connecting reproductive endocrinology with evolutionary biology,
thereby integrating proximate and ultimate levels of analysis of behavior. I was
inspired to investigate how various endocrine mechanisms may shape or
mediate the behavior of individuals within hyena society, both to determine
whether hyenas might exhibit unique mechanisms compared to other
vertebrates, and to test the generality of certain recent theories in behavioral
endocrinology‘that have been developed by studying various other mammalian
and avian species. A focus on the individual, and on the social relationships

among individuals, ultimately leads to an understanding of social systems in

138

terms of the adaptive strategies of their constituent members (Clutton-Brock and
Harvey, 1976).

The work presented in this dissertation demonstrates that spotted hyenas
can be considered a model system for the study of socioendocrinology, in
addition to behavioral ecology and development (Holekamp and Smale, 2000;
Holekamp and Smale, 1998). Due to long-term studies of known individual
hyenas in a few populations over the last twenty years, we now have enough
data from this species to be able to tie together such diverse aspects of their
lives as behavior, reproduction, life history, and ecology, in order to answer
numerous fundamental questions of interest in behavioral ecology. In addition,
with the development of both traditional and non-invasive endocrine techniques
for use in this species, we can now integrate yet another facet of hyena biology, '.
their socioendocrinology. This multifaceted approach allows us to thoroughly
understand the particular system of interest, and perhaps more importantly to put
it all together to address questions at multiple levels to provide a foundation for
behavioral ecology that constantly produces interesting new areas for further
exploration (Dugatkin, 2001).

The strength of such studies lies in direct proportion to the extent that they
both put existing concepts to the test, and generate new models and ideas; on
the one hand we can use empirical study of a model system to test well-
established theories and on the other we can develop new hypotheses
(Dugatkin, 2001). In addition to the integration of various fields and disciplines,

model systems are thus often springboards for new directions. As large

139

carnivores living a social existence more similar to that of baboons than that of
wild dogs or lions, spotted hyenas pose many interesting questions about the
factors forcing them to live in groups that are structured remarkably differently
from the other social carnivores, as well as questions about the consequences of
group living on behavior and physiology (Creel, 2001). Thus, fundamental
questions about ecology, hormones, and behavior can be addressed by
investigating spotted hyenas, with the hope of also generating new theories and
questions at the same time.

In this dissertation, I believe that l have successfully shown that spotted
hyenas are indeed both a fascinating and useful species in which to study
socioendocrinology. This is particularly true when we consider the hyena as
representative of certain patterns, but at the same time somewhat unique. For
example, the spotted hyena is a typical polygynous mammalian species in which
males live within a dominance hierarchy, mate with multiple females, and provide
no paternal care to offpsring. However, spotted hyenas are also quite unusual, in
that all adult females dominate all adult males and males do not engage in much.
direct aggressive competition over females, perhaps due to the high degree of
female control over mating. By determining whether there are similarities in
hormone-behavior relationships across closely related species that differ
remarkably in behavioral and social parameters, or across distantly related
species that are more similar, we can determine whether these parameters
indeed shape or are shaped by hormone-behavior relationships within a given

system.

140

In Chapter 3, we attempted to test the “challenge hypothesis’ (Wingfield et
al. 1993; 2000) in spotted hyenas. Although originally proposed in birds, recent
work has suggested that the challenge hypothesis may be applicable to
mammals as well. Our goals were to determine the best predictors of elevated
androgen levels in adult male spotted hyenas, and to elucidate whether sexually
motivated aggression against other males or interactions with females had a
strong relationship with fecal androgen concentrations. Whereas male-male
competition is predicted in most polygynous species, male aggression towards
other males is surprisingly low in Crocuta (Kruuk, 1972; Frank, 1986; Szykman et
al., 2003; Chapter 3). In addition, all adult males are socially subordinate to adult
females and females have extraordinary control over reproduction due to their
dominance status and masculinized external genitalia. Due to these
characteristics, spotted hyenas do not appear to fit neatly within the challenge
hypothesis framework, which has a lack of predictive power for species with
males that show low levels of male-male aggression and no paternal care
(Wingfield et al., 2000). However, if we consider interactions with attractive but
socially dominant females to be an important “challenge” for male hyenas (as
opposed to male-male competition), we are better able to predict male androgen
levels: males interact with females at high levels, and they should respond to
these challenges with elevated androgen levels during periods characterized by
high levels of association. Our results do seem to support the challenge
hypothesis if we consider interacting with female hyenas to be the key challenge

for male hyenas. Thus our work sheds light on how the challenge hypothesis can

141

be applied to complex social mammals. We have in effect tested an hypothesis
proposed to be generally applicable across taxa and found some support, but
suggest that the hypothesis may need to be altered or re-defined somewhat
depending on certain characteristics of a given social system.

The work presented in Chapter 4 also attempted to determine whether
another model developed in birds may be applicable to a social mammal,
suggesting it could be a more widespread, important phenomenon that deserves
further study. The evolution of maternal effects is a fascinating new area of
research. These effects can occur whenever variation among maternal
phenotypes provides a source of variation among offspring phenotypes
(Mousseau and Dingle, 1991). In birds, population density during the breeding
season, social rank, and aggressive interactions have all been shown to
influence the amount of androgen allocated to eggs. Nestlings hatching from
eggs with relatively high androgen concentrations exhibit increased growth rates,
accelerated embryonic development, and enhanced social rank (see Chapter 4).
We set out to determine whether maternal effects of androgens were important
mediators of behavioral development in spotted hyenas. High-ranking female
hyenas begin breeding at younger ages, are more frequently able to support
pregnancy and lactation concurrently, experience shorter intervals between
litters, and their offspring are more likely to survive to adulthood than are those of
lower-ranking females. We show here that higher-ranking females also have
greater fecal androgen concentrations during pregnancy than lower-ranking

females.

142

Many studies have shown the effects of prenatal androgens on the
organization of behavior in mammals. However, these studies have been in the
context of attempting to understand the origins of sexually dimorphic behavior
and sexual differentiation of morphology. Few studies have even suggested that
natural variation in prenatal androgen exposure could be adaptive in mammals
(vom Saal and Bronson, 1980, Zielinski et al., 1992), and no study has yet
investigated whether a mother’s social rank or her social environment might
influence prenatal androgen exposure of offspring, thus influencing behavioral
development of her offspring, in an adaptive fashion. Here we present the first
strong mammalian support for this hypothesis, and we suggest a potentially new
mechanism for transfer of social status-related traits, via prenatal hormone
exposure.

While we have contributed to an extensive body of research indicating that
the prenatal hormonal environment is extremely important, we have extended it.
By studying this phenomenon in hyenas, we have extended avian work to
another taxon, showing that the maternal effects of androgens might potentially
be having very important effects on the evolution of all sorts of behavior in a large
variety of species. This study increases our knowledge of the long-term
evolutionary processes behind the hyena’s social system, which presumably
might also be operating in other social mammals. For this species in particular,
we have shown the first definitive link between androgens and aggressive
behavior in spotted hyenas, suggesting support for the epigenetic mechanism of

female masculinization proposed by Yalcinkaya et al. (1993). The spotted hyena

143

may prove to be a particularly useful model for determining any far-reaching
effects of prenatal androgen exposure. This potential mechanism of selection is
just now being studied in a small number of vertebrate species, and we have
made a provocative contribution that will hopefully lead to more investigations.
This link between physiological mechanism and function may be a new and very
helpful way to study the evolution and maintenance of social systems.

The two chapters of this dissertation concerning the roles of androgens
within spotted hyena society thus have both tested and extended hypotheses
originally developed in other model systems. Chapters 3 and 4 are therefore
important contributions towards understanding the general applicability of broad
theories by testing their fit in this unique system. Although mainly descriptive in
nature, Chapter 5 also leads to new hypotheses to test. Our goal in Chapter 5
was to determine what factors predict glucocorticoid (GC) levels in wild spotted
hyenas, in order to determine the relative influence of social status and other
factors on the HPA axis in this species. We found no evidence of chronic social
stress or periods with extreme allostatic load in the Talek hyenas. Our results in
male hyenas thus support a growing body of evidence that GC levels are not
primarily influenced by social status in species with low frequencies and
intensities of agonistic interactions among males (reviewed in Sapolsky, 2002).

