.13.. .
. a $me .€
2,

.fi

5.. ‘1‘." 1-".1 V
2.3-1 . .5159? Vi.
Q ANNéAI
,5
A.
huh?“-
1

-
.oh§§i

E.
u. .
sunflmflhi; ,
{Lari 1‘
k .1.“ b3 .
1“... 1.. 5 h. .
. “hrs. .v2fl.i1.«r. JEN-fix. ‘3“.
1 ammna

\

3;
3%.»: . 3.

..
«.aovlé

 

> v, u . t A n a on .5353: , :21? :1? .v:
arrgfimfifig ‘ gang . 7?:

. 1.2.145-1 IxV
. y. . : ‘ . .2115.

2009

This is to certify that the
dissertation entitled

DETERMINANTS OF AGE AT DISPERSAL AND
SETTLEMENT PATTERN IN THE BELDING’S GROUND
SQUIRREL, SPERMOPHILUS BELDINGI

presented by

Eva-Maria Muecke

has been accepted towards fulfillment
of the requirements for the

 

 

PhD degree in Zoology

gum; W

Major Profe§§07’s ngfiature

December 15, 2005

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.1

DETERMINANTS OF AGE AT DISPERSAL AND SETTLMENT PATTERNS IN
THE BELDING’S GROUND SQUIRREL, SPERMOPHILUS BELDINGI

BY

Eva-Maria Muecke

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Zoology
Program in Ecology. Evolutionary Biology and Behavior

2005

ABSTRACT

DETERMINANTS OF AGE AT DISPERSAL AND SETTLMENT PATTERNS IN
THE BELDING’S GROUND SQUIRREL, SPERMOPHILUS BELDINGI

BY

Eva-Maria Muecke

Mammalian dispersal often occurs before individuals reach sexual
maturity, when they are smaller and less experienced than post-pubertal animals,
and thus more vulnerable to dispersal risks. Predominant early dispersal among
mammals suggests that dispersers gain a selective advantage by emigrating at a
young age rather than delaying emigration until they are older. For example,
early dispersal may allow emigrants to reduce the social resistance they
encounter from local residents during the immigration process. The Belding’s
ground squirrel, Spennophilus beldingi, exhibits a dichotomous dispersal pattern:
most males disperse at 7 - 10 weeks of age, as juveniles, but some males delay
emigration until they are 55 weeks old, as yearlings. I was therefore able to use
this species to assess whether young immigrants experience less social
resistance from resident adults than do older immigrants. During staged
encounters, female residents directed lower rates and intensities of aggression
towards juvenile than yearling males. During natural encounters, resident
females were less likely to aggress against juvenile than yearling males during
an encounter. These results suggest that early dispersal is adaptive for male 8.
beldingi by allowing juvenile immigrants to avoid aggression from female

residents more effectively than can males who disperse as yeanings.

Predominant early dispersal among mammals also suggests its risks may
be offset by significant fitness gains. I used long-term trapping data and
behavioral observations of S. beldingi males during the breeding season to
assess the consequences of dispersal age on male survival and mating success.
I found that age-specific survivorship was only lower among early than late
dispersers to age one, but early dispersers mated with a greater number of
different females than did late dispersers. These data illustrate that early
dispersal carries a small cost to short-term survival, but ultimately enhances
mating success for surviving males.

Because natal dispersal may be dangerous, natural selection should
shape this behavior to occur in each species at an age when emigrants are best
able to cope with its risks and challenges. I examined the role of life history,
social, and ecological factors in shaping the evolution of dispersal age in ground-
dwelling sciurid rodents. Data from 22 species were extracted from the literature,
and analyzed after computing phylogenetic independent contrasts to control for
non-independence of data points due to common ancestry. I found that the
amount of growth males were able to complete during their juvenile summer had
a significant negative effect on the age at which males dispersed. I also observed
a positive relationship between the level of social resistance immigrants were
likely to encounter and a species” dispersal age. However, I found no significant
relationship between mean dispersal distance, a measure of dispersal risk, and

the age at which a particular species generally disperses.

Copyright by
EVA-MARIA MUECKE

2005

ACKNOWLEDGEMENTS

Many people helped me throughout my graduate education, and I
would like to especially acknowledge my advisor, Dr. Kay Holekamp, and my
long-term collaborator, Dr. Scott Nunes. I have been fortunate to have such
an excellent scientist and advisor as Kay, and I appreciate all her help and
support throughout my graduate studies. Scott showed me the ropes in the
field and shared his data and other resources with me. He has been an
outstanding collaborator and friend. I also appreciate the help and guidance
of my committee members, Drs. Tom Getty, Antonio Nunez, and Laura
Smale. Chapter four could have not been done without the help of Dr. Anna
Monfils, and I would also like to thank Dr. Jim Smith for his advice on this
work.

The College of Natural Sciences, the Graduate School, the
Department of Zoology, as well as the program in Ecology, Evolutionary
Biology and Behavior, have provided me with funding and logistical support.
The Society of Mammalogists supported me with a grant-in-aid of research.
My research was also financed by fellowships to K.E.H. from the David and
Lucille Packard Foundation and the Searle Scholars Program/Chicago
Community trust. I am grateful to the Clairol Corp., a division of Proctor and
Gamble, for its generous donation of hair dye.

Students from Michigan State University and the University of San
Francisco have been instrumental in the collection of field data. I would like to

thank them and especially acknowledge the efforts of Pamela Bartholomew,

Co-Dhiem Ha, Rebecca Hoffmeier, Lesley Lancaster, Heather Ross, Keron
Greene, and Jennifer Muehlhouse. I would also like to thank Larry Ford from
the National Forest Service for his cooperation and Kerry Kellogg for logistical
support.

The EEBB community provided me with an important intellectual and
social outlet, and my time at MSU would not have been the same without the
many interesting and wonderful people who are part of this community. I
would also like to thank past and current lab mates including Sarah Benson-
Amram, Pat Bills, Erin-Boydston, Stephanie Dloniak, Anne Engh, Keron
Greene, Karen Hubbard, Joe Kolowski, Suzanne La Croix, Sarah Lansing,
Scott Nunes. Jenn Smith, Micaela Syzkman, Jaime Tanner, Kevin Theis,
Wiline Trouilloud, Russ Van Horn, Page Van Meter, Heather Watts, and Sofia
Wahaj for all their help.

Many individuals have supported me during the past years, and there
are a few people I would like to especially recognize. Russ van Horn, Cindy
Wei, and Mary Martin have been a consistent presence throughout my time at
MSU and have supported me throughout the ups and downs in my personal
life and graduate school. I also would like to thank Erin Boydston, Stacy Epp,
Vesna Gavrilovic, Elizabeth Ostrowski, Toni van Horn, and Sofia Wahaj for
their friendship. Karl Strause helped me get through the last few months of
my degree program, and I don’t know how I would have done it without him.
Markus and Uli Hecker, Amy lezzoni, and Brad and Jenny Upham were

always willing to arrange play-dates when I needed extra time for work. I

vi

would like to thank my family for all their love and support, especially my
mother, Helene Muecke. She was willing to travel thousands of miles and live
in a trailer for weeks on end at the age of 70! She is remarkable, and I am
proud to be her daughter. Lastly, I would like to thank my daughter,
Alexandra, for being, literally, a good camper and being willing to tag along

where I needed to go. I am blessed to have such a wonderful child.

vii

TABLE OF CONTENTS
LIST OF TABLESIx
LIST OF FIGURESx

CHAPTER1
GENERAL INTRODUCTION1
Overview of chapters7

CHAPTER 2
EARLY DISPERSAL REDUCES RESIDENT AGGRESSION DIRECTED
TOWARDS IMMIGRANT MALE BELDING’S GROUND SQUIRRELS
(SPERMOPHILUS BELDINGI)
Introduction... 11
Methods14
ResultsZS
Drscussron32

CHAPTER 3

EARLY DISPERSAL DOES NOT IMPROVE SURVIVORSHIP BUT IS

CORRELATED WITH ENHANCED MATING SUCCESS AMONG MALE

BELDING’ S GROUND SQUIRRELS (SPERMOPHILUS BELDINGI)
Introduction... ......... .. . . . . ...........48
MethodsSO
Result355
Discussron57

CHAPTER4

GROWTH LIMITATIONS AND AVOIDANCE OF RESIDENT AGGRESSION

INFLUENCE AGE AT DISPERSAL IN GROUND- DWELLING SCIURIDS
Introduction” . . ...69
Methods ...74
Results81
Drscussron83

CHAPTER 5
CONCLUSION .................................................................................. 92

APPENDIX ................................................................................... 100

LITERATURE CITED106

viii

LIST OF TABLES

Table 3.1. Age-specific life table for early and late dispersing male 8. beldingi,

based on disappearance from the population of members of cohorts born in
1993 -

Table 4.1. Proposed direction of relationship between predictor variables and
dispersal age based on three different hypotheses ................................ 89

Table 4.2. Results of simple bivariate regression analyses of factors proposed
to influence dispersal age in ground-dwelling sciurids ............................. 90

Table 4.3. Results of multiple regression analyses of factors proposed to
influence dispersal age in ground-dwelling sciurids ................................ 91

ix

LIST OF FIGURES

Figure 2.1: Percent of 10-min trials during which experimental females
aggressed against tethered stimulus animals. All trials shown here took place
during the lactation period. Sample sizes above bars refer to the total number
of trials conducted for each stimulus animal group. Horizontal bar with
asterisk indicates that groups differ significantly at P < 0.05 .................... 41

Figure 2.2: Mean rates of threat (a), rates of fighting behavior (b), and
aggression intensity scores (c) directed by experimental females towards
juvenile female (n = 7), juvenile male (n = 13), yearling female (n = 7), and
yearling male (n = 14) stimulus animals. All trials shown here took place
during the lactation period. Asterisk indicates significant differences between
groups at P<0.05 and double asterisks indicate P < 0.01 ....................... 42

Figure 2.3: Mean rates of threat (a), rates of fighting behavior (b), and
aggression intensity scores (0) directed by experimental females towards
tethered juvenile males (n = 13) and yearling males (n = 14) during lactation
trials and towards juvenile males (n=6) and yearling males (n = 6) during
post-weaning trials. Other notation is as in figure 2 ............................... 43

Figure 2.4: (a) Rates of naturally-occurring encounters between yearling focal
males and adult male and female residents during the breeding season (n=9)
and the post-weaning period (n = 7), and between adult residents and juvenile
focal males (n = 11) during the post-weaning period. (b) Rates of encounters
with adult resident by yearling males that had dispersed either as juveniles
(early dispersers, n = 5) or as yearlings (late dispersers, n = 5). Data in (a) but
not (b) are normally distributed, so data shown in a & b cannot be directly
compared ....................................................................................... 44

Figure 2.5: Aggressive behavior directed by female (a) and male (b) residents
towards juvenile (n = 11) and yearling (n = 10) male focal animals during
naturally-occurring encounters. Sample sizes above bars refer to the total
number of encounters observed for each focal male subject group. Other
notation is as in Figure 2. .................................................................. 45

Figure 2.6: Aggressive behavior directed by adult male and female residents
towards yearling focal males that had dispersed either as juveniles (early
dispersers, n = 5) or as yearlings (late dispersers, n = 5). Sample sizes above
the bars refer to the total number of naturally-occurring encounters we
observed involving males from each yearling group. Other notation is as in
Figure 2 ......................................................................................... 46

Figure 2.7: (a) Body mass of juvenile males (n = 62) at the time of dispersal,
juvenile stimulus males (n = 19) on the day of the experimental trial, and

juvenile focal males (n = 10) during the post-weaning period. (b) Body mass
of yearling males (n = 18) at the time of dispersal, yearling stimulus males (n
= 20) on the day of the trial, and yearling focal males that had dispersed
either as juveniles (early dispersers, n = 5) or as yearlings (late dispersers, n
= 5). Body mass of focal males weighed more than once per observation
period was averaged for this comparison. Triple asterisks indicate significant
differences at P < 0.001. Other notation is as in Figure 2 .......................... 47

Figure 3.1. Summary of sample sizes obtained from monitoring male survival
and dispersal behavior of 1993 - 1997 cohorts on Tioga pass meadow, CA.
Out of 433 tagged juvenile males, we categorized 230 individuals as early
dispersers and 90 individuals as late dispersers (see Methods: Dispersal age
assignment) .................................................................................... 63

Figure 3.2. Age-specific survivorship of early- and late—dispersing males.
Survivorship is based on 320 males, 230 early and 90 late dispersers, from
cohorts born in 1993 - 1997 ............................................................... 64

Figure 3.3. Percent of early and late dispersers surviving their first summer (to
age one) and to sexual maturity (age two). Data are based on 320 males, 230
early and 90 late dispersers, fromi cohorts born in 1993 - 1997. Asterisks

indicate significant differences between groups at p < 0.001 ..................... 65

Figure 3.4. Observed and expected mating partners of early- and late-
dispersing males. Expected values were calculated by assuming that early
and late dispersers mated with females in direct proportion to the number of
males in each dispersal age category observed mating ............................ 66

Figure 3.5. Mean number of mating partners (1 SE) acquired by early and late
dispersing males during the 1998 mating season. Asterisk indicates
significant differences between groups at p < 0.05 .................................. 67

Figure 4.1: Phylogenetic tree of 22 ground-dwelling sciurid species used in
the current study for comparative analyses. Numbers above branches
indicate branch lengths, which were used as expected units of evolutionary
change between speciation events for the Brownian motion model of
character evolution .......................................................................... 88

xi

CHAPTER ONE
GENERAL INTRODUCTION

Natal dispersal, the permanent departure of an individual from its
birthplace, occurs in most birds and mammals (Greenwood, 1980; Lidicker,
1975). This behavior represents a key event in the life history of an individual and
has critical effects on the persistence of populations and species (Colbert et al.,
2001; Greenwood, 1980). For an individual, dispersal minimizes inbreeding and
increases reproductive opportunities (Dobson, 1982; Lehmann and Perrin, 2003;
Moore and Ali, 1984; Pusey, 1987; Shields, 1982). The influx of dispersers also
enhances the viability of a population and reduces the chance of local extinction
(Hanski, 1999). Even if local extinction occurs, dispersers are frequently able to
re-colonize previously occupied areas (Roach et al., 2001 ).

Natal dispersal tends to be sex-biased in that one sex is usually more
likely to disperse or disperses over greater distances than the opposite sex
(Dobson, 1982; Greenwood, 1980). In birds, females, and in mammals males,
tend to be the predominant dispersing sex. Furthermore, the age at which
dispersal occurs tends to be highly predictable within a species. Despite the
importance of dispersal, our knowledge of this behavior is still limited, and very
little is known about the factors that determine the age at~ which dispersal occurs
among most animal species. The objective of my dissertation is to fill this gap by
examining what variables determine age at dispersal in the Belding’s ground

squirrel, Spennophilus beldingi, and other ground dwelling sciurids.

