.41.
1.... .
.4“.
. 11$...
t. .
s

.- 1:...
...»)...r‘?
....

Eu

2 s .
.x

a. x urns"... ..
. . 4.
...?! n .v I

i savanna“ he.
. .. ‘ . ‘
5...... ..5 the???) v
c! IF.| 0’ I...‘. x
‘1... :13?! .1
xi" I,‘|~mm"‘-q‘.. In
‘ I “I“, I
....r

....) .1123“! :3
«In?! . . I ...
... ismws.s.
”first: .
.17..

 

....z..;.....,2.»t.,....
.i i: .r 1...:
. than}:

”Swarunre . . .359.

 

mm LIBRARY
ZOO g Michigan State
Unwersrty

 

 

 

This is to certify that the
dissertation entitled

THREAT-SENSITIVE BEHAVIOR AND ITS ONTOGENETIC
DEVELOPMENT IN TOP MAMMALIAN CARNIVORES

presented by

WILINE MALLORY PANGLE

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Zoology
Ecology, Evolutionary Biology &
Behavior

 

 

%%

V Major Professor 596 nature
A flf

Date

MSU is an afiirmative-acfion, equal-opportunity employer

 

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

 

DATE DUE

DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Acc&PrelelRC/DateDue indd

w.— ,fliu - D

THREAT-SENSITIVE BEHAVIOR AND ITS ONTOGENETIC DEVELOPMENT IN
TOP MAMMALIAN CARNIVORES

By

Wiline Mallory Pangle

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Zoology
Ecology, Evolutionary Biology, and Behavior

2008

ABSTRACT

THREAT-SENSITIVE BEHAVIOR AND ITS ONTOGENETIC DEVELOPMENT
IN TOP MAMMALIAN CARNIVORES

By

Wiline Mallory Pangle

Animals have evolved behaviors to both detect and respond to threats in order to
maximize survivorship. In my dissertation, I describe four studies that focused on inter-
and intra-specific variation in threat-sensitive behaviors, with special attention to
vigilance, among East African mammalian carnivores.

I first evaluated the functions of vigilance behavior in adult and juvenile spotted
hyenas (Crocuta crocuta) by observing them in various behavioral contexts and over
gradients of risk. The results indicated that the main function of vigilance in hyenas is to
detect inter-specific rather than intra—specific threats, as vigilance was lower when hyenas
were in large than small groups, and was not affected by social rank. Vigilance was also
highly context-specific, with hyenas most vigilant when nursing a young litter, and least
vigilant when consuming food of high quality. Adult hyenas exhibited pronounced
individual variation in vigilance. Juvenile hyenas were less vigilant than adults in all
contexts and showed less consistency in their vigilance behavior.

I then further explored the ontogenetic variation in vigilance apparent in the first
study. I used playback experiments to evaluate age-related variation in responses by
spotted hyenas to roars of their main natural predators, lions (Panthera Ieo). Both
juvenile and adult hyenas moved in response to lion roars, but not to control sounds.

Juveniles showed a stronger response to lion sounds than did adults. Both juvenile and

adult hyenas also had stronger reactions to roars emitted by male than female lions. Thus,
juvenile hyenas appear to associate certain signals with dangers, and respond
accordingly. However, during ontogeny, young hyenas may need experience with danger
to learn the specific environmental circumstances under which they need to be vigilant
when no threat is immediately apparent.

The third study assessed the lethal and non-lethal effects of anthropogenic
activities by comparing vigilance in an undisturbed population of hyenas to that in one
experiencing human disturbance. Proportions of deaths of known causes that were
attributable to humans increased between 1988 and 2006 in the disturbed population. In
addition, hyenas from the disturbed population were more vigilant during rest than
hyenas from the undisturbed population, especially on days during which livestock were
grazing in their territory. I assessed this livestock effect using playback experiments of
cow bells. I found that hyenas from the disturbed population responded more strongly to
these bells, but also to control sounds, than did hyenas from the undisturbed populations,
indicating a heightened responsiveness to all unnatural sounds.

The last study placed the findings from the previous studies in a comparative
perspective by examining vigilance by adult members of eight sympatric East African
carnivore species. Vigilance varied strongly among the eight species, and this could be
attributed to variation in body size, mortality rate and sociality. In addition, each
carnivore species spent about the same amount of time vigilant during rest as when
feeding, but vigilance strategies differed between these activities in a way that was
consistent among species. Animals generally exhibited long, infrequent vigilance bouts

during rest, and short, frequent vigilance bouts when feeding.

Copyright by
WILINE MALLORY PANGLE

2008

ACKNOWLEDGEMENTS

I would like to thank before all my husband, Kevin Pangle, for his continuous
encouragement, patience, and love. I long ago lost count of the many fruitful discussions
we have had over a cup of tea, and later a cup of coffee, in front of boards that never
stayed white very long. I know of few partners that would have been as supportive as
Kevin was through a long 16-month field season that took me to the other side of the
ocean. Throughout this whole experience, Kevin was a rock on which I could lean, and I
am extremely grateful for his support.

I would like to thank my advisor, Kay E. Holekamp, for giving me the life-
changing opportunity to study spotted hyenas first in her lab at Michigan State
University, and then in Kenya at her field site. Few advisors offer such generous
opportunities to their students. Under Kay’s supervision, I have grown immensely as a
scientist, and I am grateful for her guidance and help throughout this process. I would
also like to thank my committee members, Tom Getty, Gary Mittelbach, and Laura
Smale, for their guidance and feedback throughout my degree.

I thank the hyena lab members, both past and present, for many fun conversations
and useful feedback over multiple lab meetings. When I joined the lab, Russ Van Horn
and Stephanie Dloniak were excellent role models. Pat Bills was always available for
help, and I am grateful for his support in using our long-term database and getting our
playback equipment. A special thanks to Jenn Smith for being such a wonderful office
mate. In the last few years, we went from being lab mates to close fiiends, and I will keep

wonderful memories of the many conversations and laughs we shared in our office.

I have also relied on our long-term data collected over the years by multiple
individuals, who I thank: Sarah Benson—Armam, Nancy Berry, Chantal Beaudoin, Erin
Boydston, Aimee Cockayne, Susan Cooper, Audrey DeRose-Wilson, Stephanie Dloniak,
Martin Durham, Anne Engh, Michelle Fisher, Josh Friedman, Paula Garrett, Isla Graham,
Tyson Harty, Cathy Katona, Karen Kapheim, Kay Holekamp, Joe Kolowski, Keith
Nelson, Karen Nutt, Gabe Ording, Laura Smale, Jenn Smith, Micaela Szykman, Jaime
Tanner, Kevin Theis, Russ Van Horn, Sofia Wahaj, Kim Weibel, Kim Wooten, and Brad
White. I also thank Michelle Fisher and Amy Kuczynski for help extracting video
sequences from archived videos.

Many people have contributed to making my time in Kenya successful. I thank
Audrey DeRose-Wilson and Aimee Cockayne for their help in the field, and especially
for their assistance in collecting data for the Mara River clan. I also thank John Keshe,
James Kerembe and Moses Sairowa for taking such great care of me in Kenya, and
Stephen and Lesingo Nairori for allowing me to sleep soundly at night under my canvas
tent. I thank the Office of the President in Kenya and the Kenya Wildlife Services for
giving me permission to carry out my field work. I also thank the Narok County Council
and the wardens of the Masai Mara National Reserve for their cooperation.

I am also thankful to many members of the academic community for their help,
only some of whom are named here. Don Kramer, my undergraduate advisor at McGill
University, has continued to provide thoughtful and generous advice to this date,
particularly on Chapter 4. I thank him for his continuous support through the years. I
thank Andrew McAdam for his statistical advice, notably on mixed-effect models, and

Ted Garland for his time and patience teaching me phylogenetic statistics. Craig Packer

vi

and Anne Engh provided recordings of lion roars and baboon wahoos respectively, which
were essential for the playback experiments presented in Chapter 2. I am grateful to Scott
Creel for generously providing videos of wild dogs feeding, which enabled me to include
an additional species in Chapter 4. I also thank Phil Perry for sending me additional
footage of leopards feeding.

My work was supported by many fellowships and grants that made my
dissertation research possible. At Michigan State University, I have received financial
support from the department of Zoology, the Ecology, Evolutionary Biology and
Behavior program, the Graduate School, and the college of Natural Sciences. I was also
supported by the Barnett Rosenberg Endowed Fellowship and the Future Academic
Scholars of Education Fellowship from Michigan State University. I am also grateful for
the external fellowships that I have received, which include the International Doctoral
Fellowship from American Association of University Women Educational Foundation,
and the Eloise Gerry Fellowship from the Graduate Women in Sciences. My research has
been part of the Michigan State University Mara Hyena Project, which was funded by the
National Science Foundation grants IBN9309805, IBN9630667, IBN9906445,
IBN0113170, IBN0343381, and IOBO618022.

And last, but not least, I would like to thank my family. My mother, Marie Oudot,
has been present and supportive throughout this entire degree, and I continue to draw
strength and learn from her advice and philosophy of life. I would not be where I am
today if it were not for her support in all my ambitious endeavors. I particularly want to
acknowledge her for not only initiating the desire to immigrate to North America, but

also for helping me, along with my father Patrick Oudot, make this dream happen despite

vii

the pain the distance inflicted. Pat has always been there, in the good and the hard
moments, and his presence by my mother’s side has softened many worries. My sister,
Lise-Helene Smith, has also been there throughout my degree, always willing to help. I
thank her for her advice and the many fun phone conversations we have had. Our
developing friendship has meant a lot to me. I also thank my biological father, Thierry
Trouilloud, for his financial help. Mike Simms, the father of the host family that
welcomed me in the United States so many years ago, has also become a family member
for me, and I thank him for his support, and for all the many thoughtful conversations we
have had over the years that challenged my thinking. And lastly, I would like to thank all
the Pangles, and notably Wendy and Chris, for being a home away from home and the

best second family I could have.

viii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... x

LIST OF FIGURES ................................................................................. xi

GENERAL INTRODUCTION .................................................................... 1
Overview of chapters ....................................................................... 5

CHAPTER 1

FUNCTIONS OF VIGILANCE BEHAVIOR IN A SOCIAL CARNIVORE, THE

SPOTTED HYENA ................................................................................. 9
Introduction .................................................................................. 9
Methods ..................................................................................... 13
Results ....................................................................................... 18
Discussion ................................................................................... 27

CHAPTER 2

DEVELOPMENT OF THREAT-SENSITIVE BEHAVIORS IN SPOTTED

HYENAS: OBSERVATIONAL AND EXPERIMENTAL APPROACHES ............... 35
Introduction ................................................................................. 35
Methods ...................................................................................... 38
Results ....................................................................................... 45
Discussion ................................................................................... 55

CHAPTER 3

LETHAL AND NON-LETHAL ANTHROPOGENIC EFFECTS ON SPOTTED

HYENAS IN THE MASAI MARA NATIONAL RESERVE ............................... 59
Introduction ................................................................................. 59
Methods ..................................................................................... 62
Results ....................................................................................... 71
Discussion ................................................................................... 83

CHAPTER 4

A COMPARATIVE STUDY OF VIGILANCE IN EAST AFRICAN

CARNIVORES ..................................................................................... 88
Introduction ................................................................................. 88
Methods ...................................................................................... 90
Results ....................................................................................... 99
Discussion ................................................................................. 106

LITERATURE CITED ........................................................................... 112

ix

LIST OF TABLES

Table 1.1. Sample sizes and average number of observations for individual juvenile
and adult hyenas in the three behavioral contexts observed (resting, feeding and
nursing) .............................................................................................. 19

Table 2.1. Effect sizes for the results of the playback experiment. All variables are
comparing the responses of juveniles (“juvs”) to those of adults after a lion roar was
played. Sample sizes represent the number of individuals tested ............................ 53

Table 3.1. Salient characteristics of the two study clans, Talek West and Mara River
clans. Unless otherwise indicated, all information from this table was obtained from
Kolowski (2007) .................................................................................. .63

Table 3.2. Number of naturalistic vigilance observations collected (N obs) on adult
hyenas (N ind) sampled in each behavioral context (resting, feeding and nursing) in

the two study clans (Talek West and Mara River clans) in the presence or absence of
lions. We never observed adult females nursing in the presence of lions .................... 74

Table 4.1. Summary of the main characteristics of the eight carnivore species
included in this study ............................................................................... 92

Table 4.2. Results from the phylogenetic analyses of eight carnivore species; NS
indicates statistically non-significant variables, * indicates marginal relationships

(p < 0.1), and ** indicates significant relationships (p < 0.05). All degrees of

freedom are 6.The column “PGLS regressions” shows the results from a

Phylogenetic Generalized Least Squares regression in which phylogeny is

controlled for (see Methods for details) ......................................................... 103

Table 4.3. Results from modeling within each species the percent time spent
vigilant while feeding in relation to group size, distance to bushes, and grass height.
Sample sizes, representing the number of individuals filmed, appear in parenthesis
after each common species name. Bold values represent statistically significant (p <
0.05) values. Group size indicates the number of conspecifics present when the focal
individual was feeding. Distance to bushes indicates the distance from the focal
animal to the closest patch of bushes. Grass height was evaluated relative to the
head of the focal animal, and was categorized as 0 if there was no grass and 6 if the
grass was taller than the head of the focal individual. This analysis was not carried
out on leopards because of insufficient sample sizes. We had no information on
distance to bushes for wild dogs ................................................................. 105

LIST OF FIGURES

Figure 1.1. Effect of behavioral context on vigilance measures in adult (a - c) and
juvenile hyenas (d - f). The line in each box represents the median, and the box
encompasses the 25th percentile (lower limit) and the 75th percentile (upper limit);
whiskers represent the 10th and 90th percentile, and data points outside the whiskers
represent outliers. N represents the number of individual hyenas sampled in each
context. Statistics for full models are available in the text .................................... 21

Figure 1.2. Effect of group size on the percent time spent vigilant by adults (a and b)

and juveniles (c and (1) when resting (top row, panels a and c) and when feeding

(bottom row, panels b and d). Group size represents the number of hyenas resting or
feeding at a given location. Each data point represents an observation of an individual
hyena. Group size was a significant factor in all four cases (mixed-effect models, all

p ’s < 0.05). .......................................................................................... 24

Figure 1.3. Average percent time spent vigilant during rest (:I: SE) by each adult

focal hyena (N = 35) ranked from least to most vigilant. Symbols indicate social

rank categories, with filled circles representing the alpha matriline, gray circles the
middle ranking matrilines, empty circles the lowest-ranking matrilines, and empty
triangles the immigrant males. Identity was a significant predictor of vigilance (p

< 0.0001), whereas social rank was not (p > 0.05). Statistics are from a multivariate
mixed-effect model in which group size was also included. Statistics for the whole
model are available in the text .................................................................... 25

Figure 1.4. Effect of food quality (low quality: black bars; high quality: white bars)

on the percent time spent vigilant by adult and juvenile hyenas. Means (3: SE) are
represented, with the number of observations indicated above each bar. Quality of

food was assessed as high if it was fresh meat, or low in other cases. Asterisks

represent a statistically significant difference (mixed-effect model, in which group

size and identity were included; **p < 0.01; ***p < 0.001). Statistics for the full

models are available in the text ................................................................... 26

Figure 1.5. Effect of age of the suckling litter on the rate of head raises by adult

female hyenas during nursing. Each data point represents a single nursing event (N =
40), and 9 different females are represented in this figure. The line represents a linear
regression performed on transformed data and plotted on untransformed data. There

was a significant effect of litter age on the rate of head raises by mothers (mixed-

effect model, p = 0.001) ........................................................................... 28

Figure 2.1. Variation in vigilance throughout ontogeny observed among hyenas resting
alone on open savannah. Means (i SE) are represented for each age category, and
numbers of individuals sampled are indicated above each bar. The asterisk (*)

indicates p < 0.05 ................................................................................... 46

Figure 2.2. Behaviors exhibited by (a) juvenile (N = 10 for wahoos, N = 12 for roars),

xi

and (b) adult hyenas (N = 12 for wahoos, N = 13 for roars) before (black bars) and

after (white bars) sound onset during playback. The sound played was either a

baboon wahoo (control sound) or a lion roar (treatment sound). Means (:t SE) are
represented ........................................................................................... 48

Figure 2.3. Mean (i SE) percent time spent avoiding, approaching or remaining
motionless in relation to the speakers by the focal hyena following a lion roar

playback. Adult responses (white bars) are compared to juvenile responses (black

bars). The asterisk indicates p < 0.05. ............................................................ 50

Figure 2.4. Mean (: SE) latency to first movement of hyenas following sound onset.

The playback sound was a roar produced either by a female lion (open circles) or a

male lion (filled circles). Numbers next to each mean represent the number of
individuals that received a given treatment. The statistics shown are results from a

2-way ANOVA ....................................................................................... 51

Figure 2.5. Mean (1- SE) rate of movement by the focal hyena following exposure to

a lion roar playback made by either a female or a male lion. Both directions of
movement (approaching and avoiding the sound source) are included here. Sample

sizes are located above each mean and indicate the numbers of hyenas tested. There

was no effect of age on the movement rate, so juveniles and adults were combined

here. The asterisk (*) indicates p < 0.05 .......................................................... 54

Figure 3.1. Proportion of hyena deaths of known causes that could be attributed to
humans in each year of the study in two clans, the Talek West clan (filled circles)

and the Mara River clan (empty circles). The line represents a linear regression
performed on transformed data but plotted on untransformed data; the regression

analysis only includes data points for the Talek West clan .................................... 72

Figure 3.2. Percent time spent vigilant by hyenas from the Talek West clan (black

bars) and the Mara River clan (white bars) during morning (AM) and evening (PM)
observation hours while observed (a) resting, (b) feeding, and (c) nursing. Means (:1:

SE) are represented, with the number of individuals sampled indicated above each

mean. Asterisks (*) represent statistically significant differences (P < 0.05). We

never observed Mara River hyenas feeding in the evening hours ............................ 75

Figure 3.3. Percent time spent vigilant by spotted hyenas of the Talek West clan

(black bars) and Mara River clan (white bars) while resting and feeding in the

presence or absence of lions. Means (i SE) are represented, with the number of
individuals sampled indicated above each bar. Asterisks represent statistically
significant differences (*p < 0.05, **p = 0.01; ***p < 0.001) ............................... 76

Figure 3.4. A comparison of the two study clans with respect to the average distance
between focal hyena nursing and the closest patch of bushes. Means (:1: SE) are
represented, with number of individuals sampled indicated above each mean.

Asterisks represent statistically significant differences (*p < 0.05) .......................... 78

xii

Figure 3.5. Effect of livestock presence on the percent time spent vigilance by Talek
West hyenas while resting. Means (:t SE) are represented, with the number of
individuals sampled indicated below each mean. The percent time vigilant was

obtained for hyenas resting on days during which livestock presence in the territory

was known. The asterisk (*) represents a statistically significant difference

(p = 0.02) ............................................................................................. 80

Figure 3.6. Effect of tourism on the percent time spent vigilant during rest by

hyenas from the Talek West and the Mara River clans during months with many or

few tourists, and during hours of the day when tour vehicles are abroad in the Reserve
(black bars) or not (white bars). Means (d: SE) are represented, with the number of
observations indicated above each mean. There were no statistically significant

effects of tourism on vigilance in either clan (mixed-effect models, in which hyena
identity was included, p > 0.05; full model results are presented in the text) ............... 81

Figure 3.7. Percent time spent vigilant by Talek West and Mara River hyenas during a
playback experiment involving use of either church bells (control sound) or cow bells
(treatment sound). Behavior during the pre-playback period (black bars) is compared

to that from the post-playback period (white bars). Means (i SE) are represented, with
the number of individuals sampled indicated below each playback type. Asterisks
represent statistically significant differences (*p < 0.05, "p < 0.01) ....................... 82

Figure 4.1. Hypothesized phylogenetic relationships and estimated times of
divergence (numbers at the nodes, indicating millions of years) for the eight species
included in the comparative analysis (adapted from Bininda-Emonds et al. (1999) ....... 91

Figure 4.2. Vigilance while feeding in eight carnivore species: (a) mean t SE percent
time spent vigilant, (b) mean i SE duration of head raises, and (c) mean :h SE rate of
head raises. Species abbreviations on the X-axis are: DM = dwarf mongooses; BM =
banded mongooses; JA = black-backed jackals; WD = wild dogs; CH = cheetahs;

SH = spotted hyenas; LF = female lions; LM = male lions. Species are presented

from left to right in order of increasing body mass ........................................... 100

Figure 4.3. Comparison of vigilance while feeding and resting in three carnivore

species: (a) mean :t SE percent time spent vigilant, (b) mean :1: SE duration of head
raises, and (c) mean :t SE rate of head raises. Black bars represent data from feeding
animals and white bars represent data from resting animals. Species abbreviations on

the X-axis are: CH = cheetahs; SH = spotted hyenas; LF = lion females; LM = lion
males. Statistics were done on transformed data, but untransformed data are presented
here. Feeding means were ignificantly different from resting means for panels b and

c (allp < 0.001) ..................................................................................... 102

xiii

GENERAL INTRODUCTION

The objectives of this general introduction are threefold. First, I will review the
study of threat-sensitive behaviors, and specifically vigilance behavior. Second, I will
explain why East African carnivores, and particularly spotted hyenas (Crocuta crocuta),
are excellent model organisms in which to further extend the study of threat-sensitive
behaviors to carnivores. Finally, I will present an overview of each chapter of my
dissertation.

Threat-sensitive behaviors can be broadly defined as behaviors used by animals to
assess risk and adjust their behavioral responses accordingly (Helfrnan 1989, Helfrnan
and Winkelman 1997). Studies of threat-sensitive behaviors have been instrumental to
understanding the risks that animals perceive and the costs they incur to avoid predation.
These costs can be substantial and have been shown to have effects on fitness as strong or
even stronger than direct, lethal predation (Lima 1998; Preissier et a1. 2005; Creel and
Christianson 2008). Because of the ubiquitous nature of threat-sensitive behaviors, their
study offers opportunities for comparisons among populations or species as well as for
answering fundamental questions in both basic and applied biology.

Many studies of threat-sensitive behaviors in mammals and birds focus on
vigilance (reviewed by Caro 2005a). Vigilance can be broadly defined as any behavior
that increases the likelihood that an animal will detect a given stimulus at a given time
(Dimond and Lazarus 1974), and is often operationally defined in birds and mammals as
any raising of the head above a certain level (Caro 2005a). Vigilance is typically studied
in animals that are engaged in fitness-dependent activities such as foraging; under these

circumstances, animals must trade off benefits gained from their current activity against

their ability to perceive information about the surrounding environment (e. g. Arenz and
Leger 1999, 2000; reviewed in Lima and Dill 1990). Variation in vigilance during
fitness-dependent activities can inform us about animals’ perceptions of risk, and about
the decision-making processes germane to these risks.

Several general patterns have emerged from the myriad past studies of vigilance
behavior in mammals and birds (reviewed by Quenette 1990 and Caro 2005a). First, as
body size increases, vigilance generally decreases; body size appears to be a good proxy
for the number of different predators that prey on a focal species (Quenette 1990).
Second, vigilance typically decreases with increasing group size; this is often referred to
as the “group-size effect” (reviewed in Elgar 1989). This group-size effect may be driven
by an increased number of eyes, allowing individuals to reduce their vigilance (Pulliam
1973), or by dilution of predation risk among larger numbers of individuals (Hamilton
1971; Bednekoff and Lima 1998; Krause and Ruxton 2002; Fairbanks and Dobson 2007).
Third, many studies have provided direct and indirect evidence that vigilance functions to
detect approaching predators (termed “antipredator vigilance”; e. g. Lima 1987; Hunter
and Skinner 1998; Lima and Bednekoff 1999; Treves 2000), but individuals can also
direct their vigilance toward conspecific group members (termed “social vigilance”; e.g.
Treves 2000; Hirsch 2002; Cameron and du Toit 2005; Lung and Childress 2007). Many
additional factors related to predation risk, such as age of the focal animal, and distance
to cover, have also been examined in studies of vigilance but no consensus has been
reached (Caro 2005a). Thus, it appears that some risk-related factors may be species- or

context-dependent.

Most earlier studies of vigilance behavior have focused on species that occupy
low trophic positions in food webs, such as passerine birds, rodents and ungulates (e. g.
Lima 1987; Hunter and Skinner 1998; Blumstein et a1. 2001; Ebensperger et al. 2006).
Studies of species at higher trophic levels generally focus on their roles as predators (e. g.
Mills and Shenk 1992; Murray et al. 1995; Cooper et al. 2007). Little research has been
done on vigilance by predators, and this has resulted in a lack of understanding of the
mechanisms these animals possess for coping with danger. Mammalian carnivores
confront many natural threats, including both intraguild predation and kleptoparasitism
by members of their own and other species (Caro and Stoner 2003). The few existing
studies on antipredator behavior exhibited by carnivorous mammals suggest that
vigilance behavior might be important for predators as well as for prey (Caro 1987; Rasa
1989; Clutton-Brock et al. 1999; Switalski 2003; Di Blanco and Hirsch 2006; Hunter et
a1. 2007; Atwood and Gese 2008). Nevertheless, general trends in vigilance behavior
among mammalian predators remain largely unexplored.

I focused most of my dissertation research specifically on threat-sensitive
behaviors of spotted hyenas. The spotted hyena offers several advantages as a model
organism in which to examine threat-sensitive behaviors. First, in contrast to most other
predators, hyenas occur in large numbers (Mills and Hofer 1998), they are easily
observable in the savannahs of sub-Saharan Africa, and they are active around dawn and
dusk (Kruuk 1972; Kolowski et al. 2007). Spotted hyenas are keystone predators in most
sub-Saharan ecosystems (Hofer and East 1995). Their social organization in large fission-
fusion societies (Kruuk 1972; Smith et al. 2008), called clans, allows for collection of

repeated measures from multiple known individuals in each of several clans. Relative to

other carnivores, spotted hyenas are extraordinarily flexible in their behavior and
ecology. For instance, they breed throughout the year, they may be either diurnal or
nocturnal, they occupy a vast array of habitat types, and they eat carrion as well as live
prey ranging in size from termites to elephants (Kruuk 1972; Mills 1990; Sillero-Zubiri
and Gottelli 1992; Holekamp et al. 1997; Holekamp et al. 1999). Therefore, their
responses to dangers presented by natural and human-related stimuli may represent
conservative indicators of how other top predators, including those that are rare and
endangered, are likely to respond to specific threats (Arcese and Sinclair 1997).

Spotted hyenas are also particularly well-suited for studies of ontogenetic
variation in behavior (e.g. Holekamp and Smale 1993). In contrast to many other
mammals, hyenas have a life history characterized by discrete stages, each of which
begins and ends with observable milestones (Holekamp and Smale 1998). Although
hyenas are top predators in the food webs of many African ecosystems, they encounter a
number of serious threats, particularly during early developmental stages, and most die
violent deaths (Kruuk 1972; Watts 2007). Indeed, fewer than 50% of cubs born survive to
adulthood, and mortality rates diminish rapidly after spotted hyenas reach reproductive
maturity (Frank et al. 1995; Watts 2007). This pattern suggests that juveniles might be
more vulnerable than adults to negative effects of competition and intraguild predation.