Although we found no effect of prey availability on GC concentrations in
adult female hyenas, our results suggest that food deprivation has an important
effect on GC concentrations in males. All adult male hyenas are subordinate to

all adult females. Therefore, when food becomes scarce, males may suffer the

144

consequences more than females. The results presented here indicate that
female dominance over males may buffer females from effects of variation in the
food supply. Due to female dominance over males, when males are spending a
lot of time courting females, they may not be eating as much. Female interactions
may directly influence GC levels in males, or food deprivation may mediate an
increase in GCs. We can now investigate these possibilities.

Studying new species is important in an obligately comparative discipline
such as socioendocrinology, for in this field variety is informative. A logical
extension of my thesis work, for example, might be to extend this line of research
to other female-dominated species in order to determine whether male-female
interactions are associated with increased androgens in these species as well as
in Crocuta, and to determine whether the challenge hypothesis is truly widely
applicable or whether female-dominated species represent an exception to the
general ‘rule’ described by this hypothesis. Lastly, little work has been done
investigating the normal roles of androgens in female mammals, and new
research should certainly investigate how widespread maternal effects of
androgens may be within this large taxonomic group. Although the female
spotted hyena is quite unique, it is surely possible that variation in androgens
could be playing subtle, but important roles in the development of offspring
and/or the influence of dominance status on reproductive success in many other
species.

Although we have been able to address many interesting questions

regarding the socioendocrinology of the spotted hyena, we have only conducted

145

observational experiments thus far, and our results are mainly correlational.
While purely observational research provides a critical test of the adaptive
significance of natural behaviors in the real world, more detailed manipulative
experiments offer another way forward. More work needs to be done to
investigate cause and effect now. For example, it would be illuminating to give
exogenous androgens to males and females to determine whether elevated
androgens do indeed cause increased rates of behavior. In order to determine
the direct effects of androgens on offspring development we could do
experiments giving androgens to pregnant females of various social ranks to
determine the extent of the effect of prenatal androgens on offspring
development. This field remains wide open, and investigating relationships
between hormones and behavior in more diverse taxa will allow us to determine
what is unique to each species and what general patterns emerge across many
species, telling us more about the evolution of these mechanisms.

Only by benefit of a long-term study program such as the Mara Hyena
Project are we able to generate the right questions in order to focus our efforts in
the most useful ways when it comes to manipulative field experiments. All of the
data presented in this dissertation represents the work of many people during 15
years of intensive research. Socioendocrinology often requires intensive field
work with known individual subject animals, so it requires a lot of time and effort.
By determining correlations and testing certain hypotheses with non-invasive
methods, we can continue in ways that will best focus our efforts. As more

studies involve scioendocrinology of diverse taxa, social systems, mating

146

systems, and behavioral repertoires, we will begin to be able to employ an even
stronger comparative method for testing whether various hormone-behavior
patterns are common to diverse species living in similar systems. The spotted
hyena is a fascinating creature, and a wonderful model for socioendocrine
research. This dissertation is but a beginning to the vast amount of work left to
do, however I hope that it is a strong foundation. We have answered some
important questions, but have generated many, many more. With luck, the hyena
project will continue for many more years and data will continue to be collected in

order to answer them.

147

LITERATURE CITED

Abbott DH, Keveme, EB, Bercovitch, FB, Shively, CA, Mendoza, SP, Saltzman,
W, Snowdon, CT. Ziegler, TE, Banjevic, M, Garland, T, Sapolsky, RM,
2003. Are subordinates always stressed? A comparative analysis of rank

differences in cortisol levels among primates. Hormones and Behavior 43,
67-82.

Abbott, DH, Saltzman, W, Shultz-Draken, NJ, Smith, TE. 1997. Specific
neuroendocrine mechanisms not involving general stress mediate social
regulation of female reproduction in cooperative breeding marmoset
monkeys. Annals of the New York Academy of Sciences 807, 219-238.

Albert, DJ, Jonik, RH, Watson, NV, Gorzalka, BB, Walsh, ML, 1990.
Horrnone-dependent aggression in male rats is proportional to serum
testosterone concentration but sexual behaviour is not. Physiology and
Behavior 48, 409-416.

Alberts, SC, Sapolsky, RM, Altmann, J, 1992. Behavioral, endocrine, and
immunological correlates of immigration by an aggressive male into a
natural primate group. Hormones and Behavior 26, 167-178.

Allolio, B, Hoffmann, J, Linton, EA, Winkelmann, W, Kusche, M, Schulte
HM, 1990. Diurnal salivary cortisol patterns during pregnancy and after
delivery: Relationship to plasma corticotrophin-releasing-hormone. Clinical
Endocrinology 33, 279-289.

Altmann, J, 1974. Observational study of behavior: Sampling methods.
Behaviour 49, 227-256.

Balthazart, J, Tlemcani, 0, Ball, GF, 1996. Do sex differences in the brain
explain sex differences in the hormonal induction of reproductive

behaviour? What 25 years of research on Japanese quail tells us.
Hormones and Behavior 30, 627-661.

Barrett, GM, Shimizu, K, Bardi, M, Asaba, S, Mori, A, 2002. Endocrine
correlates of rank, reproduction, and female-directed aggression in male

Japanese macaques (Macaca fuscata). Hormones and Behavior 42, 85-
96.

Baum, MJ, 2002. Neuroendocrinology of sexual behaviour in the male. In:
Behavioral Endocrinology (Ed. by Becker, JB, Breedlove, SM, Crews,
D, McCarthy, MM), pp. 153-203. Cambridge, Mass: MIT Press.

Baum, MJ, Kingsbury, PA, Erksine, MS, 1987. Failure of the synthetic androgen
1 7-beta-hydroxy-17-alpha-methyl-ester-4,9,1 1-triene-3-one

148

(methyltrienolone, R1881 ) to duplicate the activational effect of
testosterone on mating in castrated rats. Journal of Endocrinology. 113,
1 5-20.

Becker, JB, Breedlove, SM, 2002. Introduction to Behavioral Endocrinology. In:
Behavioral Endocrinology (Ed. by Becker, JB, Breedlove, SM, Crews,
D, McCarthy, MM), pp. 1-38. Cambridge, Mass: MIT Press.

Beletsky, LD, Gori, DF, Freeman, S, Wingfield, JC, 1995. Testosterone and
polygyny in birds. In: Current Ornithology, Vol. 12 (Ed. by Power, DM),
pp. 1-41. New York: Plenum.

Bercovitch, FB, 1991. Social stratification, social strategies and reproductive
tactics in adult male savannah baboons. Animal Behaviour 36, 47-56.

Bercovitch, FB, 1999. The physiology of male reproductive strategies. In: The
Nonhuman Primates (Ed. by Dolhinow, P, Fuentes, A), pp 237-144.
Mountain View, California: Mayfield.

Bercovitch, FB, Clarke, SA, 1995. Dominance rank, cortisol concentrations, and
reproductive maturation in male rhesus macaques. Physiology and
Behavior 58, 215-221.

Bercovitch, FB, Goy, RW, 1990. The socioendocrinology of reproductive
development and reproductive success in macaques. In:
Socioendocrinology of Primate Reproduction. (Ed. by Ziegler, TE,
Bercovitch, FB), pp 59-93. New York: Wiley-Liss.

Bercovitch, FB, Numberg, P, 1996. Socioendocrine and morphological
correlates of paternity in rhesus macaques. Journal of Reproduction and
Fertility 107, 59-68.

Bercovitch, FB, Ziegler, TE, 1990. Socioendocrinology of Primate Reproduction.
New York: Wiley-Liss.

Bercovitch, FB, Ziegler, TE, 2002. Current topics in primate socioendocrinology.
Annual Review in Anthropology 31, 45-67.

Berger, DMP, Frank, LG, Glickman, SE, 1992. Unraveling ancient mysteries:
biology, behavior, and captive management of the spotted hyena, Crocuta
crocuta. Pp 139-148 in The Proceedings of a joint meeting of the
American Association of Zoo Veterinarians and the American Association
of Wildlife Veterinarians. Oakland, California.

Bernardo, J, 1996. The particular maternal effect of propagule size, especially

egg size: Patterns, models, quality of evidence and interpretations.
American Zoologist 36, 216-236.

149

Bernstein, IS, Estep, DQ, Bruce, KE, Philips, KA, 1992. Effects of periodic
removal and reintroduction on the social and sexual behaviour of stumptail
macaques (Macaca arctoides). Folia Primatologia 59, 213-216.