Dispersal is a critical event in the life of an animal since it represents one
of the most profound environmental changes an individual is likely to experience
during its lifetime. A disperser leaves the safety of a familiar social and physical
environment and must navigate through unfamiliar habitat where it is exposed to
many mortality factors such as predation and lack of essential resources. After
the individual has located a suitable settlement site, the emigrant must rapidly
gather and assimilate knowledge of the social and physical features of its new
home to ensure successful settlement (lsbell and Van Vuren, 1996; Van Vuren
and Armitage, 1994; Waser and Jones, 1983a). The costs of dispersal can be
substantial and may include exclusion from important resources (Garrett and
Franklin, 1988), loss of reproductive opportunities (Alberts and Altmann, 1995),
injury, and even death (Byrom and Krebs, 1999). Considering its potential costs,
an individual’s decision on when to embark on this journey should have
significant consequences for a disperser. Although our knowledge of dispersal
behavior is still limited, a number of studies have succeeded in identified factors
that mediate dispersal behavior (e.g., Holekamp, 1986); see also review in Dufty
and Belthoff, 2001). These studies have also provided insights into the variables
that might determine age of dispersal in birds and mammals.

The role of body condition in mediating natal dispersal has been
documented in a number of avian and mammalian species. In the marsh tit,
individuals with better body composition disperse at younger ages than do
individuals in poorer condition (Nilsson and Smith, 1985). Larger individuals are

more likely to emigrate than smaller individuals in the greater flamingo (Barbraud

et al., 2003). Dominance is tightly coupled with body condition in screech owls,
and dominant individuals disperse at younger ages than subordinates (Ellsworth
and Belthoff, 1999). An association between body mass, body condition, and
dispersal behavior has also been observed in a number of mammalian studies
(see review by Holekamp, 1984a). For example, in the hibernating Belding’s
ground squirrel body mass and body fat regulates dispersal behavior of males.
Males that are able to reach a specific body mass/body fat threshold early during
their first active season disperse prior to their first hibernation (Nunes et al.,
1998; Nunes and Holekamp, 1996). However, males that are unable to reach this
fat threshold until later in the season postpone dispersal until after their first
hibernation. Overall, these studies suggest that body condition determines
dispersal age such that individuals with adequate energy resources disperse at
younger ages than do individuals in poor condition.

Considerable evidence suggests that endocrine secretions regulate
dispersal behavior by organizing and activating the neural circuits responsible for
the expression of this behavior. Organizational effects of gonadal steroids on
sexual differentiation of male or female behavior have been demonstrated in
birds and mammals (e.g., Adkins-Regan, 1987; Phoenix et al., 1959). Since
dispersal tends to be sexual dimorphic, gonadal steroids might therefore be
expected to play a similar role in organizing dispersal behavior in these taxa
(Holekamp and Sherman, 1989). This has shown to be the case in the grey-sided
vole, Clethn'onomys rufocanus, where females exposed to high levels of

testosterone in utero are more likely to disperse than females with low exposure

to testosterone (lms, 1989). In Belding’s ground squirrels, females treated
experimentally with testosterone exhibit the male-typical dispersal pattern,
supporting the hypothesis that testosterone organizes dispersal behavior in this
species (Holekamp et al., 1984; Nunes et al., 1999).

The correlation between sexual maturation, increases in circulating
testosterone, and the onset of dispersal suggest that testosterone may be
involved in activating this behavior in some mammalian species (reviewed in
Dufty and Belthoff 2001; but see (Holekamp and Smale. 1998). In many avian
species, however, dispersal occurs once gonads have regressed, and male
mammals often disperse before reaching puberty (Dobson, 1982). Consequently,
it is unlikely that gonadal steroid hormones have an activational affect on
dispersal behavior in a large number of avian and mammalian species. Instead,
hormones that function as primary signals in energy balance may trigger this
behavior. For example, the hormone leptin is produced by fat cells, and high
leptin titers in the blood indicate large fat stores (Caro et al., 1996). Leptin may
therefore trigger dispersal behavior in species, such as the Belding’s ground
squirrel, where a fat threshold determines the onset of dispersal behavior (Nunes
and Holekamp, 1996). However, circannual timing mechanisms appear to
override the activational signal in the Belding’s ground squirrel since males that
reach this fat threshold late in the season delay dispersal until after their first
hibernation (Nunes et al., 1998). Hence, the neural circuitry responsible for

regulating dispersal behavior is most likely similar across dispersers of different

age groups. Yet, the sensitivity of individuals to the activational signal may vary
based on seasonal or environmental factors.

External environmental factors may also mediate dispersal behavior. For
example, aggression directed towards dispersers at their natal site may trigger
emigration. This appears to be the case in species where the immigration of an
adult male leads to the eviction of young residents (red howler monkeys, Alouatta
seniculus, Crockett and Sekulic, 1984; guanacos, Lama guanicoe, Samo et al.,
2003). Avoidance of potential competitors such as littermates or lack of social
bonds with close relatives may also mediate dispersal behavior. However, only
few studies have provided support for these hypotheses (e.g., Barash, 1989)
suggesting that these factors may play a limited role in regulating dispersal
behavior (Holekamp, 1984). In the Belding’s ground squirrel, neither aggression
from conspecifics at the natal site nor avoidance of competition with relatives
mediate dispersal behavior of males (Holekamp, 1986). However, resident
aggression towards immigrants may influence dispersal success of individuals.
Social resistance towards potential emigrants has been documented in a large
number of species (e.g., lions, Panthera Ieo, Hanby and Bygott, 1987; golden lion
tamarins, Leontopithecus rosalia, Baker and Dietz, 1996; common mole rats,
Cryptomys hottentotus hottentotus, Spinks et al., 2000; coyotes, Cam's Iatrans,
Gese, 2001) and the age of the immigrant appears to influence an individual's
ability to counter or elude this aggression. In the meerkat, Sun'cata sun'catta, and
the dwarf mongoose, Helogale parvula, males that disperse when they are very

young are less likely to encounter aggression than males that immigrate into a

pack as fully grown adults (Doolan and Macdonald, 1996; Rood, 1987). In
mammals, natal dispersal typically occurs before individuals reach sexual
maturity (Dobson, 1982). Hence, early dispersal in mammals may occur because
this permits young males to better avoid social resistance during the immigration
process.

Considering its potential costs, it is reasonable to assume that natural
selection should have shaped dispersal behavior to occur at an age that
maximizes an individual’s fitness. Furthermore, selection should have acted on
this dispersal behavior to occur in species at an age when emigrants are best
able to cope with its risks and challenges. Alternatively, however, natural
selection may have had little influence on determining the age at which this
behavior occurs among individuals and species. This may be the case if
dispersers are forced from their natal site or if a developmental error alters the
neural circuitry that regulates this behavior. As a result, dispersal may occur
before individuals have reached an age when their dispersal prospects are
optimal. If this is the case, dispersal age should not influence a disperser’s
fitness and dispersal ages across related species should exhibit a random
pattern. It is possible to asses the role of natural selection in shaping this
behavior by examining the fitness consequences of age at dispersal and by
assessing the role of selective agents in shaping dispersal age across a broad

range of species.

OVERVIEW OF CHAPTERS

The Belding’s ground squirrel exhibits the sexually dimorphic dispersal
pattern typical of most mammals but also shows an age-specific bimodal
dispersal pattern; 70 - 90% of males disperse as juveniles at 7-10 weeks of age
while 10 — 30 % of males emigrate the following summer as yearlings around 55
weeks of age. All males leave home before they reach reproductive maturity at 2
years of age (Holekamp, 1986). Much is known about this species’ life history,
social, reproductive, and dispersal behavior. Consequently, S. beldingi provides
an ideal model system to examine the factors that might determine dispersal age
in mammals.

In chapter two, I used the Belding's ground squirrel to assess whether
young immigrants experience less social resistance from resident adults than do
older immigrants while searching for and settling in new homes. I used staged
and naturally occurring encounters to examine responses by residents towards
juvenile and yearling dispersers. During staged intruder experiments, lactating
females were more aggressive towards yearling than juvenile males. This
difference disappeared when the breeding season ended. Intruders in both age
groups encountered adult resident males at higher rates than resident females,
but resident females were more aggressive to them than were male residents,
particularly towards yearling intruders. Adult male residents were equally likely to
aggress against juvenile and yearling male intruders. These results indicate that

early dispersal is adaptive for male 8. beldingi by allowing juvenile immigrants to

avoid aggression from female residents more effectively than can males who
disperse as yearlings.

Predominant juvenile dispersal among mammals suggests its risks may
be offset by significant fitness gains. In chapter three, I inquired whether
dispersal at a young age enhances male fitness in the Belding’s ground squirrel.
The dichotomous dispersal pattern of this species provides an opportunity to
assess the influence of dispersal age on male fitness by comparing survival and
reproduction between early- and late-dispersing males. I used live-trapping data
collected between 1993 and 2003 to inquire whether dispersal at a young age
enhances male survival. In 1998, extensive mating observations were conducted
to assess differences in male mating success based on male dispersal age. I
found that age-specific survivorship varied significantly among early and late
dispersers, which was primarily due to significant lower juvenile survivorship for
early— than late-dispersing males. The probability of observing a male mating in
1998 did not differ between early and late dispersers, but early dispersers mated
with a greater number of different females than did late dispersers. Overall, these
data indicate that early dispersal may can'y significant survival costs early in life,
but ultimately enhances mating success for surviving males.

Because natal dispersal may be dangerous, natural selection should
shape this behavior to occur in each species at an age when emigrants are best
able to cope with its risks and challenges. In Chapter four, I examined the role of
life history, ecological, and social factors in shaping the evolution of dispersal age

in ground-dwelling sciurid rodents. Specifically, I tested three hypotheses that

have been invoked to explain variation in dispersal age among species of this
group. The growth limitation hypothesis posits that dispersal age of ground-
dwelling sciurids is determined by growth constraints imposed by life history and
environmental factors. Aggression by local residents can influence immigration
success, and the social resistance hypothesis suggests this aggression may
have acted as an important selective agent shaping dispersal age in sciurids.
Lastly, mortality during transit and early settlement can be high for dispersers,
and the risk avoidance hypothesis suggests that risks associated with increasing
dispersal distance may affect the age at which individuals emigrate. I reviewed
the literature on ground-dwelling sciurids for field studies that documented
dispersal age and reported other relevant information. I then used the extracted
data from 22 species to test the predictions of the dispersal constraint
hypotheses. Data were analyzed after computing phylogenetic independent
contrasts to control for non-independence of data points due to common
ancestry. In support of the growth limitation hypothesis, I found that the amount
of growth relative to adult body mass males were able to complete during their
juvenile summer had a significant negative effect on the age at which males
dispersed. I also observed a positive relationship between the level of social
resistance immigrants were likely to encounter and a species” dispersal age, a
result supporting the social resistance hypothesis. However, I found no
significant relationship between mean dispersal distance, a measure of dispersal

risk, and the age at which a particular species‘ generally disperses. Overall,

these results shed light on the selective forces that have shaped the evolution of

dispersal behavior in mammals.

10

CHAPTER TWO
EARLY DISPERSAL REDUCES RESIDENT AGGRESSION DIRECTED
TOWARDS IMMIGRANT MALE BELDING’S GROUND SQUIRRELS
(SPERMOPHIL US BELDINGI)
INTRODUCTION
Natal dispersal, the permanent departure of an individual from its

birthplace, occurs in most birds and mammals (Greenwood, 1980; Lidicker,
1975). Emigration from the natal area reduces the probability of inbreeding and
increases the probability of encountering potential mates (Lehmann and Perrin,
2003; Moore and All, 1984; Pusey, 1987; Shields, 1982). Dispersal is often risky
for individuals as they leave behind the safety of a familiar social and physical
environment, travel through potentially inhospitable terrain, and settle in
unfamiliar places where resources needed for survival and reproduction must be
found anew (lsbell and Van Vuren, 1996; Van Vuren and Armitage, 1994; Waser
and Jones, 1983a). Dispersers may also encounter xenophobic or hostile
conspecifics inhabiting areas in which emigrants might choose to settle (e.g.,
Garrett and Franklin, 1988). In mammals, natal dispersal tends to be male
biased, with males dispersing at higher rates and over longer distances than
females (Greenwood, 1980; Lidicker, 1975). Surprisingly, male mammals often
disperse before they reach sexual maturity (Dobson, 1982). Thus dispersers
must often cope with the challenges of dispersal even before they have
completed their own growth and maturation.

Aggression directed by local residents at prospective immigrants can

prevent dispersers from successfully settling in a new area or group (e.g., lions,

11

Panthera Ieo, Hanby and Bygott, 1987; golden lion tamarins, Leontopithecus
rosalia, Baker and Dietz, 1996; common mole rats, Cryptomys hottentotus
hottentotus, Spinks et al., 2000; coyotes, Canis Iatrans, Gese, 2001). Such
aggression can result in injury or death, and may also force dispersers to use
energy reserves that may already be severely limited (Drews, 1996; Holekamp,
1986; McNutt, 1996; Packer, 1979). Behavioral strategies adopted by some
dispersers reduce aggression directed at them by local residents. For example,
some male primates select new troops to join based on their own personal
histories of friendly interactions with members of those groups during previous
encounters (e.g., Cheney, 1983). When traversing strange habitat, dispersers
may also modify their behavior in a manner that reduces resident aggression.
Immigrant male spotted hyenas (Crocruta crocruta), for example, sometimes face
intense aggression from residents, and they consistently exhibit extreme
unsolicited appeasement behavior upon entry into the territory of a strange clan
(Smale et al., 1997). In some species, natal dispersal tends to occur during
periods each year when intraspecific aggression is infrequent, such as during the
non-breeding season (e.g., some marsupials, Cockbum et al., 1985). Thus
selection may often favor life history traits that minimize the probability that
immigrants will encounter social resistance from resident animals.