Here my primary objectives were to better understand vigilance, its functions and
its development, in spotted hyenas. I used a combination of field observations and
playback experiments on wild hyenas in the Masai Mara National Reserve, Kenya
(hereafter the Reserve). I conducted a total of four studies focused on inter- and intra-

specific variation in threat-sensitive behaviors in spotted hyenas and other carnivores of

East Africa. Throughout this dissertation, I use the term “we” instead of “I”, which
reflects the true collaborative nature of my work and the fact that all chapters were

prepared in manuscript format.

Overview of Chapters

In the first chapter, we used natural patterns of vigilance behavior by spotted
hyenas to determine functions and key predictors of vigilance. Repeated observations in
multiple behavioral contexts were collected from individual members of one clan of
hyenas over a one-year period. We examined variation in hyena vigilance attributable to
age, group size, social rank, and factors related to predation risk. We also compared
vigilance behavior exhibited by hyenas engaged in one of three different types of
activities at the time of observation: resting, feeding or nursing. Our results indicated
that: 1) the primary function of vigilance in hyenas is to detect inter-specific rather than
intra-specific threats; 2) vigilance is highly context-specific; (3) adults were more vigilant
than juveniles in all contexts sampled; and 4) adult hyenas, but not juveniles, exhibited
pronounced individual variation in vigilance within behavioral contexts. These findings
shed light on the relative importance of different risks in shaping hyena behavior, and
provide general insights into context-dependence and group-size effects on vigilance.

In the second chapter, we further explored the age-related variation in vigilance
behavior described in the first chapter. First, we compared naturally-occurring vigilance
between age groups when hyenas were resting alone, and we found that juveniles were
less vigilant than adults when no immediate threat was present. Second, we performed

playback experiments to test predictions of a hypothesis suggesting that hyenas in

different life history stages are differentially competent to deal with danger. Because
juvenile hyenas are more vulnerable than adults to mortality caused by lions, we expected
that juveniles would respond to lion roars with much more caution than adults. We also
inquired whether responses by hyenas varied with the sex of the lions, as male lions are
the leading natural mortality source for both juvenile and adults hyenas (Cooper 1991;
Frank et al. 1995). Our results indicated that both adult and juvenile hyenas responded to
lion roars more vigorously than to the control sounds (baboon “wahoo” calls), but that
juveniles showed a stronger response than adults. Both age classes also showed stronger
reactions to roars emitted by male than female lions. Thus, like adults, juvenile hyenas
appear to recognize certain signals as dangerous and respond accordingly. However,
during ontogeny, young hyenas may need experiences with danger to learn the
environmental circumstances under which they need to be vigilant when no threat is
immediately apparent.

In the third chapter, we evaluated the lethal and non-lethal effects of
anthropogenic activities on wild spotted hyenas by comparing vigilance in an undisturbed
population of hyenas to that in one experiencing human disturbance. We used three
different approaches: 1) we examined the proportion of all hyena deaths of known causes
that could be attributed to humans between 1988 and 2006; 2) we made field observations
of naturally-occurring vigilance among hyenas; and 3) we performed an experiment to
test predictions of a hypothesis suggesting that vigilance behavior in hyenas is influenced
by livestock grazing. We expected members of a clan living beside the border of the
Reserve, and experiencing frequent human disturbance in the form of livestock and

herders, would spend more time vigilant in natural contexts of resting, nursing and

feeding than hyenas living in a clan far from the Reserve boundaries and experiencing
little human disturbance. The proportions of all known deaths that could be attributed to
humans increased dramatically in the disturbed population between 1988 and 2006. In
addition, resting hyenas living in close proximity to humans were more vigilant than
hyenas from the undisturbed population, especially on days when livestock were grazing
in their territory. In southern Kenya, local pastoralists fit their livestock with metal cow
bells, so we made recordings of these sounds, and compared hyena responses to them in
playback experiments with their responses to control sounds of church bells. In these
experiments, we found that hyenas from the disturbed population responded more
strongly to these bells, but also to control sounds, than did hyenas from the undisturbed
populations, indicating a heightened responsiveness to all unnatural sounds.

In the fourth study, we broadened our approach by examining vigilance in
multiple species of mammalian carnivores. We focused on eight sympatric species that
occupy an East African savannah habitat in the northern part of the Serengeti ecosystem.
These eight species represent four different families in the mammalian order Carnivora:
Herpestidae, including dwarf mongooses (Helogale parvula) and banded mongooses
(Mungos mungos); F elidae, including Cheetahs (Acinonyxjubatus), leopards (Panthera
pardus), and lions; Hyaenidae, including spotted hyenas; and Canidae, including black-
backed jackals (Cam's mesomelas) and wild dogs (Lycaon pictus). These species vary
considerably in body size, in the intensity of the intra- and interspecific feeding
competition they experience, and in their social behavior. Our goals were to identify
variables that predict patterns of vigilance in carnivores, and to compare patterns of

vigilance among species. Vigilance varied strongly among the different species, and this

variation could be attributed to differences in body size, mortality rate and sociality.
Vigilance also varied within each species, and this could be attributed to different factors
for each species. The carnivores studied were similarly vigilant while resting and feeding.
However, they adopted alternative vigilance strategies depending on the activity in which
they were engaged: they exhibited long, infrequent vigilance bouts during rest, and short,

frequent vigilance bouts while feeding.

CHAPTER ONE

FUNCTIONS OF VIGILANCE BEHAVIOR IN A SOCIAL CARNIVORE,
THE SPOTTED HYENA

INTRODUCTION

Many studies of threat-sensitive behavior in mammals and birds focus on
vigilance because it can have important consequences for survivorship (e. g. FitzGibbon
1989), and because it is easily quantifiable. Vigilance has been defined by Dimond and
Lazarus (1974) as any behavior that increases the likelihood that an animal will detect a
given stimulus at a given time. This stimulus might be an approaching predator (e. g.
Boland 2003) or an alarm call emitted by a conspecific (e.g. Shriner 1998; Back and
Switzer 2000), but it might also be an approaching competitor or a potential mate (e.g.
Roberts 1988; Burger and Gochfield 1994). Vigilance is typically studied in animals that
are engaged in fitness-dependent activities such as feeding; under these circumstances,
the animal must trade off benefits gained from its current activity against its ability to
perceive information about the surrounding environment (e. g. Arenz and Leger 1999,
2000; reviewed in Lima and Dill 1990). Variations in vigilance during fitness-dependent
activities can inform us about animals’ perceptions of risks, and the decision-making
processes germane to these risks.

Here we studied natural patterns of vigilance behavior exhibited by spotted
hyenas (Crocuta crocuta). Hyenas live in complex, social groups, called clans, containing
up to 80 individuals; clans are rigidly structured by dominance hierarchies, and an
individual’s social rank determines its priority of access to food (Kruuk 1972; Tilson and
Hamilton 1984; Frank 1986; Mills 1990). Hyenas feed primarily on ungulates they kill

themselves, and carcasses can be monopolized by only one or a few individuals (Cooper

et al. 1999). Competition at carcasses can be fierce and kleptoparasitism among clan
members is a common event; individuals of low social rank are at particular high risk of
kleptoparasitism (Kruuk 1972; Tilson and Hamilton 1984; Frank 1986). Thus, hyenas
may need to be frequently vigilant against theft of food by conspecifics. Furthermore,
hyenas rarely die of old age, but instead succumb most often to predation by lions and
humans (Cooper 1991; Frank et al. 1995; Watts 2007). Hyenas may therefore need to be
vigilant against intraguild or human predators. The relative importance of these intra- and
inter-specific threats faced by hyenas can be best understood by relating hyena vigilance
to risk-modifying factors such as group size.

One of the most pervasive ideas in the study of vigilance is that animals reduce
their vigilance as the number of individuals per group increases (Elgar 1989). This group-
size effect may be driven by an increased number of eyes, allowing individuals to reduce
their vigilance (Pulliam 1973), or by dilution of predation risk among larger numbers of
individuals (Hamilton 1971; Krause and Ruxton 2002; Bednekoff and Lima 1998). '
However, this effect can be masked or reversed if substantial threats also come from
members of the group itself. Internal threats should increase with increasing group size,
counterbalancing the usual effect of increasing group size on antipredator vigilance (e. g.
Treves 2000; Hirsch 2002; Kutsukake 2006, 2007). Studies of primates have shown that
individuals direct a large part of their total vigilance toward monitoring conspecifics
(reviewed by Treves 2000). This behavior has been termed “social vigilance”, as opposed
to antipredator vigilance, and recent studies have also reported such vigilance in various
non-primate species (e.g., Waite 1987; Cameron and du Toit 2005; Lung and Childress

2007; Pays and J arman 2008). Comparing time spent vigilant among individuals that vary

10

in social rank, but that also occur in subgroups of varying sizes, can provide information
about the relative importance of intra- and inter-specific threats to spotted hyenas.

The function and level of vigilance may also depend on the particular behavior in
which animals are engaged. Previous studies of vigilance have focused heavily on
animals engaged in foraging, and on the factors that influence the tradeoff between food
intake and information acquisition (reviewed in Caro 2005a). Recently, there has been a
call for research that addresses factors important to animals engaged in other behaviors,
such as resting (Arenz 2003; Roberts 2003; Lima et al. 2005), to better understand the
selection pressures operating in these contexts. Increasing the number of contexts in
which vigilance is studied also enables us to validate ideas derived from observations of
foraging animals, and to tease apart confounding factors that have been identified with
foraging animals. For instance, if the group-size effect occurs in contexts other than
foraging, this would suggest that the primary mechanism driving the group-size effect is
more likely predation risk than food competition or kleptoparasitism (Arenz 2003;
Beauchamp 2003). Because the function of rest remains poorly understood, tradeoffs
animals might experience while resting are less clear than those experienced while
foraging, but animals may nevertheless be at high risk during these and other activities
(Caro 2005a). In contrast to many boreal carnivores, spotted hyenas inhabiting African
savannahs are easily observable while they are engaged in several different types of
activities.

To identify the primary functions of vigilance in spotted hyenas, we examined
hyena vigilance behavior with field observations made over gradients of group size,

social rank, and factors related to predation risk (i.e. age of subject, time of day, and

11

distance to nearest bushes). We also compared vigilance among hyenas engaged in three
behavioral contexts (resting, feeding and nursing), and between two age groups (juveniles
and adults). All else being equal, we expected that hyenas would be more vigilant during
rest than while feeding or nursing, as we assumed rest to be the behavioral context in
which the cost of vigilance would be smallest. If vigilance functions as a means to detect
intra-specific threats, we expected hyenas to be more vigilant (1) in larger than smaller
groups; (2) when they are low- rather than high-ranking; (3) when they are found on the
edge of their territory than in its center; and (4) when they are closer than further from
their nearest conspecific neighbor. If vigilance functions as a means to detect inter-
specific threats, we expected hyenas to be more vigilant (l) in smaller than larger groups;
(2) when further than closer from their nearest conspecific neighbor; (3) when they are
feeding than during rest, as their likelihood of encountering lions when food is present is
much higher than when food is absent; (4) when they are nursing younger than older
offspring; and (5) when they are closer to bushes that could potentially conceal lions. In
addition to detecting threats, vigilance may also be used to find mates or prey (e. g. Caro
1987; Quenette 1990; Pays and Jarman 2008), so these functions must be considered as
well. If vigilance functions as a means to detect potential prey, we expected hyenas to be
more vigilant (1) during rest than while feeding, and (2) during times of the year when
prey are scarce. Lastly, if vigilance functions as a means of detecting potential mates, we
expected to detect (1) sex differences in vigilance, and (2) variation based on the sex ratio
of the current group. Although all four of these functions may apply to adult hyenas, we

expected the function of vigilance among juveniles might differ from that among adults.

12

For example, juveniles neither hunt nor mate so vigilance would not function in either of
these capacities among youngsters.

Our observations focused on one clan of hyenas for over a year, and we obtained
repeated measures within and between behavioral contexts for given individuals in the
clan. This allowed us to ask questions about individual variation within and between
behavioral contexts. Our results thus shed light on the context-dependence and group-size

effects on vigilance among large carnivores.

METHODS

Study site and study animals:

This study took place in the Masai Mara National Reserve (the Reserve — 1,500
kmz) in southwestern Kenya (1°40 S, 35°50 E) between June 2005 and July 2006. The
Reserve is an area of open, rolling grassland inhabited by large numbers of resident
ungulates (Sinclair and Norton-Griffiths 1979) that support high-densities of large
carnivores (Ogutu and Dublin 2002; Dloniak 2006). Every year between June and
September, large migratory herds of wildebeest (Connochaetes taurinus) and zebra
(Equus burchelli) join resident ungulates, leading to a superabundance of prey during
these months (Sinclair and Norton-Griffiths 1979).

We monitored one clan, the Talek West clan, which contained 52 to 63
individuals during the study period, including 11 to 15 adult females, their young, and 10
to 13 immigrant males. All hyenas in the clan were known individually by their unique
spots. Ages (i 7 days) of all natal hyenas were determined using previously described

methods (Holekamp et al. 1996). The sex of each hyena was known based on the

13

dimorphic shape of its erect phallus (Frank et al. 1990). Reproductive maturity in spotted
hyenas occurs at approximately 24 months of age (Matthews 1939; Glickman et al.

1992). Females give birth to litters of 1 or 2 cubs (rarely 3) in an isolated natal den, and
they move their cubs to a communal den when their clubs reach 2 to 5 weeks of age
(Kruuk 1972; East et al. 1989). Talek cubs reside at the communal den until about 9
months of age (Boydston et al. 2005). Juveniles were considered to be den-independent
once we observed them further than 200 m from the communal den on four consecutive
occasions. Mothers nurse their litters until cubs are approximately 14 months of age
(Holekamp et al. 1996), although cubs start feeding on meat opportunistically when they
are only a few months old (Kruuk 1972; Holekamp and Smale 1990). No allonursing
occurs in this species (Mills 1985). Here, we assigned sampled individuals to one of two
age groups: juveniles were hyenas younger than 24 months of age, and adults were
hyenas older than 24 months of age. Other than observations of mothers nursing their
litters, all subjects (both juveniles and adults) were den-independent.

. We determined social ranks of hyenas in the study clan based on wins and losses
in dyadic agonistic interactions (described in Holekamp and Smale 1993; Smale et al.
1993). Cubs ‘inherit’ their social rank from their mothers (Engh et al. 2000). A clan
contains multiple matrilines of adult females and their offspring, and members of each
matriline cluster along the social rank continuum. The highest-ranking matriline was the
alpha matriline, the second through fourth were considered the mid-ranking matrilines,
and the bottom three were considered here to be the low-ranking matrilines. Each of these
three categories represented roughly a third of the females and their young in the clan.

Afier puberty, virtually all natal males disperse from their natal clan; when they join a

14

new clan as immigrant males, they assume rank positions at the very bottom of the social
hierarchy, below all other immigrant males already present (Smale et al. 1997; East and
Hofer 2001). Thus all natal animals (philopatric females and their young) are socially
dominant to all immigrant males. We ranked individual adult hyenas from highest (the
alpha female) to lowest (the immigrant male that had most recently joined the clan) as a

single hierarchy.

Observations ofnaturally-occurring vigilance:

Individual hyenas were videotaped in the field to document their vigilance
behavior. Daily observations of the study animals took place during daylight hours
between 0600 - 1200 h (the AM period) and 1200 - 1920 h (the PM period). During these
observation periods, we videotaped hyenas engaged three behavioral contexts, resting,
feeding or nursing, from a field vehicle that served as a mobile blind. During these
observations, the field vehicle was parked no closer than 15 m from each focal hyena. All
monitored hyenas were well habituated to our vehicle. Hyenas were considered to be
resting when they were lying down in the absence of food for at least five minutes (and
usually for several hours), alone or with other conspecifics, and not interacting with other
conspecifics; the eyes of resting hyenas were usually shut. Hyenas were considered to be
feeding if they were actively involved in consuming a food item for at least two minutes.
Food items were categorized as high- or low-quality, depending on whether or not they
included fresh meat. A female hyena was considered to be nursing if a litter suckled from

her for more than two minutes. Upon finding a suitable subject, we set up a Sony DCR

15

H65 video camera mounted on a window tripod, and videotaped the focal hyena for 2 - 7
min (Y = 4.01 i 0.02 min).

In conjunction with each filmed sequence, we also recorded several variables that
may affect vigilance (e. g. Elgar 1989; Devereux et al. 2006; Liley and Creel 2008;
reviewed in Caro 2005a). These variables included Julian date (used as a proxy for prey
abundance), time of day, the location of the trial (on the edge or in the center of the clan’s
territory), the strength of the wind at the time of filming (categorized from 0 if there was
no wind, up to 6 if winds were very strong), the distance in meters between the focal
individual and the closest patch of bushes (using a Bushnell range finder), and the height
of the grass where the focal individual was filmed. Grass height was evaluated relative to
the head of the focal individual, and was ranked as 0 if there was no grass and up to 6 if
the grass was higher than the raised head of the focal individual. Wind strength and grass
height are both factors that may impair hyenas’ ability to detect danger. Social variables
included the number of conspecifics present within 100 m of the focal individual
(referred to hereafter as “group size”), the distance between the focal individual and its
nearest conspecific neighbor, and the group's sex ratio (only adults were included in this
sex ratio as this factor was recorded to inquire whether vigilance was directed to detect

potential mates). Age of the suckling litter was also recorded for nursing females.

Data extraction from video tapes:
We reviewed videotaped observations using a Sony DRC TRV900 digital
camcorder that superimposed a time code on the screen with a precision of 0.33 s, with

each video frame uniquely labeled. We extracted information on all vigilance events that

16

occurred during filming. We defined vigilance behavior as occurring whenever the focal
individual lifted its head. The onset of a vigilance bout was considered the point at which
the animal lifted its head (halfway through the raising of the head), and the end was
considered the point at which the animal lowered its head again (halfway through the
lowering of the head). From the video footage, we extracted the duration of each
vigilance bout, summed all bouts, divided this sum by the total length of the filmed
sequence, and multiplied by 100 to obtain a percent time spent vigilant. We also divided
the total sum of bout durations by the number of bouts in the filmed sequence to obtain
an average bout duration for each trial. Finally, we divided the number of bouts in a
sequence by the length of the filmed sequence to obtain a rate of head raises. These three

aspects of vigilance were the response variables used in our data analyses.

Data analysis

We tested assumptions of normality and homogeneity of variance using the
Wilks-Shapiro Test and the Levene’s Test, respectively. All three response variables met
both assumptions once they were log-10 transformed. We used general linear mixed-
effect models (GLMMs) to identify the variables that best explained vigilance. We
considered each response variable separately, and we also analyzed data from adults and
juveniles separately. We ran GLMMs for each of the three behavioral contexts in which
vigilance was sampled (resting, feeding and nursing). In these models, we initially
included all measured independent variables, and removed these terms sequentially
according to the Akaike’s Information Criterion (AIC; Bumham and Anderson 1998)

until each model included only terms for which inclusion would significantly decrease

17

the AIC of the model. In all models, hyena identity was entered as a random effect,
because random terms allow for repeated sampling for the same focal individuals to
avoid pseudo-replication (Pinheiro and Bates 2000). The final variables included in each
model were not correlated with each other (highest correlation coefficient was 0.2). We
used restricted maximum likelihood (REML) methods for model estimation. We present
F and p values for the final models. Significance of random effects was evaluated using
likelihood ratio tests by comparing models with and without random effects (Pinheiro and
Bates 2000).

To compare vigilance among behavioral contexts, we averaged values for
response variables recorded multiple times from each hyena engaged in a given activity,
and analyzed these averages using an analysis of variance (ANOVA). We identified
differences among the three activities with Tukey post-hoc tests. Similarly, to compare
vigilance between juveniles and adults, we averaged the response variables for each
hyena within a given age group (juveniles vs. adults) and a given activity (resting,
feeding or nursing), and compared averages for each activity separately using t-tests. All
analyses were performed in the statistical software package R, v.2. l .1 (R Development
Core Team 2005), using two-tailed tests with an alpha level of 0.05. The GLMMs were
done using the R library ‘nlme’ (Pinheiro et al. 2005). Unless otherwise indicated,

response values are reported as means i 1 standard error.

RESULTS
The number of observations collected and the number of individuals sampled in

each behavioral context are reported in Table 1.1. We first describe differences in

18

 

Em m: each

 

E 9 _ mm; H no .v a ov I 9 g mod H SN 3 mV @532
a 9 _ wmd a K .m mm am_ 2 8— N56 M New M: 2: wagon
am 9 _ cm; H 3.: mm aov M: 8 ~ mm... H No.3 3 mg warmed
c5 \ So 2: \ So a @0383 E: \ So 9: \ Bo a BEES
mo owcwm Amm i W. .32: a Bo mo a mo owned Amm ”3 W. 565 am So mo a 38:80
3332. mozcofié

 

.Awfimbé Ea meadow .wacmob @33on 35:50
38323 con: 05 E mace»: :53 use 283% 3:239: Sm mcoumfiomno mo 598:: owes; 98 monm 038mm .2 656,—.

19

vigilance among the three behavioral contexts, and then present results within each
context. Regardless of the behavioral context, the following variables had no effect on
any vigilance variables and were therefore removed from the models: Julian date,
distance to nearest bushes, distance to nearest neighbor, strength of wind, grass height,

location within the territory, sex, sex ratio of the group, and social rank.

Vigilance in diflerent behavioral contexts

The time dedicated to vigilance by adult hyenas varied among behavioral contexts
(F176 = 13.26, p < 0.001 , r2 = 0.26; Figure 1.1 a-c). Adults spent significantly more time
vigilant while nursing than while feeding or resting (post-hoe tests, both p < 0.001;
Figure 1.1a), but vigilance did not differ significantly between feeding and resting (post-

hoc test, p = 0.15; Figure 1.1a). Rates and durations of head raises also varied
significantly among activities (rates: F 2, 76: 83.71, p < 0.001, r2 = 0.69, Figure 1b;
durations: F176 = 90.50, p < 0.001, r2 = 0.70, Figure 1.1c), but in different ways. The

highest rates were observed among feeding individuals, the lowest among resting
individuals, and intermediate rates among nursing individuals (post-hoe tests, all p <
0.01; Figure 1.1b). The shortest durations were observed while feeding (post-hoc tests, p
< 0.001), but durations did not differ significantly between nursing and resting (post-hoe
test, p = 0.89; Figure 1.1c).

As was true among adults, the time spent vigilant by juvenile hyenas varied

among behavioral contexts (F 2,47 = 62.81, p < 0.001; r2 = 0.73; Figure Lid-f). Juveniles

spent significantly less time vigilant while nursing than while feeding or resting (post-hoe

tests, both p < 0.001; Figure 1d), and behavior in the latter contexts did not differ

20

100 -

60

40

Percent time spent vigilant

Rate of head raises (#/min)
—- N b.) A VI 05

O

80*

204

 

(a)

4

1

 

$-

Adults

 

 

 

 

 

 

‘ (b)

Feeding

 

 

 

 

 

col—

Resting

@-

Nursing

#-

 

60-

50‘

40*

30*

20-

10‘

Duration of head raises (5)

 

(C)

Feeding

#5

T

Resting

 

 

 

 

 

i
0

Nursing

 

 

 

 

 

 

Feeding
(N=35)

1i

Resting
(N=35)

Nursing
(N=9)

Behavioral context

100 -

80~

60‘

40‘

20~

60-

50-

40~

30<

20‘

10‘

 

Juveniles

 

 

 

 

 

(d)
$- =$= +
Feeding Resting Nursing
(6)
O
..-
. %
_._
Feeding Resting Nursing
(0
0
$ 0 c:
Feeding Resting Nursing
(N=l 8) (N=l6) (ll/=15)

Behavioral context

Figure 1.1. Effect of behavioral context on vigilance measures in adult (a - c) and
juvenile hyenas (d - f). The line in each box represents the median, and the box
encompasses the 25‘11 percentile (lower limit) and the 75th percentile (upper limit);
whiskers represent the 10th and 90th percentile, and data points outside the whiskers
represent outliers. N represents the number of individual hyenas sampled in each
context. Statistics for full models are available in the text.

significantly (post-hoe Tukey test, p = 0.94; Figure 1.1d). Rates and durations of head
raises also varied significantly among contexts (F147: 90.47, p < 0.001; Figure 1.1), but

in different ways. As in adults, the highest rates were observed among juveniles engaged
in feeding. Rates were the lowest among nursing juveniles, and intermediate rates were

observed among juveniles at rest (post-hoe tests, all p < 0.001; Figure 1.1c). Resting
individuals kept their heads elevated longer than did feeding individuals (t3 3 = 8.43; p <

0.001; Figure 1.11). We did not compare durations of head raises when nursing to those
during other contexts due to the low number of these observations; only 2 of the 15
juveniles sampled while nursing actually ever interrupted their nursing to scan the
environment.

Juveniles spent significantly less time vigilant than adults in all behavioral

contexts sampled. While resting, juveniles were less vigilant than adults (Y juveniles =
7.66 i 1.41 %; Y am, = 16.17 i 1.55 %;149= 3.29; p = 0.002; Figure 1.1a and d), which
was driven by juveniles having shorter head raises than adults (Y juveniles = 10.01 :t 1.28
s; fadultfi 19.08 i: 1.85 s; t49= 3.28; p = 0.001; Figure 1.1c and f). Rates of head raises
did not differ significantly between juveniles and adults (fjuvennes = 0.48 i 0.07 head
raise/min; Y adults = 0.58 i 0.04 head raise/min; I49 = 1.20, p = 0.24; Figure 1.1b and e).
Juveniles spent significantly less time vigilant when feeding than did adults (NJ-”venues =
6.78 :1: 0.86 %; Yaduns = 11.50 i 1.51 %;t51= 2.75;p = 0.008; Figure 1.1a and d); this

difference was driven by juveniles raising their heads less frequently than adults

(Y juveniles = 1.63 :1: 0.20 head raise/min; Y adults = 2.43 i 0.17 head raise/min; t5, = 3.00;

22

p = 0.004; Figure 1.1b and e). Durations of head raises were similar between juveniles
and adults (Y,-,,,,,,,,,, = 2.49 :1: 0.16 s; 3? ,de = 3.11 e 0.46 3; I5] = 0.74, p = 0.46; Figure
1.10 and f). Juvenile hyenas also spent significantly less time vigilant than their mothers
when nursing (Y,,,.,,,,,,, = 0.01 e 0.01 %; Y m,,h,,,= 38.40 a: 7.28 %; :22 = 29.04; p <
0.001; Figure 1.1a and d); in contrast to cubs, mothers raised their heads frequently while

nursing their litters (Yjuvennes = 0.002 i 0.0006 head raise/min; f mothers = 1.52 i 0.16

head raise/min; t22= 19.09; p < 0.001; Figure 1.1b and e).