Bernstein, IS, Gordon, CP, Rose, RM, 1983. The interaction of hormones,
behavior, and social context in nonhuman primates. In: Hormones and
Aggressive Behavior (Ed. by Svare, BB), pp 535-561.

Bernstein, IS, Rose, RM, Gordon, CP, 1974. Behavioral and environmental
events influencing primate testosterone levels. Journal of Human
Evolution 3, 517-525.

Bernstein, IS, Rose, RM, Gordon, CP, 1977. Behavioral and hormonal responses
of male rhesus monkeys introduced to females in breeding and non-
breeding seasons. Animal Behaviour 25, 609-614.

Bottoms, GD, Roesel, OF, Rausch, FD, Atkins, EL, 1972. Circadian variation in
plasma cortisol and corticosterone in pig and mare. American Journal of
Veterinary Research 33, 785-790.

Bouissou, M-F, 1983. Androgens, aggressive behaviour and social relationships
in higher mammals. Hormone Research 18, 43-61.

Boydston, EE, Morelli, TL, Holekamp, KE, 2001. Sex differences in territorial
behavior exhibited by the spotted hyena (Hyenidae, Crocuta crocuta).
Ethology 107. 369-385.

Breedlove, SA, Hampson, E, 2002. Sexual differentiation of the brain and
behavior. In: Behavioral Endocrinology (Ed. by Becker, JB, Breedlove,
SM, Crews, D, McCarthy, MM), pp. 75-114. Cambridge, Mass: MIT
Press.

Brockman, DK, Whitten, PL, Richard, AF, Benander, B, 2001. Birth season
testosterone levels in male Verreaux's sifaka, Propithecus verreauxi:
insights into socio-demographic factors mediating seasonal testicular
function. Behavioral Ecology and Sociobiology 49, 117-127.

Brockmann, DK, Whitten, PL, Richard, AF, Schneider, A, 1998.
Reproduction in free-ranging male Propithecus verreauxi: the hormonal
correlates of mating and aggression. American Journal of Physical
Anthropology 105, 137-151.

Bronson, F H, Eleftheriou, BE, 1964. Chronic physiological effects of fighting in
mice. General and Comparative Endocrinology 4, 9-18.

150

Brown, JL, Terio, KA, Graham, LH, 1996. Fecal androgen metabolite analysis
for noninvasive monitoring of testicular steroidogenic activity in felids. Zoo
Biology 15, 425-434.

Brown, JL, Wildt, DE, Wielebnowski, N, Goodrowe, KL, Graham, LH, Wells,
8, Howard, JG, 1998. Reproductive activity in captive female Cheetahs
(Acinonyx jubatus) assessed by fecal steroids. Journal of Reproduction
and Fertility 106, 337-346.

Brown, JL, Wasser, SK, Wildt, DE, Graham, LH, 1994. Comparative aspects
of steroid hormone metabolism and ovarian activity in felids, measured
noninvasively in feces. Biology of Reproduction 51, 776-786.

Buck, CL, Barnes, BM, 2003. Androgen in free-living arctic ground squirrels:
seasonal changes and influence of staged male-male aggressive
encounters. Hormones and Behavior 43, 318-326.

Calms, SJ, Schwager, S J, 1987. A comparison of association indices. Animal
Behaviour 27, 686-698.

Casto, JM, Ward, OB, Bartke, A, 2003. Play, copulation, anatomy, and I
testosterone in gonadally intact male rats prenatally exposed to flutamide
Physiology and Behavior 79, 633-641.

Cavigelli, SA, 1999. Behavioural patterns associated with faecal cortisol levels in
free-ranging female ringtailed lemurs, Lemur catta. Animal Behaviour 57,
935-944.

Cavigelli, SA, Pereira, ME, 2000. Mating season aggression and faecal
testosterone levels in male ring-tailed lemurs (Lemur catta). Hormones
and Behavior 37, 246-255.

Cavigelli, SA, Dubovick, T, Levash, W, Jolly, A, Pitts, A, 2003. Female
dominance status and fecal corticoids in a cooperative breeder with low
reproductive skew: ring-tailed lemurs (Lemur catta). Hormones and
Behavior 43, 166-179.

Christian, JJ, Davis, DE, 1964. Endocrines, behavior and population. Science
146, 1550-1560.

Clarke, AS, Boinski, S, 1995. Temperament in nonhuman primates. American
Journal of Primatology 37, 103-125.

Clutton-Brock, TH, Harvey, PH, 1976. Evolutionary rules and primate societies.

In: Growing points in ethology (Ed. by Bateson, PPG, Hinde, RA). pp 195-
237. Cambridge: Cambridge University Press.

151

Christians, JK, 2002. Avian egg size: variation within species and inflexibility
within individuals. Biological Reviews 77, 1-26.

Clarke, AS, Wittwer, DJ, Abbott, DH, Schnider, ML, 1994. Long-term effects
of prenatal stress on HPA axis activity in juvenile rhesus monkeys.
Developmental Psychobiology 27, 257-269.

Clemens, LG, Gladue, BA, Coniglio, LP, 1978, Prenatal endogenous androgenic
influences on masculine sexul-behavior and genital morphology in male
and female rats. Hormones and Behavior 10, 40-53.

Coe, CL, Mendoza, SP, Levine, 8,1979. Social status constrains the stress
response in the squirrel monkey. Physiology and Behavior 23, 633-638.

Coe, CL, Mural, JT, Wiener, SG, Levine, S, Siiteri, PK, 1986. Rapid cortisol
and corticosteroid-binding globulin responses during pregnancy and after
estrogen administration in the squirrel monkey. Journal of Endocrinology
118, 435-440.

Cooper, SM, 1993. Denning behavior od spotted hyenas (Crocuta crocuta) in
Botswana. African Journal of Ecology 31, 178-180.

Cooper, SM, Holekamp, KE, Smale, L, 1999. A seasonal feast: long-term
analysis od feeding behaviour in the spotted hyaena (Crocuta crocuta).
African Journal of Ecology 37, 149-160.

Coulter, CL, Jaffe, RB, 1998. Functional maturation of the primate fetal adrenal in
vivo: 3. Specific zonal localization and developmental regulation of
CYP21A2 (P450c21) and CYP11B1/CYP1 1 B2 (P450c1 1/Aldosterone
synthase) lead to integrated concept of zonal and temporal steroid
biosynthesis. Endocrinology 139, 5144-51 50.

Creel, S, 2001. Social dominance and stress hormones. Trends in Ecology and
Evolution 16, 491-497.

Creel, S, 2001. Cooperative hunting and sociality in African wild dogs, Lycaon
pictus. In: Model systems in behavioral ecology: Integrating conceptual,
theoretical, and empirical approaches. (Ed. by Dugatkin, LA). pp. 466-490.
Princeton: Princeton University Press.

Creel, S, Creel, NM, Mills, MGL, Monfort, SL, 1997. Rank and reproduction
in cooperatively breeding African wild dogs: Behavioural and endocrine
correlates. Behavioral Ecology 8, 298-306.

Creel, S, Creel, N, Monfort, SL, 1996. Social stress and donimance. Nature
379, 212.

152

Creel, S, Creel, N, Wildt, DE, and Monfort, SL, 1992. Behavioural and
endocrine mechanisms of reproductive suppression in Serengeti dwarf
mongooses. Animal Behaviour 43, 231-245.

Creel, S, Sands, J, 2003. Is social stress a consequence of subordination or a
cost of dominance? In: Animal Social Complexity (Ed. by de Waal, FBM,
Tyack, PL), pp 153-169. Cambridge, Mass: Harvard University Press.

Creel, SR, Wildt, DE, Monfort, SL, 1993. Androgens, aggression and
reproduction in wild dwarf mongooses: a test of the challenge hypothesis.
American Naturalist 141, 816-825.

Cunha, GR, Wang, YZ, Place, NJ, Liu, WH, Baskin, L, Glickman, SE, 2003.
Urogenital system of the spotted hyena (Crocuta crocuta Erxleben): A
functional histological study. Journal of Morphology 256, 205-218.

Curtis, D, Zaramody, A, Di, G, Pickard, A, 2000. Non-invasive monitoring of
reproductive status in wild mongoose lemurs (Eulemur mongoz).
Reproduction, Fertility, and Development 12, 21-29.

Dixson, AF, 1998. Primate sexuality: comparative studies of the prosimians,
monkeys, apes, and human beings. Oxford: Oxford University Press.

Dixson, AF, Bossi, T, Wickings, EJ, 1993. Male dominance and genetically
determined reproductive success in the mandrill (Mandn‘llus sphinx).
Primates 34, 525-532.