Here, I studied Belding’s ground squirrels, Sperrnophilus beldingi, to
inquire whether resident aggression might have acted as a selective agent
affecting age at dispersal in gregarious mammals. S. beldingi offer an excellent

model system in which to examine variation in resident aggression encountered

12

by dispersing males. This species not only exhibits the sexually dimorphic
dispersal pattern characteristic of most mammals, but also an age-specific
bimodal dispersal pattern; 70-90% of males disperse as juveniles at 7-10 weeks
of age whereas 10—30% of males emigrate the following year as yearlings, at
around 51-55 weeks of age (Holekamp, 1984b). All males leave home before
they reach adult male body size, at the end of their second active season, and
before they start mating at 2 years of age, and before they achieve adult body
mass (Holekamp, 1986; Sherman, 1976). Juvenile males disperse at the end of
their first summer, when agonistic interactions among adults typically occur at
low rates. By contrast, yearling males disperse early in the season, shortly after
emerging from their first hibernation, during the breeding season. Thus, yearling
dispersers may attempt to settle in new areas when adult males exhibit high
rates of aggressive behavior while competing for mates, or immediately after
mating when aggression peaks among females as they begin to set-up and
defend territories (Nunes et al., 2000; Sherman, 1976; Wolff and Peterson,
1998)

My objective was to test the hypothesis that the age at which dispersal
occurs influences the social resistance encountered by male immigrants. I
compared resident responses towards juvenile and yearling male 8. beldingi
during staged experimental encounters between tethered “intruders" and resident
adult females, and also during naturally-occurring encounters between residents
and intruding males in each age class. Since most male 8. beldingi disperse as

juveniles, my hypothesis predicted in staged encounters that adult females would

13

 

be more likely to aggress against, and direct higher rates or intensities of
aggression towards, yearling than juvenile males. My hypothesis also predicted
that residents would be more likely to aggress against yearling than juvenile
intruders during naturally-occurring encounters. Because males of the two
dispersal groups leave home at different ages, and during different portions of
the species’ active season, I performed three additional comparisons to tease
apart the influence of age and season on responses by residents towards male
intruders. In staged encounters, I compared responses of adult females towards
intruders between breeding and post-breeding seasons. In addition, I compared
resident responses towards yearlings during naturally-occurring encounters that
took place during the breeding and post-breeding seasons. Lastly, during
naturally-occurring encounters, I compared resident responses towards yearling
males grouped on the basis of the age at which they had dispersed. That is, I
compared responses by residents towards yearling males that had dispersed
early (as juveniles) to their responses towards yearling males that had dispersed
late (as yearlings). Since body mass varies with age in immature animals of this
species, I also examined possible effects of intruder body mass on responses by
residents towards intruding males.
METHODS

Study population

I studied a population of S. beldingi in a 20 hectare alpine meadow from
31 May to 5 August 1999, 17 May to 8 August 2000, and 5 May to 20 August

2001 in the Sierra Nevada mountains at Tioga Pass in Mono County, CA. The

14

study site has been described in detail by Sherman and Morton (1984). In S.
beldingi, a four month active season during late spring and summer is followed
by eight months of hibernation (Morton et al., 1974; Sherman and Morton, 1984).
I defined individuals as juveniles if they were born during the current active
season, and as yearlings if they were born the previous year. I trapped squirrels
using tomahawk live-traps (Tomahawk Live-trap Co., Tomahawk, WI) baited with
peanut butter, and released individuals at their original capture sites. At each
capture, the squirrel's age, sex, reproductive condition, and trapping location
were recorded. I also weighed individuals to the nearest lg using spring balance
scales (Avinet, Dryden, NY). Yearling and juvenile males were repeatedly
trapped and weighed during the course of the study period.

Trapped individuals were marked with monel metal ear tags (National
Band & Tag 00., Newport, KY) and fur dye (Nice’n Easy #124, Clairol lnc.,
Stamford, CT) to facilitate visual identification. I considered females to be
reproductively active if they were either pregnant or lactating. I located and
marked the nest burrow of each female by observing her activity during the
breeding season. The breeding season included mating, gestation, and lactation
periods, whereas the post-weaning period began when juveniles first emerged
from their natal burrows at weaning, 25-28 days after birth (Holekamp, 1984b). l
trapped and tagged all juveniles within 4 days of their first emergence above
ground, during which time they remain within a few meters of their natal burrows
and can be unambiguously assigned to a litter (Sherman, 1976). Based on the

maximum size of a maternal territory at the time when litters are weaned, l

15

defined a male’s natal area as a circle with a 80 m radius centered on his natal
burrow (Holekamp, 1984b). I considered an individual as having dispersed if he
had permanently departed from his mother’s territory, and if he was later
consistently observed and trapped at least 80 m from his birth site. The dispersal
history of all yearlings was known from this and earlier work (e.g., Nunes and

Holekamp, 1996).

Staged Encounters

I performed one set of staged intruder experiments in 1999, and another in
2001, to compare responses by adult females towards tethered intruders
(“stimulus animals’) that were either juveniles or yearlings at the time of testing. In
1999, I conducted intruder experiments during the lactation period (26 June — 27
July 1999) using both male and female stimulus animals that were either
juveniles or yearlings at the time of testing. These experiments allowed me to
evaluate effects of sex and age of the stimulus animal on aggressive responses
by lactating females towards intruders. I used as female stimulus animals only
animals that were not yet reproductively active, as indicated by their small nipple
size. In 2001, using only males as stimulus animals, I conducted additional
intruder experiments during two different periods, the lactation period (21 June ——
05 July 2001) and the post-weaning period (11 - 23 July 2001), to examine
seasonal influences on responses of females towards male intruders. In each

experiment, I tethered a single stimulus animal near the nest burrow entrance of

16

an adult female conspecific, and then recorded events ensuing when the resident
female encountered the tethered intruder.

I assigned each experimental female to one of two periods based on when
her particular trial took place. Females lactating at the time of testing were
assigned to the “lactation period,’ and experimental females that had already
weaned their young of the year were assigned to the ‘post-weaning period’. I
trapped pregnant females frequently and estimated a female’s parturition date
based on when I observed a significant drop in body mass in conjunction with
nipple enlargement. We used estimated parturition date as the criterion for
selecting females as subjects for trials during the lactation period, and females
were used approximately 14 days after they gave birth. I later calculated a more
precise parturition date after each female weaned her litter by subtracting 27
days, the average length of the lactation interval in this species, from the date I
first observed her litter above ground (Holekamp, 1984b). For two experimental
females whose litters had died before weaning, l was unable to determine the
number of days the females had been lactating at the time they were tested. For
the remaining 39 lactation trials, experimental females had been lactating for an
average of 16.46 days on the day of the trial. Post-weaning trials were conducted
10.67 days after litters were first observed above ground.

Each yearling stimulus animal was drawn from an area far from the test
female's territory, and was therefore unfamiliar to her. Juvenile stimulus animals
were trapped from a second 8. beldingi population located 9 km east of Tioga

Pass in Lee Vining Canyon (LVC). Holekamp (1983) describes this second site in

17

detail. Reproductive events in LVC occurred 34 weeks ahead of those in our
main study population because LVC was located almost 1000 m lower in
elevation. Thus, litters from the LVC population emerged above ground
approximately 3-4 weeks before litters from our main study site. By using
stimulus animals from LVC population we were able to select large juveniles for
our lactation trials, which allowed us to better match the body sizes of yearling
and juvenile stimulus animals. Each experimental female (n=53) and each
stimulus animal (n=50) was used in only one experimental trial, except for two
yearling males and one yearling female that were each used twice as stimulus
animals. Stimulus animals were captured one day prior to the trial and kept in a
tent near the study site overnight. These animals were housed individually in
cages (60 x 40 x 15 cm) containing cotton nest material and ad lib mouse chow
and water. Each stimulus animal was randomly assigned to an experimental
female the next morning. The body mass of each stimulus animal was measured
immediately after the trial and was returned to and released at its capture site.
The morning of the trial, I captured the experimental female and briefly
removed her from the immediate vicinity of her nest burrow to avoid any
premature exposure to the stimulus animal. I attached to the stimulus animal a
harness made of light nylon line, similar in design to the harness developed by
(Bakker et al., 1994). I then tethered the stimulus animal to a stake placed one
meter downwind of the entrance of the experimental female’s nest burrow. The
tether consisted of a string permitting the animal to move within a circle of 50 cm

diameter around the concealed stake to which the tether was attached. The

18

experimental female was then released directly into her burrow and her behavior
was recorded for 10 minutes, starting when she re-emerged and clearly detected
the stimulus animal, as indicated by orienting her head and body towards it.
Before and after the trial we carefully examined each stimulus animal to
determine whether it had received any wounds during the trial, but I never
observed any injuries being inflicted during any of the experimental trials.

I used the ethogram for Belding’s ground squirrels developed by
Holekamp (1983), and defined aggressive behavior by each experimental female
as threatening, lunging at, or fighting with the stimulus animal. I recorded all
aggressive behaviors directed by the female toward the stimulus animal during
each trial. Each female was assigned an aggression intensity score based on the
highest intensity of aggression she directed towards the stimulus animal: a
female received a score of 0 if she showed no aggressive behavior, a score of 1
if she only threatened the stimulus animal, a score of 2 if she lunged at the
stimulus animal, or a score of 3 if she initiated a fight with the tethered animal
during the trial. I assessed the likelihood of experimental females aggressing
against stimulus animals drawn from each sex and age group based on the
proportion of trials during which females were observed to aggress against
stimulus animals. I used the number of times a female emitted aggressive
behaviors during the 10 min trial to calculate the rate at which the behavior was
emitted per minute. I evaluated the propensity of females to aggress against

stimulus animals, and the rates and intensities of their aggressive behavior,

19

based on the sex and age of stimulus animals, and on whether trials were

conducted during lactation or post-weaning periods.

Naturally-occurring encounters

In 2001, I conducted focal animal observations (Altmann, 1974) on free-
ranging juvenile and yearling male intruders to evaluate responses by adult
resident males and females towards these intruders in the context of naturally-
occurring encounters. Behavioral observations were conducted between 0800
and 1900 hours Pacific daylight time, and I conducted 59 focal animal surveys
during which I observed 10 yearling and 11 juvenile focal males for an average of
32.30 (1:187) minutes during each survey. Focal animal observations of yearling
males were conducted both during the breeding season and during the post-
weaning period. Juvenile males were only observed during the post-weaning
period, which was the earliest I observed them venturing away from their natal
areas, approximately 2 weeks after their initial emergence from the natal burrow.
Juvenile males were observed outside of their natal areas between 31 July and
12 August 2001. Yearling males were observed between 6 June and 13 August
2001 in areas at least 80 m away from their sleeping burrows, and outside of
their natal areas. Yearling males whose ages at dispersal were known could be
classified as either ‘early’ or ‘late’ dispersers. ‘Early’ dispersers were yearlings
that had dispersed at 7-10 weeks of age, and ‘Iate’ dispersers were yearlings that
had dispersed at 51-55 weeks of age. We radio-collared (SOM-2190

transmitters, AVM Instrument Company, Ltd., Colfax, CA) most males, which

20

allowed us to locate individuals regularly during exploratory forays. All yearling
males and 9 of the 11 juvenile males wore radio-collars during the observation
penod.

During each focal animal survey, I recorded all cases in which a focal
animal encountered another squirrel. I defined an encounter as a focal and non-
focal squirrel coming within 20 m of each other, regardless of whether an
interaction occurred between the squirrels. When an encounter occurred, I
recorded the identity of the encountered squirrel and the behavior of the resident
squirrel towards the intruding male. I considered an interaction to occur during an
encounter when the two animals came within 1 m of each other or when one of
the squirrels clearly changed its behavior in response to the presence of the
other squirrel by either orienting to or approaching it. When an interaction
occurred, it was identified as aggressive or non-aggressive. That is, in an
aggressive interaction a non-focal animal threatened and/or chased a focal
animal, whereas during a non-aggressive interaction a conspecific came within 1
m of the focal animal but showed no signs of aggression.

From focal animal data, I calculated hourly encounter rates at which
intruding focal males came within 20 m of adult male and female residents.
Adults were considered to be residents of a particular area if they were known to
sleep in burrows there. A focal male’s encounter rate with adult residents of one
sex was its total number of encounters with adults of that sex divided by the total
number of hours that focal animal was observed. Because early dispersers leave

home during the post-weaning period whereas late dispersers leave home during

21

the breeding season, I compared rates of encounters between residents and
juveniles during the post-weaning season with encounter rates between
residents and yearlings during the breeding season. I also compared rates of
encounters of yearling male focal animals and adult residents of each sex
between breeding and post-weaning seasons.

Because intraspecifc aggression is common in S. beldingi during the
breeding season when late dispersers emigrate, but rare when early dispersers
leave home after weaning, I predicted that local residents would be more likely to
aggress against late than early dispersers during immigration. To test this
prediction I compared probabilities of experiencing aggression from resident
adults between juvenile focal males when they dispersed during the post-
breeding season and yearling focal males when they dispersed during the
breeding season. For this comparison, I only included yearling male focal data
collected during the breeding season but all juvenile focal data, since juvenile
males were only observed during the post-weaning period. I examined seasonal
influences on resident responses towards yearling intruders by comparing
breeding and post-weaning seasons with respect to the probability of adult
residents of both sexes aggressing against yearling focal males. In addition, I
also performed pair-wise comparisons of male and female responses towards
focal males to determine whether aggressive tendencies differed between male
and female adult residents. Finally, I examined encounter rates of yearling males
and responses of residents towards these males based on yearling male

dispersal history. That is, I evaluated the rates of encounters by yearlings with

22

residents of each sex, as well as the probability of experiencing aggression from
residents during encounters, based on whether the yearling males had dispersed

early or late.

Analysis of body mass

The body mass of male 8. beldingi varies with age, so resident animals
may perceive large intruders as a greater threat than small intruders,
independent of intruder age, and modulate their aggressive responses
accordingly. Although I attempted to select subject animals of similar body size
within each age group, I also performed several analyses to systematically
compare body mass among: early and late dispersers at the time of dispersal,
stimulus males tested in tethering experiments, and males watched as focal
animals. I also compared the body mass of focal and stimulus males within each
age group (e.g., juvenile focal vs. juvenile stimulus animals), and I compared
body mass between age groups within sampling groups (e.g., juvenile focal vs.
yearling focal animals). Body mass at dispersal was estimated for ‘early’ (n=62)
and ‘Iate’ dispersers (n=18) between 1993 and 1997 by trapping dispersers
within 10 days of their settlement in new areas after emigration from their natal
sites. Body mass of each stimulus animal was obtained on the day of its tethering
trial. Yearling focal males were weighed on average 7.7 (i056) times during the
period in which we observed them, so I calculated a mean body mass for each
focal yearling during the observation period. Most juvenile focal males were

weighed only once during the observation period, with the exception of two males

23

that were each weighed three times. For each of these two males, I calculated a
mean body mass for the observation period. One juvenile focal male was not
weighed during the observation period, and was therefore not included in the
body mass analysis. On average, yearling males were weighed within 10.4
(11.64) days and juvenile males 3.6 (i164) days of the dates they were

observed.