Determinants of vigilance within each behavioral context
Three explanatory factors remained in the final model for vigilance among resting

adult hyenas: group size, time of day, and hyena identity. Adult hyenas spent less time

vigilant when resting in large than small groups (F 1, 372 = 9.88, p = 0.002; Figure 1.2a)
and more time vigilant in the AM period than in the PM period ( 3(— AM = 18.29 :1: 1.71 %,

N: 196; 3513M = 11.36 i 0.89 %, N = 213; F,,372= 19.67,p < 0.0001). Hyena identity
was also a strong predictor of the percent time spent vigilant while resting (likelihood
ratio test statistic distributed as X2 = 22.78, p < 0.0001; Figure 1.3). In other words,

individual hyenas exhibited either consistently high or low vigilance over multiple
observations in which they were resting. We inquired whether the same individuals that
were highly vigilant in one behavioral context were also highly vigilant in other contexts,
but we found no statistically significant correlations (all p > 0.3). Among juveniles, we

did not detect any effect of time of day (F 1’ 149 = 0.38, p = 0.54) or identity (12 = 1.34, p =

0.25) on vigilance, so these variables were removed from the model. Like adults,

23

 

 

 

 

 

 

Adults Juveniles
(a) (C)
E" l00<u - 100
‘3 o '
u 0 o
a --
g 30<. , g 30
3 o
“ I
g 60 l". 60 .
's a". . '
E 3:“ . .
§ 40‘.:: =0 40-:-.
E 'l 0.: . 0 .'... O
.5 O
E 20 "I! ' 20‘! :' o .
’ ul- .
8 o' o o
3 II' i'! . o ' .. 3 . o
n- 0< 0| ...! ° 0 0«."'o". I o o
0 5 10 15 20 25 5 10 15 20 25
(d
g1 100W”) 100 )
'U
0
c9.
5 80‘ 80
.=
3
E; 60" 60
.99 o
>
5g 40« . o . 4o-
: -' -
.E . . ' '
g 20« :3: . 20-:'o . ..
§ li':-s:'3' 33.. ' 333:“: i
O O
:33 ol.o 3.8.l' "t c 0 .. 0,‘o!-o.-Ig 3|. . I
0 5 10 15 20 25 5 10 15 20 25
Group size Group size

Figure 1.2. Effect of group size on the percent time spent vigilant by adults (a and b) and
juveniles (c and d) when resting (top row, panels a and c) and when feeding (bottom
row, panels b and (1). Group size represents the number of hyenas resting or feeding at a
given location. Each data point represents an observation of an individual hyena. Group
size was a significant factor in all four cases (mixed-effect models, all ps < 0.05).

24

50
0 Alpha rmtriline

45 Q Mid-ranking matrilines

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O Low-rankingnntiilines [-

a 40 , A Immigrant males '
d)
“2:"
- 35 4
3% ‘ AO
E 30 € .- 0o
'61) ‘ ..
'5 l O A
... 25 j A
g l AQA
in 0
g 20 -
E .- T z> C) _
g 15 4 A.
9 O

10 %{

11M -- “

,hl __

Least vigilant y Mostvigilant

 

Hyena Identity

Figure 1.3. Average percent time spent vigilant during rest (:t SE) by each adult focal
hyena (N = 35) ranked from least to most vigilant. Symbols indicate social rank
categories, with filled circles representing the alpha matriline, gray circles the middle-
ranking matrilines, empty circles the lowest-ranking matrilines, and empty triangles
the immigrant males. Identity was a significant predictor of vigilance (p < 0.0001),
whereas social rank was not (p > 0.05). Statistics are from a multivariate mixed-effect
model in which group size was also included. Statistics for the whole model are
available in the text.

25

juveniles spent less time vigilant while resting in large than small groups (FL 151 = 10.69,

p = 0.0013; Figure 1.2c).
Three explanatory factors remained in the final model for the percent time spent

vigilant while feeding among adult hyenas: group size, food quality and hyena identity.
Adult hyenas spent less time vigilant when feeding in large than small groups (F 1,92 =
16.03, p = 0.0001; Figure 12b) and when feeding on food of high- than low-quality
(F192 = 6.19, p = 0.01; Figure 1.4). Hyena identity was also a good predictor of the
percent time spent vigilant while feeding (12 = 4.96, p = 0.026). We did not detect any
effect of hyena identity on juvenile vigilance ([2 = 0.10, p = 0.75). Like adults, juveniles
spent less time vigilant when feeding in large than small groups (F131: 6.38, p = 0.01;

Figure 1.2d) and when feeding on food of high- than low-quality (F 1,8 1 = 5.82, p = 0.02;

Figure 1.4).
Only one factor remained in the final model explaining variation in vigilance of

adult female hyenas nursing their cubs: age of the litter being nursed. As the litter being
suckled grew older, mothers spent less time vigilant when nursing (F 1, 30 = 9.98, p =

0.004; Figure 1.5). In contrast to results obtained when adult females were engaged in

other behaviors, adult female identity was not a significant predictor of vigilance while
nursing (12 = 0.002, p = 0.97). In contrast to adults, juvenile hyenas were seldom vigilant
while suckling from their mothers: juveniles raised their heads during only 2 of the 43

trials, resulting in very low percent time spent vigilant ( 3(— = 0.01 :t 0.001 %).

26

Percent time spent vigilant when feeding

*4“?

 

 

l6~

14-

12 ‘ an:

 

- Low food quality

‘0 ‘ 1:] High food quality

73

 

 

 

 

 

 

 

 

Adults Juveniles

Age class

Figure 1.4. Effect of food quality (low quality: black bars; high quality: white bars) on
the percent time spent vigilant by adult and juvenile hyenas. Means (:1: SE) are
represented, with the number of feeding observations indicated above each bar.
Quality of food was assigned as high if it was fresh meat, or low in other cases.
Asterisks represent a statistically significant difference (mixed-effect model, in which
group size and identity were included; "p < 0.01; ***p < 0.001). Statistics for the full
models are available in the text.

27

4‘ o
o
3~ o

 

Rate of head raises (#/min) by mothers when nursing
N

 

 

I I I I I 1

0 2 4 6 8 10 12 l4 l6

_
d

Age (mo) of litter being nursed

Figure 1.5. Effect of age of the suckling litter on the rate of head raises by adult female
hyenas during nursing. Each data point represents a single nursing event (N = 40), and 9
different females are represented in this figure. The line represents a linear regression
performed on transformed data and plotted on untransformed data. There was a significant
effect of litter age on the rate of head raises by mothers (mixed-effect model, p = 0.001).

28

DISCUSSION

Our results supported the hypothesis that hyena vigilance functions more
importantly to detect inter-specific than intra-specific threats. One line of evidence
supporting this idea was that hyenas were more vigilant in small than large groups
regardless of the behavioral activity in which they were engaged. If hyenas were directing
their vigilance toward potential intra-specific threats as when avoiding aggression, we
would not expect to see such a decrease in vigilance as group size increases. Another line
of evidence was that we observed the same group-size effect among both feeding and
resting individuals, indicating that vigilance was not used as a means to detect potential
foraging competition or kleptoparasitism (Lendrem 1983; Arenz 2003; Beauchamp 2003;
Roberts 2003). Other lines of evidence confirmed that vigilance was not used to detect
intra-specific threats: there were no effects of social rank on vigilance, nor were hyenas
more vigilant on the edge than in the center of their territory, or when closer than further
to conspecific neighbors. Hyenas were also more vigilant when nursing younger than
older litters, but this may be a response to either inter- or intraspecific threats, as
infanticide, although rare, is known to occur in this species (Kruuk 1972; East and Hofer
2002; White 2005).

We found no support for the idea that vigilance functions as a means to detect
potential mates in spotted hyenas. We did not detect any differences in vigilance between
sexes, nor any effect of sex-ratio in the current group. Thus, adults were similarly vigilant
whether they were resting with only members of their sex, or with varying combinations
of male and female adult hyenas. Similarly, we did not find any support for the

hypothesis that vigilance by hyenas may function to detect potential prey, as vigilance

29

was not affected by Julian date, and hyenas were equally vigilant during rest and while
feeding. It is not surprising that hyenas did not use vigilance to monitor prey, as spotted
hyenas are not stalking predators; instead, they are coursing predators (Kruuk 1972) that
rely on endurance. Hyenas looking for animals to hunt typically run toward a herd of
ungulates, then stop to watch the prey animals running before picking a specific
individual to chase (Kruuk 1972; Holekamp et al. 1997).

Although vigilance has been reported to detect inter-specific threats in many
species (reviewed in Elgar 1989; Caro 2005a), we were surprised to witness this effect in
hyenas because of the very high levels of intra-specific competition observed in this
species (Kruuk 1972; F rank 1986). Many social species, including primates, are known to
use vigilance to monitor conspecifics (e. g. Waite 1987; Treves 2000; Hirsch 2002;
Cameron and du Toit 2005; Nutsukake 2006; Lung and Childress 2007; Pays and Jarman
2008). It is interesting to note that, although hyena society resembles those of
cercopithecine primates more than those of other carnivores (Drea and Frank 2003;
Holekamp et al. 2007), hyenas still displayed a group-size effect, whereas many primates
do not (reviewed by Treves 2000). In addition, the group-size effect we observed was
similar for adult and juvenile hyenas. We are only aware of two studies, both of birds,
that have evaluated the group-size effect for adults and juveniles in a single species
(Boukhriss et al. 2007; Avilés and Bednekoff 2007). Contrary to our results, in these
studies, a group-size effect was detected for adults but not for juveniles when birds were
foraging in mixed-age flocks.

We did not detect any evidence for social vigilance in hyenas; however, hyenas

may be monitoring conspecifics without adjusting vigilance time, or by using modalities

30

other than sight. Indeed, we noticed that resting hyenas exhibited multiple ear-twitching
movements even in the absence of insects, suggesting that they are using auditory cues to
monitor the environment. In addition, it has been suggested in studies of birds (Bednekoff
and Lima 2005; Femandez-Juricic et al. 2005) that animals might be using peripheral
vigilance to monitor conspecifics, and overt vigilance to monitor predatory threats. The
avian species used in such studies have broad fields of vision (e.g. Guillemain et al.
2002), allowing individuals to be vigilant even with their heads down. However, it is
unlikely that spotted hyenas could monitor conspecifics with such peripheral vigilance. In
contrast to birds, rodents or ungulates, hyenas rely heavily on binocular vision, and their
eyes are located on the front rather than the sides of the head (Bradbury and Wehrencarnp
1998), so their fields of view are much more restricted. Further, contrary to recent studies
showing that animals can be vigilant while handling food (e.g. Lima and Bednekoff 1999,
Makowska and Kramer 2007), the behavioral contexts in which we sampled hyenas
restricted their visual fields: hyenas often had their heads deep in a carcass while feeding,
or had their eyes closed while resting and nursing. Thus, it is unlikely that hyenas used
peripheral vigilance while engaged in the behavioral activities we observed.

Spotted hyenas varied their vigilance based on the behaviors in which they were
currently engaged, indicating that different activities may involve different tradeoffs. We
assume that hyenas should attempt to optimize the relationship between the benefits of a
given activity and its perceived risk. This was suggested, for instance, by our observation
that hyenas were less vigilant when feeding on high- than low-quality meat; when
feeding on high-quality meat, the cost of vigilance would rise by limiting intake of richer

food. Similarly, female hyenas spent the most time vigilant when they were nursing the

31

youngest cubs: young cubs are more vulnerable to predation than older cubs, resulting in
a enhanced need for the mother to monitor the environment against threats (Caro 1987).
We expected adult hyenas to be most vigilant when resting, a behavioral activity in which
we assumed the cost of vigilance would be minimal. However, vigilance did not differ
significantly between feeding and resting hyenas, indicating that either vigilance while
resting has a higher cost than we expected, or resting is as risky a behavior as feeding. It
is interesting to note that, although hyenas spent similar time vigilant during both feeding
and resting, this was obtained by using different strategies: resting hyenas exhibited
infrequent long vigilance bouts, whereas feeding hyenas exhibited frequent short
vigilance bouts. This was true for both adult and juvenile hyenas in our study, and this
has also been documented in other carnivore species (Chapter 4). The use of different
strategies may be due to varying tradeoffs associated with resting and feeding, or to
individuals monitoring different aspects of their environment when engaged in these two
activities.

Few previous studies, all done on primates, have compared vigilance in a single
species across several behavioral contexts (Cords 1995; Cowlishaw 1998; Treves et al.
2001; Hirtch 2002; Kutsukake 2006), although some studies have focused specifically on
sleeping birds (Lendrem 1983; Lendrem 1984; Gauthier-Clerc et al. 1994, 1998, 2002;
Gauthier-Clerc and Tamisier 2000). Contrary to the hyenas observed here, primates were
found to be less vigilant while foraging than while resting in all these studies. It is
difficult to assess whether this difference is based on biological differences between
carnivores and primates, or whether it is based on methodological differences regarding

what is considered ‘resting behavior’: hyenas were considered to be resting in this study

32

when lying down with their eyes closed, whereas primates were often considered to be
resting when not actively engaged in any specific activity.

We found a strong effect of individual hyena identity on adult vigilance during
rest and feeding. Other studies have included the identity of focal individuals in statistical
models as a random factor (e. g. Treves et al. 2001; Kutsukake 2007; Atwood and Gese
2008), but the results regarding this identity effect were not reported or discussed; its
inclusion in statistical models was used strictly to control for individual variation. Our
results indicate that some adult hyenas were consistently ‘relaxed’ while others were
consistently ‘worried’ about their environments. This finding is similar to those from
studies examining animals that vary along the shy-bold continuum (e.g. Coleman and
Wilson 1988; Lopez et al. 2005; Quinn and Cresswell 2005; van Oers et al. 2005; Réale
et a1. 2007). We did not find that trends in hyena vigilance were consistent between
behavioral contexts; for example, hyenas that were consistently vigilant while resting
were not necessarily those that were consistently vigilant while feeding. Such variation
should caution against generalizing results from one behavioral context to another.
Interestingly, we did not find any effect of hyena identity among juveniles, suggesting
that experience may play a role in determining the extent to which a particular individual
is vigilant as an adult.

Juvenile hyenas were less vigilant than adults, regardless of the activity in which
they were engaged. Similar results regarding variation among age classes have been
found in other animals including birds (e. g. Avilés and Bednekoff 2007), ungulates (e. g.
Alados 1985; Burger and Gochfeld 1994), rodents (e.g. Loughry 1993; Arenz and Leger

1997; Monclus et al. 2006) and primates (e.g. De Ruiter 1986; Fragaszy 1990). However,

33

a few studies have reported mixed effects or no effects of age on vigilance (e. g. Alberts
1994; McDonough and Loughry 1995; Blumstein 1996). The lower vigilance we
observed among juvenile than adult hyenas can be explained by several hypotheses that
are not mutually exclusive. Juvenile hyenas might require more rest or higher food ration
than adults, they might rely on nearby adults to detect approaching threats, or they might
need more time during ontogeny to correctly identify circumstances under which they
need to be vigilant (Caro 2005a).

Ecological factors that have been found to affect vigilance in many species of
prey were not found to affect vigilance in hyenas (Caro 2005a). For instance, despite
large sample sizes and high resolution, we did not detect any effect of distance to nearest
bushes on vigilance, nor of grass height or wind strength. This might be due to
differences in tradeoffs experienced by prey and predators. However, resting adult hyenas
were more vigilant in the morning than in the afiemoon in all seasons. Hyenas resting in
the morning may be more alert after the night of activities, whereas hyenas sampled in
the afternoon, which have just spent the majority of the day resting, may be less diligent.
Another possibility is that hyenas perceive their environment as riskier in the morning.

In conclusion, we have shown that spotted hyenas relied on vigilance to detect
inter-specific threats as demonstrated by group-size effects in adults and juveniles, that
vigilance varied among behavioral context, that adult hyenas exhibited consistent
vigilance across multiple trials within a given context, but not between contexts, and that
juveniles exhibited less vigilance than adults. Our study thus sheds light on the relative

importance of ecological functions of vigilance in a large, gregarious carnivore.

34

CHAPTER TWO

DEVELOPMENT OF THREAT-SENSITIVE BEHAVIORS IN SPOTTED HYENAS:
OBSERVATIONAL AND EXPERIMENTAL APPROACHES

INTRODUCTION

Compared to adults, juvenile mammals appear poorly equipped to deal with
danger (Janson and van Schaik 1993). Juveniles often fail to recognize or detect dangers
as rapidly as adults do, and juveniles sometimes respond inappropriately to alarm signals
produced by conspecifics (e.g., Seyfarth and Cheney 1986; Janson and van Schaik 1993;
Holle’n and Manser 2006). Perhaps because juveniles are often tacitly assumed to be
imperfect adults, they are often ignored in studies of threat-sensitive behaviors. Indeed,
most such studies focus exclusively on adult responses to threat stimuli (reviewed in Caro
2005a). However, as noted by Janson and van Schaik (1993), juvenile and adult
conspecifics may confront different sets of selection pressures, and juvenile behavior may
therefore reflect responses to age-specific challenges (e. g. Holmes 1984; Janson and van
Schaik 1993; Hoogland et al. 2006; Boukriss et al. 2007). Thus, juveniles might be
expected to adopt different strategies for dealing with danger than do adults. Indeed, age-
specific adaptive strategies in antipredator behavior have now been documented in
primates, ground squirrels, lizards and birds (e. g. Janson and van Schaik 1993; Hersek
and Owings 1994; Mateo 1996; Ramakrishnan and C033 2000; McCowan et al. 2001;
Martin and Lopez 2003; Platzen and Magrath 2004; Stone 2007). One of our goals here
was to extend our knowledge about the ontogeny of threat-sensitive behaviors to include
large mammalian carnivores. We focused specifically on spotted hyenas (Crocuta

crocuta).

35

Spotted hyenas are particularly well-suited for studies of ontogenetic variation in
behavior (e. g. Holekamp and Smale 1993). In contrast to many other mammals, hyenas
have a life history characterized by discrete stages, each of which begins and ends with
observable milestones (Holekamp and Smale 1998). Although hyenas are top predators in
the food webs of many African ecosystems, they encounter a number of serious threats,
particularly during early developmental stages, and most die violent deaths (Kruuk 1972;
Watts 2007). Indeed, fewer than 50 % of cubs born survive to adulthood, and mortality
rates diminish rapidly after spotted hyenas reach reproductive maturity (Frank et a1. 1995;
Watts 2007). This pattern suggests that juvenile hyenas might be more vulnerable than
adults to negative effects of competition and intraguild predation.

Most hyena mortality is caused by lions, especially male lions (Cooper 1991;
Frank et al. 1995; Trinkel and Kastberger 2005). Furthermore, young hyenas are
substantially more vulnerable to lion-caused mortality than adults, particularly after they
become independent of dens (Kruuk 1972; Frank et al. 1995; Watts 2007). Interestingly,
however, in addition to representing a potential source of injury or mortality for hyenas,
lions may also offer benefits by providing hyenas with opportunities to acquire food by I
scavenging (Kruuk 1972; Cooper 1991; Watts and Holekamp 2008). Hyenas have been
shown to be highly successful at scavenging from kills abandoned by lions, and at
stealing kills from lions if the hyenas are present together in sufficient numbers (Kruuk
1972; Cooper 1991; Honer et al. 2002). Thus, hyenas might potentially either avoid lions
to minimize risk of injury and death, or approach them to explore the possibility of
acquiring food. Furthermore, the nature of the relationship betweeen lions and hyenas

might be expected to change during ontogenetic development as hyenas gain motor

36

coordination, strength and experience. Previous work on carnivores has examined inter-
specific competition, intra-specific threats, and avoidance of intraguild competitors (e. g.
McComb et al. 1993, 1994; Durant 1998, 2000), but age-specific variation in stimulus
valence has not, to our knowledge, previously been eXplored in studies of threat-sensitive
behaviors in carnivores.

Here we used two different approaches to evaluate ontogenetic variation in the
threat-sensitive behavior of spotted hyenas. First, we evaluated naturally-occurring
vigilance in the absence of any immediate threats when hyenas were resting alone. We
expected that juvenile hyenas would be less vigilant than adults, because juveniles may
need more rest than adults, or because they are relatively naive about how to correctly
identify circumstances under which to be vigilant (Kruuk 1971; Caro 2005a). Second, we
performed playback experiments of lion roars to juvenile and adult hyenas. Because
juvenile hyenas are more vulnerable than adults to mortality caused by lions, and because
juvenile hyenas usually have poorer motor coordination as well as shorter limbs and thus
more limited escape capabilities than adult hyenas (Frank et al. 1991), we expected that
juveniles would respond to lions with much more caution than adults. Finally we inquired
whether responses by hyenas varied with the sex of the lions producing acoustic stimuli.
Male lions are the leading natural mortality source for both juvenile and adults hyenas
(Cooper 1991; Frank et al. 1995). Therefore, we expected adult hyenas to respond
differently to acoustic stimuli produced by male and female lions, but we expected that
juvenile hyenas would be unable to make this discrimination if the distinction between

male and female lions was learned during the juvenile period.

37

METHODS

Study site and study animals

All work was conducted in the Masai Mara National Reserve (henceforth the
Reserve — 1,500 kmz) in southwestern Kenya (1°40' S, 35°50' E) between August 2005

and June 2006. The Reserve is an area of open, rolling grasslands inhabited by large

numbers of ungulates (Sinclair and Norton-Griffiths 1979); this habitat therefore supports

high densities of both lions and hyenas (0.439 lions / km2, Ogutu and Dublin 2002; and

0.95 hyenas / km2, Watts 2007).

Spotted hyenas are gregarious carnivores that live in social groups, called clans,
containing up to 80 individuals (Kruuk 1972). Three clans that inhabit the Reserve were
monitored during data collection here, and playback experiments were carried out on
individuals from all three clans. All hyenas were known individually by their unique
spots. Birth dates (3: 7 days) and ages of all studied hyenas were determined using
previously described methods (Holekamp et al. 1996). Reproductive maturity in spotted
hyenas occurs at approximately 24 months of age (Glickman et al. 1992). Here, juvenile
subjects were those between 9 and 24 months of age, and subjects older than 24 months
were considered to be adults. All tested or observed subjects were independent of dens.
Social ranks were known for all hyenas based on wins and losses in dyadic agonistic
interactions (described in Holekamp et al. 1996). We classified individuals as high-
ranking if they were in the upper half of the social dominance hierarchy, and as low-
ranking if they were in the lower half of the hierarchy. Note that juveniles were assigned

the same ranks as their mothers.

38

Naturalistic observations of vigilance in the absence of threats

Hyenas were videotaped in the field to document their natural vigilance behavior.
Daily observations of the study animals took place between 0600 — 1000 h and 1600 -
1900 h from a field vehicle that served as a mobile blind. During these observations, the
field vehicle was parked no closer than 15 m from the focal hyena. Hyenas were well
habituated to our vehicle as all natal animals were exposed to it since infancy. Vigilance
data were only taken when we found a solitary hyena resting, defined as lying down in
the absence of food for at least five minutes, and at least 400 m from the nearest

conspecific. Upon finding a suitable subject, we set up a Sony DCR H65 video camera

mounted on a window tripod, and videotaped the focal hyena for 2 - 7 min (E = 4.00 i
0.07 min). We divided hyenas into 4 age groups: 8 - 16 months, 17 - 24 months, 25 - 48
months, and more than 48 months. If the same hyena was sampled repeatedly within an
age group, we averaged the data points for that individual and used this average in

subsequent analyses.

Playback stimuli

Hyenas were exposed to two types of sounds during playback experiments: lion
roars were treatment sounds, and baboon contest ‘wahoos’ were used as control sounds.
Lion roars should signal potential danger to hyenas and, because lions often roar at their
kills (C. Packer, pers. com.), they should also potentially indicate the presence of food.
Baboons (Papio anubis) abound in the Reserve and their contest wahoo calls are similar

in intensity and frequency to lion roars (Fischer et al. 2002, 2004; Kitchen et al. 2003).

39

Male baboons produce these calls during aggressive interactions with conspecifics
(Fischer et al. 2002). Baboons represent no threat to hyenas, nor do hyenas from the
Reserve hunt baboons. Wahoos therefore represent sounds with which hyenas are
familiar, but which should neither attract nor repel them.

Prior to the start of our study, we created a library of sound stimuli for use in the
playback experiments. Recordings of lion roars were obtained from Craig Packer (N = 8)

and from the Borror Laboratory of Bioacoustics (N = 5). Each recording consisted of a
Single bom 0f roars (7(- number of roars per bout = 19; } roar duration = 39-23 5) emitted by a

single adult female (N = 5) or male (N = 8) lion. Roars produced by male lions are
acoustically different from roars produced by females in that males produce roars of
lower pitch than females (Pferfferle et al. 2007). Lion sounds were not manipulated here.

Recordings of baboon wahoos (N = 15) were obtained from A. L. Engh. Each wahoo

recording consisted of a series of ‘wahoos’ (X numberofwahoos per bout = 8.53; X duration 0mg

stimulus = 25.3 s) emitted by a single adult male. Six of the fifteen wahoo recordings were

manipulated using the software Pratt either to lengthen or shorten the call, or to remove

vocalizations produced by additional baboons.

Playback experiment set-up

Experimental and control sounds were played back from our research vehicle to
solitary hyenas resting on open savannah. Here, a hyena was considered to be resting if it
was lying down for more than ten minutes. All hyenas were tested under similar

conditions, near the centers of their clan territories. Upon finding a suitable subject, the

car was positioned approximately 100 in (range: 73 to 130 m; Ydistance = 91 :1: 2.01 m)

40

from the subject hyena, at an angle of approximately 45 degrees to the plane formed by a
straight line between the subject and the observer, who was videotaping the hyena’s
response.

Recordings used in the field were picked randOmly from the sound library, but no
recording was used more than three times to minimize pseudo-replication (McGregor et
al. 1992). The recordings were played from a Creative Nomad Jukebox 3 linked to two
loud speakers (Fender Passport P-150) placed on window mounts facing away from the
focal hyena, and thus hidden by the car. The two speakers were separated only by about
20 cm. Both roars and wahoos are loud, low frequency calls (Fischer et al. 2002;
Pferfferle et al. 2007), requiring large speakers and high amplification when played back.
Sound amplitude was calibrated by ear to match natural sounds. The playbacks were
carried out between 0600 - 0900 hours and 1700 - 1900 hours when both lion roars and
baboon wahoos are heard naturally in this environment.

Before playing back any sound, all equipment was set-up and turned on, and a
window-mounted digital video camera aimed at the focal hyena. The observer (WMP)
started filming at least three minutes before sound onset; filming continued throughout
the duration of the playback, and for at least five minutes after sound onset. Once sound
onset occurred, a second observer in the same vehicle recorded the distance between the
focal hyena and the car every five seconds using a Bushnell laser range finder.