Dloniak, SM, French, JA, Place, NJ, Weldele, ML, Glickman, SE, and
Holekamp, KE, 2004. Non-invasive monitoring of fecal androgens in
spotted hyenas (Crocuta crocuta). General and Comparative
Endocrinology 135, 51-61.

Drea, CM, Place, NJ, Weldele ML, Coscia EM, Licht, P, Glickman, SE,
2002. Exposure to naturally circulating androgens during foetal life incurs
direct reproductive costs in female spotted hyenas, but is prerequisite for
male mating. Proceedings of the Royal Society of London Series B 269
(1504), 1981-1987.

Drea, CM, Weldele, ML, Forger, NG, Coscia, EM, Frank, LG, Licht, P,
Glickman, SE, 1998. Androgens and masculinization of genitalia in the
spotted hyaena (Crocuta crocuta). 2. Effects of prenatal anti-androgens.
Journal of Reproduction and Fertility 113, 117-127.

Drickamer, LC, 1996. Intra-uterine position and anogenital distance in house
mice: Consequences under field conditions. Animal Behaviour 52, 223-

153

223.

Dugatkin, LA, 2001. Model Systems in Behavioral Ecology: Integrating
Conceptual, Theoretical, and Empirical Approaches. Princeton: Princeton
University Press.

East, ML, Burke, T, Wilhelm, K, Greig, C, Hofer, H, 2003. Sexual conflicts in
spotted hyaenas: male and female mating tactics and their reproductive
outcome with respect to age, social status, and tenure. Proceedings of the
Royal Society of London Series B 270, 1247-1254.

East, ML, Hofer, H, 2001. Male spotted hyenas (Crocuta crocuta) queue for
status in social groups dominated by females. Behavioral Ecology 12,
558-568.

East, ML, Hofer, H, 2002. Conflict and cooperation in a female-dominated
society: A reassessment of the “hyperaggressive” image of spotted
hyenas. Advances in the Study of Behavior 31, 1-30.

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

Eaton, GG, Resko, JA, 1974. Plasma testosterone and male dominance in a
Japanese macaque (Macaca fuscata) troop compared with repeated
measures of testosterone in laboratory males. Hormones and Behavior 5,
251-259.

Eising, CM, Eikenaar, C, Schwbl, H, Groothuis, TGG, 2001. Maternal androgens
in black-headed gull (Larus ridibundus) eggs: consequences for chick
development. Proceedings of the Royal Society of London Series B
268 (1469): 839-846.

Engh, AL, Esch, K, Smale, L, Holekamp, KE, 2000. Mechanisms of maternal
rank “inheritance” in the spotted hyaena. Animal Behaviour 60, 323-332.

Engh, AL, Funk, SM, Van Horn, RC, Scribner, KT, Bruford, MW, Libants,
S, Szykman, M, Smale, L, Holekamp, KE, 2002. Reproductive skew
among males in a female-dominated mammalian society. Behavioral
Ecology 13, 193-200.

Farabollini, F, 1987. Behavioral and endocrine aspects of dominance and
submission in male rabbits. Aggressive Behavior 13, 247-258.

Faulkes, CG, Abbott, DH, Jarvis, JUM, 1990. Social suppression of ovarian

cyclicity in captive and wild colonies of naked mole-rats, Heterocephalus
glaber. Journal of Reproduction and Fertility 88, 559-568.

154

Faulkes, CG, Abbott, DH, Jarvis, JUM, 1991. Social suppression of
reproduction in male naked mole-rats, Heterocephalus glaber. Journal of
Reproduction and Fertility 91, 593-604.

Faulkes, CG, Abbott, DH, 1991. Social control of reproduction in both breeding
and non-breeding male naked mole-rats, Heterocephalus glaber. Journal
of Reproduction and Fertility 93, 427-435.

Faulkes, CG, Abbott, DH 1997. The physiology of a reproductive dictatorship:
Regulation of male and female reproduction by a single breeding female in
colonies of naked mole rats. In: Cooperative Breeding in Mammals (Ed. by
Solomon, NG, French, JA). Cambridge University Press, Cambridge.

Feldhammer, GA, Drickamer, LC, Vessey, SH, Merritt, JF, 1999. Mammalogy:
Adaptation, Diversity, and Ecology. McGraw-Hill: Boston.

Fensterrnaker, SB, Zup, SL, Frank, LG, Glickman, SE, Forger, NG, 1999. A sex
difference in the hypothalamus of the spotted hyena. Nature Neuroscience
2, 943-945.

Foley, CAH, Papageorge, S, Wasser, SK, 2001. Noninvasive stress and
reproductive measures of social and ecological pressures in free-ranging
elephants. Conservation Biology 15, 1134-1142.

Frank, LG, 1986. Social organisation of the spotted hyaena (Crocuta crocuta). ll.
Dominance and reproduction. Animal Behaviour 35, 1510-1527.

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

Frank, LG, Davidson, JM, Smith, ER, 1985. Androgen levels in the spotted
hyaena Crocuta crocuta: The influence of social factors. Journal of
Zoology London A 206, 525-531.

Frank, LG, Glickman, SE, Licht, P, 1991. Fatal sibling aggression, precocial
development, and androgens in neonatal spotted hyaenas. Science 252,
702-704.

Frank, LG, Glickman, SG, Pouch, l, 1990. Sexual dimorphism in the spotted
hyaena (Crocuta crocuta). Journal of Zoology 221, 308-313.

Frank, LG, Weldele, ML, Glickman, SE, 1995. Masculinization costs in hyaenas.
Nature 377, 584-585.

Fulkerson, WJ, Sawyer, GJ, Gow, CB, 1980. Investigations of ultradian and

155

circadian rhytms in the concentration of cortisol and prolactin in the
plasma of dairy cattle. Australian Journal of Biological Science 33, 557-
561.

Glickman, SE, Coscia, EM, Frank, LG, Licht, P, Weldele, ML, Drea, CM, 1998.
Androgens and masculinization of genitalia in spotted hyaena (Crocuta
crocuta). 3. Effects of juvenile gonadectomy. Journal of Reproduction and
Fertility 113, 129-135.

Glickman, SE, Frank, LG, Davidson, JM, Smith, ER, Siiteri, PK, 1987.
Androstenedione may organize or activate sex-reversed traits in female
spotted hyenas. Proceedings of the National Academy of Sciences USA
84, 3444-3447.

Glickman, SE, Frank, LG, Pavgi, S, and Licht, P, 1992. Hormonal correlates of
“masculinization” in female spotted hyaenas (Crocuta crocuta). 1. Infancy
to sexual maturity. Journal of Reproduction and Fertility 95, 451-462.

Gonzalez, G, Sorci, G, Smith, LC, Lope, F de, 2001. Testosterone and sexual
signaling in male house sparrows (Passer domesticus). Behavioural
Ecology and Sociobiology 50, 557-562.

Gordon, TP, Rose, EL & Bernstein, IS, 1976. Seasonal rhythm in plasma
testosterone levels in the rhesus monkey (Macaca mulatta): A three year
study. Hormones and Behavior 7, 229-243.

Gould, SJ, 1981. Hyena myths and realities. Natural History 90, 16-24.

Gould, SJ, Vrba, ES, 1982. Exaptation — A missing term in the science of form.
Paleobiology 8, 4-15.

Goy, RW, 1997. Patterns of juvenile behavior following early hormonal
interventions. In: The Lifespan Development of Individuals: Behavioral,
Neurobiological, and Psychosocial Perspectives. (Ed. by Magnusson, D)
pp 297-314. Cambridge.

Goy, RW, Bercovitch, FB, McBrair, MC, 1988. Behavioral masculinization is
independent of genital masculinization in prenatally androgenized female
rhesus macaques. Hormones and Behavior 22, 552-571.

Goymann, W, 2000. Hormone Physiology and Life History of Spotted Hyaenas.
Doctoral Dissertation. Shaker Verlag, Aachen.

Goymann, W, East, ML, Hofer, H, 2001. Androgens and the role of female

“hyperaggressiveness” in spotted hyenas (Crocuta crocuta). Hormones
and Behavior 39, 83-92.

156

Goymann, W, East, ML, Hofer, H, 2003. Defense of females, but not social
status, predicts plasma androgen levels in male spotted hyenas.
Physiological and Biochemical Zoology, 76, 586-593.