Statistical Analyses

Since rates and intensities of aggression emitted by experimental females
in intruder experiments were not normally distributed (Lilliefors test, P<0.05), I
used Mann-Whitney U tests to assess these measures. Similarly, naturally-
occurring encounter rates between residents and yearling males grouped on the
basis of dispersal age were not normally distributed, so we used Mann-Whitney
U tests to compare encounter rates between early and late dispersers observed
as yearlings. To assess naturally-occurring encounter rates between focal males
and residents, I used a two-way ANOVA to examine effects of resident sex (male
vs. female) and current age (juvenile vs. yearling) of the focal animal. Chi-square
tests and two-tailed Fisher exact tests were used to evaluate the probability of
aggression occurring during staged and naturally-occurring encounters. I used a
t-test to evaluate differences in body mass among male and female juvenile and
among male and female yearling stimulus animals used for lactation trials. I also
used a t-test to compare the mean body mass of juveniles and yearlings within

each subject group (disperser, stimulus animal, or focal animal). For males in

24

each age group, we used a one-way ANOVA to determine whether body mass
varied among dispersers, stimulus animals, and focal animals. I further divided
yearling male focal animals into early and late dispersal groups for this analysis.
Differences among groups were considered significant when P<0.05. I adjusted
significance levels of multiple pair-wise comparisons using the Bonferroni
correction. All statistical tests are described in (Sokal and Rohlf, 1995). Rates of

behavior are presented as means :85.

RESULTS

Staged Encounters

In 1999, I conducted 30 staged intruder experiments during the lactation
period using seven juvenile females, seven yearling females, eight juvenile
males, and eight yearling males as stimulus animals. In 2001, I conducted 11
additional experiments during the lactation period with five juvenile and six
yearling stimulus males, and 12 experiments during the post-weaning period with
six juvenile and six yearling stimulus males. I combined data from lactation trials
conducted with male stimulus animals in 1999 and 2001 because I observed no
differences between years with respect to aggression emitted by experimental
females (towards juvenile males: threat: U = 25.00, P = 0.245, fight: U = 20.50, P
= 0.921, aggression intensity scores: U = 25.00, P = 0.401; towards yearling
males: threat: U = 37.50, P = 0.076, fight: U = 22.50, P = 0.842, aggression

intensity scores: U = 25.00, P = 0.880).

25

Lactating females were equally likely to aggress against males and
females when juvenile stimulus animals were tethered (X2 = 0.20, df = 1, P =
0.658; Fig. 1). Lactating females tended to be more likely to aggress against
yearling males than yearling females, but this difference was not statistically
significant ()8 = 2.68, df = 1, P = 0.102). Lactating females were no more likely to
aggress against yearling than juvenile females (Fisher exact test, p = 1.00), but
were significantly more likely to aggress against yearling than juvenile males (x2
= 4.49, df= 1, P = 0.034).

Sex of stimulus animal did not appear to influence rate or intensity of
aggression directed by experimental females towards tethered juveniles (Fig. 2).
Lactating females threatened and fought with juvenile male and female stimulus
animals at similar rates, and with similar intensity scores (threat: U = 45.00, P =
0.949, fight: U = 45.00, P = 0.958, intensity score: U = 48.00, P = 0.814; Fig. 2). I
also observed no significant differences between yearling males and yearling
females with respect to rate or intensity of aggression experimental females
directed at them (threat: U = 71.00, P = 0.086, fight: U = 60.50, P = 0.370,
intensity score: U = 63.5, P = 0.220).

Aggressive responses of lactating females varied with age of the tethered
stimulus animal. Experimental females tended to direct higher rates and
intensities of aggression towards yearling than juvenile stimulus animals of both
sexes (Fig. 2), but this tendency was only statistically significant among male
stimulus animals (juvenile females vs. yearling females: threat: U = 21.00. P =

0.534, fight: U = 18.0, P=0.332, aggression intensity scores: U=21.00, P=0.591;

26

 

juvenile male vs. yearling male: threat: U=36.50, P=0.004, fight: U=43.50,
P=0.011, aggression intensity scores: U=49.50, P=0.028).

The seasonal comparison of aggressive tendencies of females showed
that experimental females were just as likely to aggress against juvenile males
(x2=0.05, df=1, P=0.830) or yearling males (x2=1.63, df==1, P=0.201) during both
test periods. Whereas male age influenced the probability experimental females
directed aggression towards intruders in trials conducted during the lactation
period (Fig. 1), this age effect was not observed in post-weaning trials. During
post-weaning trials, experimental females were equally likely to show aggression
towards juvenile and yearling males (Fisher exact test, P=1.00).

Aggression rates and intensity scores of experimental females were
similar forjuvenile males between lactation and post-weaning trials (threat:
U=45.00, P=0.324, fight: U=48.0, P=0.214, intensity score: U=48.50, P=0.311;
Fig. 3). In contrast, lactating females showed significantly higher rates of
aggressive behavior, and received higher intensity scores, during their
interactions with yearling male stimulus animals than did females used in post-
weaning trials (threat: U=66.00, P=0.037, fight: U=66.00, P=0.034, intensity
score: U=64.50, P=0.042). During post-weaning trials, rates and intensities of
aggression by experimental females did not differ between juvenile and yearling
males (threat: U=15.00, P=0.317, fight: U=15.00, P=0.317, intensity score:
U=15.00, P=0.528).

27

Naturally-occurring encounters between focal males and resident adults

I observed 11 juvenile males for an average of 1.03 h (i021) during which
we recorded 17 naturally-occurring encounters with resident females and 14
encounters with resident males. I also observed 10 yearling males, including five
males that had dispersed early and five that had dispersed late, for an average of
2.11 h ($0.46) during which we recorded 67 encounters with resident females,
and 27 encounters with resident males. Yearlings that had dispersed early were
observed on average for 2.36 h (220.82), and yearlings that had dispersed late for
1.86 h ($0.52). I conducted focal animal surveys on nine of the 10 yearling males
during the breeding season (1.56 $0.82 h/male) and on seven of the 10 yearling
males during the post-breeding season (1.01 h :033 h/male). Only six of the 10

yearling males were observed during both seasons.

Rates at which focal animals came within 20 m of male and female
residents varied significantly overall (F1,48=58.63, P< 0.000; Fig. 4a). Focal males
encountered male residents at significantly higher rates than they encountered
female residents (Fug: 4.81, p = 0.033). However, encounter rates with
residents did not differ significantly between juvenile and yearling focal males
(F1l48=0.42, P=0.659), and I observed no interaction effect between focal animal
age and sex of resident on encounter rates (F2,43=0.68, P=0.510). Since juvenile
males did not begin making forays out of their natal areas until the post-weaning
period, this analysis did not include encounter rates of juveniles collected during
the breeding season. Yearling males that had dispersed late tended to encounter

male residents at higher rates than female residents (Z=1.83, P=0.068; Fig. 4b),

28

but I observed no such pattern for yearlings that had dispersed early (Z=1.46,
P=0.144). Encounter rates with female residents did not differ significantly
between early and late dispersers (U=9.50, P=0.531), but late dispersers tended
to encounter male residents at higher rates than did early dispersers (U=4.00,
P=0.076). All interactions (n=22) between yearling focal animals with adult male
and female residents were aggressive in nature as were 71% (n=5) of
interactions between juvenile focal animals and adult male and female residents.

Intruder age influenced responses by female but not male residents
towards focal males during encounters. Adult female residents were significantly
more likely to aggress against yearling focal males they encountered during the
breeding season than against juvenile focal males they encountered during the
post-weaning period (x2=3.96, df=1, P=0.047; Fig. 5a). Adult males residents,
however, were just as likely to aggress against yearling males they encountered
during the breeding season as against juvenile males they encountered during
the post-weaning period (78:001. df=1, P=0.919; Fig. 5b).

The seasonal comparison of aggressive responses by residents during
encounters with yearling focal males showed that the likelihood of residents
aggressing against focal yearlings tended to be higher during the breeding
season than the post-weaning period. However, this seasonal difference in
resident behavior towards yearling intruders was only statistically significant for
female residents (female residents: x2=7.22, df=1, P=0.007, Fig. 5a; male
residents: x2=1.36, df=1, P=0.244, Fig. 5b). During the post-weaning period,

male and female residents were equally likely to aggress against juvenile males

29

during encounters (x2=0.46, df=1, P=0.497). I also observed no significant
differences between aggressive tendencies of male and female residents
towards yearling males during either the breeding season (x2=0.99, df=1,
P=0.320) or the post-weaning period (x2=0.01, df=1, P=0.923).

Although both early and late dispersers were observed simultaneously,
female residents were significantly more likely to aggress against focal yearlings
that had dispersed late than against those that had dispersed early (x2=8.89,
df=1, P=0.003; Fig. 6). In contrast, male residents were equally likely to aggress
against yearlings in both groups (x2=0.48, df=1, P=0.488). Male residents and
female residents were equally likely to aggress against early dispersers during
encounters (x2=0.50, df=1, P=0.481). Female residents, however, were
significantly more likely than male residents to aggress against late dispersers
(x2=4.33, df=1, P=0.038). These data suggest that the age at which a yearling
male dispersed influences aggressive responses by female but not male

residents towards focal males during naturally-occurring intrusion events.

Body mass comparison

I observed marked variation among the body masses of the various
subject groups of juvenile and yearling males (juvenile males: F2,88=5.88,
P=0.004; yearling males: F3,44=8.08, P<0.001; Fig. 7). Juvenile focal males were
significantly lighter than juvenile males at the time of dispersal (I130 =3.63,
P<0.001). The body mass of juvenile stimulus males did not differ significantly

from that of either juvenile focal males (t1,7o=1.34, P=0.192) or juvenile dispersers

30

 

at the time of dispersal (t1_79=1.74, P=0.085). Yearling stimulus males and
yearling focal males of both dispersal age groups were significantly heavier than
yearling dispersers at the time of emigration, even after Bonferroni correction
(dispersal vs. stimulus: t1_35 =3.65, P=0.001, aadj; 0.013; dispersal vs. early focal:
t1,21=3.90, P=0.001, aadj,=0.01; dispersal vs. late focal: t1,21=3.13, P=0.005,
aaag,=0.017). Yearling focal and stimulus males did not differ significantly in body
mass (stimulus vs. early focal: t1,23 =1.700, P=0.103, aad): 0.025; stimulus vs.
late focal: t1,23=1.04, P=0.307, dam-F005). Yearling males in each male category
were significantly heavier than juvenile males in the same categories (dispersal:
t1,73=2.06, P=0.042, stimulus: t1,37=5.18, P<0.001; focal: t1,18=9.31, P<0.001). We
observed no significant sex difference in body mass among male and female
juvenile stimulus animals used for lactation trials (tug =1.46, P=0.161). However,
yearling males stimulus animals were significantly heavier than their female
counterparts (t1,19=3.27, P=0.004)

Despite my best efforts to equalize their body masses, yearling subject
animals were consistently heavier than juvenile subjects. Therefore, I performed
two post-hoc comparisons to inquire whether aggressive responses of female
residents varied with intruder body mass rather than with intruder age. First, I
used regression analysis to determine whether there was a relationship between
stimulus male body mass and observed female behavior during lactation trials.
For theses analyses, I used the body mass of stimulus males as the independent
variable, and aggression rates and intensities emitted by lactating females as

dependent measures. I found that stimulus animal body mass explained less

31

 

than 4% of the observed variance in female aggression towards males for all
three measures of aggression (threat: R2=0.03; fight: R2 = 0.01; intensity scores:
R2=0.04). Furthermore, no relationship between stimulus male mass and any
measure of female aggression towards stimulus males was statistically
significant (threat: F125 =0.82, #0374; fight: F125 =0.21, P=0.651; intensity
score: F1,25=1.03, P=0.320). Second, I compared the mean body mass of
yearling focal males that had dispersed early to that of yearling focal males that
had dispersed late. Although resident females were significantly more likely to
aggress against late than early yearling males (Fig. 6), body mass of focal
yearlings did not differ between early and late dispersers (t1,9 =0.61, P=0.557).
These results suggest that aggressive responses of resident females towards

intruding males are not determined by intruder body mass.

DISCUSSION

In the Belding’s ground squirrel, intruder age significantly influenced
aggressive responses by resident females towards intruding males during both
tethering experiments and naturally-occurring encounters. Lactating females
were more likely to aggress against, and directed higher rates and intensities of
aggression towards, yearling than juvenile males, although these differences
disappeared after the breeding season ended. Intruders in both age groups were
more likely to encounter adult resident males than resident females, but resident
females were far more aggressive to intruders than were male residents,

particularly to yearling intruders. Adult male residents were equally likely to

32

aggress against juvenile and yearling male intruders. These findings are
consistent with previous accounts of social interactions reported for this species.
For example, Sherman (1976) observed that 83% of encounters between female
residents and yearling male immigrants lead to yearlings being chased by adult
females. In addition, Holekamp (1983) noted that 91% of observed interactions
between yearling males and unrelated adults involved adult females, whereas
only 9% occurred between yearlings and adult males. Considering that all adult-
yearling interactions we observed were aggressive in nature, Holekamp’s data
are consistent with our observation that yearling males are primarily targeted for
aggression by adult females. Thus, by dispersing as juveniles, males searching
for a new home appear to avoid aggression from female residents.

Variation in female responses towards intruders reflects variation in
infanticidal tendencies among various types of intruders. lnfanticide is common in
this species, and territorial aggression by female 8. beldingi throughout the
breeding season helps them defend their young from infanticidal conspecifics
(Nunes et al., 2000; Sherman, 1976). Yearling male immigrants have been
observed to commit about one third of observed infanticides, whereas non-
reproductive yearling females have rarely, and juvenile males and females have
never, been observed killing infants (Sherman, 1981 b). Here, we found that
aggressive responses of experimental females during lactation trials were
greatest towards yearling males, intermediate towards yearling females, and
lowest towards juveniles of both sexes (Figures1 & 2). Differential treatment by

experimental females of juvenile and yearling stimulus males ceased after young

33

were weaned, when pups were mature enough to flee and defend themselves
from conspecifics. In mammalian species where immigrants commit lnfanticide,
the need to protect young from such intruders has been invoked to explain
female aggression toward immigrant males (Agrell et al., 1998). Female
aggression may be most effective for excluding immigrants in species where
males and females are relatively monomorphic as in S. beldingi, and males are
not able to use larger body size to overcome aggression by female residents.
Hence, the need to avoid female aggression may have been an especially
important selection pressure shaping age of dispersal in such mammals.