Most hyenas were exposed to both a lion roar and a baboon wahoo in random
order, with playbacks to any given hyena separated by at least two weeks (and, on
average, by 67 d: 10 days) to avoid habituation (McGregor et a1. 1992). Not all hyenas

were exposed to both sounds due to lack of opportunities in the field. We obtained data

41

for both types of playbacks from 18 hyenas, but 11 other hyenas were exposed to only

one of the two sound stimuli.

Data extraction from video tapes

We reviewed videotapes using a Sony DRC TRV900 digital camcorder that
superimposed a time code on the screen with a precision of 0.33 s, with each video frame
uniquely labeled. From tapes of naturally-occurring vigilance, we extracted information
on all vigilance events that occurred during filming. We defined vigilance behavior as
occurring whenever the focal individual lifted its head. The onset of a vigilance bout was
considered the point at which the animal lifted its head (halfway through the raising of
the head), and the end was considered the point at which the animal lowered its head
again (halfway through the lowering of the head). From the video footage, we extracted
the duration of each vigilance bout, summed all bouts, divided this sum by the total
length of the filmed sequence, and multiplied by 100 to obtain a percent time spent
vigilant. In our experimental work, we extracted the following data from each video tape
for both pre- and post-playback periods: 1) seconds spent orienting toward the speakers;
2) seconds spent resting (when individuals were lying motionless with their heads down);
3) seconds spent approaching the speakers, avoiding the speakers, or not moving in
relation to the speakers; and 4) the rate of movement (in meters per minute) of the focal
hyena. We also extracted the latency to first movement by the hyena after sound onset.
This included such events as the hyena standing up, but we did not count grooming

behavior as a first movement. Hyenas that never moved toward or away from the

42

speakers were assigned a latency to first movement of five minutes, as filming usually

stopped five minutes after sound onset.

Data analysis

We tested assumptions of normality and homogeneity of variance in our
observational data using the Wilks-Shapiro Test and the Levene’s Test, respectively. The
percent time spent vigilant while resting alone met both assumptions once it was square-
root transformed. We used a one-way ANOVA to test the effect of age group on the
percent time spent vigilant, and we assessed differences between age groups using a post-
hoc Tukey test. To compare overall juvenile vigilance to adult vigilance, we used a
Student t-test on the transformed percent time spent vigilant. In addition, we used linear
regressions to examine the effect of social rank on hyena vigilance.

We tested assumptions of normality and homogeneity of variance in our
experimental playback data. Only latency to first movement met both assumptions once it
was log-10 transformed. All other variables were non-normal, even after transformations.
Therefore, we used non-parametric tests in the analyses of all variables except latency to
first movement. In addition, we used linear regressions to examine the effect of age on
responses to playbacks within juvenile and adult age groups separately. We examined the
effect of social rank on all response variables, first among all individuals using linear
regression, and second among juveniles or adults separately, using a Fisher exact
probability test to compare low- and high-ranking individuals.

We then evaluated the effect of age group (juveniles vs. adults) and playback type

(baboon or lion) on hyena responses. We began by comparing the percent time hyenas

43

spent resting, orienting and moving before and after each playback, using Wilcoxon
signed rank tests for paired samples. We then compared responses between lions roars
and baboon wahoos within each age group using Wilcoxon-Mann-Whitney tests for non-
paired data (Siegel and Castellan 1988). In the remainder of our analyses, we only
considered responses to lion roars. In this subset, we used chi-square tests, applying
Yates’ correction for small sample sizes as necessary, to examine the effect of age group
on whether or not subjects moved in response to playback sounds (Zar 1999). We then
compared juveniles and adults with respect to the percent time spent avoiding, not
moving, or approaching the sound source after a lion roar using Wilcoxon-Mann-
Whitney tests for non-paired data. Here, we used a one-tailed test, as the hypothesis being
tested predicted that juveniles would react more negatively than adults to a lion roar. We
used a two-way ANOVA to test the effects of the sex of the lion roaring, and the age
group of the responding hyena, on latency to first movement. Finally, we compared rates
of movement made by hyenas after roars emitted by male and female lions using a
Wilcoxon-Mann-Whitney test. Here again we used a one-tailed test, as the hypothesis
being tested predicted that hyenas would move at higher rates after hearing roars emitted
by male than female lions. We performed retrospective power analysis if we found no
significant differences between juveniles and adults after roar onset to determine the
effect size detectable by our design (Thomas 1997; Orrock et al. 2004).

All analyses were performed in the statistical software package R, v.2.1.0 (R
Development Core Team 2005), using an alpha level of 0.05. All tests were two-tailed
except the two indicated above. Unless otherwise indicated, means i 1 standard error are

represented.

44

RESULTS

Naturalistic observations of vigilance in the absence of immediate threat
We obtained vigilance observations for 9 hyenas between 8 and 16 months of age,
12 hyenas between 17 and 24 months, 17 hyenas between 25 and 48 months, and 15

hyenas 49 months and older. When resting alone in open savannah, age groups varied in
the percent time spent vigilant (F 3, 49 = 4.14, r2= 0.20, p = 0.01; Figure 2.1), with the

youngest age group showing less vigilance than the oldest age group (post-hoc test, p =
0.012). The youngest hyenas sampled (between 8 and 16 months of age) spent an average
of 5.88 i 3.0 % of their time vigilant, whereas adults older than 48 months of age spent
an average of 18.16 i 3.52 % of their time vigilant. Hyenas in the two age groups older
than 24 months showed nearly identical levels of vigilance, but hyenas in the 17 — 24
months age group showed intermediate values between the two older groups and the

youngest group (Figure 2.1). Overall, we observed that hyenas younger than 25 months
differed from those 25 months and older (t5, = 3.12, p = 0.003). Thus, it appears that

hyenas reach adult-like vigilance after 24 months of age. There were no effects of

hyena’s social rank on the percent time spent vigilant (p > 0.1).

Common responses to playbacks

We carried out a total of 22 playbacks on juvenile hyenas (10 baboon playbacks
and 12 lion roar playbacks), and a total of 25 playbacks on adult hyena (12 baboon
playbacks and 13 lion roar playbacks). Although juveniles tested here included hyenas

from 9 to 24 months of age, age did not significantly affect any of the response variables

45

Percent time spent vigilant while resting alone

30

25

20

15

10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

" l l
17
_l
T
T

7 9

T l I I

8to 16 I7t024 25 t048 49andup

Age categories (in mo)

Figure 2.1. Variation in vigilance throughout ontogeny observed among
hyenas resting alone on open savannah. Means (t SE) are represented for
each age category, and numbers of individuals sampled are indicated
above each bar. The asterisk (*) indicates p < 0.05.

46

documented among juveniles (all R2 < 0.1, all p > 0.05), and the same was true among
animals classified as adults. There was also no detectable effect of clan membership on
any of the dependent variables (Krustal-Wallis tests, all p > 0.05), so we combined data
from all clans. There were no effects of the focal hyena’s social rank on its reaction to the
playback stimulus, regardless of the response variable examined (all r2 < 0.06, all p >
0.25).

The behavior of hyenas observed before sound onset differed significantly from
that observed after sound onset, and this was true for both juveniles and adults (Figure
2.2). Before sound onset, hyenas spent the majority of their time resting, which was
occasionally interrupted by bouts of vigilance oriented in various directions, but never
involved movement to a new location. After sound onset, hyenas increased their time
spent orienting directly toward the hidden speakers by an average of 27.4 i 3.2 %,
regardless of which sound was played back (Figure 2.2). This increase in time spent
orienting after sound onset was statistically significant for both juveniles (wahoo: Z =
2.09,p = 0.037; roar: Z = 3.06,p = 0.002; Figure 2.2a) and adults (wahoo: Z = 2.84,p =
0.004; roar: Z = 2.97, p = 0.003; Figure 2.2b). This increase in orientation was not
significantly different between control sounds and lion roars, regardless of whether we
looked at juveniles (W = 81, p = 0.166) or adults (W = 70, p = 0.663). Juveniles only
decreased their time spent resting after roar onset (wahoo: Z = 2.09, p = 0.37; roar: Z = -
3.06, p = 0.002; Figure 2.2a), whereas adult hyenas significantly decreased their time
spent resting after onset of both wahoo and roar sounds (wahoo: Z = 0.2.84, p = 0.004;

roar: Z = 3.18,p = 0.001; Figure 2.2b).

47

 

 

 

 

 

 

 

 

 

 

 

 

 

'00 l (a) — Before playback
:1 After playback
E 80 4
o
a.
m
a.) 60 r
E
g 40 .
o
L:
a)
e 20’ j j
0 . , .
Wahoo Roar Wahoo Wahoo Roar
100 1
(b)
E 80 ~
to
c.
m
o 60 -
.§
5 40 ‘
8
a)
On 20 4
0
Wahoo Roar Wahoo Roar

Orienting Resting Moving

Figure 2.2. Behaviors exhibited by (a) juvenile (N = 10 for wahoos, N = 12 for roars),
and (b) adult hyenas (N = 12 for wahoos, N = 13 for roars) before (black bars) and

after (white bars) sound onset during playback. The sound played was either a baboon
wahoo (control sound) or a lion roar (treatment sound). Means (:1: SE) are represented.

48

Among both juvenile and adult hyenas, movement away from the animal’s
original location only occurred after lion roars (Figure 2.2). Most hyenas initiated
movement after hearing lion roars (83 % of juveniles and 54 % of adults), regardless of
age (percent hyenas that moved after sound onset, Fisher exact test: juveniles: p < 0.001;
adults: p = 0.005). After hearing a baboon wahoo, hyenas never moved, and always

continued to lie in their original positions (Figure 2.2).

Differences between juveniles and adults exposed to lion roars

Differences between juveniles and adults in their responses to playbacks of lion
roars were surprisingly subtle. Juveniles and adults spent about the same percent time
motionless (Figure 3: around 30 %; W= 59; p = 0.16), and moved at similar rates after
lion roar playbacks (if-juveniles = 38.2 i 8.4 m / min; Y adults = 25.2 i 7.9 m / min; W=
60.5; p = 0.17). However, more juveniles (10 out of 12) than adults (7 out of 13) initiated
movement after roar onset, although this age difference was not statistically significant
(Yates’ x2 = 1.32, p = 0.25). Out of the 10 juvenile hyenas that initiated movement, 8
moved away from the sound source. In contrast, 3 adult hyenas approached the sound
source, while 4 avoided it. When considering the percent time spent moving in specific
directions, juveniles were significantly more likely than adults to avoid speakers after
roars (percent time spent avoiding speakers: fjuvennes = 42 i 9.7 %; Y adults = 12.7 i 5.0

%; W= 44; p = 0.03; Figure 2.3). In contrast, adults tended to approach the speakers

slightly more than juveniles (percent time spent approaching speakers: fjuvennes = 2.7 :1:

1.8 %; X adults = 13.8 i 6.8 %; Figure 2.3), but this trend was not statistically significant

49

60a *

50-

40 r - Juveniles (N=12)

1:1 Adults (N=13)

30-

Percent time spent

20-

 

 

 

 

 

 

 

Avoid No movement Approach

Figure 2.3. Mean (3: SE) percent time spent avoiding, approaching or remaining
motionless in relation to the speakers by the focal hyena following a lion roar playback.
Adult responses (white bars) are compared to juvenile responses (black bars). The
asterisk (*) indicates p < 0.05.

50

(W = 87.5; p = 0.27). Juveniles exhibited a shorter latency to first response after roars

than did adults, although this difference was not significant (Bf—juveniles = 108.2 i 34.6 s;

Y adults = 189.54 i 33.8 5; F122 = 2.69;p = 0.11; Figure 2.4). Thus, ofthe six variables

used to compare responses of juvenile and adult hyenas, only one (percent time spent
avoiding speakers) was statistically significant. However, five of these six variables show
relatively large effect sizes that are in the direction predicted by our hypothesis (Table
2.1). Results from our power analyses showed that sample sizes of 29 to 64 (instead of
the 12 and 13 we obtained) would have been required to statistically detect differences
between juveniles and adults. Thus, it appears that differences in responses to roars
between adults and juveniles are subtle, but may nonetheless have biological

significance.

Differential responses to roars emitted by male and female lions
We found that hyenas reacted more strongly to roars emitted by male than female

lions. Hyenas exhibited a lower mean latency to first movement after male than female

roars that was marginally significant (Y male roars = 108.94 i 27.44 s; 2? female was =

238.75 i 38.27 5; F122 = 3.83;p = 0.06; Figure 2.4), and this response did not differ

between juvenile and adult hyenas (non-significant age x roar type interaction). In
contrast to our expectation based on hyena mortality patterns, no hyena ever approached
speakers after female roars, whereas adults spent 12.5 % of their time, on average,
approaching speakers after roars emitted by male lions (U = 40; p = 0.04). Hyenas also

exhibited a higher mean rate of movement after male than female roars (U = 36.5; p =

51

350 -

 

 

 

 

300 ~ -—
’8
3”: 250 - 5
E
g C LionF
g 200 - 3 Q __ O LionM
E
E 150 u 8 Q Age:p=0.11
9, Roar Type: p = 0.06
3 ==
g 100 a -L
.3 9 j)

50 -

0
Juveniles Adults

Figure 2.4. Mean (:t SE) latency to first movement of hyenas following sound
onset. The playback sound was a roar produced either by a female lion (open
circles) or a male lion (filled circles). Numbers next to each mean represent the
number of individuals that received a given treatment. The statistics shown are
results from a 2-way ANOVA.

52

Aeoom

use—m0d 0wfic020m 0 35 m_ can 2: ... 3:60 5033 \ 9.560 5002 I warm 5003:3104 .mm “00330—00 83 050 000%.? 02028 0": ......
.8160 55 55:00 088 5:5 0:0: 8 “0:808

2:03 8:52: 35 “008090 03 £8: .3 E3008 8 3:60 :05 03805.3 088 0.8 3:0»: 03.02: 0.30005 .mu0wcmv fies 30¢
3 0:809:00 bfimfiaflfiv 0B $me @053 0.:— E0u0b=0 E 3:0»: «05 @5033. £35093 :0 .3 “000235 20:00.50 05 mm WEE ...

 

 

mg n a. N2 u x as 53 em mm 03% A 05.. £0 8.68 as: as»; aeeau
S u m ”cm 1 a: as 8.? $.62 :2: 2306 v 62.. E 868668 e_ 9 6:83
2.0 n m 0% n a as 3.: 8.8 8% seen A we: ESE 6686.68 06 sea
nmd u m 05 H x: X. odw flam— mnd 3160 V mafia Ae\ev 28—00% wcfiomoamm 08: E083
0:0 1 a 6m nu: as. ea 3% S. _ m 22a v 95.. £0 muses an 2e: unease
3o 1 m 3. n a s. 38 6.2 q? seen A use.“ axe unseen unease use 385
tens seam A2120 A2 is 1.35.6.5 2.3:; 3.5%:—

mu—Smfih Twomummuawm “Frans—0N— m==n=w flaw—Z mar—um. £602 £6300th

 

B0300 £0323on mo 0038:: 05 E0m0a0u 005m 2ng 00er 003 .88 so: 0 Ste 833 mo 08%
8 AntS 8:20.580 m0mcoam0u 05 @5388 08 m050te> :< $256090 033.33 05 mo £38 05 BM 35m 80am AN 050%

53

0.03; Figure 2.5), regardless of the direction in which they were moving. There were no
effects of sex of the roaring lion on the percent time afier playback during which hyenas

oriented, remained motionless, or avoided speakers (all U > 50, all p > 0.3).

DISCUSSION

Our results indicate that hyenas exhibit age-related variation in threat-sensitive
behaviors. First, young hyenas exhibited less naturally-occurring vigilance than adults
when resting alone, and they reached adult-like levels after puberty, which occurs at
around 24 months of age. Second, juvenile hyenas responded more strongly than adults in
response to lion roars in experimental playback. This was indicated by the fact that
juveniles started to move more quickly than adults afier sound onset, and spent more time
avoiding the speakers than did adults. In addition, there were several suggestive, but not
significant, trends in the data: juveniles moved away from speakers at slightly higher
rates than adults, and spent slightly less time approaching the sound source after sound
onset. Some adults also spent time avoiding the sound source, but adults showed much
more variability in their responses to the roars than did juveniles; whereas most juveniles
(8 out of 10) that changed positions after sound onset moved away from the hidden
speakers, some adults did not move at all and some moved toward the speakers.

Young hyenas exhibited low vigilance when resting alone in the absence of any
imminent threats, a pattern also observed in other species in which both juveniles and
adults were studied (e. g. Fragaszy 1990; Loughry and McDonough 1999; Monclus et al.

2006; Avilés and Bednekoff 2007). This pattern of low vigilance among juveniles can be

54

 

 

 

 

 

 

 

 

 

50 -
17

4o - I
.3
§
Ta“ 30 -
Q)
E 8
G)
>
O
E 20 -
Q—
o
O)
*5
a:

10 -

0

Female Male

Sex of lion roaring in the playback stimulus

Figure 2.5. Mean (1 SE) rate of movement by the focal hyena following exposure to a
lion roar playback made by either a female or a male lion. Both directions of movement
(approaching and avoiding the sound source) are included here. Sample sizes are located
above each mean and indicate the numbers of hyenas tested. There was no effect of age
on the movement rate, so juveniles and adults were combined here. The asterisk (*)
indicates p < 0.05.

55

explained by several hypotheses that are not mutually exclusive. Young animals may
need more rest than do adults due to growth requirements (e. g. Arenz and Leger 2000).
Earlier studies (e. g. Avilés and Bednekoff 2007) have proposed that young may rely on
nearby adults to compensate for their lower vigilance behavior, although in our case,
juvenile hyenas were sampled alone. In addition, young animals might need time during
ontogeny to learn to identify the circumstances under which they should be vigilant.

In the presence of an immediate threat indicated by lion roars, juvenile hyenas
showed slightly stronger negative responses than did adults. As the escape capabilities of
juveniles may be limited relative to those of adults, lions might represent a larger threat
for juveniles than for adults. Indeed, this notion is supported by the observation that
juveniles experience much higher rates of lion-induced mortality (Watts 2007). That
juveniles react more strongly to threats than do adults has also been shown in a number of
other species (e. g. Ramakrishnan and Coss 2000; Pongracz and Altbéicker 2000), and our
results suggest that juvenile hyenas appear to be somewhat less nai've than juvenilesof
many other species (e. g. Hollén and Manser 2006; Hanson and Coss 1997).

Our study is quite simplistic compared to those done on primates in which more
complex arrays of threat stimuli were presented to juveniles (e. g. vervet monkeys,
Cercopithecus aethiops: Seyfarth et al. 1980). Nevertheless, we observed differences
between juvenile and adult hyenas, albeit not as strong as those reported in monkeys.
This suggests that perhaps certain antipredator responses emitted by hyenas require less
learning than do those of monkeys. This idea is supported by an anecdotal observation
reported by Kruuk (1972) regarding a hand-reared juvenile hyena who showed a very

strong antipredator response to lion odor before having ever been exposed to any lions.

56

Although this response was elicited via a different sensory modality, it may indicate
nonetheless that hyenas need little or no direct experience with lions in order to respond
to them appropriately.

We found that both juvenile and adult hyenas responded more strongly to roars
emitted by male than female lions, suggesting that both age groups of hyenas can
distinguish lion sex based on roars. Lions have acoustically different roars based on their
sex, which can be entirely explained by their large size dimorphism (Pferfferle et al.
2007). However, these differences are not readily audible to naive human observers.
Nevertheless, hyenas responded to roars emitted by males and females in accordance
with our predictions, in that they showed higher rates of movement afier male than
female roars. However, contrary to our expectations, all occasions on which adult hyenas
approached the hidden speakers occurred after male roars. This might be either because
male roars provide a stronger indication of scavenging opportunities, or because they
induce a heightened need to acquire information about the potential threat. In general,
studies of threat-sensitive behavior do not evaluate the effect of the predator’s sex
because the two sexes are assumed to represent similar risks to prey. The fact that we
found several response variables to vary with lion sex suggests perhaps that prey animals
may be more sensitive than we had previously assumed, and able to make finer
discriminations than anticipated.

We did not find an effect of hyena social rank on the response to lion roars,
although our small sample sizes might be responsible for this. Indeed, two of the three
adult hyenas that spent time approaching the speakers after lion roars were low-ranking

hyenas. These hyenas might have been looking for opportunities to scavenge food from

57

lions, and lions may have appeared more attractive to low- than high-ranking hyenas.
Similar results, in which low-ranking individuals were willing to take more risks than
high-ranking individuals after exposure to a threat, have been obtained in studies of a
number of different types of animals (e.g. Reinhardt 1999; Hegner 1985; Slotow and
Paxinos 1997). For instance, subordinate blue tits (Parus caeruleus) returned sooner to a
food patch than did dominant individuals after exposure to a threat stimulus (Hegner
1985). To more fully explore the effect of social rank on the behavior of hyenas in
response to lions, more extensive experimental work will be necessary with larger
samples of adults holding various rank positions.

In conclusion, our results suggest that hyenas early in life appear to recognize
certain signals (in our case, lion roars) as dangerous, and respond adaptively by avoiding
the source of potential danger. Nevertheless, juvenile hyenas appear to be learning about
the environmental circumstances under which they need to be vigilant, and they appear to
require many months to fine tune their behavior. Our study extends our knowledge about
the ontogeny of threat sensitive behaviors to include a large, gregarious carnivore. Such
ontogenetic studies can provide useful insights into the potential variability among

selective forces acting on animals during different stages of their life history.

58

CHAPTER THREE

LETHAL AND NON-LETHAL ANTHROPOGENIC EFFECTS ON SPOTTED
HYENAS IN THE MASAI MARA NATIONAL RESERVE

INTRODUCTION

Persecution by humans often represents a leading source of mortality for wild
carnivores, even inside protected areas (Woodroffe and Ginsberg 1998, 2000), and
contributes to reduced population size (Woodroffe 2000, 2001; Creel and Creel 2002).
This mortality is frequently associated with human-carnivore conflicts in areas where
carnivores persist but human populations are expanding (Woodroffe 2000; Ogutu et al.
2005; Johnson et al. 2006). In addition to such lethal effects, humans can potentially also
have non-lethal effects on carnivores by modifying their behavior and physiology (Lima
1998; Creel et al. 2002). Recent studies have shown that carnivores display both spatial
and temporal avoidance in response to increasing human activity (e. g. Boydston et al.
2003; Kolowski et al. 2007), and they also undergo changes in their stress physiology
(Van Meter et al., in review). Anthropogenic activity has also been linked to elevated
concentrations of excreted stress hormones, reduced time spent feeding, and more time
spent alert in various other carnivores (e.g. Creel et al. 2002; Dyck and Baydack 2004;
Nevin and Gilbert 2005; reviewed by F rid and Dill 2002). Such findings have led to a call
for a better understanding of both lethal and non-lethal effects of human activity on large
mammalian carnivores in order to conserve them (e. g., Caro and Durant 1995; Berger
1998, 2000; Gittleman et al. 2001).

Threat-sensitive behaviors, such as vigilance, are of particular significance to
conservation in that they have the potential to influence foraging and time-budgets (Lima

and Dill 1990; Caro 1999), which in turn can affect population growth rates (Dobson and

59

Poole 1998). Currently, most studies on mammalian carnivores focus on the roles played
by these animals as predators rather than prey (e.g. Mills and Shenk 1992; Murray et al.
1995; Cooper et al. 2007), resulting in a lack of understanding of the mechanisms these
animals possess for coping with danger. Only a few studies have examined threat-
sensitive behaviors in mammalian carnivores (Caro 1987; Rasa 1989; Clutton-Brock et
al. 1999; Switalski 2003; Di Blanco and Hirsch 2006; Hunter et al. 2007; Atwood and
Gese 2008). Because vigilance is ubiquitous and easily quantifiable, it may provide a
non-invasive indicator of how much risk a given animal perceives, especially in
populations disturbed by human activity. Indeed, vigilance has been used as such a tool
with polar bears (Ursus maritimus), which increase their vigilance and energetic costs in
the presence of wildlife-viewing vehicles (Dyck and Baydack 2004).

Conducting behavioral studies on most carnivores can be quite challenging
because they occur at low densities, they are generally nocturnal, and they are wary of
humans (Sargeant et a1. 1998). However, spotted hyenas (Crocuta crocuta) are well-
suited for such investigations because they are easily observable in the open habitats of
sub-Saharan Africa, and they are active around dawn and dusk (Kruuk 1972; Kolowski et
al. 2007). Their social organization in large fission-fusion societies, called clans,
containing up to 80 individuals (Kruuk 1972; Mills 1990), allows for repeated
observations of many known individuals. Furthermore, although spotted hyenas are
currently considered a species of Lower Risk by the World Conservation Union (IUCN
2006), they may represent conservative indicators of how threatened and endangered
carnivores inhabiting the same ecosystems might react to human disturbance. In contrast

to most other carnivores, spotted hyenas breed year-round, can be active day or night,

60

occur in a diverse array of habitat types, and can survive on food ranging from insects to
elephants (Kruuk 1972; Mills 1990; Sillero-Zubiri and Gottelli 1992; Cooper et al.1999).
In parts of the Masai Mara National Reserve (henceforth the Reserve), spotted hyenas
must cope not only with lions (Panthera leo), which are their only natural predators
(Cooper 1991; Frank et al. 2005), but also with intensive disturbance by humans,
particularly along Reserve borders (Boydston et al. 2003; Kolowski 2007).

Here we compared two hyena clans that differed dramatically in their exposure to
anthropogenic disturbance in order to evaluate the effects of human activity. We
compared behavior of hyenas in the Mara River clan, which was located in the center of
the Reserve, with that of hyenas in the Talek West clan, which was located at the edge of
the Reserve. The hyenas in these clans experienced different levels of human disturbance
from both tourists in vehicles and pastoralists on foot guarding livestock inside the
Reserve (Kolowski et al. 2007). We used three different approaches in this study. First,
we assessed long-term trends in human-caused mortality among hyenas in the two clans.
Second, we documented naturally-occurring vigilance behavior exhibited by wild hyenas
in three different behavioral contexts, both when lions were present with hyenas and
when lions were absent. Finally, we performed an experiment to test predictions of an
hypothesis suggesting that vigilance behavior in hyenas is influenced by the presence of
livestock and herders. Because of their frequent exposure to human disturbance, we
expected that Talek West hyenas would spend more time vigilant than Mara River
hyenas. In southern Kenya, local pastoralists always accompany their livestock, which
wear metal cow bells; herders sometimes kill or harass indigenous wildlife while

watching their herds, suggesting that hyenas might associate the sound of cow bells with

61

danger. In playback experiments, we compared responses of hyenas in both study clans to
cow bells (treatment sounds) and church bells (control sounds). We predicted that hyenas
from the clan experiencing little human disturbance would react similarly to sounds of
cow bells and church bells, as both should represent novel sounds, whereas we expected
that hyenas from the clan experiencing cow bell sounds daily would show greater

vigilance in response to these sounds than to sounds of church bells.