Goymann, W., East, M.L., Wachter, B., aner, O.P., M6stl, E, and Hofer, H.,
2003. Social status does not predict corticosteroid levels in postdispersal
male spotted hyenas. Hormones and Behavior 43, 474-479.

Goymann, W, East, ML, Wachter, B, Honer, OP, Mestl, E, Van’t Hof, TJ, Hofer,
H, 2001. Social, state-dependent and environmental modulation of faecal
corticosteroid levels in free-ranging female spotted hyaenas. Proceedings
of the Royal Society of London Series B 268, 2453-2459.

Goymann, W, Mbstl, E, Van’t Hof, T, East, ML, Hofer, H 1999. Noninvasive
fecal monitoring of glucocorticoids in spotted hyenas, Crocuta crocuta.
General and Comparative Endocrinology 114, 340-348.

Goymann, W, Wingfield, JC, 2004. Allostatic load, social status and stress
hormones: the costs of social status matter. Animal Behaviour In Press.

Graham, CE, Nadler, RD, 1990. Socioendocrine interactions in Great Ape
reproduction. In: Socioendocrinology of Primate Reproduction. (Ed. by
Ziegler, TE, Bercovitch, FB), pp 33-58. New York: Wiley-Liss.

Hamilton, RA, Stanton, PG, O’Donnell, L, Steele, VR, Taggart, DA, Temple-
Smith, PD, 2000. Determination of seasonality in southern hairy-nosed
wombats (Lasiorhinus Iatifrons) by analysis of fecal androgens. Biology of
Reproduction 63, 526-531.

Hamilton lll, WJ, Tilson, RL, Frank, LG, 1986. Sexual monomorphism in spotted
hyaenas, Crocuta crocuta. Ethology 71, 63-73.

Harding, CF, 1981. Social modulation of circulating hormone levels in the male.
American Zoologist 21, 223-231.

Henschel, JR, Skinner, JD, 1987. Social relationships and dispersal patterns in a
clan of spotted hyaenas Crocuta crocuta in the Kruger National Park.
South African Journal of Zoology 22, 18-24.

Herrenkohl, LR, 1979. Prenatal stress reduces fertility and fecundity in female
offspring. Science 206,1097-1099.

Hirschenhauser, K, Winkler, H, Oliveira, RF, 2003. Comparative analysis of

male androgen responsiveness to social environment in birds: the effects
of mating system and paternal incubation. Hormones and Behavior 43,

157

508-51 9.

Hofer, H, East, ML, 2003. Behavioral processes and costs of co-existence in
female spotted hyenas: a life history perspective. Evolutionary Ecology
1 7, 31 5-331 .

Hoffsis, GF, Philip, MS, Murdick, W, Tharp, VL, Adult, K, 1970. Plasma
concentrations of cortisol and corticosterone in the normal horse.
American Journal of Veterinary Research 31, 1379-1387.

Holekamp, KE, Sisk, CL, 2003. Effects of dispersal status on pituitary and
gonadal function in the male spotted hyaena. Hormones and Behavior 44,
385-394.

Holekamp, KE, Smale, L, 1990. Provisioning and food sharing by lactating
spotted hyaenas (Crocuta crocuta). Ethology 86, 191-202.

Holekamp, KE, Smale, L, 1991. Dominance acquisition during mammalian
social development: the ‘inheritance’ of maternal rank. American Zoologist
31, 306-317.

Holekamp, KE, Smale, L, 1993. Ontogeny of dominance in free-living spotted
hyaenas: juvenile rank relations with other immature individuals. Animal
Behaviour 46, 451 -466.

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

Holekamp, KE, Smale, L, 1998. Behavioral development in the spotted hyena.
Bioscience 48, 997-1005.

Holekamp, KE, Smale, L, 2000. Feisty females and meek males: reproductive
strategies cm the spotted hyena. In: Reproduction in context (Ed. by
Wallen, K, Schneider, JE), pp 257-285. Cambridge, Mass: MIT Press.

Holekamp, KE, Smale, L, Simpson, HB, Holekamp, NA, 1984. Hormonal
influences on natal dispersal in free-living Belding’s ground squirrels
(Spermophilus beldingl). Hormones and Behavior 18, 465-483.

Holekamp, KE, Smale, L, Szykman, M, 1996. Rank and reproduction in female
spotted hyenas. Journal of Reproduction and Fertility 108, 229-237.

Holekamp, KE, Szykman, M, Boydston, EE, Smale, L, 1999. Association of
seasonal reproductive patterns with changing food availability in an
equatorial carnivore, the spotted hyena (Crocuta crocuta). Journal of
Reproduction and Fertility 116, 87-93.

158

Hudson, S, Mullord, M, Whittlestone, WG, Payne, E, 1975. Diurnal variations in
blood cortisol in the dairy cow. Journal of Dairy Science 58, 30-33.

Ingram, JR, Crockford, JN, Matthews, LR. 1999. Ultradian, circadian and
seasonal rhythms in cortisol secretion and adrenal responsiveness to
ACTH and yarding in unrestrained red deer (Cervus elephus) stags.
Journal of Endocrinology 163, 289-300.

Irvine, CHG, Alexander, SL, 1994. Factors affecting the circadian rhythm in
plasma cortisol concentrations in the horse. Domestic Animal
Endocrinology 11, 227-238.

Jainudeen, MR, Katongole, CB, Short, RV, 1972. Plasma testosterone levels in
relation to musth and sexual activity in the male asiatic elephant, Elephas
maximus. Journal of Reproduction and Fertility 29, 99-103.

Jarvis, JUM, 1991. Reproduction of naked mole-rates. In: The biology of the
naked mole-rat. (Ed. by Sherman, PW, Jarvis, JUM, Alexander), pp 66-97.
Princeton: Princeton University Press.

Jolly, A, 1984. The puzzle of female feeding priority. In: Female Primates:
Studies by woman primatologists. (Ed. by Small, M), pp 197-215. New
York: Alan R. Liss.

Kappeler, PM, 1990. Female dominance in Lemur catta: more than just female
feeding priority? Folia Primatologia 55, 92-103.

Katongole, CB, Naftolin, F, Short, RV, 1971. Relationship between blood levels
of luteinizing hormone and testosterone in bulls, the effects of sexual
stimulation. Journal of Endocrinology 50, 457-466.

Ketterson, ED, Nolan, V, 1992. Hormones and life histories: An integrative
approach. American Naturalist 140, Suppl. 33-62.

Khan, MZ, Altmann, J, lsani, SS, Yu, J, 2002. A matter of time: evaluating the
storage of fecal samples for steroid analysis. General and Comparative
Endocrinology 128, 57-64.

Kingsley, S, 1982. Causes of non-breeding and the development of the
secondary sexual characteristics in the male orangutan: a hormonal study.
In: The Orangutan: Its Biology and Conservation. (Ed. by de Boer, LEM),
pp 215-229. The Hague: W. Junk. 3

Kinsley, CH, Bridges, RS, 1988. Prenatal stress and maternal-behavior in intact
virgin rats — response latencies are decreased in males and increased in ‘

159

females. Hormones and Behavior 22, 76-89.

Knutsson, U, Dahlgren, J, Macus, C, Rosberg, S, Bronnegard, M, 1997.
Circadian cortisol rhythms in healthy boys and girls: relationships with age,
growth, body composition, and pubertal development. Journal of Clinical
Endocrinology and Metabolism 82, 536-540.

Kriesten, K, Murawski, U, 1988. Concentrations of serum cortisol, progesterone,
estradiol, cholesterol ester in the doe during the reproductive stadium, in
the fetal serum, in the amniotic fluid and in the milk of rabbits, as well as
correlations between these parameters. Comparative Biochemistry and
Physiology A 90, 413-420.

Kruuk, H, 1972. The Spotted Hyena. University of Chicago Press, Chicago.

Licht, P, Frank, LG, Pavgi, S, Yalcinkaya, TM, Siiteri, PK, Glickman, SE, 1992.
Hormonal correlates of “masculinization” in female spotted hyaenas
(Crocuta crocuta). 2. Maternal and fetal steroids. Journal of Reproduction
and Fertility 95, 463-474.

Licht, P, Hayes, T, Tsai, P, Cunha, G, Kim, H, Golbus, M, Hayward, 8, Martin,
MC, Jaffe, RB, Glickman, SE, 1998. Androgens and masculinization of
genitalia in the spotted hyaena (Crocuta crocuta). 1. Urogenital
morphology and placental androgen production during fetal life. Journal of
Reproduction and Fertility 113, 105-116.