In S. beldingi. adult male aggression peaks during the breeding season,
when adult males actively compete for mates, and subsequently drops to its
lowest levels as males begin fattening for hibernation (Sherman, 1976). I
therefore predicted that adult male residents would be more likely to aggress
against yearling males intruding during the breeding season than against
juveniles intruding during the post-weaning period. However, aggression by
resident males towards yearling intruders was relatively uncommon, and the
proportion of interactions between adult and yearling males that were aggressive
was only slightly higher during the breeding season than during the post-weaning
period. Similarly, Sherman (1976) found significantly fewer wounds from male
combat on yearling than older male 8. beldingi during the breeding season. The
relatively low frequency of adult aggression towards yearling males suggests that
adult males do not indiscriminately target potential competitors, but rather appear

to direct their aggression selectively based on the reproductive status and

34

competitive ability of other males. This may also be true in European ground
squirrels, S. cite/Ius, in which only 11.4% of male-male aggression observed
during the mating period occurs between adult males and immature yearling
males (Millesi et al., 1998). In S. beldingi, immature males pose little threat to the
future reproductive success of adult resident males because male mortality is
high across all ages, and there is only a small chance that an adult male will later
compete with males he once encountered as immature individuals (Sherman and
Morton, 1984).

In addition to S. beldingi, juvenile dispersal coincides with reduced
resident aggression in a number of other ground-dwelling sciurids. Resident
aggression towards juvenile immigrants has only been documented in two of the
13 species showing predominant juvenile dispersal. In the Uinta ground squirrel,
S. armatus, adult males aggressively target juvenile males during the post-
weaning period and this aggression appears to reduce the likelihood of the
immigrants becoming sexually mature as yearlings (Slade and Balph, 1974).
Suitable hibernation sites are scarce in the Arctic ground squirrel, S. parryii, and
adult males and females aggressively defend territories surrounding their
hibemacula during the post-weaning period from other conspecifics, including
juvenile immigrants (Carl, 1971 ). Interestingly, both early and late dispersing S.
columbianus males appear to avoid resident aggression by dispersing towards
the end of the species’ active season when resident aggression is low
(Waterman, 1992; Wiggett and Boag, 1993). In most prairie dog (black-tailed

prairie dog, C. Iudvicanus; Gunnison’s prairie dog, C. gunnisonr) and marmot

35

species (M. flaviventn's; hoary marmot, M. caligata; golden marmot, M. caudata
aurea; Alpine marmot, M. mannota; Olympic marmot, M. Olympus) the resident
male or all members of a social group aggressively defend a territory throughout
the species’ active season, and early dispersal therefore does not reduce
resident aggression towards prospective immigrants (Armitage and Downhower,
1974; Barash, 1973; Blumstein and Arnold, 1998; Garrett and Franklin, 1988;
Hoogland, 1999; Michener, 1979a; Michener, 1979b). Hence, in species where
territorial defense tends to be continuous, males may benefit from delaying
dispersal until they are larger and have greater chances of immigrating
successfully. However, lack of tolerance by local residents towards immature
males may make it impossible for males to delay emigration until a time when
immigration prospects are optimal. This appears to be the case in S. panyii
where juvenile males are evicted by their mothers (Carl, 1971). Although, the
timing of dispersal in other ground squirrel species relative to seasonal
fluctuations in resident aggression supports the hypothesis that early dispersal
reduces social resistance encountered by immigrant males, our results must be
interpreted cautiously. A more rigorous comparative approach and the inclusion
of other characters, such as body size, are necessary to assure independence
among data points (Harvey and Page, 1991) and to assess the importance of
other variables in the evolution of dispersal age in these species (Armitage,
1981; Barash, 1974).

In our study, intruder age rather than intruder body mass appeared to

influence aggressive responses by resident females. However, age and body

36

size tend to be highly correlated in mammals, and the combination of age and
competitive ability, rather than age alone, may affect immigration patterns in
some species. For example, in both Japanese macaques (Macaca fuscata) and
savannah baboons (Papio cynophalus), reproductive potential of males is highest
when they are physically and socially at their prime, but the timing of dispersal
relative to this prime differs between the species, as does the relative size and
age at which males disperse. Male savannah baboons emigrate shortly after
reaching adult body size, and tend to rise quickly to the top of the dominance
hierarchy in the new troop (Alberts and Altmann, 1995). In contrast, male
Japanese macaques disperse when they are not yet fully grown, and enter the
new troop at the bottom of the hierarchy (reviewed by (Sprague, 1992). In
meerkats, Sun'cata suricatta, and dwarf mongoose, Helogale parvula, males that
immigrate into a pack as subadults do so as subordinate non-breeders, whereas
males that disperse once they are near or at adult size can attain breeding status
if they can able to aggressively oust the dominant male(s) (Doolan and
Macdonald, 1996; Rood, 1987). In many mammals, dispersers often exhibit
specific behaviors that allow them to minimize aggression from residents. For
example, dispersers may try to avoid residents by being silent and inconspicuous
while traversing or entering foreign habitat (e.g., red squirrels, Tamiasciurus
hudsonicus, Larsen and Boutin, 1994; gibbons, Hylobates Iar, Brockelman et al.,
1998).

Although the adult sex ratio is female-biased in our 8. beldingi study

population (Sherman and Morton 1982, Muecke and Nunes unpublished data),

37

intruding males encountered resident males at significantly higher rates than
resident females. Males thus appeared to minimize conflict by avoiding female
residents. Sherman (1977) observed that adult and immigrant yearling male 8.
beldingi tend to limit their activity to areas between female breeding sites or to
areas of low female density. Similar observations have been made for other
sciurid species. For example, in Richardson’s ground squirrels, S. richardsonii
(Michener, 1979a), and S. columbianus (Murie and Harris, 1978), adult and
yearling males are either absent from, or occupy areas peripheral to, female
nesting sites. In addition, 8. beldingi yearlings may also avoid female aggression
by altering their activity schedule to minimize temporal overlap with territorial
females. This was suggested by Sherman (1976), who reported that immigrant
males with sleeping burrows on or adjacent to female territories delay or advance
their entry and exit from their sleeping burrows to avoid the resident females. We
observed similar movement and activity patterns among yearling males here, and
we also noted that males on exploratory excursions tended to utilize non-
breeding areas as travel routes. Hence, it appears that by avoiding females in
time and space, male 8. beldingi are able to reduce the likelihood of
encountering aggression by female residents during exploratory excursions.
Here the age at which males dispersed influenced their behavior as
yearlings as well as the responses directed by female residents towards them as
yearlings. Yearling males that had dispersed late tended to encounter male
residents at higher rates than female residents, and also tended to have higher

encounter rates with male residents than did yearlings that had dispersed early.

38

During encounters with female residents, these males were also significantly
more likely to be targets of aggression than were early dispersers. Familiarity
with intruders can reduce aggressive responses by residents (French et al.,
1995), and repeated intrusions by the same individual into a territory might tend
to reduce aggressive responses by the resident towards the intruder (Switzer et
al., 2001). The relatively higher male than female encounter rates of late
dispersers suggest that late dispersers tend to intrude less frequently into areas
that are occupied by females, which appears to influence how these yearlings
are treated by resident females. Hence, age at the time of dispersal has
significant social consequences for males that extend beyond the dispersal
period and perhaps even persist into adulthood. Familiarity with a male’s odor
increases sexual behavior of female golden hamster, Mesocrr’tecus auratus,
during mating trials (Tang-Martinez et al., 1993), and female giant kangaroo rats,
Dipodomys ingens, prefer to mate with familiar males (Randall et al., 2002). S.
beldingi males roam widely in their search for receptive mates, and if familiarity
plays a similar role in female mate choice in this species, early dispersal may
allow males to introduce themselves to females as prospective mates, and
subsequently enhances their own mating success. This might offer an additional
ultimate explanation for why early dispersal predominates in this species.
However, overall our results support the hypothesis that juvenile dispersal is
adaptive for male 8. beldingi by allowing male immigrants to avoid aggression
from local residents. Aggression by residents towards immigrants has been

observed in many mammals, and this aggression can influence immigration

39

 

patterns (Brandt, 1992). The need to avoid resident aggression may thus also

affect age at dispersal of other mammals.

40

100 -

_ Juvenile stimulus animals *
I: Yearling stimulus animals
r:
.2
U)
3
I51, 60 ~
2?
i
3 7
U)
75 40 -
'C
l-
9*”
20 -
0 - _

 

 

 

   

Females Males

Sex of stimulus animal

Figure 1: Percent of 10-min trials during which experimental females aggressed
against tethered stimulus animals. All trials shown here took place during the
lactation period. Sample sizes above bars refer to the total number of trials
conducted for each stimulus animal group. Horizontal bar with asterisk indicates

that groups differ significantly at P < 0.05.

41

0'6 ‘ (a) ..

_ I ..
0-5 ' — Juvenile stimulus animals
0 4 (:1 Yearling stimulus animals

 

 

0.3 ‘

0.2 ‘

Threats/minute

 

0.0 - --»A»—
0.6

R.

0.2 <

Fights/minute
O
b)

0.1 <

 

3.0 ~ (C)
2.5 «

2.0 ~

0.5 -

 

Aggression Intensity Score
M

 

 

0.0 - —
Females Males
Sex of stimulus animal
Figure 2: Mean rates of threat (a), rates of fighting behavior (b), and aggression
intensity scores (c) directed by experimental females towards juvenile female (n
= 7), juvenile male (n = 13), yearling female (n = 7), and yearling male (n = 14)
stimulus animals. All trials shown here took place during the lactation period.

Asterisk indicates significant differences between groups at P<0.05 and double

asterisks indicate P < 0.01.

42

0.6 «

 

 

 

 

 

 

 

 

 

 

 

 

#0 g
(a) l j I 1
0‘5 l 1— Juvenile stimulus males
0 . .
E 04 j E2223 Yearllng stimulus males
g 0.3 ~
8
.E 0.2 ‘
l—
0.1 -
0,0 ._fiE [—1—]
0.6 ‘ (b) .‘ r t 1
l———"l
0.5 .
3 l
:3 0.4 I
.E
E
E 0.3 -
'50
i: 0.2 ~
r—l—a
t
j

 

 

Aggression Intensity Score

 

 

 

 

Lactation Post-weanin g
Experimental period

Figure 3: Mean rates of threat (a), rates of fighting behavior (b), and aggression
intensity scores (0) directed by experimental females towards tethered juvenile
males (n = 13) and yearling males (n = 14) during lactation trials and towards
juvenile males (n = 6) and yearling males (n = 6) during post-weaning trials.

Other notation is as in figure 2.

43

 

1.2 -
(a) - Encounters with adult females (b)

Encounters with adult males

1.0 -

 

 

 

 

Hourly encounter rates (individuals/hour)

 

 

 

 

 

 

 

 

Yearling Juvenile Yearling Early disperser Late disperser

 

 

 

 

 

 

 

 

Breeding season Post-weaning period Yearling males

 

 

Figure 4: (a) Rates of naturally-occurring encounters between yearling focal
males and adult male and female residents during the breeding season (n=9)
and the post-weaning period (n = 7), and between adult residents and juvenile
focal males (n = 11) during the post-weaning period. (b) Rates of encounters with
adult resident by yearling males that had dispersed either as juveniles (early
dispersers, n = 5) or as yearlings (late dispersers, n = 5). Data in (a) but not (b)

are normally distributed, so data shown in a & b cannot be directly compared.

LII
O
r

**

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

s: I * — Juvenile focal males

% I Yearling focal males

8 40 ‘ .42 .

5b . .

on

:6

g 30 .

3

i’l

g 20 -

a

s

o

E 10 n

LL]

0\""

0 0 , .. ..

Breeding Post-weaning Breeding Post-weaning
LResident females I I Resident males

 

 

Figure 5: Aggressive behavior directed by female (a) and male (b) residents
towards juvenile (n = 11) and yearling (n = 10) male focal animals during
naturally-occurring encounters. Sample sizes above bars refer to the total
number of encounters observed for each focal male subject group. Other

notation is as in Figure 2.

45

— Encounters with adult females
l:l Encounters with adult males

50-

**

 

40~

301

10a

% Encounters with aggression

 

 

 

 

 

Early dispersers Late dispersers

Yearling male dispersal history

Figure 6: Aggressive behavior directed by adult male and female residents
towards yearling focal males that had dispersed either as juveniles (early
dispersers, n = 5) or as yearlings (late dispersers, n = 5). Sample sizes above the
bars refer to the total number of naturally-occurring encounters we observed

involving males from each yearling group. Other notation is as in Figure 2.

46

 

 

 

 

 

 

 

 

 

a b it
300 - ( ) ( ) I m I
an [—H—j
250 - Q
30 200 -
m .
E
>, 150 - '
8
a: 3
100 ‘ . I an I
J til! I
50 4
l t l
O - r I I r I r T
Dispersal Stimulus Focal Dispersal Stimulus Early disp. Late disp.

 

[Yearling males I lYearling focal males]

Figure 7: (a) Body mass of juvenile males (n = 62) at the time of dispersal,
juvenile stimulus males (n = 19) on the day of the experimental trial, and juvenile
focal males (n = 10) during the post-weaning period. (b) Body mass of yearling
males (n = 18) at the time of dispersal, yearling stimulus males (n = 20) on the
day of the trial, and yearling focal males that had dispersed either as juveniles
(early dispersers, n = 5) or as yearlings (late dispersers, n = 5). Body mass of
focal males weighed more than once per observation period was averaged for
this comparison. Triple asterisks indicate significant differences at P < 0.001.

Other notation is as in Figure 2.

47

CHAPTER THREE
EARLY DISPERSAL DOES NOT IMPROVE SURVIVORSHIP BUT IS

CORRELATED WITH ENHANCED MATING SUCCESS AMONG MALE
BELDING’S GROUND SQUIRRELS (SPERMOPHILUS BELDINGI)

INTRODUCTION

Natal dispersal, the permanent departure of an individual from its
birthplace, is sexually dimorphic in most mammals, with males dispersing while
females remain at their birth site throughout their lives (Greenwood, 1980;
Lidicker, 1975). Dispersal represents one of the most profound environmental
changes an individual is likely to experience during its lifetime. A disperser leaves
behind the safety of a familiar social and physical environment, and must
navigate through unfamiliar habitat where he is exposed to many mortality factors
such as predation and lack of essential resources. After he has located a suitable
habitat, the emigrant must rapidly assimilate knowledge about the social and
physical features of his new home to ensure successful settlement (lsbell and
Van Vuren, 1996; Van Vuren and Armitage, 1994; Waser and Jones, 1983a).
The costs of dispersal can be substantial and may include exclusion from
important resources (Garrett and Franklin, 1988), loss of reproductive
opportunities (Alberts and Altmann, 1995), injury, and even death (Byrom and
Krebs, 1999; Larsen and Boutin, 1994). In mammals, natal dispersal often occurs
before individuals reach sexual maturity (Dobson, 1982), yet because of their
small body size and inexperience, young emigrants may be particularly
vulnerable to risks associated with dispersal. This apparent paradox suggests

that dispersal age has a strong impact on male fitness. Despite the fact that

48

considerable anecdotal evidence in the literature supports this notion, few studies
have evaluated the effect of age at dispersal on male fitness.