METHODS

Study site and study animals:

The Reserve (1,500km2) is located in southwestern Kenya (1°40 S, 35°50 E), in
an area of Open, rolling grasslands. The Reserve is inhabited by large numbers of
ungulates (Sinclair and Norton-Griffiths 1979) and high densities of the various large
carnivores that prey on them (Ogutu and Dublin 2002; Dloniak 2006).

The two clans monitored in the Reserve, the Talek West and the Mara River
clans, are ecologically very similar except for exposure to human activity (Kolowski
2007). Both clans experience similar rainfall, prey density, lion density and vegetation
cover (Table 3.1; Dloniak 2006; Kolowski et al. 2007). Because their territory lies along
the border of the Reserve, Talk West hyenas live in close proximity to Masai villages,
and experience daily human disturbances in the forms of livestock herds guarded by
herdsmen, and in the form of vehicles carrying tourists engaged in wildlife viewing. In
contrast, the territory defended by the Mara River clan is located over 6 km from the

nearest Reserve border, which is too deep in the Reserve for livestock to travel for daily

62

coca x0805 ...

 

80: :0: 00:03

 

w _ .o 56 82000850 003200, 88 .00 .098: Z
m6 o A80: 8008: 3.500% 9 00:8me
mm H DSEB 05 :o
o a: 0 mm? 30: wfinfiw 03883— 90 038:: :002
one 0 am 000 a 3. mm a :8: 00 a: 0:80 03. .8: :82
8.0 A 2 m2 0. 2.2 ...mm A 000 000: so: =82
m m 0:00:80: wfim: 002.8 :0: 00 83:82
0. _ M 3m €50 0% bets:
.2. - mm mm I t. was: 00 000. .06
:05 00>;— 0002 :05 0003 30.0%

 

.Anoomv EmBBOM 80¢ 008050 003 058 0:: 80¢ 8:08.508
:0 80:00:08 0.0650050 3.0—:3 .803 003% 0:02 0:0 303 0.20:. .0:03 3080. 025 05 mo 00580800050 80=0m 8m 030:.

63

grazing; Mara River hyenas experience some visitation by tourists, but no livestock herds
or pastoralists (Table 3.1). All hyenas from both study clans were known individually. by
their unique spots. Ages (3: 7 days) of all hyenas were determined using previously
described methods (Holekamp et al. 1996). Only adults (older than 2 years old; Matthews

1939; Glickman et al. 1992) were included in this study.

Determination of causes of death:

The Talek West clan was observed between July 15‘, 1988 and September 15‘,
2006 as part of a long-term monitoring program. During this period, we recorded all
known deaths, and attributed a cause of death when this could be reliably determined
from observer reports or marks on the body of each dead hyena. If the body of a hyena
revealed evidence of snaring, spearing, or poisoning, or if resident pastoralists informed
us they had killed the hyena, we attributed the cause of death to humans. Other sources of
mortality included lions, disease, starvation, den flooding, infanticide and siblicide. Most
deaths could not be reliably attributed to one particular source, so we restricted our
dataset to 83 deaths of known causes. From this subset, we calculated the proportion of
deaths that could be attributed to humans in each of the 18 years of the study. We
recorded the same information for Mara River hyenas, which were monitored between
January 15‘, 2001 and September 15‘, 2006. For this clan, eight deaths could be reliably

attributed to known causes.

64

Naturalistic observations of vigilance:

Individual hyenas in both study clans were videotaped in the field between June
2005 and July 2006 to document variation in their naturally-occurring vigilance behavior.
Daily observations of the study animals took place between 0600 - 1100 h (the AM
period), and between 1500 - 1900 h (the PM period) from a field vehicle that served as a
mobile blind. During observations, the field vehicle was parked no closer than 15 m from
each focal hyena. Hyenas are well habituated to our vehicle as all natal animals were
exposed to it since infancy. We documented vigilance in three different behavioral
contexts: while resting, feeding at kills and nursing cubs. Hyenas were considered to be
resting when they were lying down in the absence of food for at least 5 min, and not
interacting with conspecifics; the eyes of resting hyenas were usually shut. Hyenas were
considered to be feeding at kills when we found them consuming a fresh ungulate
carcass; hyenas feeding on low quality scraps were not included in this study. Female
hyenas were considered to be nursing when they suckled their litters while lying down.
Upon finding a hyena engaged in one of these three activities, we set up a Sony DCR-
H65video camera mounted on a window tripod and videotaped the focal animal for 2 - 7
min (f: 4.25 i 0.04 3).

At the time of filming, we recorded various types of ecological and social
variables, including the number of conspecifics present within 100 m of the focal
individual (referred to hereafier as the “group size”). “Clan size” refers to the size of the
entire social unit, although all members are seldom present concurrently. Spotted hyenas
live in fission-fusion societies in which clan members are found in groups of varying size

and composition (Smith et al. 2008). We also recorded the distance between the focal

65

individual and the closest patch of bushes (using a Bushnell range finder), as hyenas may
view bushes as refugia or bushes might conceal lions. Lions were considered to be
present with the focal hyena if they were less than 200 m from it. During daily
observations, we also recorded the presence, and estimated the size, of any livestock
herds seen in the territory of the Talek West clan; no livestock were ever seen in the Mara
River territory (Table 3.1). Only herds containing cows were recorded, although herds of
sheep and goats are also seen occasionally in the Talek West territory.

Tourist visitation to the Reserve varied seasonally, peaking during western
holiday periods. To evaluate the potential effect of tourist presence on vigilance in
hyenas, we noted whether our filming was taking place during months of heavy tourism,
which we determined using data available from Heath (2007). The number of tourists
visiting the Reserve each month ranged from 5,336 to 23,847; heavy tourism months
were considered those during which the number of tourists exceeded 10,000 (N = 6):
June, July, August, September, October and December. As tour vehicles are abroad in the
Reserve during only two specific periods each day (0630 - 0900 h and 1630 - 1900 h;
Kolowski et al. 2007), we also noted whether each filming event occurred during tourism

hours or non-tourism hours.

Playback experiment:

Hyenas were exposed to two types of sounds during playback experiments:
treatment sounds were cow bells ringing and control sounds were church bells ringing.
Masai pastoralists generally hand bells on collars around the necks of several cattle in

each herd. These metal bells ring whenever the cattle move their heads, and this ringing

66

can be heard from over one kilometer away. Like cow bells, church bells are also made of

metal, but their frequencies are much lower than those produced by cow bells (mean

dominant frequencies calculated using Adobe Audition 1.5: -X— church be”, = 738 Hz; Yew

be”, = 1,121 Hz), resulting in very different sounds.

Cow bell sounds were recorded from moving herds of cattle in and around the
Talek West clan territory, using a Marantz PMD-22 portable cassette recorder and a
Sennheiser ME66 shotgun microphone. All recordings (N = 11) were made from herds
that had approximately ten individuals fitted with bells. Church bell sounds were
recorded with the same equipment from eight different churches located around the city
of La Rochelle, France. The recordings were obtained when bells rang in a carillon-style
15 minutes prior to mass (this specific ringing is called “l’appel de messe”). We
manipulated all recordings using Adobe Audition 1.5 to shorten each recording to 60 s in
duration.

Experimental and control sounds were played from our research vehicle to
solitary hyenas resting on open savannah. A hyena was considered solitary if it was
separated by at least 400 m from the nearest conspecific. Upon finding a suitable subject,
the car was positioned approximately 100 m (range: 80 to 115 m; fan-game = 91 i 2.84
m) from the hyena, at an angle of approximately 45 degrees to the plane formed by a
straight line between the subject and the observer videotaping the hyena’s response.
Recordings used in the field were picked randomly from the sound library, and no
recording was used more than twice to minimize pseudo-replication (McGregor et a1.
1992). The recordings were played from a Creative Nomad Jukebox 3 linked to two loud

speakers (Fender Passport P-150) placed on window mounts facing away from the focal

67

hyena and thus hidden by the car. The two speakers were separated only by about 20 cm.
Sound amplitude was calibrated by ear to match natural sounds. The playbacks were
carried out between 0700 and 1830 h when cow bells are typically heard in the Talek
West territory.

Before playing back any sound, all equipment was set up and turned on, and a
digital video camera was aimed at the focal hyena. The observer (WMP) started filming
at least three minutes before sound onset; filming continued throughout the playback, and
for at least four minutes after sound onset. Fourteen individual hyenas were included in
the playback experiment. Seven hyenas were exposed to both cow and church bell sounds
in random order, with playbacks to any given hyena separated by at least two weeks (and,
on average, by 28 days) to avoid habituation (McGregor et al. 1992). Seven hyenas were
exposed to only one or the other of the sounds due to lack of additional opportunities in

the field.

Data extraction:

We watched videotapes made during naturalistic observations and experiments
using a Sony DRC TRV900 digital camcorder that superimposed a time code on the
screen with a precision of 0.033 s, with each video frame uniquely labeled. We identified
vigilance behavior whenever a focal individual lifted its head. Onset of a vigilance bout
was considered to be the point at which the individual lifted its head (halfway through the
raising of the head), and the end was considered to be the point at which the individual
lowered its head again (halfway through the lowering of the head). From the video

footage, we extracted the duration of each vigilance bout, and summed all bout durations

68

before dividing this by the number of bouts to obtain an average bout duration. The total
number of bouts in the filmed sequence was also used to calculate the number of bouts
per minute filmed. From the average duration and the rate of vigilance bouts, we
calculated the total percent of its time that an animal spent vigilant during the filmed
sequence. Percent time spent vigilant was calculated for both naturalistic observations
and experiments, although hyenas that lified their heads following sound onset in the
experiments typically oriented toward the sound source. In experimental trials, we also
extracted the following data from each video tape for both the pre- and the post-playback
periods: 1) latency to first movement by the focal hyena afier sound onset; 2) direction of

movement —away or toward speakers; and 3) percent time spent moving after sound

onset.

Data analyses:

We tested assumptions of normality and homogeneity of variance in our response
variables using the Wilks-Shapiro Test and the Levene’s Test, respectively. All response
variables met both assumptions once they were log-10, square-root, or arcsine-square-
root transformed.

We analyzed data from our long-term monitoring program to test whether the
proportion of known-cause deaths attributable to humans changed between 1988 and
2006 in the Talek West clan, and between 2001 and 2006 in the Mara River clan. To do
this, we used linear regression to evaluate the relationship between year of the study and

the proportion of hyena deaths of known causes each year that could be attributed to

humans.

69

We analyzed data from the naturalistic observations to determine whether
vigilance differed between the two clans, and whether human-related activities could
explain variation in vigilance. We first examined the effect of clan membership on
vigilance using a general linear mixed-effect model (GLMM), in which hyena identity
was entered as a random effect, nested within clan membership, to avoid potential
pseudo-replication (Pinheiro and Bates 2000). Significance of the random identity effect
in this model and subsequent models was tested using likelihood ratio tests, comparing
models with and without random effects (Pinheiro and Bates 2000). In this GLMM, we
also included the time at which each observations was made (AM vs. PM), lion presence
or absence, group size, and the interaction between these factors and clan membership as
fixed effects, because these factors are known to affect hyena vigilance (Chapter 1). We
ran this analysis on the percent time spent vigilant in each of the three behavioral
contexts (resting, feeding and nursing) separately. In addition, we also evaluated the
effect of clan membership on distance between nursing females and the nearest bushes
using a GLMM in which female identity was entered as a random effect. We also
included group size in this model, but not lion presence, as we never observed nursing in
the presence of lions.

Whenever differences between clans were revealed in our analyses, we then used
GLMMs to inquire whether this effect could be attributed to anthropogenic activity
associated with tourism or livestock grazing. This analysis included livestock presence
(yes or no) and tourist presence (2 variables: tourism months or non-tourism months,
tourism hours or non-tourism hours, and their interaction) as fixed effects, as well as time

of day (AM vs. PM), group size, lion presence or absence, clan membership and hyena

70

identity, which was nested within clan membership and entered as a random effect. For
this analysis, we only included observations of resting hyenas (N = 633, on 34 Talek
West hyenas and 24 Mara River hyenas), and evaluated effects on the percent time spent
vigilant.

We analyzed the response variables recorded during playback experiments in two
different ways. To test the effect of the playback sounds on the percent time spent
vigilant, we used a 3-way analysis of variance with clan membership, playback type, and
pre vs. post playback period as the three factors. We also used a chi-square analysis,
applying Yates’ correction for small sample sizes as necessary, to examine effect of clan
membership and playback type on whether or not subjects moved in response to playback
sounds.

All analyses were performed in the statistical sofiware package R, v.2. l .0 (R
Development Core Team 2005), using two-tailed tests with an alpha level of 0.05. The
GLMMs were done using the R library ‘nlme’ (Pinheiro et al. 2005). Unless otherwise

indicated, means 3: 1 standard error are presented.

RESULTS

Lethal effects of humans on spotted hyenas:
Between 1988 and 2006, 20 of 83 hyena deaths of known causes in the Talek

West clan could unambiguously be attributed to humans. The relative proportion of

human-caused deaths increased significantly during the years of our study (F1, 17 = 11.28,
p = 0.004, r2 = 0.40; Figure 3.1). We observed a similar increase when we examined the

actual number of deaths attributable to humans during the years of our study (F 1, ,7 =

71

L0 - 1.

 

 

 

O Talek West clan: F117= 11.28 p= 0.004
O Mara River clan
U)
a 0.8 -
:3
Cd
0
g, o..-
'4...
0
U)
5 0.4 -
«3
d)
u
H
O
S 0.2 -
t
o
o.
8
24 01)-
l l I l
1985 1990 1995 2000 2005

Year of the study

Figure 3.1. Proportion of hyena deaths of known causes that could be attributed to
humans in each year of the study in two clans, the Talek West clan (filled circles)
and the Mara River clan (empty circles). The Mara River clan was not monitored
before 2001. The line represents a linear regression performed on transformed data
but plotted on untransformed data; the regression analysis only includes data points
for the Talek West clan.

72

6.76, p = 0.02, r2 = 0.28). In the Mara River clan, only 8 hyena deaths of known causes

were observed, and none could be attributed to humans (Figure 3.1).

Naturalistic observations of vigilance in hyenas from bath clans:
The number of observations collected, and the number of individuals sampled in

each behavioral context in each clan, are reported in Table 3.2. When resting, Talek West
hyenas were more vigilant than Mara River hyenas (F 1, 58 = 6.53, p = 0.01; Figure 3.2a),
with Talek West hyenas spending more than twice as much time vigilant as Mara River

hyenas (Y Mek= 16.17 i 1.55 %; f Mm Rm: 7.21 :1: 0.88 %; Figure 3.2a). Although

both clans were significantly more vigilant in the morning than in the evening (F1576 =

29.47, p < 0.001; Figure 3.2a), vigilance of resting Talek West hyenas was consistently
greater than that of Mara River hyenas regardless of time of day (the clan x time
interaction was not significant). The interaction between clan and group size was also not

significant, despite the fact that vigilance was significantly less in both clans when
hyenas rested in large than small groups (F1576 = 7.29, p = 0.007). Interestingly, the
difference in vigilance between clans was apparent in the absence of lions, but not in the
presence of lions, resulting in a significant clan x lion interaction (F1577 = 3.63, p = 0.05;
Figure 3.3). Specifically, when lions were present, resting hyenas approximately doubled
their percent time spent vigilant over that when lions were absent (F 1, 576 = 57.70, p <

0.001; Figure 3.3); the two clans did not differ significantly on this measure.

73

Table 3.2. Number of naturalistic vigilance observations collected (N obs) on adult
hyenas (N ind) sampled in each behavioral context (resting, feeding and nursing) in the
two study clans (Talek West and Mara River clans) in the presence or absence of lions.
We never observed adult females nursing in the presence of lions.

 

Talek West I Mara River

Context without lions with lions without lions with lions

 

Nobs N ind N obs N ind Nobs N ind Nobs N ind

 

Resting 412 34 42 22 170 22 15 9
Feeding 75 29 15 1 l 26 21 5 4
Nursing 42 9 - - 39 7 - -

 

74

.930: w5:0>0 05 5 05600.: 00:03 03¢ 0.02 003030

:0>0: 0 3 .God v 5 000:0:0bm0 500$:me 300303000 0:000:00: AL mxmt0gm< .508 £000 0>0n0 000006E
003800 20:03:05 .«0 0098:: 05 53> 68:08:00: 00 Amm “3 0:002 .300 :05 365: A8 0:0 .w:_000.«
By “@3000 A3 00?:030 0:03 800: 53020000 A23 w:E0>0 0:0 @243 wEEoE wfibfi $00 0:85 :03
03¢ 002 05 20 A003 023 :03 0003 030,—. 05 80¢ 00:0»: 3 0:0=w_> 0:000 08: 0:00:0m .N.m 053m

 

  

 

 

 

 

 

 

@652 @600: w::m0m
SE $2 $5 E< SE Z<
-. . u -
. 3 mm . 2 mm
—N _N
. ON . ON
0m
. Om . Om
[* .IlL
*
. ow . ow .
*
3 V
a. A 0 om 3 0on

00>;— 002 U
>2 320,—. I

3

 

rueufi 1A ruads awn ruaorad

75

**

 

60 - aura: uni:

 

 

50‘ 9

- Talek West
[:1 Mara River

30-1

20:

Percent time spent vigilant

10‘

 

 

 

 

 

 

 

No lions Lions No lions Lions

 

 

Resting Feeding

Figure 3.3. Percent time spent vigilant by spotted hyenas of the Talek West clan
(black bars) and Mara River clan (white bars) while resting and feeding in the
presence or absence of lions. Means (i SE) are represented, with the number of
individuals sampled indicated above each mean. Asterisks represent statistically
significant differences (*p < 0.05, **p = 0.01; ***p < 0.001).

76

We did not observe differences in vigilance between clans when hyenas were

feeding or nursing (Figure 3.2b, c). Talek West and Mara River hyenas spent roughly the
same percent time vigilant while feeding (Y Talek = 9.39 i 1.00 %, 2? Mam Rive, = 10.43 i
1.22 %; [7,47 = 1.55,p = 0.22; Figure 3.2b). Hyenas frOm both clans spent a larger
percent time vigilant when feeding in the presence of lions than when lions were absent
(F 1,69 = 9.83, p = 0.002; Figure 3.3), and when feeding in large than small groups (171.69 =
5.18, p = 0.026). Vigilance while feeding did not differ between AM and PM observation
periods (F 1,69 = 1.73, p = 0.19; Figure 3.2b). Hyenas from both clans spent similar
percentages of their time vigilant while nursing their litters (f Talek = 38.40 i 7.28 %;

Y Mam Rive, = 28.94 :t 4.57 %; F,” = 0.32, p = 0.32; Figure 3.2c). Nursing mothers in
both clans were more vigilant in the morning than in the evening (F 1,63 = 5.07, p = 0.03).
There was no effect of group size on vigilance while nursing (F 1,6 3 = 2.91, p = 0.09).

Female hyenas choose where to suckle their litters, so we evaluated differences

between clans in the distance that mothers nursed from bushes. Talek West females

nursed significantly closer to bushes than did Mara River females (Y Talek = 77.47 i

23.65 m; EM“, River= 267.68 0 38.97; FL” = 6.18,p = 0.026; Figure 3.4). Hyenas from
both clans were observed further from bushes if they were nursing in large than small
groups (F 1,60 = 4.54, p = 0.037), but the difference between clans was apparent regardless

of group size (i.e., non-significant clan x group size interaction). Distance to bushes did

not vary between clans when hyenas were resting or feeding (p > 0.1).

77

 

350 ‘ l 1

300 - T

250 -

 

200 -
150 -+

100 - 'l'

50-

 

Average distance to closest bushes when nursing (m)

 

 

 

 

 

 

o . . '
Talek West Mara River

Figure 3.4. A comparison of the two study clans with respect to the average
distance between focal hyena nursing and the closest patch of bushes. Means
(:1: SE) are represented, with the number of individuals sampled indicated
above each mean. Asterisks represent statistically significant differences (*p
< 0.05).

78

We next focused on how different human activities affected vigilance among

resting hyenas. Talek West hyenas were more vigilant on days when livestock were
present in their territory than when livestock were absent (F 1, 4 1 3 = 3.79, p = 0.02; Figure
3.5). This effect was of the same magnitude regardless. of whether there were 30 or 2200
livestock present in the territory (Figure 3.5). In contrast to pastoralist activity, there were
no effects of tourist activity on vigilance by either Talek West or Mara River hyenas.

Resting hyenas spent the same percent time vigilant in months when many tourists were
visiting the Reserve as in months when few tourists were visiting the Reserve (F 1, 565 =
0.05, p = 0.82; Figure 3.6), and during hours when tour vehicles were present or absent

from the Reserve (F 1, 565 = 3.74, p = 0.06; Figure 3.6).

Playback experiment:

We carried out a total of 13 playbacks on hyenas from the Talek West clan (7
cows bell playbacks and 6 church bell playbacks), and a total of '8 playbacks on hyenas
from the Mara River clan (4 cow bell playbacks and 4 church bell playbacks). Hyenas

from both clans spent more time vigilant after a sound stimulus was played than before,
regardless of the stimulus type (F 1, 38 = 10.16, p = 0.003, Figure 3.7). However,
unexpectedly, the increase in vigilance of hyenas after sound onset did not differ
significantly between church and cow bells within each clan (17134 = 0.65, p = 0.42,
Figure 3.7). Talek West hyenas were more vigilant than Mara River hyenas, regardless of

the stimulus played (F 1, 34 = 7.31, p = 0.01, Figure 3.7). Specifically, Talek West hyenas

79

 

 

——1
——1

 

 

 

 

Percent time spent vigilant when resting
0. S

 

 

 

 

 

 

 

 

 

 

 

 

0 I I l I I
None 30 to 200 201 to 500 501 to 1000 1001 to 2200
N=19 N=30 N=24 N=22 N=21

Numbers of livestock in Talek West territory

Figure 3.5. Effect of livestock presence on the percent time spent vigilance by Talek
West hyenas while resting. Means (i SE) are represented, with the number of
individuals sampled indicated below each mean. The percent time vigilant was
obtained for hyenas resting on days during which livestock presence in the territory
was known. The asterisk (*) represents a statistically significant difference (p =
0.02).

80

Percent time spent vigilant during rest

_ Tourism hours
I: Non-tourism hours

 

 

 

 

   

 

 

 

25 - 62
20 -‘
100
15 - 13 24' 10
5 .
4
0 - . .
Tourism Non-tourism Tourrsm Non-tourism
months months months months
Talek West clan Mara River clan

Figure 3.6. Effect of tourism on the percent time spent vigilant during rest by
hyenas from the Talek West and the Mara River clans during months with
many or few tourists, and during hours of the day when tour vehicles are
abroad in the Reserve (black bars) or not (white bars). Means (:t SE) are
represented, with the number of observations indicated above each bar. There
were no statistically significant effects of tourism on vigilance in either clan
(mixed-effect models, in which hyena identity was included, p > 0.05; full
model results are presented in the text).

81

Percent time spent vigilant

70-

60-

50:

40~

30-

20~

10*

 

— Pre-playback
:1 Post-playback

at:

 

 

 

 

 

 

 

 

 

 

 

 

NS
1 1
=1: :1:
1—1
:1: :1:
[fl
Church bells Cow bells Church bells Cow bells

N=6 =7 N=4 N=4
Talek West Clan Mara River Clan

Figure 3.7. Percent time spent vigilant by Talek West and Mara River hyenas during a
playback experiment involving use of either church bells (control sound) or cow bells
(treatment sound). Behavior during the pre-playback period (black bars) is compared
to that from the post-playback period (white bars). Means (:1: SE) are represented, with
the number of individuals sampled indicated below each playback type. Asterisks
represent statistically significant differences (*p < 0.05, **p < 0.01).

82

spent roughly 25 % more time vigilant than did Mara River hyenas (Figure 3.7) after
sounds were played back from either type of bells.

The only hyenas that moved after being exposed to a playback sound were Talek
West hyenas that had just heard cow bells. Two out of the 7 (28 %) Talek West hyenas
that received the cow bell treatment moved, whereas no Mara River hyenas moved afier

either sound was played to them. However, this difference between the two clans was not
statistically significant (Yates’ X2 = 0.136, p = 0.71). In the case of the two Talek West

hyenas that moved after hearing cow bells, both hyenas started to move about 1 min after
hearing the first cow bells, avoided the speakers, and spent about 22 % of their filmed
time after sound onset moving away from the speaker; one of these hyenas moved only a

few meters away, while the other moved about 200 m away.

DISCUSSION

Our results suggest that humans are having both lethal and non-lethal effects on
spotted hyenas. The former was inferred from shifts in deaths of known causes, and the
latter from behavioral data. We found that human-caused hyena mortality observed
during long-term monitoring of the disturbed clan has comprised an increasingly greater
proportion of known—caused hyena deaths since 1998. Hyenas alive today appear to be
perceiving this increased risk, and modifying their behavior accordingly. Hyenas from
the heavily disturbed Talek West clan spend about twice as much time vigilant during
naturalistic observations at rest, and nursed closer to bushes, than do hyenas from the
undisturbed Mara River clan. When we sought factors that might account for these

differences, we found that hyenas from the disturbed clan spent more time vigilant when

83

resting on days when livestock and herders were present in their territory than on the rare
days when livestock and herders were absent. We experimentally tested for an effect of
livestock presence on hyena vigilance, and we found that, although Talek West hyenas
did not increase their vigilance after hearing cow bells (the treatment sounds) any more
than after hearing church bells (the control sounds), their responses to both sounds were
stronger than those of the Mara River hyenas. To put our experimental results in a
broader perspective, Talek West hyenas spent a larger percentage of their time vigilant
after cow bell playbacks than Mara River hyenas spent in the physical presence of lions.
This suggests that Talek West hyenas showed greater responsiveness to unnatural sounds
than did hyenas living in an undisturbed area.

In our behavioral observations, we detected an effect of livestock presence on
hyena vigilance, but not an effect of tourist presence. The latter finding was encouraging
because it suggests that the Reserve’s regulations are sufficient to ensure that large
carnivores such as hyenas do not perceive tour vehicles as threats. This was not the case
in two recent studies that reported an increase in mammalian vigilance associated with
tourist presence (e. g. polar bear: Dyck and Baydack 2004; marmots: Griffin et al. 2007).
On the other hand, hyenas were more vigilant on days when livestock were present on
their territory. Herdsmen always accompany their livestock when herds graze inside the
Reserve, and pastoralists occasionally kill hyenas by spearing them. Therefore Talek
West hyenas may associate livestock presence with potential danger, and be more
vigilant in the presence of livestock than in their absence. Human-caused mortality is
quite unpredictable, as pastoralists do not consume hyena meat; instead they comer and

spear hyenas opportunistically or when hyenas attack their livestock. Behavioral changes

84

have been observed in other animals in response to anthropogenic activity, but these were
caused by more systematic poaching and hunting (Caro 1999; Brashares et al. 2001; Caro
2005b; Setsaas et al. 2007). Thus, our results suggest that sporadic killings are sufficient
to induce detectable shifts in behavior.