Lincoln, GA, Guiness, F, Short, RV, 1972. The way in which testosterone
controls the social and sexual behavior of the red deer stag (Cervus
elephus). Hormones and Behavior 3, 375-396.

Lindeque, M, Skinner, J D, 1982. Fetal androgens and sexual mimicry in spotted
hyaenas (Crocuta crocuta). Journal of Reproduction and Fertility 65, 405-
410.

Lipar, JL, Ketterson, ED, 2000. Matemally derived yolk testosterone enhances
the development of the hatching muscle in the red-winged blackbird
Agelaius phoeniceus. Proceedings of the Royal Scoiety of London Series
B. 267, 2005-2010.

Lockwood, CJ, Radunovic, N, Nastic, D, Petkovic, S, Aigner, S, Berkowitz, GS,
1996. Corticotropin-releasing hormone and related pituitary-adrenal axis
hormones in fetal and maternal blood during the second half of pregnancy.
Journal of Perinatal Medicine 24, 243-251.

Louch, CD, Higginbotham, M, 1967. The relation between social rank and
plasma corticosterone levels in mice. General and Comparative

160

Endocrinology 8, 441 -444.

Lunn, SF, Recio, R, Morris, K, Fraser, HM, 1994. Blockade of the neonatal rise in
testosterone by a gonadotrophin-releasing hormone antagonist: effects on
timing of puberty and sexual behaviour in the male marmoset monkey.
Journal of Endocrinology 141, 439-447.

Lunstra, DD, Ford, JJ, Echtemkamp, SE, 1978. Puberty of beef bulls: hormone -
concentrations, growth, testicular development, sperm production and

sexual aggressiveness in bulls of different breeds. Journal of Animal
Science 46, 1054-1062.

Lynch, JW, Ziegler TE, Strier KB, 2002. Individual and seasonal variation in
fecal testosterone and cortisol levels of wild male tufted capuchin
monkeys, Cebus ape/Ia nigritus. Hormones and Behavior 41, 275-287.

Macrides, F, Bartke, A, Dalterio, S, 1975. Strange females increase plasma
testosterone levels in male mice. Science 189, 1104-1106.

Maggioncalda, AN, Sapolsky, RM, Czekala, NM, 1999. Reproductive hormone
profiles in captive male orangutans: implications for understanding
developmental arrest. American Journal of Physical Anthropology, 109,
1 9-32.

Maggioncalda, AN, Czekala, NM, Sapolsky, RM, 2000. Growth hormone and
thyroid stimulating hormone concentrations in captive male orangutans:

implications for understanding developmental arrest. American Journal of
Primatology 50, 67-76.

Manogue, KR, Leshner, Al, Candland, DK, 1975. Dominance status and
adrenocortical reactivity to stress in squirrel monkeys (Saimiri sciureus).
Primates 16, 457-463.

Marchlewska-Koj, A, Kapusta, J, Kruczek, M, 2003 Prenatal stress modifies
behavior in offspring of bank voles (Clethn'onomys glareolus). Physiology
and Behavior 79, 671-678.

Marie, M, Anouassi, A. 1986. Mating-induced luteinizing surge and ovulation in
the female camel (Camelus dromadarius). Biology of Reproduction 35,
792-798.

Martin, P, Bateson, P, 1988. Measuring Behavior: An Introductory Guide.
Cambridge: Cambridge University Press.

Matthews, LH, 1939. Reproduction in the spotted hyena, Crocuta crocuta
(Erxleben). Philosophical Transactions of the Royal Society of London

161

Series B 230, 1-78.

McComb, K, 1987. Roaring by red deer stage advances the date of exstrous in
hinds. Nature 330, 648-649.

McEwen, BS, 1998. Protective and damaging effects of stress mediators. New
England Journal of Medecine 338, 171-179.

McEwen, BS, Wingfield, JC, 2003. The concept of allostasis in biology and
biomedicine. Hormones and Behavior 43, 2-15.

Mills, MGL, 1990. Kalahari Hyenas: The Behavioral Ecology of Two Species.
London: Unwin Hyman.

Moore, IT, Wada, H, Perfito, N, Busch, DS, Hahn, TP, Wingfield, JC, 2004.
Territoriality and testosterone in an equatorial population of rufous-
collared sparrows, Zonotn’chia capensis. Animal Behaviour 61, 411-420.

Mbhle, U, Heistermann, M, Palme, R, Hodges, J K, 2002. Characterization of
urinary and fecal metabolites of testosterone and their measurement for
assessing gonadal endocrine function in male nonhuman primates.
General and Comparative Endocrinology 129, 135-145.

Monaghan, EP, Glickman, SE, 1992. Hormones and aggressive behavior. In:
Behavioral Endocrinology (Ed. by Becker, JB, Breedlove, SM, Crews, D),
pp. 261-285. Cambridge, Mass: MIT Press.

Monfort, SL, Wasser, SK, Mashburn, KL, Burke, M, Brewer, BA, Creel, SR, 1997.
Steroid metabolism and validation of noninvasive monitoring in the African
wild dog (Lycaon pictus). Zoo Biology 16, 533-548.

Moore, MC, 1983. Effect of female sexual displays on the endocrine physiology
and behavior of male white-crowned sparrows, Zonotn’chia Ieucophyrs.
Journal of Zoology 199, 137-148.

Moss, AM, CIutton-Brock, TH, Monfort, SL, 2001. Longitudinal gonadal steroid
excretion in free-living male and female meerkats (Suricata sun'catta).
General and Comparative Endocrinology 122, 158-171.

Mousseau, TA, Dingle, H, 1991. Maternal effects in insect life histories. Annual
Review of Entomology 36, 511-534.

Mousseau, TA, Fox, CW, 1998. Maternal Effects as Adaptations. Oxford: Oxford
University Press.

Milller, W, Eising, CM, Dijkstra, C, Groothuis, TGG, 2002. Sex differences in yolk

162

 

hormones depend on maternal social status in Leghorn chickens (Gallus
gallus domesticus) Proceedings of the royal society of London Series B
269, 2249-2255.

Muller, MN, Lipson, SF 2003. Diurnal patterns of urinary steroid excretion in wild
chimpanzees. American Journal of Primatology 60, 161-166.

Muller, MN, Wrangham, RW 2003. Dominance, cortisol, and stress in wild
chimpanzees (Pan troglodytes schweinfurthil). Behavioral Ecology and
Sociobiology. 55, 332-340.

Muller, MN, Wrangham, RW 2004. Dominance, aggression and testosterone in
wild chimpanzees: a test of the ‘challenge hypothesis.’ Animal
Behaviour 76, 113-123.

Munck, A, Guyre, PM, Holbrook, NJ, 1984. Physiological functions of
glucocorticoids in stress and their relation to pharmacological actions.
Endocrine Reviews 5, 25-44.

Neaves, WB, Griffin, JE, Wilson, JD, 1980. Sexual dimorphism of the phallus in
spotted hyenas (Crocuta crocuta). Journal of Reproduction and Fertility
59, 509-513.

Norman, A, Litwack, J, 1997. Hormones. New York: Academic Press.

Nunes, S, Fite, JE, French, JA, 2000. Variation in steroid hormones associated
with infant care behaviour and experience in male marmosets (Callithrix
kuhll). Animal Behaviour 60, 857-865.

Olioff, M, Stewart, J, 1978. Sex-differences in play-behavior of prepubescent
rats. Physiology and Behavior 20, 113-115.

Orgeur, P, 1995. Sexual play-behavior in lambs androgenized in-utero.
Physiology and Behavior. 57, 185-187.

Ostner, J, Heisterrnann, M, Kappeler, PM. 2003. lntersexual dominance,
masculinized genitals and prenatal steroids: comparative data from
lemurid primates. Natunlvissenschaften 90, 141-144.

Ostner, J, Kappeler, PM, Heisterrnann, M, 2002. Seasonal variation and social
correlates of androgen excretion in male redfronted lemurs (Eulemur
fulvus rufus). Behavioral Ecology and Sociobiology 52, 485-495.

Pellis, SM, Pellis, VC, Kolb, B, 1992. Neonatal testosterone augmentation
increases juvenile play fighting but does not influence the adult dominance

163

relationships of male rats. Aggressive Behavior 18, 437-447.

Pepe, GJ, Johnson, DK, Albrecht, ED 1982. The effects of estrogen on cortisol
metabolism in female baboons. Steroids 39, 471-477.