My objective here is to assess the influence of dispersal age on male
fitness in Belding’s ground squirrels (Sperrnophilus beldingi). Dispersal behavior
has been studied extensively in this species, which exhibits the male-biased
pattern of emigration typical of most mammals. S. beldingi are diurnal, easily
trapped and observed, and individuals can be followed throughout most of the
reproductive lifespan. In addition, the reproductive and social behavior of this
species has been described in detail, providing the necessary foundation on
which to evaluate male fitness. In S. beldingi, a four-month active season is
followed each year by eight months of hibernation (Morton et al., 1974; Sherman
and Morton, 1984). Although approximately 70 - 90 % of dispersers emigrate as
juveniles (early dispersers), 10 - 30 % depart the following summer as yearlings
(late dispersers), and all males leave home before they are 55 weeks old
(Holekamp, 1986). Juvenile dispersers typically leave their natal site when they
are 7 — 10 weeks old and weigh only 183 g (Nunes and Holekamp, 1996).
Yearling dispersers are roughly one year old and weigh 250 9 (Chapter 2).
Mortality is high throughout a male’s lifespan, and a male has a less than 50%
chance of surviving from one year to the next (Sherman and Morton, 1984).
Males become reproductively mature at age two. 8. beldingi are highly
polygynous, and reproductive success is highly skewed among adult males

(Sherman 1976).

49

Considering the persistent bias towards early dispersal and a male’s
limited prospects for survival and reproduction it is likely that selection has
favored juvenile dispersal in Belding’s ground squirrels because males that
disperses early gain significant fitness benefits over males that disperse as
yearlings. Here, I test predictions of this hypothesis using extensive live-trapping
and observations of matings to evaluate the effects of age at dispersal on male
survival and reproductive success. If early dispersal improves male survival, then
early dispersers should be more likely to survive to reproductive maturity, and
should be over-represented in all age classes. lf early dispersers accrue
reproductive benefits, they should be more likely to be observed mating than late
dispersers, and they should be able to mate with greater numbers of different

mating partners than males that delay emigration.

METHODS

Study population

This study was conducted in a 20-hectare meadow located in the Sierra
Nevada Mountains near Tioga Pass in Mono County, CA. Sherman and Morton
(1984) describe the study site in detail. Between May and September from 1993
to 2003, individual squirrels on the study meadow were trapped at regular
intervals using Tomahawk live-traps baited with peanut butter and released at
their original capture sites. Each squirrel was weight to the nearest 1 9 using
spring balances, and the capture location was recorded. An individual’s

reproductive condition was evaluated, and males were considered reproductively

50

mature if they showed descended testes during the mating period at the onset of
the active season. Males of this species reach sexual maturity at age 2. At first
capture, individuals were marked with metal ear tags and hair dye to facilitate
visual identification (Sherman 1976). The location of each adult female’s nest
burrow was identified by observing the female’s activity daily throughout the
summer. Juveniles were first trapped within 4 days of their initial emergence
above ground at weaning, during which time they could be unambiguously
assigned to a mother (Sherman, 1976).

The dispersal age of each male was established using methods
developed by (Holekamp, 1983; Holekamp, 1984a). Marked litters were observed
at dawn and dusk to determine if they were still sleeping in their natal areas.
Based on the maximum radius of 80 m of a maternal territory, a male’s natal area
was defined as a circle of 160 m in diameter centered on the male’s natal burrow
(Holekamp 1983). Males were considered absent from their natal area if they
were not seen exiting their natal burrow at dawn or entering the burrow at dusk. If
a male did not spend the night in his natal burrow, extensive trapping and visual
observations were performed on the study meadow and environs in an attempt to
relocate him. Males were considered to have dispersed during theirjuvenile
summer if they were not observed in their natal areas late in the summer but
were subsequently relocated elsewhere. These males were considered early
dispersers whereas males that over-wintered in their natal area were considered
late dispersers. We were unable to assign dispersal age to some males since

they had disappeared from their natal site, and we were unable to relocate them

51

on our study site or the surrounding areas. However, some of these males
disappeared before they had reached eight weeks of age, while others
disappeared after eight weeks of age. Since previous work by Nunes et al.
(1998) has shown that only a small proportion of males disperse at seven weeks
of age, I made the assumption that males of the first group had died while still at
their natal site while males that disappeared after eight weeks of age were early
dispersers that died during dispersal. As a result, males that disappeared after 8
weeks of age could have died at their natal site, died after dispersing or
dispersed and not been recovered, it is possible that l slightly overestimated the

juvenile mortality of early dispersers using this criteria.

Fitness parameters 8 data analysis

Trapping data collected between 1993 and 2003 were used to determine
survivorship of early- and late-dispersing males. Data for survivorship analyses
came from the five cohorts (1993 - 1997) for which there were complete records
of survivorship as all males from these cohorts had died by 2003. For each age
class, a male was considered to be alive if he was captured at least once during
a given year and was othen/vise considered to have died. Horizontal life tables
were constructed separately for early- and late-dispersing males following the
methods of Sherman and Morton (1984). First, the number of individuals that
were alive at age x (nx) and the number of individuals that did not survive to the
next age (disappearance, d,) were tabulated. Survival rate (Ix), disappearance

rate (q,) and future life expectancy (ex) were calculated following Deevey (1947)

52

and Ricklefs (2000). The Cox’s F test was used to compare age-specific
survivorship (Ix) between early- and late-dispersing males. Chi-squared analyses
were used to assess whether early dispersers were more likely to survive to age
one (juvenile survival) or to reproductive maturity (survival to age two) than were
late dispersers. These analyses were based on the total number of males in each
dispersal age group and the number of individuals that were still alive as one- or
two-year olds, respectively.

In 1998, 6 observers monitored the study area for matings throughout the
mating period (June 4 — 25) to assess differences in male mating success
between early and late dispersers. Critical incident sampling and focal animal
observations (Altmann, 1974) of adult males and females were used to monitor
male mating behavior. During the mating season, critical incident sampling was
conducted throughout the species” daily activity period (0730 to 1930), totaling
264 daytime hours. During this time, we also conducted 71 h of focal animal
observations. When copulations were observed, the date, time, location, and
identity of mating partners were recorded. During each active season, female 8.
beldingi are in estrus on only one day, for 3-4 h, and remain in a specific area of
the study meadow throughout this period (Sherman, 1976; Sherman, 1977).
Female estrus behavior could therefore be used to determine if a male had
mated with the same or different female in three cases in which we were unable
to determine the identity of a female mating partner. For example, if a male was
observed mating with an unidentified female during the morning hours and a

known female the following evening, we could be certain that these matings

53

involved two different females. Using this method, we were able to determine that
the identity of a male’s unknown mate must have differed from that of his
previous mating partner(s) in all three cases. Male and female 8. beldingi may
mate multiple times with the same mating partner (Sherman, 1976). I therefore
chose to use “number of different mating partners” rather than “number of
copulations” as our measure of male mating success. Dispersal ages were
assigned to adult males based on previous trapping and observation records.
Body mass is an important determinant of male mating success in Belding’s
ground squirrels (Sherman, 1976). As a result, differences in male mating
success may be a consequence of differences in body mass rather than
dispersal age per se. To account for this possibility, I compared the body mass of
early- and late dispersers taken within 1 4 days of the middle of the mating
season (June 14‘").

Fisher’s exact test was used to determine whether the probability of being
observed mating differed between adult males in the two dispersal age groups.
This comparison was based on the number of males from each dispersal age
group observed mating, relative to their representation in the overall population of
males trapped during the 1998 mating season. A chi-square test was used to
assess whether the number of different mates acquired by males of known
dispersal age varied in proportion to the number of males that mated in each
dispersal class. Since data were not normally distributed (Lilliefors test, p < 0.01),
a Mann-Whitney U test was used to compare the mean number of different

mates acquired by males in each dispersal age group. T-tests were used to

54

compare the mid-mating season body mass of early- and late dispersers. The
statistical package STATISTICA 6.1 (StatSoft Inc.) was used for all data
analyses. Two-tailed p-values are reported for statistical tests, and results were

considered significant if p < 0.05. Means are reported t SE.

RESULTS
Survivorship ‘
Between 1993 - 1997, 433 juvenile males from 153 different litters were trapped
and tagged (Fig. 1). Dispersal ages could be unambiguously assigned to 239
males, including 149 early and 90 late dispersers. For the remaining 194 males,
113 of them disappeared before they had reached age week 8 while 81 males
disappeared after age week 8. We added the number of individuals that had
disappeared after age week 8 (n = 81) to the number of males (n = 149) we were
able to unambiguously categorize as early dispersers. Data from these males (n
= 230) and from late dispersers (n = 90) were used to construct life tables for
males in each dispersal age group (table 1). Age-specific survivorship (Ix) varied
significantly between early and late dispersers (Fig. 2; F 450,180 = 1.57, p < 0.000).
Early dispersers were significantly less likely to survive to age one than late
dispersers (Fig. 3; x2 = 42.22, df = 1, p < 0.000). However, the probability of
surviving to reproductive maturity (age two) did not differ significantly between
dispersal age groups (X2 = 0.74, df = 1, p = 0.390). Survivorship analyses
including only the subset of early dispersers with unambiguous dispersal age (n =

149) produced similar results, which are not reported here.

55

Mating success

During the 1998 mating season, 34 different adult males were trapped. Dispersal
ages were known for 16 of these 34 males since some males had immigrated
into the meadow from outside the study area as yearlings (age one; n = 1) or as
adults (_>_ age two, n = 16). Of the 16 males with known dispersal ages (8 early
and 8 late dispersers), 11 individuals (6 early and 5 late disperser) mated with a
total of 22 different females (early dispersers mated with 16 females; late
dispersers mated with 6 females). Dispersal age did not influence the likelihood
of observing a male mating (Fisher’s exact test, p = 1.00). Overall, however, the
observed and expected number of female mating partners differed significantly
between early and late disperses (Fig. 4; x2 = 6.60, df= 1, p < 0.010).
Furthermore, early dispersers mated on average with a greater number of
different females than did late dispersers (Fig. 5; Z = 2.02, p = 0.044). Early- and
late dispersers were similar in body mass when considering all males of known
dispersal age (early dispersers: mean = 231 i 12.01 9, late dispersers: mean =

229 3; 12.69 9; t1,“ = 0.11, p = 0.911) or only males that were observed mating

(early dispersers: mean = 239.00 i 14.47 9, late dispersers: mean = 232.40 i
20.27 g; t1,g = 0.272, p = 0.792). These results suggest that differences in mating
success between early and late dispersers are not merely an artifact of male

body mass.

56

DISCUSSION

My results suggest that early dispersal in Belding”s ground squirrels is
correlated with an increase in male fitness. Early dispersers acquired more, and
late dispersers acquired fewer, mates than expected. Furthermore, early
dispersers mated on average with a greater number of different females than did
late dispersers. Anecdotal evidence suggests that dispersal age also influences
mating behavior in the black-tailed prairie dog, Cynomys Iudovicianus (Garrett
and Franklin, 1988). In this species, males most frequently emigrate in the spring
of their yearling active season, and males become reproductively mature as two-
year olds. Yet, a few unusually large individuals may emigrate when they are
very young, prior to their first winter. These large males also reach sexual
maturity at age one and are able to acquire mates at this age. Although a rare
event, this observation suggests that male prairie dogs that are able to disperse
at an early age may also enjoy a reproductive advantage.

However, male mating success may not necessarily provide an accurate
measure of male reproductive success since in many ground-dwelling sciurids,
including S. beldingi, females mate with multiple mates and produce mixed
paternity litters (reviewed by Travis et al., 1996; Lacey et al., 1997). In ground-
dwelling sciurids, the frequency of multiple paternity ranges from 3 % in C.
Iudovicianus (Hoogland and Foltz, 1982) to 88.9 % in the California ground
squirrel, Sperrnophilus beecheyi (Boellstorff et al., 1994). In S. beldingi, multiple
paternity estimates range from 32 to 66 %, and similar to the Arctic

(Spemiophilus parryii) and thirteen-lined ground squirrel (Spermophilus

57

tn'decemlineatus), a female’s first mating partner sires the greatest proportion of
offspring in a female’s litter (Foltz and Schwagmeyer, 1989; Lacey et al., 1997;
Sherman, 1989). Unfortunately, our data were insufficient to determine the
mating sequences of males. However based on his data and extensive
behavioral observations, Sherman (1989) proposed that the evolutionary stable
mating strategy for a male 8. beldingi should involve mating with as many
females as possible. If a male‘s mating success reflects the number of pups he
sires, early dispersers should therefore sire a greater number of pups than late
dispersers. Clearly, future genetic analysis of paternity comparing males of
different dispersal ages will be helpful to confirm the differential mating success
of early and late dispersers I observed in this study.

The influence of dispersal age on male mating success has also been
documented in some species of primates. For example, in both savannah
baboons (Papio cynophalus) and long-tailed macaques (Macaca fascularis), the
reproductive potential of males is highest when they are physically and socially at
their prime, but the timing of dispersal relative to this prime differs between the
species, as does the relative size and age at which males disperse. Male P.
cynophalus emigrate shortly after reaching adult body size, and tend to rise
quickly to the top of the dominance hierarchy in the new troop (Alberts and
Altmann, 1995). By contrast, male M. fascularis disperse when they are not yet
fully grown, and enter the new troop at the bottom of the dominance hierarchy
(Noordwijk and Schaik, 2001; van Noordwijk and van Schaik, 1985).

Occasionally, male M. fascularis attempt to immigrate into a new troop when they

58

are fully grown, but these males are almost always rebuffed by the local alpha
male. This suggests that factors other than a male’s physical condition influence
the immigration success of male M. fascularis. Van Noordwijk and Schaik (2001)
proposed that prolonged residence in the new troop allows a young male to
monitor and judge the stability of the dominance hierarchy and use this
knowledge to determine when the alpha male is most vulnerable to rank
challenges. Hence, physical prowess is critical for the outcome of rank
challenges in P. cynophalus, whereas M. fascularis males must gather and apply
new social knowledge in order to maximize their take-over prospects.

In S. beldingi, male-male competition for mates is intense, and a male’s
ability to win fights influences his mating success (Sherman, 1976). Thus,
enhanced competitive abilities reflected in greater fighting success of early than
late dispersers may have been responsible for the enhanced mating success of
early dispersers in 1998. Although S. beldingi males do not defend exclusive
territories, early dispersers may have also been more effective than late
dispersers at excluding other males from receptive females (Lacey and
Wieczorek, 2001). Receptive female 8. beldingi appear to stimulate male-male
competition, observe male fights, and then solicit some males while rejecting
others, suggesting that females in this species actively choose male mating
partners (Sherman 1976). Consequently, the differential mating success we
observed might have been due to females preferentially mating with early

dispersers. Further work will be necessary to assess the relative contribution of

59

male competitive ability and female mate choice in determining the differential
mating success between males of the two dispersal groups.