Although factors other than human activities can certainly affect vigilance in
hyenas, these factors appear unlikely to have caused the differences we observed between
our two clans. Lion density may affect vigilance of spotted hyenas because lions are their
main natural predators (Cooper 1991; Frank et al. 1995). However, the territories of both
clans are used by similar numbers of lion prides, and lion subgroups observed on either
territory are of similar sizes (Table 3.1; Dloniak 2006; Kolowski et al. 2007). The
availability of refugia may also affect hyena vigilance, but there are no substantial
differences in habitat structure or vegetation cover between the territories of the two clans
(Kolowski 2007). Finally, the size of a clan may affect vigilance as well, as this in turn
might affect the number of hyenas observed concurrently. The Mara River clan was
smaller at the time of our study than the Talek West clan (Table 3.1). In our naturalistic
observations, we found that hyenas in larger groups were less vigilant than in smaller
groups, so Talek West hyenas should have been less vigilant if the observed differences
were driven by clan size.

Since the late 19803, we have accumulated evidence of the importance and
magnitude of non-lethal anthropogenic effects on hyenas from the Talek West clan. In
addition to the non-lethal effects observed in this study, Talek West hyenas have been
shown to exhibit spatial avoidance of areas of their territory where the majority of the

livestock grazing takes place (Boydston et al. 2003), and they have been found to exhibit

85

temporal avoidance of interactions with pastoralists by becoming more strictly nocturnal
(Kolowski et al. 2007). Talek West hyenas also excrete higher concentrations of
glucocorticoid stress hormones than do Mara River hyenas, as with vigilance, and this is
driven by pastoralist activity rather than tourism (VanMeter et al., in review). Little is
known about how such non-lethal effects caused by humans ultimately affect fitness in
spotted hyenas. However, recent studies of natural predator-prey interactions in other
systems have shown that reductions in prey fitness caused by predator-induced non-lethal
effects can be as large as, or larger than, reductions due to lethal predator effects (Nelson
et al. 2004; Pangle et al. 2007; Creel and Christianson 2008). Animals may perceive
human disturbance as a predation risk (Frid and Dill 1990), so non-lethal anthropogenic
effects may comprise a substantial component of the net effect of humans on fitness of
animals in disturbed populations. In addition, behavioral changes detected in earlier
studies of Talek West hyenas predicted the proportional changes in mortality sources
reported here, with a time lag of about three years (Watts 2007). Our study is the first to
detect a significant increase over time in the lethal impact of humans on hyenas, most
likely because of the statistical power available with 18 years of observations. Although
we have no evidence that overall mortality rates among Talek West hyenas increased
between 1988 and 2006, our data nevertheless strongly suggest that the proportion of
total known-cause deaths attributable to humans has increased dramatically during this
l8-year period.

Information on the effects of human disturbance on threat-sensitive behavior of
mammalian carnivores is potentially useful as we attempt to conserve their populations.

Relative to other carnivores, spotted hyenas are extraordinarily flexible in their behavior

86

and ecology. Hyenas’ reactions to human disturbance may therefore represent
conservative indicators of how other mammalian carnivores exhibiting less behavioral
plasticity might respond to such challenges, including species that are rare and

endangered.

87

CHAPTER FOUR
A COMPARATIVE STUDY OF VIGILANCE IN EAST AFRICAN CARNIVORES
INTRODUCTION
Vigilance is a behavior that increases the likelihood that an animal will detect a
given stimulus at a given time (Dimond and Lazarus 1974). The stimulus in question
might be an approaching predator or an alarm signal emitted by a conspecific, but it
might also be an approaching competitor or a potential mate (reviewed in Quenette 1990;
e.g. Knight and Knight 1986; Roberts 1988; Coolen 2002; Trouilloud et al. 2004).
Vigilance is most often studied in animals engaged in foraging, as this may produce a
tradeoff between food intake and the ability to perceive danger (Lima and Dill 1990).
More recently, there has been a call for studies of vigilance exhibited by resting
individuals (Arenz 2003; Roberts 2003; Lima et al. 2005; Lima and Rattenborg 2007), as
these are rarely examined. Tradeoffs animals might experience when resting are less clear
than those experienced when foraging, because the function of (rest is still poorly
understood, but animals can nevertheless be at high risk during both activities. Vigilance
is an easily quantifiable behavior, with important effects on survivorship (e. g. F itzGibbon
1989; Krause and Godin 1996; reviewed in Caro 2005a), so vigilance is often the focus of
studies of antipredator behavior. In addition, vigilance in many mammals involves a
common set of motor patterns; this facilitates interspecific comparisons, and allows us to
gain insight into how vigilance behavior contributes to survival among species
confronting different suites of challenges in the wild.
Several general patterns have emerged from the myriad past studies of vigilance

in mammals and birds (reviewed by Quenette 1990 and Caro 2005a). First, as body size

88

increases, vigilance generally decreases; body size appears to be a good proxy for the
number of different predators that prey on a focal species (Quenette 1990), and thus the
need for vigilance. Second, vigilance is found to decrease with increasing group size; this
is often referred to as the “group-size effect” (reviewed in Elgar 1989). This can be due to
a risk dilution effect such that, when larger numbers of individuals are present, the
probability of one individual being attacked by a predator diminishes (Hamilton 1971).
The group-size effect can also be due to an increase in overall vigilance at a particular
moment in time as the number of conspecifics present increases; this is referred to as the
“many eyes hypothesis” (Pulliam 1973; Elgar 1989; Fairbanks and Dobson 2007). Most
studies documenting those general trends have focused on species that occupy low
trophic positions in food webs, such as passerine birds, rodents or ungulates (e. g. Lima
1987; Blumstein et al. 2001; Hunter and Skinner 1998). Studies of species at higher
trophic levels generally focus on their roles as predators (e. g. Mills and Shenk 1992;
Murray et al. 1995; Cooper et al. 2007). Relatively little research has been done on
vigilance behavior by predators, resulting in a lack of understanding of the mechanisms
these animals possess for coping with danger. Mammalian carnivores confront many
natural threats, including both intraguild predation and kleptoparasitism by members of
their own and other species (Palomares and Caro 1999; Caro and Stoner 2003). The few
past studies on antipredator behavior exhibited by carnivorous mammals suggest that
vigilance might be important for predators as well as for prey (Caro 1987; Rasa 1989;
Clutton-Brock et al. 1999; Switalski 2003; Di Blanco and Hirsch 2006; Hunter et al.
2007; Atwood and Gese 2008). Nevertheless, general trends in regards to vigilance

behavior among mammalian predators remain largely unexplored.

89

Here we examined vigilance in adult mammalian carnivores in the wild while
they were feeding or resting. We focused on eight sympatric species that occupy an East
African savannah habitat in the northern part of the Serengeti ecosystem. These eight
species represent four different families in the mammalian order Carnivora (see
phylogeny in Figure 4.1): Herpestidae, including dwarf mongooses (Helogale parvula)
and banded mongooses (Mungos mungos); F elidae, including cheetahs (Acinonyx
jubatus), leopards (Panthera pardus), and lions (Panthera Ieo); Hyaenidae, including
spotted hyenas (Crocuta crocuta); and Canidae, including black-backed jackals (Canis
mesomelas) and wild dogs (Lycaon pictus). These species vary considerably in body size,
in the intensity of the intra- and inter-specific feeding competition they experience, and in
their social behavior (Table 4.1). This variation allowed us to inquire whether vigilance
behavior by mammalian carnivores follows the same general patterns as those
documented in other mammals and birds. A broad, comparative approach, involving
multiple species concurrently, is rarely adopted in studies of vigilance, which usually
focus on only one or two species (but see Blumstein 1996; Boinski et a1. 2003;
Beauchamp 2007). Our goals were to identify variables that predict patterns of vigilance

in carnivores, and to compare patterns of vigilance between predators and prey.

METHODS

Study sites and study periods:
Our study was conducted in the Masai Mara National Reserve (MMNR — 1,500
kmz) in southwestern Kenya (1°40'S, 35°50'E), and in the northern portion of the Selous

Game Reserve (SGR - 43,600 kmz) in Tanzania (7° 60’ S, 38° 10’ E). The MMNR

90

88: .0 8 8860
00:85 80¢ 008008 803080 0300000800 05 8 000208 000000 580 05 :0.“ A800» m0 80:08 800808
.0000: 05 00 800805 0808030 no 008: 00:08:00 0:0 0020:0020: 0005808: 00500509»: .38 0.83:

 

 

as: .8 05:02

0 N0: 0.0m w.mm

 

 

«00005:: 00000005 00.00808 000:0m J
L

«Satan 8828*» 0000808 .8030 0d

 

 

A 0500:.380503 £80000 0.5m

 

A 050.8% 0.0885 00:00.: llll‘ NE

A03 0:05:05 :05 Ill— wd: 1

mm w.mm

 

 

 

 

 

8300.0 8000va 0:08 00085

«0803 2000005 w00 0:3 |.||

 

 

 

930800.05 0.80Uv _00—00n 000.000-0005 l. 0.x.

 

 

91

~3— :0__0:0m .0 M000: :08280 .0 :00: 00mm .: £00: :0 :0 000000 .8 Avoom 800N180_:m 0:0 20800006 000
0::000w00w 0:: 8 00800 0050 00 000: 00:09:08 80¢ :0008000 .0. ”30m 000:0:0 0:0 :05 ..a 600: 000% .: ”N03 0.8.0— .w
$00: 0:00 ¢ $.00: Kn0:0m .0 mmoon _0000 0:0 _0000 .0 m000m E::N-80_:m 0:0 2080002 .0 H3: 000: .9 ”>3: :00w8M .0

$3 :08035 80¢ :< 3......

.080 :0>_w 0:0 00 0080 :000 ¢0 b80800 000:0 85:3 0800 0:03 0:0:w 0 00 0000808 :0 .: m 00 0:0 $0830

003 0:0:w :03w 0 00 0000808 .: _ 0 008.5300 003 00:0000 0 ”080000 :000 00¢ 3:880 :0 8:08:88 w80: 0008 0::
00880000 0 3 .080 :8: 0000 08 3 0008000 00: 0: 3:808 00 80580 0:3. 00:00:00 003 0:0:w 05 00 8:00:08 0:0
:05» 000% :0>:w 0 ¢0 0000808 :0 000 0: 003 0080000 :0 300:— 30: 0.8000 0: 00880000 03 x08 000:0>:00:00 0:0. I...

.00::0>:.a 0:0 :00» 05 00 8:00 000808 :8800 05 8 :0>:w 20880 00 0:88 0:9 ......
5:0 3800 000808 0080 000. 0080 .000: 0080080 0: 0:0: 3:00:08 2:00 3:0 ...

 

 

. . 0 _d ”2 03 “202
3 ..2 N. ...2 0 m 0.: cm 3 :0 0 E E 00~ 00052000 :05
Z 0 0. .0 0.000% :0_0:0-:0:00E 83.0 ”0 E mi ”20800
0 0: 0
00 0: mm m w on 0 0 0 N. _ .o w on 030000 800000 0:08 00:00m
0000a 80808005
00 08 0 0:0:— 000
w m _ ~ N 0 _ 0 . ._ hmmd ow 00:00::88050V 880:0
:080 8 :0 5:8 0 0
- v _ _ 0 b0::0m 0 w : .o 0.0 ow 000.80 0:08:00 000000.:
mm 0: . w00
- 0 0N o 0.0. 0503 2000:.
mm m m 0: 00 0x000 0300.000 0 0.0 . q 0:3 :00_¢<
man—Ohw —Nv_ofl_.
- 0 .: N6 2 00 050005 080 .
mm N 0 0:80.: = 0 E0 0 0 0:00 0. 0.0 N . D 0000002005
00 0. . . 0000w:08
- 0 .: m3 0 m _ 00338 EMS:
_m E m 0: 00 00.000 0:00:00 .0 0.0 E 000:0m
. cm 9 . . 83:00 0000w:08
- 0 c m o m 0
cm 3 m 0: 00 00.000 0:00:00 : 0 03840.20: 0030—
w w
530“ EUQOH— «0000N_m .13 036—: 00 b=fiocw 03w“ 5:306: va—v 0552 Band OENZ
Z Z :0: _0_00m 0005:0200 3:04. .88.: 0002 :00: 008800

 

080 0:8 8 000808 00:0000 0003800 0:80 08 00 00000—08005 808 08 00 >088:m .30 030,—.

92

consists primarily of rolling grassland and scattered bushland, with riparian forest along
the major watercourses, whereas the SGR is dominated by miombo woodlands. African
wild dogs were studied in the SGR between 1991 and 1996 by Scott Creel (Montana
State University) who provided us with video tapes of wild dogs when they were feeding.
All other carnivore species included here were monitored directly in the MMNR between
1988 and 2006, mainly in 2005-2006. Although a few wild dogs still occur sympatrically
with the other MMNR carnivores, their population was decimated by a rabies epidemic in

1989 (Kat et a1. 1995), and they are seldom seen today.

Study species:

Our eight study species occur sympatrically throughout much of Eastern Africa.
Although other carnivores are present in this region, we restricted our observations to
these eight species because all of them can be observed and filmed easily. The eight focal
species have different diets, mating systems, and body masses, and they occur in social
groups that vary in size and cohesiveness. Table 4.1 summarizes salient characteristics

for each species. All subjects used here were adult individuals.

Data collection and data extraction from videotapes:

Individual members of each focal species were filmed while feeding or resting,
and behavioral measures of vigilance were later extracted from the video footage. All
data collection in the field was done between 0600 and 1900 h when there was sufficient
daylight to film. All species sampled except leopards were regularly found feeding during

daylight hours. Because we rarely witnessed leopards feeding in the MMNR during

93

daylight, we increased our sample size by including two male leopards filmed while
feeding in South Afi'ica by Phil Perry, a wildlife photographer. We considered an
individual to be feeding when it was consuming a fresh ungulate carcass (jackals, wild
dogs, cheetahs, leopards, spotted hyenas and lions) or foraging on smaller food items
(dwarf and banded mongooses, which feed on invertebrates and small vertebrates; Estes
1991) for more than five consecutive minutes. Thus, mongooses were sampled during
both search for and handling of prey, while all other carnivores were sampled only during
prey handling. We considered an individual to be resting when it was lying down for
more than ten minutes in the absence of food, and not interacting with conspecifics. Each
individual was filmed only once in all species except hyenas; we avoided re-sampling
either by knowing the identities of all focal animals (wild dogs, cheetahs, leopards,
spotted hyenas and lions), or by using footage acquired in areas that were distant from
one another in space (at least three km apart) or time (at least two years apart) for jackals
and both mongoose species. Individual adult spotted hyenas were sampled multiple
times, but animals were known individually so an average was calculated for each
individual across its different trials, and this mean was used in all analyses. Multiple
conspecifics were sometimes filmed simultaneously while feeding on a single kill
(jackals, wild dogs, cheetahs, hyenas and lions). Under these circumstances we extracted
data for all animals fully visible throughout filming. Because data points derived from
animals feeding on the same kill are not independent, we initially chose one individual at
random at each kill and used its data in analyses. However, as no result generated by

using a single individual per kill differed from that generated using all individuals

94

feeding at each kill (t-tests, all p > 0.2), we present results here based on the full dataset
to increase our sample sizes.

All filming in Kenya was done using a digital Sony camcorder (Sony DCR
HC65). We used vehicles as mobile blinds for filming all species except mongooses, and
most footage was obtained when our vehicle was parked no closer than 15 meters from
the subject animal. All subject animals were routinely watched by tourists, so they were
well habituated to cars. Mongooses were filmed from the car, as with other species, but
also in camp grounds or grounds of tourist lodges, and were well habituated to the
presence of humans on foot. Therefore, most footage of mongooses was obtained by

following foraging individuals on foot. For each individual from each species, we

extracted data from 1 to 8 min (Y = 4.04 min) of naturally occurring feeding at kills (all

species except mongooses) or while foraging 030th mongooses species). For individual

cheetahs, lions and hyenas, we also extracted data from 4 to 10 min (f = 4.64 min) of
footage taken when the focal animal was resting undisturbed. Because we filmed some
species for longer durations than others, we inquired whether vigilance behavior varied
with the length of the filmed sequence.

Videotapes were watched in slow-motion with a Sony DCR TRV900 digital
camcorder that superimposed a time-code on the screen with a precision of 0.033 s. We
identified vigilance behavior whenever a focal individual lifted its head. The onset of a
vigilance bout was considered the point at which the animal lifted its head (midway
through the raising of the head), and the end was considered the point at which the animal
lowered its head again (midway through the lowering of the head). From the video

footage, we extracted the duration of each vigilance bout, as well as the total number of

95

bouts in the filmed sequence. The latter number was used to calculate a rate of bouts per
minute filmed. A mean bout length was also calculated for each filmed sequence. From

the average duration and the rate of vigilance bouts, we calculated the total percent time
each animal spent vigilant in the filmed sequence.

We recorded several factors at the time of filming that are known to affect
vigilance in other species. In conjunction with each filmed sequence, we recorded the
number of conspecifics present within 100 m of the focal individual (referred to hereafter
as “group size”), the distance between the focal individual and the closest conspecific, the
distance between the focal animal and the car or observer on foot, and the distance to the
closest patch of bushes (using aiBushnell range finder). We also noted the time of day at
which the film sequence began, as well as the strength of the wind at the time of filming
and the height of the grass where the focal animal was filmed. Grass height was evaluated
relative to the head of the focal animal, and ranked from 0, if there was no grass, to 6, if
the grass was higher than the raised head of the focal individual.

In order to evaluate variation in vigilance behavior among species, we also
obtained other variables from the literature; all relevant sources of information are
indicated in Table 4.1. First, we obtained mean body mass of each carnivore species. In
the case of lions, we determined mass separately for adults of each sex, as this was the
only species included here that exhibited strong sexual dimorphism in body size. From
the literature, we also gathered data on the mean adult mortality rate for each species.
Third, for each species we obtained information on mean “social unit size”, defined as the
average number of individuals sharing a common territory; note that this differs from the

“group size”, as defined above as the number of conspecifics present with the focal

96

animal at the time of filming. Finally, we calculated a “cohesion index” to reflect the
degree of social cohesion characteristic of each focal species. Species with high cohesion,
those in which all members of a group are always found together, were ranked as 5 (this
included wild dogs and both species of mongooses), Whereas species rarely seen with

conspecifics except when mating (such as leopards) were ranked as 1 (see Table 4.1).

Data Analysis."

We conducted separate analyses for three measures of vigilance: duration of head
raises, rate of head raises, and percent time spent vigilant. We tested assumptions of
normality and homogeneity of variance in each measure using Wilks-Shapiro and
Levene’s Tests, respectively. Duration of head raises, rate of head raises, and mortality
rate of all species met both assumptions once they were square-root transformed. Body
mass of each species and social unit size variables also met both assumptions after being
logl O-transformed, and percent time spent vigilant for each focal animal was arcsine-
square-root transformed to meet both assumptions. Prior to analysis, we tested and found
no effect on any dependent measure of time of day, wind, or distance between subjects
and observer (multiple linear regressions, all p > 0.05). We therefore pooled our data
across these conditions to increase our statistical power. Sample sizes are shown in Table
4.1.

Interspecific comparisons:

Data from feeding individuals were analyzed for all eight species. Considering
individuals as the statistical units of analysis would be inappropriate when comparing

species, as this would inflate degrees of freedom (Harvey and Pagel 1991). Thus, each

97

species in the following analyses is represented by a single data point. Because our
analyses included comparisons among multiple species with shared evolutionary history,
the data points could not be considered to be independent of each other (Harvey and
Pagel 1991; F reckleton et a1. 2002). To determine whether phylogenetic relationships
influenced clustering in the vigilance data, we used Generalized Least Squares linear
regressions through the origin of phylogenetically independent contrasts (F elsenstein
1985; Harvey and Page] 1991; Garland and Ives 2000; Garland et a1. 1992). To carry out
those analyses, we used two Phenotypic Diversity Analysis Programs (PDAP), available
upon request from T. Garland (Garland et al. 1993, 1999). Using the phylogenetic
hypothesis for the order Carnivora produced by Bininda-Emonds et al. (1999; Figure
4.1), we first entered the phylogenetic tree into the program PDTREE. Using this tree, we
then created an across-species variance-covariance matrix with program PDDIST
(Garland et al. 1993; Garland et al. 1999; Garland and Ives 2000; Ashton 2004) and used
this matrix in the MATLAB (vR2006a) Regressionv2.m (Spoor et al. 2007) to run
Phylogenetic Generalized Least Square (PGLS) regressions. We evaluated the strength of
the various variables by comparing Akaike Information Criteria (AIC), and considering
the model with the lowest AIC value as the strongest model (Burham and Anderson
2002)

Data from resting individuals were compared to data from feeding individuals
using 2-way ANOVAs that evaluated the effects of species, activity and the interaction

between species and activity.

98

Intraspecific comparisons:

In addition to comparing vigilance among species, we examined variation within
each species separately using a multivariate approach. All species except leopards
provided a large enough sample to permit such analyses. We included group size,
distance to nearest conspecific, distance to nearest patch of bushes, and grass height as
independent variables in a multiple linear regression, and examined their relationship to
the percent time spent vigilant by feeding individuals. From this full model, we
eliminated distance to nearest conspecific, since it was a non-significant (p > 0.05)
explanatory variable for all species. For purposes of brevity, we report results only for the
percent time spent vigilant; however, results for duration and rate of head raises were
similar to those for percent time spent vigilant.

All analyses other than the phylogenetic analyses were carried out using a
significance level of alpha = 0.05 and two-tailed tests. Unless otherwise indicated, means

i: 1 standard error are presented.

RESULTS

Interspecific comparisons: Vigilance during feeding and resting

Vigilance in the carnivores observed here varied among species and between
activities. Percent time spent vigilant while feeding ranged from 3.6 % in wild dogs to 21
% in dwarf mongooses (Figure 4.2a), with a mean of 12 :t 2.16 %. The two mongoose

species and wild dogs had the shortest head raises (1.9 3), whereas cheetahs had the

99

30

 

 

 

 

 

 

 

(a)
“ 25~
E
ZED
> 20-
H
G
8.
(1)15-
O
s
E 10-
d.)
E
0 5—
0
A10—
3
(I)
.2 8«
8
'U
(H
O
.5 4-
‘0'
E
D 2-
0
8-

 

   

Rate of head raises (#/min)

  

 

 

 

 

 

 

0 . i—L-i ‘. fl . iJ—i " ;* m
DM BM JA WD LE CH SH LF LM
L

7

Increasing body mass

Figure 4.2. Vigilance while feeding in eight carnivore species: (a) mean :1: SE percent time
spent vigilant, (b) mean t SE duration of head raises, and (c) mean :1: SE rate of head raises.
Species abbreviations on the X-axis are: DM = dwarf mongooses; BM = banded
mongooses; JA = black-backed jackals; WD = wild dogs; CH = cheetahs; SH = spotted
hyenas; LF = female lions; LM = male lions. Species are presented from left to right in
order of increasing body mass.

100

longest (8.7 s), and overall, carnivores had a mean head raise duration of 3.8 :I: 0.8 s
(Figure 4.2b). The largest animals, male lions, raised their heads least oflen (0.8 head
raises / min), whereas the smallest species, dwarf mongooses, raised their heads most
often (7 head raises / min); overall, carnivores had a mean rate of 2.9 :1: 0.6 head raises /
min (Figure 4.2c). The three carnivores sampled while resting (cheetahs, spotted hyenas
and lions) spent an average of 10.5 i 2.2 % of their time vigilant while resting (range: 1.8
% to 14 % - Figure 4.3a). This vigilance level was achieved with a mean head raise
duration of 23.5 i 5.8 3 (range: 11.6 to 35.1 s — Figure 4.3b) at an average rate ofl i 0.4
head raise every 3 min (range: 0.05 to 0.58 head raises / min - Figure 4.3c).

The components of vigilance varied with the activity in which animals were
engaged when sampled, and this variation was consistent among the three species
examined, but the species x activity interaction was not significant (Figure 4.3).

Camivores sampled when feeding exhibited head raises that were shorter in duration
(F,_,42 = 30.56,p < 0.001; Figure 4.3b) and greater in frequency (FL/62 = 61 .61,p <
0.001; Figure 4.3c) than when they were resting. However, the overall percent time spent
vigilant did not differ significantly between feeding and resting (F 1, 175 = 0.001, p = 0.97,

Figure 4.3a).

Interspecific comparisons: Predictors of vigilance during feeding

Body mass was a significant predictor of rate of head raises in the carnivores
sampled (r = -O.68, p = 0.03, Table 4.2), with larger carnivores raising their heads less
frequently than smaller carnivores. Body mass was also a significant predictor of the

percent time carnivores spent vigilant (r = -0.78, p = 0.02, Table 4. 2), with larger

lOl

25

 

(a)
20 «

15*

10*

Percent time spent vigilant

 

 

 

 

 

 

 

 

(b)
40 .

30~

20*

10‘ jj—lj
, i. i,

Duration of head raises (5)

 

 

 

 

 

 

 

 

(C)

N b»)

Rate of head raises (#/min)

 

 

.7 1m

CH SH LF LM

 

 

Figure 4.3. Comparison of vigilance while feeding and resting in three carnivore
species: (a) mean :1: SE percent time spent vigilant, (b) mean :1: SE duration of head
raises, and (c) mean :t SE rate of head raises. Black bars represent data from feeding
animals and white bars represent data from resting animals. Species abbreviations on
the X-axis are: CH = cheetahs; SH = spotted hyenas; LF = lion females; LM = lion
males. Statistics were done on transformed data, but untransformed data are presented
here. Feeding means were significantly different from resting means in panels b and c
(all p < 0.001).

102

 

103

 

m2 3 .3. :5- 53 gauge
mz 3.8 23. saw as 368 8a.: .32.
...... MK. 5 :3 33— €1.52 2 35—
...... E. a :3- 3.2 38m
... w: 3o- 63 c2880
.. 6: 3o- 35 :5 Boom 8a.:
. . b_ a o .52. 2.
m2 3 : 2 o 05 ._ e 2 8:35
m2 2.: 26 £2 boom
mz 8a- ”mo- x03 8328 a
. - . - «:3. 3
m2 3. m 3 o cum :5 38m 2.2: 2.5
...... £2- :3 sum 2552 2.8.5..
...... OS. ”3- 222 room
a 9:. a
8232on mqwm

 

Amzfiow Sm $5532 83 how 3:05:00 fl game—Eu £033

5 mcommmocwoc moumscm ammo: woNzfiocoO ouocowoinm 80¢ 33%: 2: @597. 23238on qum: 5:28 2:. .c 0.8 8039a
mo 32on =< .Amod v 3 33823—8 Emu—mama 3:863 .I. new .26 v $ 822232 _meBE 3829: ... weigh;
Eoncwmméo: bfioumufim 828me m2 $20on 80388 Ems mo women—mam. ouocowoinm 2: 80¢ $38M .Né Bash

carnivores spending less time vigilant than smaller carnivores. However, body mass was
not a strong predictor of duration of head raises (Table 4.2). Mortality rate was a
significant predictor of rate of head raises (r = 0.81, p = 0.015, Table 4.2), with species
experiencing higher mortality rates raising their heads more frequently during feeding.
Mortality rate was also a significant predictor of the percent time the carnivores spent
vigilant (r = 0.81, p = 0.015, Table 4.2). Mortality rate was not a good predictor of
duration of head raises.