Peters, A, Astheimer, LB, Cockbum, A, 2001. The annual testosterone profile
in cooperatively breeding superb fairy-wrens, Malums cyaneus, reflects
their extreme infidelity. Behavioral Ecology and Sociobiology 50, 519-527.

Pinxten, R, de Ridder, E, Eens, M, 2003. Female presence affects male behavior
and testosterone levels in the European starling (Stumus vulgan's).
Hormones and Behavior 44, 103-109.

Place, NJ, Holekamp, KE, Sisk, CL, Weldele, ML, Coscia, EM, Drea, CM,
Glickmann, SE, 2002. Effects of prenatal androgens on LH secretion and
sex steroid concentrations in adult spotted hyenas, Crocuta crocuta.
Biology of Reproduction 67, 1405-1413.

Pottinger, TG, 1999. The impact of stress on animal reproductive activities. In:
Stress Physiology in Animals. (Ed. by Baum, PHM), pp130-177. Boca
Raton, Florida: CRC Press.

Pusey, AE, Packer, C, 1997. The ecology of relationships. In: Behavioural
Ecology, 4th ed. (Ed. by Kreds, J, Davies, NB), pp 254-283. Blackwell
Science.

Quinn, GP, Keough, MJ, 2002. Experimental Design and Data Analysis for
Biologists. Cambridge: Cambridge University Press.

Racey, PA, Skinner, JD, 1979. Endocrine aspects of sexual mimicry in spotted
hyaenas Crocuta crocuta. Journal of Zoology London 187, 315-326.

Reed, WL, Vleck, CM, 2001. Functional significance of variation in egg-yolk
androgens in the American coot. Oecologia 128, 164-171.

Rose, RM, Holaday, JW, Bernstein, IS, 1971. Plasma testosterone, dominance
rank and aggressive behavior in male rhesus monkeys. Nature 231, 366-
368.

Rose, RM, Gordon, TP, Bernstein, IS, 1972. Plasma testosterone levels in the
male rhesus: Influence of sexual and social stimuli. Science 178, 643-
645.

Rice, WR, 1989. Analyzing tables of statistical tests. Evolution 43, 223-225.

Saltz, D, White, GC, 1991. Urinary cortisol and urea nitrogen responses in

164

irreversibly undernourished mule deer fawns. Journal of Wildlife Disease
27, 41-46.

Saltzman, W, Schultz-Darken, NJ. Scheffler, G, Wegner, FH, and Abbott, DH,
1994. Social and reproductive influences on plasma cortisol in female
marmoset monkeys. Physiology and Behavior 56, 801-810.

Saltzman, W, Schultz-Darken, NJ, Wegner, FH, Wittwer, DJ, and Abbott, DH,
1998. Suppression of cortisol levels in subordinate female marmosets:
reproductive and social contributions. Hormones and Behavior 33,58-74.

Sands, J, Creel, S, 2004. Social dominance, aggression, and faecal
glucocorticoid levels in a wild population of wolves, Canis lupus. 67, 387-
396.

Sannen, A, Heisterrnann, M, van Elsacker, L, Mohle, U, Eens, M, 2003. Urinary
testosterone metabolite levels on bonobos: A comparison with
chimpanzees in relation to social system. Behaviour 140, 683-696.

Sapolsky, RM 1982. The endocrine stress-response and social status in the wild
baboon. Hormones and Behavior 16, 279-292.

Sapolsky, RM, 1983. Endocrine aspects of social instability in the olive baboon
(Papio anubis). American Journal of Primatology 43, 51-64.

Sapolsky, RM 1985. Stress-induced suppression of testicular function in the wild
baboon: Role of glucocorticoids. Endocrinology 116, 2273-2278.

Sapolsky, RM 1986. Endocrine and behavioral correlates of drought in the wild
baboon. American Journal of Primatology 11, 217-228.

Sapolsky, RM, 1992. Do glucocorticoid concentrations rise with age in the rat?
Neurobiology of ageing 13, 171-174.

Sapolsky, RM, 1993. The physiology of dominance in stable versus unstable
social hierarchies. In: Pn’mate Social Conflict (Ed. by Mason, W.A. &
Mendoza, S.P.), pp. 171-204 Albany: State University of New York Press.

Sapolsky, RM, 2002. Endocrinology of the stress response. In: Behavioral
Endocrinology. (Ed. by Becker, J, Breedlove, SM, Crews, D, McCarthy,
M), pp 409-450. MIT Press: Cambridge, MA.

Sapolsky, RM, Altmann, J, 1991. Incidence of hypercortisolism and

dexamethasone resistance increases with age among wild baboons.
Biological Psychiatry 30, 1008-1016.

165

Sapolsky, RM, Romero, LM, Munck, AU, 2000. How do glucocorticoids influence
stress responses? Integrating permissive, stimulatory, and preparative
actions. Endocrine Reviews 21, 55-89.

Schneider, KM, 1926. Uber Hyaenenzucht. Die Pelztierzucht 2,1-14.

Schwabl, 1993. Yolk is a source of maternal testosterone for developing birds.
Proceedings of the National Academy of Sciences USA. 90, 11446-
11450.

Schwabl, 1996. Maternal androgen in the avian egg affects postnatal growth.
Comparative Biochemistry and Physiology A 114, 271 -276.

Schwabl, 1997. The contents of maternal testosterone in house sparrow Passer
domesticus eggs vary with breeding conditions. Natunrvissenschafen 84,
406-408.

Seeman, TE, McEwen, BS, Rowe, JW, Singer, BH, 2001. Allostatic load as a
marker of cumulative biological risk: MacArthur studies of successful
ageing. Proceedings of the National Academy of Sciences USA 98, 4770-
4775.

Setchell, JM, Dixson, AF, 2001. Changes in the secondary sexual adomments of
male mandrills (Mandn'llus sphinx) are associated with gain and loss of
alpha status. Hormones and Behavior 39, 177-184.

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

Smale, L, Nunes, S, Holekamp, KE 1997. Sexually dimorphic dispersal in
mammals: Patterns, causes, and consequences. Advances in the Study of
Behavior 26, 180-250.

Smith, TF, French, JA, 1997. Social and reproductive conditions modulate
urinary cortisol excretion in black tufted-ear marmosets (Callithrix kuhlr).
American Journal of Primatology 42, 253-267.

Smith, CJ, Norman, RL, 1987. Influence of the gonads on cortisol secretion in
female rhesus macaques. Endocrinology 121, 2192-2198.

Smith, R, Wickings, EJ, Bowman, ME, Belleoud ,A, Dubreuil, G, Davies, JJ,
Madsen, G, 1999. Corticotropin-releasing hormone in chimpanzee and
gorilla pregnancies. Journal of Clinical Endocrinology and Metabolism 84,
2820-2825.

166

Solomon, NG, French, JA, 1997. Cooperative Breeding in Mammals. Cambridge
University Press: Cambridge.

Sousa, MB, Ziegler, TE, 1998. Diurnal variation in the excretion patterns of fecal
steroids in common marmoset (Callithrixjacchus) females. American
Journal of Primatology 46, 105-1 17.

Stacey, Koenig, W, 1990. Cooperative breeding in birds: long-term studies of
ecology and behavior. Cambridge University Press: Cambridge.

Stavisky, RC, Adams, MR, Watson, 8., Kaplan, JR, 2001. Dominance, cortisol,
and behavior in small groups of female cynomolgus monkeys (Macaca
fasciculan's). Hormones and Behavior 39, 232-238.

Stavisky, RC, Watson, SL, Anthony, MS, Manuck, SB, Adams, MR, Kaplan, JR,
2003. Influence of estradiol on cortisol secretion in ovariectomized
cynomolgus macaques (Macaca fasciculan's). American Journal of
Primatology 60, 17-22.

Steklis, HD, Raleigh, MJ, Kling, AS, Tachiki, K, 1986. Biochemical and
hormonal correlates of dominance and social behavior in all-male groups
of squirrel monkeys. American Journal of Primatology 11, 133-145.

Steklis, HD, Brammer, GL, Raleigh, MJ, McGuire, MT, 1996. Serum
testosterone, male dominance, and aggression in captive groups of vervet
monkeys (Cercopithecus aethiops sabaeus). Hormones and Behavior 19,
154-163.

Stockman, ER, Callaghan, RS, Gallagher, CA, Baum, MJ, 1986. Sexual-
differentation of play-behavior in the ferret. Behavioral Neuroscience 100,
563-568.