Early dispersers were significantly less likely to survive their first year of
life than late dispersers, suggesting a tradeoff between early dispersal and
survival for male 8. beldingi. Exceptionally high losses of juveniles relative to
adults have been observed in many ground-dwelling sciurids and have been
attributed to greater vulnerability of young squirrels to predation and over-winter
mortality (yellow-bellied marmot, Marmota flaviventn’s, Armitage and Downhower,
1974; Richardson’s ground squirrel, S. richardsonii, Michener and Michener,
1971; S. parryii, Byrom and Krebs, 1999). In addition, young dispersers may be
especially vulnerable to these mortality factors since extensive locomotor
behavior can increase predation risk (Metzgar, 1967) and reduce or limit the
amount of energy stores young animals have available for hibernation (Buck and
Barnes, 19993; Nunes et al., 1999). In S. panyii, juveniles are exposed to high
predation rates during dispersal but their survival rate remains constant after
settlement (Byrom and Krebs, 1999). Our data suggest that about 44% of early
dispersers (81 of 183 juvenile losses) died during transit indicating that predation
risk may be an important source of mortality for young dispersers in S. beldingi.
However, a significant number of juvenile losses occurred post-settlement
suggesting that other factors, such as over-wintering mortality, are responsible
for juvenile losses as well.

Available energy stores are crucial for over-winter survival, and individuals

with low body mass and small stores of body fat should therefore experience

60

high mortality rates during hibernation (Lenihan and Van Vuren, 1996). The
amount of energy individuals can store is directly proportional to their body size.
Higher losses of early dispersers may thus be due to the smaller size and lower
energy stores of early than late dispersers prior to their entry into hibernation.
Placement and other characteristics of hibemacula might also influence over-
winter survival (Buck and Barnes, 1999b). Young dispersers may not select
appropriate hibernation sites or may not have sufficient time or energy available
to build or modify a hibernaculum that provides sufficient protection from low
temperatures during the hibernation period (Arnold, 1993).

The survival costs associated with early dispersal suggest a tradeoff that
may be responsible for the maintenance of the dispersal age dichotomy in S.
beldingi. In this species, circannual timing mechanisms tightly regulate dispersal
behavior of males and inhibit dispersal if individuals fail to reach a specific body
mass/body fat threshold early enough in the season to maximize their chances of
surviving both dispersal and hibernation (Nunes et al., 1998). This is especially
evident in years when a prolonged snow cover significantly shortens the species’
active season and most males born during such short seasons disperse late
(Nunes et al., 1998). The extent of growth males are able to complete during
their juvenile summer also appears to affect dispersal age among males of other
ground-dwelling sciurids (Chapter four). In most marmot species, males are only
able to acquire less than 50% of adult body mass by the end of theirjuvenile
summer, and in these species males consistently delay dispersal until they are

older (Blumstein and Armitage, 1999). In the Columbian (S. columbianus) and

61

Richardson’s ground squirrels exceptionally small males disperse several months
after most males emigrate (Festa-Bianchet and King, 1984; Michener and
Michener, 1977; Wiggett and Boag, 1991). In these species, as in Belding”s
ground squirrels, the need for juveniles to accumulate sufficient energy stores to
survive hibernation appears to constrain dispersal, suggesting that energetic
constraints have played an important role in shaping dispersal age among many

ground-dwelling sciurids.

62

 

Tagged juveniles:
n = 433

 

 

 

 

 

I I

Losses Dispersal ages
n = 194 known: n = 239

 

 

 

 

 

 

 

 

 

 

 

 

 

I I I I,

 

Losses Losses Early Late
before 8 after 8 weeks: dispersers: dispersers:
weeks: n=113 n=81 n=149 n=90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I

Total early
dispersers:
n = 228

Figure 1. Summary of sample sizes obtained from monitoring male survival and
dispersal behavior of 1993 - 1997 cohorts on Tioga pass meadow, CA. Out of
433 tagged juvenile males, I categorized 230 individuals as early dispersers and

90 individuals as late dispersers (see Methods: Dispersal age assignment).

63

 

 

 

1.0 q
A + Early dispersers
g; 0.8 . ---O-- Late dispersers
.9- '
.2
E
g 0.6 -
Z
3
II)
5:32 0.4 -
O
o
O.
U)
0': 0.2 .
O)
<

0.0 ~

0 1 2 3 4 5
Age (years)

Figure 2. Age-specific survivorship of early- and late-dispersing males.
Survivorship is based on 320 males, 230 early and 90 late dispersers, from

cohorts born in 1993 - 1997.

64

N
O
I

***
_ Early dispersers
E:l Late dispersers

O)
O
l

 

01
O
I

A
O
1

(A)
O
J

 

N
O
l

 

Probability of surviving age interval

 

 

 

 

0 - 1 0 - 2
Age interval (years)
Figure 3. Percent of early and late dispersers surviving their first summer (to age
one) and to sexual maturity (age two). Data are based on 320 males, 230 early
and 90 late dispersers, from cohorts born in 1993 - 1997. Asterisks indicate

significant differences between groups at p < 0.001.

65

- Observed numbers
[:3 Expected numbers

 
  
 

No. of different mating partners

 

 

 

 

Early dispersers Late dispersers

Figure 4. Observed and expected mating partners of early- and late-dispersing
males. Expected values were calculated by assuming that early and late
dispersers mated with females in direct proportion to the number of males in

each dispersal age category observed mating.

66

3.0 -

2.5 .

2.0 -

Mean # of mating partners (i SE)
8 a:

 

 

Early dispersers Late dispersers

Figure 5. Mean number of mating partners (i SE) acquired by early and late
dispersing males during the 1998 mating season. Asterisk indicates significant

differences between groups at p < 0.05.

67

Table 1. Age-specific life table for early and late dispersing male S. beldingi,

based on disappearance from the population of members of cohorts born in 1993

 

 

 

-1997.
Early dispersers Late dispersers
(ytgfs) nx dx 'x qx 9x In dx 'x qx ex
0 230 183 1 0.511 1.627 90 38 1 0.769 2.169
1 47 24 0.204 0.478 0.979 52 40 0.575 0.500 0.650
2 23 11 0.100 0.500 1.083 12 6 0.133 0.500 1
3 12 0.052 0.833 1 6 3 0.066 0.333 1
4 6 0.026 0.500 0.600 3 1 0.033 0.750 1
5 1 0.004 1 2 2 0.022 1
6 0 0

 

68

CHAPTER FOUR

GROWTH LIMITATIONS AND AVOIDANCE OF RESIDENT
AGGRESSION INFLUENCE AGE AT DISPERSAL IN GROUND-DWELLING
SCIURIDS

INTRODUCTION
Natal dispersal, the permanent departure of an individual from its birth site, is an
important event inthe life history of most mammals and birds (Greenwood, 1980;
Lidicker, 1975). In natal dispersal, the disperser leaves behind the safety of a
familiar social and physical environment and ventures into unfamiliar terrain in its
search for a new home (lsbell and Van Vuren, 1996; Waser and Jones, 1983b).
This can be costly since the increase in locomotor activity associated with this
behavior can deplete a disperser‘s energy resources (Holekamp, 1986; Rood,
1987), and make the individual more vulnerable to predation (Metzgar, 1967).
Compounding this effect, the threat of attack by hostile resident conspecifics may
force a disperser to reduce foraging time in order to increase vigilant or defensive
behavior, or prevent an individual from immigrating into a new site in suitable
habitat (e.g., Garrett and Franklin, 1988). The prevalence of natal dispersal
among birds and mammals suggests that dispersal costs are offset by significant
fitness gains such as avoiding inbreeding and increasing reproductive
opportunities (Dobson, 1982; Lehmann and Perrin, 2003; Moore and All, 1984;
Pusey, 1987; Shields, 1982). Yet, the need to minimize dispersal costs relative to
fitness gains theoretically should have played a crucial role in shaping the

evolution of this behavior.

69

Evidence from a wide range of species suggests selection has favored
evolution of behavior patterns that minimize the risks associated with dispersal.
Frequently dispersers will use their birth site as a secure base from which they
undertake extensive exploratory forays prior to dispersal (wolves, Canis lupus,
Van Ballenberghe, 1983; Belding’s ground squirrels, Spennophilus beldingi,
Holekamp, 1986). In addition, dispersers may emigrate during periods when food
is abundant, when vegetation provides cover from predators (black-tailed prairie
dogs, Cynomys Iudovicianus, Garrett and Franklin, 1988), or when safe travel
routes are most likely to be available (beavers, Castor canadensis, Sun et al.,
2000). Individuals may also disperse at an age when they are able to effectively
counter aggression directed at them by local residents (savannah baboons,
Papio cynophalus, Alberts and Altmann, 1995) or at an age when they are likely
to confront low levels of aggression from local residents (Japanese macaques,
Macaca fuscata; reviewed by Sprague, 1992).

In mammals, dispersal tends to be male-biased; males usually disperse at
either higher rates or over longer distances than females (Greenwood, 1980;
Lidicker, 1975). Males of many mammalian species emigrate before they have
completed growth and maturation, but dispersal ages vary among species
(Dobson, 1982) suggesting that selection may have shaped this behavior to
occur at an age when an individual is best able to cope with the risks and
challenges of dispersal. Ground-dwelling sciurid rodents show the dispersal

patterns characteristic of most mammals, exhibit a range of dispersal ages, and,

70

as a consequence, are an ideal group in which to examine the role of selective
agents in shaping the evolution of dispersal age in mammals.

Ground-dwelling sciurid rodents includes the antelope squirrels
(Amospemrophilus spp.), ground squirrels (Spermophilus spp.), marmots
(Marmota spp.), and prairie dogs (Cynomys spp.) and, with the exception of
Amospermophilus, much is known about the life history of the members in this
group (Murie and Michener, 1984). All species share a common life style in that
they are all diurnal vegetarians, and most live in underground burrows (Nowak,
1999). Most species also hibernate during the winter or aestivate during the
summer months, during which time they rely primarily on stored fat as an energy
source. Females of most species produce only one litter each year. Mating
occurs at the onset of the active season, followed by a 3-5 week gestation
period. Litters emerge above ground towards the end of the lactation period,
when adults begin fattening in preparation for dormancy. In most species, male
dispersal occurs prior to first reproduction (see Blumstein and Armitage, 1998)
for exceptions). Although males of each species predictably emigrate at a
specific age, dispersal age varies significantly among sciurid species. That is,
males of some species emigrate as juveniles, during the year in which they are
born, whereas males of other species consistently tend to delay dispersal until
they are 1, 2, or 3 years of age (reviewed by Holekamp, 1984a; Blumstein and
Armitage, 1998; Holekamp, 1984a).

Ground-dwelling sciurids show also differences in a variety of life history,

ecological, and social factors. For example, adult body mass varies greatly

71

among ground—dwelling sciurids (reviewed by Armitage, 1981), ranging from 135
g (thirteen-lined ground squirrel, S. tridecemlineatus) to 5000 9 (yellow bellied
marmots, M. flaviventn’s). In addition, the duration of the active season varies
among species from 4.5 months (Olympic marmot, M. olympus) to 9 months
(round-tailed ground squirrel, S. tereticaudus), and one species, the black-tailed
prairie dog (C. Iudovicianus), is active throughout the year. Ground-dwelling
sciurids with short active seasons tend to live at high elevations or high latitudes,
where winters are long and severe. Species with longer active seasons are found
at lower elevations and in regions where climatic conditions are relatively mild.
Evidence from several species suggests that growth of immature sciurids is
limited by the duration of the active season and by adult body size because
juveniles must balance the amount of energy they allocate towards somatic
growth with the need to accumulate adequate fat stores to survive dormancy
(Barash, 1973; Webb, 1981; Armitage, 1981; Zammuto and Millar, 1985; Van
Vuren, 1990; Dobson, 1992; Arnold, 1993). Ground-dwelling sciurids occupy a
variety of different habitats ranging from vast expanses of continuous grassy
terrain (Columbian ground squirrel, S. columbianus; Arctic ground squirrel, S.
pan'yir) to regions in which suitable habitat patches are small and separated by
rocky outcroppings or mountains (M. flaviventn's, Armitage, 1981). The social
organization of ground-dwelling sciurids is also variable and ranges from solitary
(groundhog, M. monax) to complex stable social groups (M. Olympus, C.
Iudovicianus; reviewed by Michener, 1983). Depending on the social system,

adult males, adult females, or adults of both sexes defend territories against

72

intruding conspecifics during specific periods or throughout the active season.
Overall, it seems reasonable to expect that variation in life history, ecology, and
sociality have played important roles in shaping the evolution of dispersal age in
ground-dwelling sciurids.

Three hypotheses have been put fonivard to explain the variation observed
in the timing of natal dispersal among ground-dwelling sciurids. The growth
limitation hypothesis posits that males delay dispersal until they have acquired
sufficient energetic resources to support the demands of dispersal and
hibernation (Armitage, 1981; Barash, 1974). This hypothesis predicts that the
extent of growth males are able to complete during their first summer determines
dispersal age. The risk avoidance hypothesis states that males delay emigration
until they are older and better able to cope with the risks of dispersal, or until the
costs of remaining at the natal site exceed the costs of dispersing (reviewed by
(Arnold, 1993; Emlen, 1991 ). Travel over long distances and across unsuitable
habitat increases mortality risk for dispersers (Arnold, 1993; Van Vuren, 1990;
Webb, 1981). This hypothesis predicts that males should delay dispersal in
species where dispersers have to travel long distances to locate suitable
settlement sites. Lastly, the social resistance hypothesis suggests that
aggression by local residents creates a significant barrier to potential immigrants,
and dispersers that are unable to avoid or counter this aggression effectively may
be forced into marginal habitat where an individual’s prospects for survival and
reproduction are significantly reduced (Armitage, 1999; Garrett and Franklin,

1988). Since older males should be better able to counter aggression by local

73

residents, males from species in which immigrants are likely to be targets of
aggression should disperse at an older age than males from species in which

immigrants encounter little aggression.

The goal of this study was examine dispersal patterns of ground-dwelling
sciurids to further investigate factors affecting the evolution of dispersal age
among mammals. Specifically, I conducted a review of existing literature on
ground-dwelling sciurids to identify field studies documenting ecology and natural
history relevant to dispersal age. I combined these data with a phylogenetic
hypothesis developed for ground-dwelling sciurids (Herron et al. 2004) to

evaluate the growth limitation, risk avoidance and social resistance hypotheses.