Mean social unit size tended to predict the duration of head raises, although this
remained marginally non-significant (r = -0.63, p = 0.09, Table 4.2): species living in
larger social units tend to have shorter head raises than those living in smaller units.
Similarly, highly cohesive species tended to have shorter head raises than less cohesive
species (r = -O.61, p = 0.11, Table 4.2). However, neither social unit size nor degree of
cohesiveness among group members significantly predicted rates of head raises, or

percent time spent vigilant.

Intraspecific analyses: Variation in vigilance within species
We did not find any factors that could explain vvithin—species variations in the
percent time spent vigilant while feeding in dwarf and banded mongooses, wild dogs,

cheetahs, and female lions. Group size was a significant factor only in spotted hyenas
(F 1,133 = 7.69, p = 0.006, Table 4.3), which spent less time vigilant as group size
increased. Distance to bushes was only a significant factor in explaining variation in
vigilance for male lions (F 1,7 = 11.22, p = 0.01, Table 4.3), with male lions spending

more time vigilant when bushes were closer. This relationship explained a large part of

104

 

 

. 6-: . Sow: . c 7:
. . no no
me a 8 .0 05 o 8.0 a SN 8 0 a a 00m mm o 8.0 a 9% a: as S
. . . 3-: . 8%-: . 9-:
mo .0 o co—
omo as as 0333 2 o 832 a; 0303:. a: .60 .0
S... 3... N3 6-: Rd 62-: 02:. cm: a: 352 38%
no.0 a 02 2 a N: ”no a 03
. . . 6-: . 82-: . 01:
m 00
:5 So 2 o 3.00:0 NZ 2:"va 02 «mean: a: 8 :0
. . . 2-: . 8-:
- - m
80 So So :33 oo 2.03% am: 8 83
. . . a: . 88-: . a:
moo ”mo 2:. 3.0:; :3 2 g 33 0:103 8:883.
. . . 2-: . 83-: . AWN-o: 2:
2 o 9. o S o 20 a mod _N o 3. a 02 3 o 23 a 8.2 380088 80.30
. . . as: . a: - a: a:
8 o S 0 mm o a; a 3 3 o 8.0 a 2: o a om 380088 830
A088: A098: A080:
& a & mm HR 0. mm HR 0. mm 0%- 9: 0300.5
.0002 h$880: 0.8-.0 A8: 00:8: 0. 008305 080 9:080

 

.80: 0:3 80¢ 00:85 0“ 008.808 :0 8288888 0: :0: 0 >7 .0080 0:880 806F808 .8 0.8003 08800—
80 80 00:80 88 003 803080 8:- 883:08 808 05 .8 :00: 05 82.: 80:8 003 80% 05 .2 o 80 000.8 08 003 0005

t o 00 008.8380 003 :80 “8880 80¢ 05 .8 :00: 05 00 030.308 0080—«>0 003 830: 0086 00nt .8 :38 000020 05 00
8880 88¢ 06. 80¢ 008308 05 008088 00:25 00 008905 .w8000¢ 00>» 820388 808 05 8083 80008 mo¢_00%800 ¢0
80888 05 008068 080 8080 .003? God v S 80058me 3808830 8000808 0.02? 20m .088 006080 808800 8000
80¢0 $888088 8 8080 .0085 £82308 .8 80088 08 98880808 .0080- 0880m 830; $08 80 £0nt 9 00830-8
.080 0:08 00 80:20.— 8 8808 0:55 80wa 80% 08:. 80800 05 006000 8000 8:85 mam—0:08 80¢ $83M .mé 033-

105

the variation in this measure among male lions (r = 0.80). Lastly, jackals were found to

be less vigilant as grass height increased (F 1,16 = 8.06, p = 0.01, Table 4.3).

DISCUSSION

The carnivores sampled in this study spent an average of 12 % of their time
vigilant while feeding, achieved by having brief head raises (3.8 s on average) at a high
rate (3 head raises/min). Those levels of vigilance are surprisingly similar to the levels
reported for species occupying lower trophic positions. For example, the average percent
time spent vigilance among multiple species of ungulates sampled in the Serengeti
ecosystem is 12 % (average compiled from Burger and Gochfeld 1994; Hunter and
Skinner 1998). Other studies have found levels of vigilance in small carnivores such as
meerkats (e. g. Moran 1984) to be similar in magnitude to those in prey species, but our
observations suggest that the need for larger carnivores to monitor their environment is
also high. The maintenance of these levels of vigilance in the carnivore species sampled
here might be explained by the high levels of competition and intra-guild predation that
occur among Serengeti carnivores (e.g. Caro and Stoner 2003). It therefore seems likely
that the vigilance of larger carnivores might serve more directly to detect conspecifics or
intraguild competitors than does vigilance of species occupying lower trophic positions.
Similar patterns of vigilance have been found in several species of primates that direct
more of their vigilance toward conspecifics than toward predators (reviewed in Treves
2000)

Although all species sampled in the current study exhibited vigilance while

feeding, we found significant differences among species. For instance, the two canid

106

species had very different levels of vigilance, with jackals spending far more time
vigilant than wild dogs. Such differences within families might be partly explained by the
ecology of each species. For example, wild dogs live in larger and more cohesive social
units than jackals, and have also much larger body masses (Table 4.1; Creel and Creel
2002), both of which may lead to wild dogs being less vulnerable to predation.
Substantial differences were also observed between the two mongoose species; however,
in this case, both dwarf mongooses and banded mongooses are small bodied and highly
gregarious. The variation here might be explained instead by differences in their
mortality rates: dwarf mongooses experience much higher mortality rates than do banded
mongooses (Table 4.1; Rood 1990; Otali and Gilchrist 2004). Lastly, cheetahs were the
most vigilant felids in our study. Cheetahs experience high levels of kleptoparasitism, and
also have a high mortality rate for their body mass (Caro 1994), which might explain
their high levels of vigilance. It is interesting to note that, overall, phylogenetic
relationships offered little help in explaining carnivore vigilance behavior.

The three carnivore species sampled while resting spent about 10 % of their time
vigilant. Resting is a necessary activity that is potentially dangerous, because animals are
reducing their attention to environmental stimuli and therefore to potential danger
(Rattenborg et al. 1999; Lima et al. 2005; Lima and Rattenborg 2007). Whereas vigilance
during rest has gained attention only recently (Lima et a1. 2005), the existing studies,
mainly conducted with birds, suggest that resting animals continue to maintain vigilance
behavior similar to that reported for feeding animals (Lendrem 1983, 1984; Gauthier-

Clerc and Tamisier 2000; Gauthier-Clerc et al. 1998, 2002; Dominguez 2003).

107

Although we found that all monitored carnivore species spent roughly the same
overall percent time engaged in vigilance during feeding and resting (means of 12 % and
10 %, respectively), these similar percentages were achieved differently. That is, when
feeding, carnivores raised their heads briefly but at a'high rate; when resting, they used
prolonged head raises that occurred infrequently. Some have compared vigilance within a
single species across several different activities (Cords 1995; Cowlishaw 1998; Hirtch
2002; Treves et a1. 2001; Kutsukake 2006), but results from all these studies differed
from those reported here in that the primates were less vigilant while foraging than while
resting. Because vigilance in these primate studies was not broken down into duration
and rate of head raises, we cannot inquire whether the two different vigilance strategies
we observed in carnivores were also used by other species.

There are several possible explanations for the different vigilance strategies we
observed during feeding and resting. One is that targets of vigilance might vary according
_ to the activity in which individuals are engaged. For example, resting animals might be
searching the environment for approaching predators, and this might require long
durations of head raises. On the other hand, feeding animals might be monitoring their
environment to detect competitors approaching to steal food, thus demanding more
frequent, shorter head raises. A second possible explanation is that, when resting, animals
are looking for distant targets, which might require long, infrequent head raises that need
not necessarily be repeated often, whereas feeding animals are looking for very close
targets, which might require short, frequent head raises. Lastly, it may be that the two
activities, resting and feeding, occur in environments of differential complexity: resting

animals here were typically lying down on open savannah, with little to monitor on a

108

short-term basis, whereas feeding animals were often monitoring a shifting social
environment containing multiple conspecifics.

In contrast to recent studies showing that animals can be vigilant while handling
food (e. g. Lima and Bednekoff 1999; Makowska and Kramer 2007), the carnivores we
studied often had restricted visual inputs because they had their heads deep in a carcass
while feeding or had their eyes closed while resting. Although this study considered only
vigilance through visual orientation, our subjects were most likely using multiple
modalities to detect potential danger (e.g. Radford and Ridley 2007). For instance, we
noticed that resting carnivores exhibited multiple ear-twitching movements even in the
absence of insects, suggesting that they are using auditory cues to monitor the
environment when resting. Similarly, olfaction is most likely a particularly important
sensory modality for danger detection among carnivores.

When considering variation in vigilance among species, the carnivores sampled
here exhibited the same general trends as those observed in other mammals and birds.
Body size is often an important determinant of vigilance in other mammals, with
vigilance increasing with decreasing body size (Quenette 1990). Our study confirms that
this relationship holds for mammalian carnivores as well, at least in regard to percent
time spent vigilant and rates of head raises. We did not find that duration of vigilance
bouts increased significantly with body mass. We found that species experiencing higher
mortality rates engaged in more vigilance by increasing the rate of head raises rather than
their duration. Even though carnivores occupy high trophic positions, they nevertheless
experience high adult mortality rates (e. g. Caro 1994). Lastly, we found sociality

measures (both social unit size and cohesiveness) to be important predictor of vigilance

109

across multiple species of carnivores, but they only affected the duration of vigilance
bouts, not the bout rate or the overall percent time spent vigilant. More gregarious species
(i.e. species that lived in larger social units or were more cohesive) engaged in shorter
vigilance bouts than did gregarious social species. This is similar to what has been found
in other mammals and birds (as reviewed by Caro 2005a).

When examining vigilance within each species, we found a group size effect only
among spotted hyenas. The presence of a group size effect has been found in the majority
of vertebrate species examined (reviewed in Elgar 1989 and Caro 2005a), so it is
surprising that only hyenas exhibited a decrease in vigilance as group size increased here.
It may be that hyenas are the only species for which we had both large sample sizes and
considerable variation in the size of the groups in which animals were feeding.

Other factors previously shown to influence vigilance in species occupying lower
trophic positions also helped explain within-species variation in the carnivores we
observed. Male lions decreased their vigilance when feeding further from bushes. Other
species, particularly birds and rodents, have been shown either to increase or decrease
their vigilance in relation to distance to cover (e. g. Lazarus and Symonds 1992; Hughes
and Ward 1993; reviewed in Caro 2005a). In the case of male lions, which defend
territories and kills against intruders (Schaller 1972), bushes might conceal immigrant
male lions that could pose a potential threat. We also observed decreasing vigilance by
jackals with increasing grass height. Other studies done on rodents have reported
vigilance changes in relation to grass height, especially in rodents species that use grass
as protective cover and decrease their vigilance when grass is high (e. g. Tchabovsky et al.

2001), similarly to what we observed among jackals.

110

Our study points to the need for more broad scale investigations of vigilance that
allow comparisons within species among activities, as well as among species, to generate
robust understanding of the ways in which animals cope with danger. There is currently
little information on vigilance behavior in carnivores due in part to the underlying
assumption that large carnivores are vigilant to locate prey rather than to detect threats
(e. g., Leyhausen 1979). However, carnivores rarely die of old age (e.g. Kruuk 1972;
Rood 1975; Bailey 1993; Caro 1994; Packer et al. 1998; Creel and Creel 2002; Otali and
Gilchrist 2004), and they regularly face intraguild predation and competition (Caro and
Stoner 2003). We found that the carnivores studied here devoted roughly the same
proportion of their time to vigilance as do ungulates, rodents and birds. Whether
vigilance in carnivores serves to locate predators or to prevent kleptoparasitism will be
difficult to assess until specific experiments are conducted (e. g. Hunter et a1. 2007).
Future studies should attempt to tease apart the relative importance of antipredator
vigilance and vigilance toward potential competitors, as in recent studies of primates (e.g.

Treves 2000; Hirsch 2002; Kutsukake 2007).

111

LITERATURE CITED

Alados, C. L. (1985). An analysis of vigilance in the Spanish ibex (Capra pyrenaica).
Zeitschriftfu'r T ierpsychologie, 68, 58-64.

Alberts, S. C. (1994). Vigilance in young baboons: effects of habitat, age, sex and
maternal rank on glance rate. Animal BehaviOur, 47, 749-755.

Arcese, P., and Sinclair, A. R. E. (1997). The role of protected areas as ecological
baselines. Journal of Wildlife Management, 61, 587-602.

Arenz, C. L., and Leger, D. W. (1997). The antipredator vigilance of adult and juvenile
thirteen-lined ground squirrels (Sciuridae: Spermophilus tridecemlineatus): visual
obstruction and simulated hawk attacks. Ethology, 103, 945-953.

Arenz, C. L., and Leger, D. W. (1999). Thirteen-lined ground squirrel (Sciuridae:
Spermophilus tridecemlineatus) antipredator vigilance decreases as vigilance cost
increases. Animal Behaviour, 57, 97-103.

Arenz, C. L., and Leger, D. W. (2000). Antipredator vigilance of juvenile and adult
thirteen-lined ground squirrels and the role of nutritional need. Animal Behaviour,
59, 535-541.

Arenz, C. L. (2003). The group size effect on vigilance: many unanswered questions.
Behavioural Processes, 63, 123- 1 24.

Ashton, K. G. (2004). Comparing phylogenetic signal in intraspecific and interspecific
body size datasets. Journal of Evolutionary Biology, 17, 1 157-1 161.

Atwood, T. C., and Gese, E. M. (2008). Coyotes and recolonizing wolves: social rank
mediates risk-conditional behaviour at ungulate carcasses. Animal Behaviour, 75,
753-762.

Avilés, J. M., and Bednekoff, P. A. (2007). How do vigilance and feeding by common
cranes Grus grus depend on age, habitat, and flock size? Journal of Avian
Biology, 38, 690-697.

Baack, J. K., and Switzer, P. V. (2000). Alarm calls affect foraging behavior in eastern
chipmunks (T amias striatus, Rodentia: Sciuridae). Ethology, 106, 1057-1066.

Bailey, T. N. (1993). The African Leopard: Ecology and Behavior of a Solitary F elid.
New York: Columbia University Press.

Beauchamp, G. (2003). Group-size effects on vigilance: a search for mechanisms.
Behavioural Processes, 63, 111-121.

112

Beauchamp, G. (2007). Exploring the role of vision in social foraging: what happens to
group size, vigilance, spacing, aggression and habitat use in birds and mammals
that forage at night? Biological Reviews, 82, 511-525.

Bednekoff, P. A., and Lima, S. L. (1998). Re-examining safety in numbers: interactions
between risk dilution and collective detection depend upon predator targeting
behaviour. Proceedings of the Royal Society in London, Series B, 265, 2021-2026.

Bednekoff, P. A., and Lima, S. L. (2005). Testing for peripheral vigilance: do birds value
what they see when not overtly vigilant? Animal Behaviour, 69, 1165-1171.

Berger, J. (1998). Future prey: some consequences of the loss and restoration of large
carnivores. In: Behavioral Ecology and Conservation Biology (Ed. by T. M.
Caro), pp. 80-100. Oxford: Oxford University Press.

Berger, J. (2000). Anthropogenic extinction of top carnivores and interspecific animal
behaviour: implications of the rapid decoupling of a web involving wolves, bears,
moose and ravens. Proceedings of the Royal Society in London, Series B, 226,
2261 -2267.

Bininda-Emonds, O. R. P., Gittleman, J. L., and Purvis, A. (1999). Building large trees by
combing phylogenetic information: a complete phylogeny of the extant Carnivora
(Mammalia). Biological Reviews, 74, 143-175.

Blumstein, D. T. (1996). How much does social group size influence golden marmot
vigilance? Behaviour, 133, 1133-1151.

Blumstein, D. T., Daniel, J. C., and Bryant, A. A. (2001). Anti-predator behavior of
Vancouver Island marmots: using congeners to evaluate abilities of a critically
endangered mammal. Ethology, 107, 1-14.

Boinski, S., Kauffman, L., Westoll, A., Stickler, C. M., Cropp, S., and Ehmke, E. (2003).
Are vigilance, risk from avian predators and group size consequences of habitat
structure? A comparison of three species of squirrel monkey (Saimiri oerstedii, S.
boliviensis, and S. sciureus). Behaviour, 140, 1421-1467.

Boland, C. R. J. (2003). An experimental test of predator detection rates using groups of
free-living emus. Ethology, 109, 209-222.

Boukhriss, J ., Slaheddine, S., Béchet, A., and Nouira, S. (2007). Vigilance in greater
flamingos wintering in southern Tunisia: age-dependent flock size effect.
Ethology, 113, 377-385.

Boydston, E. E., Kapheim, K. M., Watts, H. E., Szychman, M., and K. E. Holekamp.

(2003). Altered behaviour in spotted hyenas associated with increased human
activity. Animal Conservation, 6, 207-219.

113

Boydston, E. E., Kapheim, K. M., Van Horn, R. C., Smale, L., and Holekamp, K. E.
(2005). Sexually dimorphic patterns of space use throughout ontogeny in the
spotted hyena (Crocuta crocuta). Journal of Zoology, London, 267, 271-281.

Bradbury, J. W., and Vehrencamp, S. L. (1998). Principles of Animal Communication.
Sunderland, Massachusetts: Sinauer Associates Inc.

Brashares, J. S., Arcese, P., and Sam, M. K. (2001). Hunting demography and reserve
size predict wildlife extinction in West Afiica. Proceedings of the Royal Society
of London, Series B, 268, 2473-2478.

Burger, J., and Gochfeld, M. (1994). Vigilance in Afi'ican mammals: differences among
mothers, other females, and males. Behaviour, 131, 153-169.

Bumham, K. P., and Anderson, D. R. (1998). Model Selection and Inference: A Practical
Information-theoretic Approach. New York: Springer-Verlag.

Cameron, E. Z., and du Toit, J. T. (2005). Social influences on vigilance behaviour in
giraffes, Giraffa camelopardalis. Animal Behaviour, 69, 1337-1344.

Caro, T. M. (1987). Cheetah mothers' vigilance: looking out for prey or for predators?
Behavioral Ecology and Sociobiology, 20, 351-361.

Caro, T. M. (1994). Cheetahs of the Serengeti Plains: Group Living in an Asocial
Species. Chicago, Illinois: The University of Chicago Press.

Caro, T. M., and Durant, S. M. (1995). The importance of behavioral ecology for
conservation biology: examples fiom Serengeti carnivores. In: Serengeti 1]:
Dynamics, Management, and Conservation of an Ecosystem (Ed. by A. R. E.
Sinclair and P. Arcese), pp. 451-472. Ohio: University of Chicago Press.

Caro, T. M. (1999). Demography and behaviour of African mammals subject to
exploitation. Biological Conservation, 91, 91-97.

Caro, T. M., and Stoner, C. J. (2003). The potential for interspecific competition among
Afiican carnivores. Biological Conservation, 110, 67-75.

Caro, T. M. (2005a). Antipredator Defenses in Birds and Mammals. Chicago and
London: The University of Chicago Press.

Caro, T. M. (2005b). Behavioural indicators of exploitation. Ethology, Ecology and
Evolution, 17, 189-194.

114

Clutton-Brock, T. H., O' Riain, M. J ., Brotherton, P. N. M., Gaynor, D., Kansky, R.,
Griffin, A. S., and Manser, M. (1999). Selfish sentinels in cooperative mammals.
Science, 284, 1640-1644.

Coleman, K., and Wilson, D. S. (1998). Shyness and boldness in pumpkinseed sunfish:
individual differences are context-specific. Animal Behaviour, 56, 927-936.

Coolen, I. (2002). Increasing foraging group size increases scrounger use and reduces
searching efficiency in nutmeg mannikins (Lonchura punctulata). Behavioral
Ecology and Sociobiology, 52, 232-238.

Cooper, S. M. (1991). Optimal hunting group size: the need for lions to defend their kills
against loss to spotted hyaenas. African Journal of Ecology, 29, 130-136.

Cooper, S. M., Holekamp, K. E., and Smale, L. (1999). A seasonal feat: long-term
analysis of feeding behaviour in the spotted hyaena (Crocuta crocuta). African
Journal of Ecology, 37, 149-160.

Cooper, A. B., Pettorelli, N., and Durant, S. M. (2007). Large carnivore menus: factors
affecting hunting decisions by cheetahs in the Serengeti. Animal Behaviour, 73,
651-659.

Cords, M. (1995). Predator vigilance costs of allogrooming in wild blue monkeys.
Behaviour, 132, 559-569.

Cowlishaw, G. (1998). The role of vigilance in the survival and reproductive strategies of
desert baboons. Behaviour, 135, 431-452.

Creel, S., and Creel, N. M. (2002). The African Wild Dog: Behavior, Ecology, and
Conservation. Princeton, New Jersey: Princeton University Press.

Creel, S., Fox, J. E., Hardy, A., Sands, J ., Garrott, B., and Peterson, R. O. (2002).
Snowmobile activity and glucocorticoid stress responses in wolves and elk.
Conservation Biology, 16, 809-814.

Creel, S., and Christianson, D. (2008). Relationships between direct predation and risk
effects. Trends in Ecology and Evolution, 23, 194-201.

Di Blanco, Y., and Hirsch, B. T. (2006). Determinants of vigilance behavior in the ring-
tailed coati (Nasua nasua): the importance of within-group spatial position.
Behavioral Ecology and Sociobiology, 61, 173-182.

Dimond, S., and Lazarus, J. (1974). The problem of vigilance in animal life. Brain,
Behavior and Evolution, 9, 60-79.

115

de Ruiter, J. R. (1986). The influence of group size on predator scanning and foraging
behaviour of wedgecapped capuchin monkeys (Cebus olivaceus). Behaviour, 98,
240-258.

Devereux, C. L., Whittingham, M. J ., Fernandez-Jtu'icic, E., and Vickery, J. A. (2006).
Predator detection and avoidance by starlings under differing scenarios of
predation risk. Behavioral Ecology, 17, 303-309.

Dloniak, S. M. (2006). Annual report of research activities, 2005. Masai Mara Predator
Research Project, Masai Mara National Reserve, Kenya. Unpublished report.

Dobson, A., and Poole, J. (1998). Conspecific aggregation and conservation biology. In:
Behavioral Ecology and Conservation Biology (Ed. T. Caro), pp. 193-208. New
York: Oxford University Press.

Dominguez, J. (2003). Sleeping and vigilance in black-tailed godwit. Journal of
Ethology, 21 ,57-60.

Drea, C. M., and Frank, L. G. (2003). The social complexity of spotted hyenas. In:
Animal Social Complexity: Intelligence, Culture, and Individualized societies (Ed.
by F. B. M. de Waal and P. L. Tyack), pp. 121-148. Cambridge, Massachusetts:
Harvard University Press.

Durant, S. M. (1998). Competition refuges and coexistence: an example from Serengeti
carnivores. Journal of Animal Ecology, 67, 370-386.

Durant, S. M. (2000). Living with the enemy: avoidance of hyenas and lions by cheetahs
in the Serengeti. Behavioral Ecology, 11, 624-632.

Dyck, M. G., and Baydack, R. K. (2004). Vigilance behaviour of polar bears (Ursus
maritimus) in the context of wildlife-viewing activities at Churchill, Manitoba,
Canada. Biological Conservation, 116, 343-350.

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

East, M. L., and Hofer, H. (2001). Male spotted hyenas (Crocuta crocuta) queue for
status in social groups dominated by females. Behavioral Ecology, 12, 55 8-568.

East, M. L., and 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.

Ebensperger, L. A., Hurtado, M. J ., and Ramos-Jiliberto, R. (2006). Vigilance and

collective detection of predators in degus (Octodon degus). Ethology, 112, 879-
887.

116

Elgar, M. A. (1989). Predator vigilance and group size in mammals and birds: a critical
review of the empirical evidence. Biological Reviews, 64, 13-33.

Engh, A. L., Esch, K., Smale, L. ,and Holekamp, K. E. (2000). Mechanisms of maternal
rank 'inheritance' in the spotted hyaena, Crocuta crocuta. Animal Behaviour, 60,
232- 332.

Estes, R. D. (1991). The Behavior Guide to African Mammals, Including Hoofed
Mammals, Camivores, Primates. Berkeley and Los Angeles, California: The
University of California Press.

Fairbanks, B., and Dobson, F. S. (2007). Mechanisms of the group-size effect on
vigilance in Columbian ground squirrels: dilution versus detection. Animal
Behaviour, 73, 115-123.

Felsenstein, J. (1985). Phylogenies and the comparative method. American Naturalist,
125, 1-15.

F emandez-Juricic, E., Smith, R., and Kacelnik, A. (2005). Increasing the costs of
conspecific scanning in socially foraging starlings affects vigilance and foraging
behaviour. Animal Behaviour, 69, 73-81.

FitzGibbon, C. D. (1989). A cost to individuals with reduced vigilance in groups of
Thomson's gazelles hunted by cheetahs. Animal Behaviour, 37, 508-510.

Fischer, J., Hammerschrnidt, K. Cheney, D. L. and Seyfarth, R. M. (2002). Acoustic
features of male baboon loud calls. influences of context, age, and individuality.
Journal of Acoustical Society of America 111, 1465- 1474.

Fischer, J., Kitchen, D. M., Seyfarth, R. M., and Cheney, D. L. (2004). Baboon loud calls
advertise male quality: acoustic features and their relation to rank, age, and
exhaustion. Behavioral Ecology and Sociobiology, 56, 140-148.

Fragaszy, D. M. (1990). Sex and age differences in the organization of behavior in
wedge-capped capuchins, Cebus olicaceus. Behavioral Ecology, 1, 81-94.

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

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

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

117

Frank, L. G., Holekamp, K. E., and Smale, L. (1995). Dominance, demography, and
reproductive success of female spotted hyenas. In: Serengeti II: Dynamics,
Management, and Conservation of an Ecosystem (Ed. by A. R. Sinclair and
Arcese P.), pp. 364-384. Chicago: The University of Chicago Press.