Strier, KB, Ziegler, TE, Wittwer, DJ, 1999. Seasonal and social correlates of fecal
testosterone and cortisol levels in wild male muriquis (Brachyteles
arachnoides). Hormones and Behavior 35, 125-134.

Szykman, M, 2001. Sexual behavior and male mate choice in the spotted hyena,
Crocuta crocuta. Ph.D. thesis, Michigan State University.

Szykman, M, Engh, AL, Van Horn, RC, Funk, SM, Scribner, KT, Holekamp, KE,
2001. Association patterns among male and female spotted hyenas
(Crocuta crocuta) reflect male mate choice. Behavioral Ecology and
Sociobiology, 50, 231-238.

Szykman, Engh, AL, Van Horn, RC, Boydston, EE, Scribner, KT, Holekamp, KE,

167

2003. Rare male aggression directed toward females in a female-
dominated society: Baiting behavior in the spotted hyena. Aggressive
Behavior 29, 457-474.

Terio, KA, Brown, JL, Morelund, R, Munson, L, 2002. Comparison of different
drying and storage methods on quantifiable concentrations of fecal
steroids in the cheetah. Zoo Biology 21, 215-222.

Tilson, RT, Hamilton, WJ, 1984. Social dominance and feeding patterns of
spotted hyaenas. Animal Behaviour 32, 715-724.

Umezaki, H, Hess, DL, Valenzuela, GJ, Ducsay, CA, 2001. Fetectomy alters
maternal pituitary-adrenal function in pregnant rhesus macaques. Biology
of Reproduction. 65 (5): 1616-1621.

Van Horn, RC, McElhinney, TL, Holekamp, KE, 2003. Age estimation and
dispersal in the spotted hyena (Crocuta crocuta). Journal of Mammalogy
84, 1019-1030.

Van Horn, RC, Wahaj, SA, Holekamp, KE, 2004. Role-reversed nepotism
among cubs and sires in the spotted hyena (Crocuta crocuta). Ethology In
Press.

van Jaarsveld, AS, Skinner, JD, 1992. Adrenocortical responsiveness to
immobilization stress in spotted hyenas (Crocuta crocuta). Comparative
Biochemistry and Physiology 103A, 73-79.

van Schaik, CP, van Noordwijk, MA, van Bragt, T, Blankenstein, MA, 1991. A
pilot study of the social correlated of levels of urinary cortisol, prolactin,
and testosterone in wild long-tailed macaques (Macaca fasciculan’s).
Primates 345-356.

Velloso, AL, Wasser, SK Monfort, SL, Dietz, JM, 1998. Longitudinal fecal steroid
excretion in maned wolves (Chrysocyon brachyurus). General and
Comparative Endocrinology 1 12, 96-107.

Virgin, CE, Sapolsky, RM, 1997. Styles of male social behavior and their
endocrine correlates among low-ranking baboons. American Journal of
Primatology 42, 25-39.

vom Saal, FS, Bronson, FH, 1980. Variation in the length of the estrous cycle in
mice due to former intrauterine proximity to male fetuses. Biology of
Reproduction 22, 777-780.

Walker, SL, Waddell, WT, Goodrowe, KL, 2002. Reproductive endocrine patterns
in captive female and male red wolves (Canis rufus) assessed by fecal

168

I .

and serum hormone analysis. Zoo Biology 21, 321-335.

Wang, MH, vom Saal, FS, 2000. Maternal age and traits in offspring - The timing
of a mouse's first litter influences the development of her pups. Nature
407, 469-470.

Ward, IL, Reed, J, 1985. Prenatal stress and prepubertal social rearing
conditions interact to determine sexual-behavior in male rats. Behavioral
Neuroscience 99, 301-309.

Wasser, SK, Thomas, R, Nair, PP, Guidry, C, Southers, J, Lucas, J, Wildt, DE,
Monfort, SL, 1993. Effects of dietary fiber on faecal steroid measurements
in baboons (Papio cynocephalus cynocephalus). Journal of Reproduction
and Fertility 101, 213-220.

Wasser, SK, Hunt, KE, Brown, JL, Cooper, K, Crockett, CM, Bechert, U,
Millspaugh, JJ, Larson, S, Monfort ,SL, 2000. A generalized fecal
glucocorticoid assay for use in a diverse array of nondomestic mammalian
and avian species. General and Comparative Endocrinology 120, 260-
275.

Weingrill, T, Gray, DA, Barrett, L, Henzi, SP, 2004. Fecal cortisol levels in free-
ranging female chacma baboons: relationship to dominance, reproductive
state and environmental factors. Hormones and Behavior 45, 259-269.

Welberg, LAM, Seckl, JR, 2001. Prenatal stress, glucocorticoids and the
programming of the brain. Journal of Neuroendocrinology 13, 113-128.

Whitten, PL, 2000. Evolutionary endocrinology of the cercopithecoids. In: Old
World Monkeys (Ed. by Whitehead, PF, Jolly, CJ), pp 269-297.
Cambridge: Cambridge University Press.

Whittingham, LA, Schwabl, H, 2002. Maternal testosterone in tree swallow eggs
varies with female aggression. Animal Behaviour 63, 73-67.

Wickings, EJ, Marshall, GR, Nieschlag, E, 1986. Endocrine regulation of male
reproduction. In: Comparative Primate Biology (Ed. by Dukelow, WR,
Enlvin, J), pp149-170. New York: Liss.

Wickings, EJ, Dixson, AF, 1992. Testicular function, secondary sexual
development, and social status in male mandrills (Mandrillus sphinx).
Physiology and Behavior 52, 909-916.

Williams, T, 1994. lntraspecific variation in egg size and egg composition in birds:
effects on offspring fitness. Biology Reviews 68, 35-59.

169

Wingfield, JC, Ball GF, Dufty Jr., AM, Hegner, RE, Ramenofsky, M, 1987.
Testosterone and aggression in birds. American Scientist 75, 602-608.

Wingfield, JC, Hegner, RE, Dufty Jr., AM, Ball, GF, 1990. The “challenge
hypothesis”: theoretical implications for patterns of testosterone secretion,
mating systems, and breeding strategies. American Naturalist 136, 829-
846.

Wingfield, JC, Whaling, CS, Marler, P, 1994. Communication in vertebrate
aggression and reproduction: The role of hormones. In: The Physiology of
Reproduction, 2nd ed. (Ed. by Knobil, E, and Neill, JD), p 303-342. Raven
Press, New York.

Wingfield, JC, Maney, DL, Breuner, CW, Jacobs, JD, Lynn, 8, Ramenofsky, M,
Richardson, RD,1998. Ecological bases of hormone-behavior interactions:
The “emergency life history stage." American Zoologist 38, 191-206.

Wingfield, JC, Jacobs, JD, Tramontin, AD, Perfito, N, Meddle, S, Maney, DL,
Soma, K, 2000. Toward an ecological basis of hormone-behavior
interactions in reproduction of birds. In: Reproduction in Context: Social
and Environmental Influences on Reproduction (Ed. by Wallen, K,
Schneider, JE), pp. 85-127. Cambridge, Mass: MIT Press.

Yalcinkaya, TM, Siiteri, PK, Vigne, J Licht, P, Pavgi, S, Frank, LG, Glickman, SE,
1993. A mechanism for virilization of female spotted hyenas in utero.
Science 260, 1929-1931.

Young, AL, Richard, AF, Aiello, LC, 1990. Female dominance and maternal
investment in strepsirhine primates. American Naturalist 135, 473-488.

Zar, JH, 1996. Biostatistical Analysis. Upper Saddle River, New Jersey: Prentice-
Hall.

Zielinski, WJ, Vandenbergh, JG, 1993. Testosterone and competitive ability in
male house mice, Mus musculus — laboratory and field studies. Animal
Behaviour 45, 873-891.

Zielinski, WJ, vom Saal, FS, Vandenbergh, JG, 1992. The effect of intrauterine
position on the survival, reproduction and home range size of female
house mice (Mus musculus). Behavioral Ecology and Sociobiology. 30,
185-191.

Zumpe, D, Michael, RP, 1996. Social factors modulate the effects of hormones

on sexual and aggressive behavior of macaques. American Journal of
Primatology 38, 233-261.

170

 

 

 

 

 

m
L
Y
H
S
R
E
N
N
U
E
T
A
l
S
N
A
m
H
m
M