Freckleton, R. P., Harvey, P. H., and Pagel, M. (2002). Phylogenetic analyses and
comparative data: a test and review of the evidence. American Naturalist, 160,
712-726.

Frid, A., and Dill, L. (2002). Human-caused disturbance stimuli as a form of predation
risk. Conservation Ecology 6, 11.[online] URL:
http://www.consecol.org/vol6/issl/artl 1

Garland, T. Jr., Harvey, P. H., and Ives, A. R. (1992). Procedures for the analysis of
comparative data using phylogenetically independent contrasts. Systematic
Biology, 41, 18-31.

Garland, T. Jr., Dickerman, A. W., Janis, C. M., and Jones, J. A. (1993). Phylogenetic
analysis of covariance by computer simulation. Systematic Biology, 42, 265-292.

Garland, T. Jr., Midford, P. E., and Ives, A. R. (1999). An introduction to
phylogenetically based statistical methods, with a new method for confidence
intervals on ancestral states. American Zoologist, 39, 374-388.

Garland, T. Jr., and Ives, A. R. (2000). Using the past to predict the present: confidence
intervals for regression equations in phylogenetic comparative methods. American
Naturalist, 155, 346-364.

Gauthier-Clerc, M., Tamisier, A., and Cezilly, F. (1994). Sleeping and vigilance in the
white-faced whistling duck. Wilson Bulletin, 106, 759-762.

Gauthier-Clerc, M., Tamisier, A., and Cezilly, F. (1998). Sleep-vigilance trade-off in
green-winged teals (Anas crecca crecca). Canadian Journal of Zoology, 76,
2214-2218.

Gauthier-Clerc, M., and Tamisier, A. (2000). Sleep-vigilance trade-off in gadwall during
the winter period. Condor, 102, 307-313.

Gauthier-Clerc, M., Tamisier, A., and Cezilly, F. (2002). Vigilance while sleeping in the
breeding pochard A ythya ferina according to sex and age. Bird Study, 49, 300-
303.

Gittleman, J. L. (1989). Carnivore group living: comparative trends. In: Carnivore
Behavior, Ecology, and Evolution I (Ed. J. L. Gittleman), p. 183-207. Ithica, New
York: Cornell University Press.

118

Gittleman, J. L., Funk, S. M., Macdonald, D., and Wayne, R. K. (2001). Carnivore
Conservation. Cambridge: Cambridge University Press.

Glickman, S. E., Frank, L. G., 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.

Griffin, S. C., Valois, T., Taper, M. L., and Mills, L. S. (2007). Effects of tourists on
behavior and demography of Olympic marmots. Conservation Biology, 21, 1070-
1081.

Guillemain, M., Martin, G. R., and Fritz, H. (2002). Feeding methods, visual fields and
vigilance in dabbling ducks (Anatidae). Functional Ecology, 16, 522-529.

Hamilton, W. D. (1971). Geometry for the selfish herd. Journal of Theoretical Biology,
31, 295—31 1.

Hanson, M. T., and C033, R. G. (1997). Age differences in the response of California
ground squirrels (Spermophilus beecheyi) to avian and mammalian predators.
Journal of Comparative Psychology, 111, 174-184.

Hanson, M. T., and Coss, R. G. (2001). Age differences in the response of California
ground squirrels (Spermophilus beecheyi) to conspecific alarm calls. Ethology,
107, 259-275.

Harvey, P. H., and Pagel, M. D. (1991). The Comparative Method in Evolutionary
Biology. Oxford: Oxford University Press.

Heath, B. (2007). Mara Conservancy Achievement Reports. Unpublished data available
at: www.maraconservancy.com/achievements.html

Helfman, G. S. (1989). Threat-sensitive predator avoidance in dameselfish-trumpetfish
interactions. Behavioral Ecology and Sociobiology, 24, 47-58.

Helfman, G. S, and Winkelman, D. L. (1997). Threat sensitivity in bicolor damselfish:
effects of sociality and body size. Ethology, 103, 369—383

Hegner, R. E. (1985). Dominance and anti-predator behaviour in blue tits (Parus
caeruleus). Animal Behaviour, 33, 762-768.

Hersek, M. J., and Owings, D. H. (1994). Tail flagging by young California ground
squirrels, Spermophilus beecheyi: age-specific participation in a tonic

communicative system. Animal Behaviour, 48, 803-811.

Hirsch, B. T. (2002). Social monitoring and vigilance behavior in brown capuchin
monkeys (Cebus apella). Behavioral Ecology and Sociobiology, 52, 45 8-464.

119

Hofer, H., and East, ML. (1995). Population dynamics, population size, and the
commuting system of Serengeti spotted hyenas. In: Serengeti II: Dynamics,
Management, and Conservation of an Ecosystem (Ed. by A.R.E. Sinclair and P.
Arcese), pp. 332-363. Chicago: University of Chicago Press.

Holekamp, K. E., and Smale, L. (1990). Provisioning and food sharing by lactating
spotted hyenas, C rocuta crocuta (Mammalia: Hyaenidae). Ethology, 86,191-202.

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

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

Holekamp, K. E., Smale, L., Berg, R., and Cooper, S. M. (1997). Hunting rates and
hunting success in the spotted hyena (Crocuta crocuta). Journal of Zoology,
London, 242, 1-15.

Holekamp, K. E., and Smale, L. (1998). Behavioral development in the spotted hyena.
Bioscience, 48, 997-1005 .

Holekamp, K. E., Szykman, M., Boydston, E. E., and Smale, L. (1999). Association of
seasonal reproductive patterns with changing food availability in an equatorial
carnivore, the spotted hyaena (Crocuta crocuta). Journal of Reproductive Fertility
116: 87-93.

Holekamp, K. E., Sakai, S. T., and Lundrigan, B. L. (2007). The spotted hyena (Crocuta
crocuta) as a model system for study of the evolution of intelligence. Journal of
Mammalogy, 88, 545-554.

Hollén, L. I., and Manser, M. B. (2006). Ontogeny of alarm call responses in meerkats,
Suricata suricatta: the roles of age, sex and nearby conspecifics. Animal
Behaviour, 72, 1345-1353.

Holmes, W. G. (1984). Predation risk and foraging behavior of the hoary marmot in
Alaska. Behavioral Ecology and Sociobiology, 15, 293-301.

Honer, O. P., Wachter, 8., East, M. L., and Hofer, H. (2002). The response of spotted
hyaenas to long-term changes in prey populations: functional response and
interspecific kleptoparasitism. Journal of Animal Ecology, 71, 236-246.

Hoogland, J. L., Cannon, K. E., DeBarbieri, L. M., and Manno, T. G. (2006). Selective
predation on Utah prairie dogs. American Naturalist, 168, 546-552.

120

Hughes, J. J, and Ward, D. (1993). Predation risk and distance to cover affect foraging
behaviour in Namib Desert gerbils. Animal Behaviour, 46, 1243-1245.

Hunter, L. T. B., and Skinner, J. D. (1998). Vigilance behaviour in Afiican ungulates: the
role of predation pressure. Behaviour, 135, 195-211.

Hunter, J. S., Durant, S. M., and Caro, T. M. (2007)". To flee or not to flee: predator
avoidance by cheetahs at kills. Behavioral Ecology and Sociobiology, 61, 1033-
1042.

IUCN. (2006). IUCN Red list of threatened species. <iucnredlist.org> Downloaded on
February 14 2007.

Janson, C. H., and van Schaik, C. P. (1993). Ecological risk aversion in juvenile primates:
slow and steady wins the race. In: Juvenile Primates: Lifie History, Development
and Behavior (Ed. by M. E. Pereira and L. A. Fairbanks), pp. 57-76. Oxford:
Oxford University Press.

Johnson, A., Vongkhamheng, C., Hedemark, M., and Saithongdam, T. (2006). Effects of
human-carnivore conflict on tiger (Panthera tigris) and prey populations in Lao
PDR. Animal Conservation, 9, 421-430.

Kat, P. W., Alexander, K. A., Smith, J. S., and Munson, L. (1995). Rabies and wild dogs
in Kenya. Proceedings of the Royal Society of London, Series B, 262, 229-233.

Kingdon, J. (1977). East African Mammals Volume III Part A: Camivores. Chicago,
Illinois: University of Chicago Press.

Kitchen, D. M., Seyfarth, R. M., Fischer, J ., and Cheney, D. L. (2003). Loud calls as
indicators of dominance in male baboons (Papio cynocephalus ursinus).
Behavioral Ecologz and Sociobiology, 53, 374-384.

Knight, 8., and Knight, R. L. (1986). Vigilance patterns of bald eagles feeding in groups.
Auk, 103, 263-272.

Kolowski, J. M. (2007). Anthropogenic influences on the behavior of large carnivores in
the northern Serengeti ecosystem (Doctoral dissertation). East Lansing: Michigan
State University.

Kolowski, J. M., Katan, D., Theis, K. R., and Holekamp, K. E. (2007). Daily patterns of
activity in the spotted hyena. Journal of Mammalogy, 88, 1017-1028.

Krause, J ., and Godin, J .-G. J. (1996). Influence of prey foraging posture on flight and

predation risk: predators take advantage of unwary prey. Behavioral Ecology, 7,
264-271.

121

Krause, A. P., and Ruxton, G. D. (2002). Living in Groups. Oxford: Oxford University
Press.

Kruuk, H. (1972). The Spotted Hyena: A Study of Predation and Social Behavior.
Chicago: University of Chicago Press.

Kutsukake, N. (2006). The context and quality of sOcial relationships affect vigilance
behaviour in wild chimpanzees. Ethology, 112, 581-591.

Kutsukake, N. (2007). Conspecific influences on vigilance behavior in wild chimpanzees.
International Journal of Primatology, 28, 907-918.

Lazarus, J., and Symonds, M. (1992). Contrasting effects of protective and obstructive
cover on avian vigilance. Animal Behaviour, 43, 519-521.

LeBlanc, D. C. (2004). Statistics: Concepts and Applications for Science. Sudbury: Jones
and Bartlett Publishers.

Lendrem, D. W. (1983). Sleeping and vigilance in birds. 1. Field observations of the
mallard (Anas platyrhynchos). Animal Behaviour, 31, 532-538.

Lendrem, D. W. (1984). Sleeping and vigilance in birds, 11. An experimental study of the
barbary dove (Streptopelia risoria). Animal Behaviour, 32, 243-248.

Leyhausen, P. (1979). Cat Behavior. New York, New York: Garland STPM Press.

Liley, S., and Creel, S. (2008). What best explains vigilance in elk: characteristics of
prey, predators, or the environment? Behavioral Ecology (advanced online
doi:10.1093/beheco/arm1 16).

Lima, S. L. (1987). Vigilance while feeding and its relation to the risk of predation.
Journal of Theoretical Biology, 124, 303-316.

Lima, S. L., and Dill, L. (1990). Behavioral decisions made under the risk of predation: a
review and prospectus. Canadian Journal of Zoology, 68, 619-640.

Lima, S. L. (1998). Nonlethal effects in the ecology of predator-prey interactions.
Bioscience, 48, 25-34.

Lima, S. L., and Bednekoff, P. A. (1999). Back to the basics of antipredatory vigilance:
can nonVigilant animals detect attack? Animal Behaviour, 58, 537-543.

Lima, S. L., Rattenborg, N. C., Lesku, J. A., and Amlaner, C. J. (2005). Sleeping under
the risk of predation. Animal Behaviour, 70, 723-736.

122

Lima, S. L., and Rattenborg, N. C. (2007). A behavioural shutdown can make sleeping
safer: a strategic perspective on the function of sleep. Animal Behaviour, 74, 189-
197.

Lopez, P., Hawlena, D., Polo, V., Amo, L., and Martin, J. (2005). Sources of individual
shy-bold variations in antipredator behaviour of male Iberian rock lizards. Animal
Behaviour, 69, 1-9. '

Loughry, W. J. (1993). Determinants of time allocation by adults and yearling black-
tailed prairie dogs. Behaviour, 124, 23-43.

Loughry, W. J ., and McDonough, C. M. (1999). Calling and vigilance in California
ground squirrels: age, sex and seasonal differences in responses to calls. American
Midland Naturalist, 121, 312-321.

Lung, M. A., and Childress, M. J. (2007). The influence of conspecifics and predation
risk on the vigilance of elk (Cervus elaphus) in Yellowstone National Park.
Behavioral Ecology, 18, 12-20.

MacDonald, D. W, and Sillero-Zubiri, C. (2004). The Biology and Conservation of Wild
Canids. Oxford: Oxford University Press.

Makowska, I. J ., and Kramer, D. L. (2007). Vigilance during food handling in grey
squirrels, Sciurus carolinensis. Animal Behaviour, 74, 153-158.

Martin, J ., and Lopez, P. (2003). Ontogenetic variation in antipredator behavior of Iberian
rock lizards (Lacerta monticola): effects of body-size (dependent therrnal-
exchange rates and costs of refuge use. Canadian Journal of Zoology, 81, 1131-
1137. .

Mateo, J. M. (1996). The development of alarm-call response behaviour in free-living
juvenile Belding's ground squirrels. Animal Behaviour, 52, 498-505.

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

McComb, K., Pusey, A. E., Packer, C., and Grinnell, J. (1993). Female lions can identify
potential infanticidal males from their roars. Proceedings of the Royal Society in
London, Series B, 252, 59-64.

McComb, K., Packer, C. and Pusey, A. E. (1994). Roaring and numerical assessment in

contests between groups of female lions, Panthera leo. Animal Behaviour, 47,
379-387.

123

McCowan, B., F ranceschini, N. V., and Vicino, G. A. (2001). Age differences and
developmental trends in alarm peep responses by squirrel monkeys (Saimiri
sciureus). American Journal of Primatology, 53, 19-31.

McDonough, C. M., and Loughry, W. J. (1995). Influences on vigilance in nine-banded
armadillos. Ethology, 100, 50-60.

McGregor, P. K., Catchpole, C. K., Dabelsteen, T., Falls, J. B., Fusani, L., Gerhardt, H.
C., Gilbert, F., Horn, A. G., Klump, G. M., Kroodsma, D. E., Lambrechts, M. M.,
McComb, K. E., Nelson, D. A., Pepperberg, I. M., Ratliffe, L., Searcy, W. A., and
Weary, D. M. (1992). Design of playback experiments: the Thombridge Hall Nato
ARW consensus. In: Playback and Studies of Animal Communication (Ed. by P.
K. McGregor), pp. 1-9. New York: Plenum Press.

Mills, M. G. L. (1985). Related spotted hyaenas forage together but do not cooperate in
rearing young. Nature, 316, 61-62.

Mills, M.G. L. (1990). Kalahari Hyenas: Comparative Behavioural Ecology of Two
Species. London: Unwin Hyman.

Mills, M. G., and Shenk, T. M. (1992). Predator-prey relationships: the impact of lion
predation on wildebeest and zebra populations. Journal of Animal Ecology, 61 ,
693-702.

Mills, M. G. L. ,and Hofer, H. (1998). Hyaenas. Status Survey and Conservation Action

Plan. IUCN/SSC Hyaena Specialist Group. IUCN, Gland, Switzerland and
Cambridge.

Monclus, R., Rodel, H. G., and von Holst, D. (2006). Fox odour increases vigilance in
European rabbits: a study under semi-natural conditions. Ethology, 112, 1186-
1193.

Moran, G. (1984). Vigilance behaviour and alarm calls in a captive group of meerkats,
Suricata suricatta. Zeitschriftfuer T ierpsychologie, 65, 228-240.

Murray, D. L., Boutin, S., O'Donoghue, M., and Nams, V. O. (1995). Hunting behaviour
of a sympatric felid and canid in relation to vegetative cover. Animal Behaviour,
50, 1203-1210.

Nelson, E. H., Matthews, C. E., and Rosenheim, J. A. (2004). Predators reduce prey
populations growth by inducing changes in prey behavior. Ecology, 85, 1853-
1 85 8.

Nevin, O. T., and Gilbert, B. K. (2005). Measuring the cost of risk avoidance in brown

bears: further evidence of positive impacts of ecotourism. Biological
Conservation, 123, 453-460.

124

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

Ogutu, J. O., Bhola, N., and Reid, R. (2005). The effects of pastoralism and protection on
the density and distribution of carnivores and their prey in the Mara ecosystem of
Kenya. Journal of Zoology London, 265, 281-293.

Orrock, J. L., Danielson, B. J ., and Brinkerhoff, R. J. (2004). Rodent foraging is affected
by indirect, but not by direct, cues of predation risk. Behavioral Ecology, 15, 433-
437.

Otali, E., and Gilchrist, J. S. (2004). The effects of refuse feeding on body condition,
reproduction, and survival of banded mongooses. Journal of Mammalogy, 85,
491-497.

Packer, C., Herbst, L., Pusey, A. E, Bygott, J. D., Hanby, J. P, Cairns, S. J, and Mulder,
M. B. (1988). Reproductive success of lions. In: Reproductive Success: Studies of
Individual Variation in Contrasting Breeding Systems (Ed. by T. H. Clutton-
Brock), pp. 363-383. Chicago and London: University of Chicago Press.

Pangle, K. L., Peacor, S. D., and Johannsson, O. R. (2007). Large nonlethal effects of an
invasive invertebrate predator on zooplankton population growth rate. Ecology,
88, 402-412.

Palomares, F., and Caro, T. M. (1999). Interspecific killing among mammalian
carnivores. American Naturalist, 153, 492-508.

Pays, O., and J arman, P. J. (2008). Does sex affect both individual and collective
vigilance in social mammalian herbivores: the case of the eastern gray kangaroo?
Behavioral Ecology and Sociobiology, 62, 757-767.

Pinheiro, J. C., and Bates, D. M. (2000). Mixed-ejfects Models in S and S—PL US.
Springer-Verlag, New York.

Pinheiro, J. C., Bates, D. M., DebRoy, S., and Sarkar, D. (2005). nlme: linear and
nonlinear mixed effects models. R package version 3.1-66.

Pferfferle, D., West, P. M., Grinnell, J ., Packer, C., and Fischer, J. (2007). Do acoustic
features of lion, Panthera leo, roars reflect sex and male condition? Journal of
Acoustical Society of America, 121, 3947-3953.

Platzen, D., and Magrath, R. D. (2004). Parental alarm calls suppress nestling

vocalization. Proceedings of the Royal Society in London, Series B, 271, 1271-
1276.

125

Pongracz, P., and Altbttcker, V. (2000). Ontogeny of the responses of European rabbits
(Oryctolagus cuniculus) to aerial and ground predators. Canadian Journal of
Zoology, 78, 655-665.

Preissier, E.L., Bolnick, D. 1., and Benard, M. F. (2005). Scared to death? The effects of
intimidation and consumption in predator-prey interactions. Ecology, 86, 501-509.

Pulliam, H. R. (1973). On the advantages of flocking. Journal of Theoretical Biology, 38,
419—422.

Quenette, P.-Y. (1990). Functions of vigilance behaviour in mammals: a review. Acta
0ecologica, 1 1, 801—818.

Quinn, J. F., and Cresswell, W. (2005). Personality, anti-predation behaviour and
behavioural plasticity in the chaffmch F ringilla coelebs. Behaviour, 142, 1377-
1402.

Radford, A. N, and Ridley, A. R. (2007). Individuals in foraging groups may use vocal
cues when assessing their need for anti-predator vigilance. Biology Letters, 3,
249-252.

Ramakrishnan, U., and Coss, R. G. (2000). Age differences in the responses to adult and
juvenile alarm calls by bonnet macaques (Macaca radiata). Ethology, 106, 131-
144.

Rasa, O. A. (1989). Behavioural parameters of vigilance in the dwarf mongoose: social
acquisition of a sex-biased role. Behaviour, 110, 125-145.

Rattenborg, N. C, Lima, S. L, and Amlaner, C. J. (1999). Half-awake to the risk of
predation. Nature, 397, 397-398.

R Development Core Team. (2005). R: A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-
900051-07-0, URL http://wwwR-proiectog

Re’ale, D., Reader, S. M., Sol, D., McDougall, P. T., and Dingemanse, N. J. (2007).
Integrating animal temperament within ecology and evolution. Biological
Reviews, 82, 291-318.

Reinhardt, U. G. (1999). Predation risk breaks size-dependent dominance in juvenile
coho salmon (Oncorhynchus kisutch) and provides growth opportunities for risk-
prone individuals. Canadian Journal of Fisheries and Aquatic Sciences, 56, 1206-
1212.

126

Roberts, S. C. (1988). Social influences on vigilance in rabbits. Animal Behaviour, 36,
905-913.

Roberts, G. (2003). The group-size effect in non-feeding animals. Behavioural Processes,
63, 127-128.

Rood, J. P. (1975). Population dynamics and food habitats of the banded mongoose. East
African Wildlife Journal, 13, 89-111.

Rood, J. P. (1990). Group size, survival, reproduction, and routes to breeding in dwarf
mongooses. Animal Behaviour, 39, 566-572.

Sargeant, G. A., Johnson, D. H., and Berg, W. E. (1998). Interpreting carnivore scent-
station surveys. Journal of Wildlife Management, 62, 1235-1245.

Schaller, G. B. (1972). The Serengeti Lion: A Study of Predator-prey Relations. Chicago
and London: University of Chicago Press.

Setsaas, T. H., Holmem, T., Mwakalebe, G., Stokke, S., and Roskaft, E. (2007). How
does human exploitation affect impala populations in protected and partially

protected areas? A case study from the Serengeti Ecosystem, Tanzania. Biological
Conservation, 136, 563-570.

Seyfarth, R. M., Cheney, D. L., and Marler, P. (1980). Monkey responses to three
different alarm calls: evidence of predator classification and semantic
communication. Science, 210, 801-803.

Seyfarth, R. M., and Cheney, D. L. (1986). Vocal development in vervet monkeys.
Animal Behaviour, 34, 1640-165 8.

Shriner, W. M. (1998).Yellow-bellied marmot and golden-mantled ground squirrel
responses to heterospecific alarm calls. Animal Behaviour, 55, 529-536.

Siegel, S., and Castellan, N. J. J. (1988). Nonparametric Statistics for the Behavioral
Sciences. McGraw-Hill, Inc.

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

Sinclair, A. R., and Norton-Griffiths, M. (1979). Serengeti: Dynamics of an Ecosystem.
University of Chicago Press, Chicago and London.

Slotow, R., and Paxinos, E. (1997). Intraspecific competition influences food retum-
predation risk trade-off by white-crowned sparrows. Condor, 99, 642-650.

127

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

Smale, L., Nunes, S., and Holekamp, K. E. (1997). Sexually dimorphic dispersal in
mammals: patterns, causes, and consequences. Advances in the Study of Behavior,
26, 181-250. '

Smith, J. E., Kolowski, J. M., Graham, K. E., Dawes, S. E., and Holekamp, K. E. (2008).
Social and ecological determinants of fission-fusion dynamics in the spotted
hyaena. Animal Behaviour. In press.

Spoor, F., Garland, T. Jr., Krovitz, G., Ryan, T. M, Zilcox, M. T, and Walker, A. (2007).
The primate semicircular canal system and locomotion. Proceedings of the
National Academy of Sciences, 104, 10808-10812.

Stone, A. I. (2007). Age and seasonal effects on predator-sensitive foraging in squirrel
monkeys (Saimiri sciureus): a field experiment. American Journal of
Primatology, 69, 127-141

Switalski, TA. (2003). Coyote foraging ecology and vigilance in response to gray wolf
reintroduction in Yellowstone National Park. Canadian Journal of Zoology, 81 ,
985-993.

Tchabovsky, A. V., Krasnov, B. R., Khokhlova, I. S., and Shenbrot, G. I. (2001). The
effect of vegetation cover on vigilance and foraging tactics in the fat sand rat
Psammomys obesus. Journal of Ethology, 19, 105-113.

Thomas, L. (1997). Retrospective power analysis. Conservation Biology, 11, 276-280.

Tilson, R. T., and Hamilton, W. J. 111. ( 1984). Social dominance and feeding patterns of
spotted hyaenas. Animal Behaviour, 32, 715-724.

Treves, A. (2000). Theory and method in studies of vigilance and aggregation. Animal
Behaviour, 60, 711-722.

Treves, A., Drescher, A., and Ingrisano, N. (2001). Vigilance and aggregation in black
howler monkeys (Alouatta pigra). Behavioral Ecology and Sociobiology, 50, 90-
95.

Trinkel, M., and Kastberger, G. (2005). Competitive interactions between spotted hyenas

and lions in the Etosha National Park, Namibia. African Journal of Ecology, 43,
220-224.

128

Trouilloud, W., Delisle, A., and Kramer, D. L. (2004). Head raising during foraging and
pausing during intermittent locomotion as components of antipredator vigilance in
chipmunks. Animal Behaviour, 67, 789-797.

Van Meter, P. E., French, J. A., Dloniak, S. M., Watts, H. E., and Holekamp, H. E. (In
review). Fecal glucocorticoids reflect socio-ecological and anthropogenic
stressors in the lives of wild spotted hyena. Hormones and Behavior.

van Oers, K., Klunder, M., and Drent, P. J. (2005). Context dependence of personalities:
risk-taking behavior in a social and a nonsocial situation. Behavioral Ecology, 16,
716-723.

Waite, T. A. (1987). Vigilance in the white-breasted nuthatch: effects of dominance and
sociality. The Auk, 104, 429-434.

Watts, H. E. (2007). Social and ecological influences on survival and reproduction in the
spotted hyena, Crocuta crocuta (PhD dissertation). East Lansing: Michigan State
University; 166 p.

Watts, H. E., and Holekamp, H. E. (2008). Interspecific competition influences
reproduction in spotted hyenas. Journal of Zoology, London. In press.

White, P. A. (2005). Maternal rank is not correlated with cub survival in the spotted
hyena, Crocuta crocuta. Behavioral Ecology, 16, 606-613.

Woodroffe, R. (2000). Predators and people: using human densities to interpret declines
of large carnivores. Animal Conservation, 3, 165-173.

Woodroffe, R. (2001). Strategies for carnivore conservation: lessons from contemporary
extinctions. In: Carnivore Conservation (Ed. by J. L. Gittleman, R. K. Wayne, D.
W. Macdonald, and S. M. Funk), pp. 61-92. Cambridge: Cambridge University

Press.

Woodroffe, R., and Ginsberg, J. R. (1998). Edge effects and the extinction of populations
inside protected areas. Science, 280, 2126-2128.

Woodroffe, R., and Ginsberg, J. R. (2000). Ranging behaviour and vulnerability to
extinction in carnivores. In: Behaviour and Conservation (Ed. by L. M. Gosling
and W. J. Sutherland), pp. 125-140. Cambridge: Cambridge University Press.

Zar, J. H. (1999). Biostatistical Analysis. 4th edn. U er Saddle River: Prentice Hall.
PP

129

111”‘”'“

SlT‘t

STATE UNI‘IER

MICHIGAN

1W

1’

Hill 111111111

1

l

6
1
1.
7
6
5
9
2
o
3
9
2
1

3

 

 

 

 

 

 

. ...u u.