METHODS
Data set
A thorough review of the ground dwelling sciurid literature was conducted. When
possible, variables were recorded from the identical study and/or study
populations to minimize variance in the data set due to differences in research
methods and/or variation in behavior and ecology among populations.

1. Dispersal age (wks). Dispersal age was determined by calculating the
number of weeks elapsing between the date pups were first observed
above ground and the date when males were initially observed emigrating
from their natal site. This measure was selected because number of days
between parturition and the date on which pups first emerge above ground

is unknown for many ground-dwelling sciurids. For 8. armatus, S. cite/Ius,

74

C. Ieucurus and S. franklinii authors reported males dispersed as juveniles
but did not specify the exact age. I assigned dispersal ages to these four
species assuming that males dispersed midway through their juvenile
active season. This was a reasonable assumption considering that in most
species where juvenile dispersal occurs males emigrate during the middle
of the species’ juvenile active phase (Byrom and Krebs, 1999; Holekamp,
1983; Olson and Home, 1998).

2. ggvenile maturity index (Ml; ). Growth juvenile males were able to
complete during their first active season relative to the species” adult body
size was calculated using the juvenile maturity index (Mlj) developed by
(Barash, 1973). This index provides a measure of the total somatic growth
occurring during the juvenile summer relative to a species’ body size at
maturity. Mlj was calculated by dividing the post-emergence body mass of
yearling males by the minimum adult male body mass. Minimum adult
body mass was determined using the early season body mass of adult
males when reported, or the minimum body mass of adult males from the
range of values provided (Armitage, 1981; Blumstein and Armitage, 1999).

3. Age at agglt body mass lwks). Following previous investigators (Armitage,
1981; Barash, 1973), I used age at adult body mass as an indicator of how
much time males require to complete growth and sexual maturation. I
defined the age at which males reached adult body mass to be the year in
which they had achieved minimum adult body mass by the onset of the

mating season that year.

75

4. Effective dispersal distance. The effective dispersal distance for each
species was calculated by dividing the mean dispersal distance of males
by the mean home range diameter reported for the same species. This
corrects for differences in home range size among species (Shields 1982).
Mean dispersal distance is the mean straight-line distance (In) between a
male”s natal burrow and a male”s settlement site after dispersal. Mean
home range diameter was defined as the mean diameter (d) of the home
range reported for individual females or social groups. When only the

mean home range size (m2) was reported, home range diameter was

calculated as follows: d = *1 (Area / 1t) * 2.

5. Social agression score. A social aggression score was determined for
each species based on levels of resident aggression during the dispersal
period. Dispersal period was defined as the time between the first and last
date males were observed emigrating from their natal area. To determine
levels of resident aggression, we noted when aggression was reported
during a species” active season, as well as the sex of the main
aggressor(s). Presence of territory defense, wounding patterns, and
aggressive interactions among conspecifics were used as indicators of
high levels of resident aggression; these events were previously reported
as good measures of resident aggression in ground-dwelling sciurids (e.g.,
Garrett and Franklin, 1988; Millesi et al., 1998; Sherman, 1976). A score
of ‘1’ was assigned if the dispersal period did not coincide with high levels

of resident aggression. If adult female, adult male, or adult male and

76

female aggression was reported, a score of ‘2,”‘3,’ or ‘4” was assigned,
respectively. Male aggression was scored higher than female aggression,
since male aggression is more intense and can result in serious injury or

death (Garrett and Franklin, 1988; Huber et al., 2002; Sherman, 1976).

Comparative methods

Closely related species may share common traits due to shared ancestry rather
than independent evolutionary events (reviewed in Martins, 2000). As a result,
comparative data are not statistically independent and violate one of the basic
assumptions of statistical tests (Felsenstein, 1985; Purvis and Rambaut, 1995).
To remove the issue of non-independence in data sets, comparative methods
have been developed which examine relationships among pairs of traits using
phylogenetic hypotheses to accounting for potential effects of shared
evolutionary history. To assess the relationships among pairs of behavioral traits
the independent contrast method of Felsenstein (1985) was implemented using
the Comparative Analysis of Independent Contrasts (CAIC, version 2.6.9, Purvis
and Rambaut, 1995). In CAIC, we selected the CRUNCH algorithm, which

generates phylogenetic independent contrasts (PICs) for continuous traits.

Ple assume a fully re-solved and accurate phylogeny and branch lengths
can be used as an estimate of expected variance in character evolution. l
approximate the phylogenetic relationships of the species for which I were able to

extract data from the literature using Herron et al.’s (Herron et al., 2004)

77

hypothesis (Fig. 3 in Herron et al., 2004) of phylogenetic relationships within the
sciurid family of rodents. Herron et al.’s (2004) hypothesis was generated using
Bayesian phylogenetic analysis based on mitochondrial cytochrome b sequences
from 114 species in 21 genera, and currently provides the best approximation of
the evolutionary relationships among species of ground-dwelling sciurids. l
constructed a modified phylogenetic hypothesis (Fig. 1) including the 22 species
from our study by creating a pruned tree constructed in McClade 4.06 (Maddison
and Maddison, 2000). To generate branch lengths, I linked this tree with a
matching data set of mitochondrial cytochrome b gene sequences from Herron et
al. (2004; see appendix for GenBank accession numbers). Branch lengths were
based on differences in base pair substitutions of the cytochrome b sequences.
Branch lengths were generated in PAUP 4.0.b10 (Swofford, 2002) with our
optimality criterion set to minimum evolution distance using maximum likelihood
parameters from the GTR + G + l model of Herron et al. (2004) and limited to

nonnegative branch length values (Fig. 1).

Following other comparative studies (Hagman and Forsman, 2003;
Merrnoz and Omelas, 2004; Summers and Clough, 2001), I examined the
relationships among pairs of related traits under three different models of
character evolution: the Brownian motion (BM) model, the punctuated equilibrium
(PE) model, and the star phylogeny model. The first two methods could be
utilized in CAIC. The BM model states that character changes occur along
branches between speciation events and variations among species are

contingent upon when species diverged from their common ancestor (Diaz-Uriate

78

and Garland, 1996; Felsenstein, 1985). For this model, Ple were calculated
using branch lengths as a proxy for time since evolutionary divergence. Since
CAIC does not accept branch lengths less than 2 units, we multiplied all branch
lengths by 10,000 and added 2 units to all branch lengths (Purvis and Rambault,
1995). The PE model of character evolution states that changes in characters are
associated with speciation events and, as a result, all character state changes
occur at the nodes of the phylogeny. For this model, PICs were calculated by
setting branch lengths to unity (Diaz-Uriarte and Garland, 1996). Lastly, the star
phylogeny assumes that taxa diverged rapidly from a common ancestor, and the
resulting phylogeny can be viewed as a star since all species radiate from a
single point and all branch lengths are equal (Martins, 2000). As a consequence,
species can be treated as independent units, and raw species” data were used to

examine the relationship among traits of interest.

Regression analyses

Ple generated under the BM and PE model and the raw data were used
to conduct regression analyses to address the predictions of the growth
limitation, risk avoidance, and social resistance hypotheses (Table 1). In these
analyses, I used dispersal age as the dependent variable and variables 2-5 as
the independent variables:

1) To test the growth limitation hypothesis, I examined the relationship
between dispersal age and two growth parameters, juvenile maturity index and

age at adult body mass. If growth limitations influence dispersal age, little total

79

growth during a male’s first summer should lead to a delay in dispersal resulting
in a negative relationship between dispersal age and juvenile maturity index. In
species' in which males complete growth early in life, males should be able to
disperse at a young age, resulting in a positive relationship between dispersal
age and age at adult body mass.

2) To test the risk avoidance hypothesis, I examined the relationship
between dispersal age and effective dispersal distance. I expected to observe a
positive relationship between these two variables since dispersal should be
delayed in species in which males have to travel great distance before they are
able to settle in a new place.

3) Lastly, I examine the social resistance hypothesis to assess the
relationship between dispersal age and social aggression score. I expected to
observe a positive relationship between these two variables since males should
delay dispersal in species where resident aggression is high during the dispersal
period and until they are older and better able to counter this aggression.

To examine these hypotheses simple and multiple regression analyses
were performed using independent contrasts generated in CAIC (BM and PE
model) and raw data (star phylogeny model). Simple linear regression analyses
for contrast data were performed in CAIC; simple linear regression analyses for
raw data and multiple regression analyses for contrast and raw data were
performed in STATISTICA for windows version 6.1 (StatSoft). For independent
contrast data, regression analyses were forced through the origin as required for

PICs (Garland et al., 1992). Prior to generating contrasts or analysis, data were

80

log-transformed (Mlj, age at adult body mass, and effective dispersal distance) or
square-root transformed (social aggression score) to best meet statistical
assumptions associated with regression analyses. Data were inspected for
significant outliers (studentized residuals greater than 3; Jones and Purvis,
1997), and homogeneity of variance with no significant violations. I did not
include effective dispersal distance in the multiple regression analyses since
inclusion of this variable in the model would have substantially reduced sample
sizes. Sample sizes refer to the number of species used in regression analyses;
these vary due to missing values of some variables for some species. Two-tailed
P - values, and results were considered significant if P < 0.05. Means are

reported i standard error.

RESULTS

I were able to determine dispersal ages of males for a total of 22 ground-
dwelling sciurids; including 6 of the 14 extant marmot species (Marmota spp.), 3
of the 5 extant prairie dog species (Cynomys spp.), and 13 of the 38 extant
ground squirrel species (Spermophilus spp.; Appendix). Dispersal ages ranged
from 3 to 152 wks (mean = 33.8 i 10.14 wks). I was able to calculate juvenile
maturity indices for 20 species, which ranged from 0.37 to 1 (mean = 0.74 i
0.052). Age at adult body mass varied from 35 to 146 wks (mean = 90.55, i 7.82
wks, n = 22), and effective dispersal distances ranged from 0.45 to 25.8 home
range diameters (mean = 6.16, 3: 2.39, n = 13). Dispersal period did not coincide

with resident aggression in 54 % (n = 12; aggression score one) of species, co-

81

occurred with adult female aggression in 5 % (n = 1; aggression score two), adult
male aggression in 18 % (n = 4; aggression score three), and adult male and
female aggression in 23 % (n = 5; aggression score four) of species.

Results of the bivariate regression analyses under the three evolutionary
models were summarized in Table 2. Contrasts were adequately standardized
except for effective dispersal distance in the PE model where I observed a
violation of the evolutionary assumption that standardized contrasts are
independent of the estimated value of the character for the node at which the
contrast was taken. As predicted by the growth limitation hypothesis, the amount
of growth relative to their adult body mass males were able to complete during
their juvenile summer (M'j), showed a significant negative effect on the age at
which males dispersed under all three evolutionary models. I observed a
significant positive relationship between age at adult body mass and dispersal
age under the star phylogeny, but not under the PE and BM evolutionary models.
The relationship between effective dispersal distance and dispersal age was
significant under the PE model, marginally significant under the star phylogeny,
and non-significant under the BM evolutionary model providing inconsistent
support for the risk avoidance hypothesis. I consistently observed the expected
positive relationship between a species’ social aggression score and dispersal
age under all three evolutionary models, supporting the prediction of the social
resistance hypothesis.

The results of the multiple regression analyses were consistent with those

from bivariate regression analyses (Table 3). Across all three models, MI, and

82

 

social aggression scores significantly predicted dispersal age even after
controlling for other variables in the model. Greater B coefficients for Mlj than
social aggression scores in these analyses indicate that Mlj contributes to a
greater extent to the observed variation in dispersal age.

In summary, our results provide compelling evidence that early growth
patterns and level of social resistance males encounter during the dispersal
process were influential in shaping dispersal age in ground-dwelling sciurids. The
relationship between effective dispersal distance and dispersal age were
inconsistent under alternative evolutionary models and evolutionary assumptions
of contrasts were violated under the PE model, making it difficult to evaluate the
risk avoidance hypothesis. Further work with larger sample sizes and different

measures of dispersal risk will be necessary to re-examine this hypothesis.

DISCUSSION

Our study provided does not allow us to reject the risk avoidance
hypothesis, but provides compelling support for the growth limitation and social
resistance hypothesis. No consistent relationship between a species’ effective
dispersal distance and dispersal age was detected. This result was unexpected
since Byrom and Krebs (1999) documented an increase in mortality risk
associated with greater dispersal distances in the Arctic ground squirrel, S.
parryii. My measure of dispersal risk may have been too crude or have failed to
encompass enough of the risks young males are exposed to during transit.

Habitat patchiness, availability of refugia, habitat familiarity, and duration of

83

transit may contribute significantly to risks males are exposed to during transit.
Also, dispersal events reported in the literature are often biased towards short
distance dispersal (Koenig et al., 1996) and dispersal distance is often measured
as a straight line while a disperser’s transit route rarely follows a direct pattern
(Boydston et al., in press; Wiggett et al., 1989). In addition, my results may be
due to insufficient sample sizes, making it difficult to detect a significant
relationship between the variables. More work will be necessary before I can rule
out the possibility that risk aversion has shaped survival of males and
consequently the evolution of dispersal age in ground-dwelling sciurids and other
mammals.

In support of the growth limitation hypothesis, a significant negative
relationship between juvenile male growth and dispersal age was observed. This
is consistent with the intraspecific variation in dispersal ages of males reported
for several ground-dwelling sciurids. For example, in the Belding”s (S. beldingi),
Columbian, and Richardson’s ground squirrel (S. richardsonii) exceptionally small
males disperse several months after most males emigrate (Festa-Bianchet and
King, 1984; Holekamp, 1986; Michener and Michener, 1977; Wiggett and Boag,
1991). In the black-tailed prairie dog (C. Iudovicianus), unusually large males
disperse several months before their same-aged, smaller counterparts (Garrett
and Franklin, 1988). Body mass and body fat is the most important predictor of
dispersal age in Belding”s ground squirrels, and males of this species only
emigrate as juveniles if they have reached a specific lean body mass/body fat

ratio early in the species” active season (Nunes et al., 1998). Overall, results of

84

interspecific studies are consistent with our findings implying that energy
availability determines optimum dispersal age for survival and reproduction.

I observed a positive relationship between a species’ dispersal age and
social aggression score in support of the social resistance hypothesis. Immigrant
males rarely encounter high levels of aggression in species in which males
emigrate when they are only a few weeks old. At this age, males are still
relatively small and immature, do not commit lnfanticide and, as a consequence,
do not pose a threat to adult male or female residents. Furthermore, these males
disperse after reproductive activities have ceased and adults are fattening for
hibernation. Over-winter survival in ground-dwelling sciurids is critically
dependent on the amount of energy stores individuals are able to acquire prior to
hibernation (Lenihan and Van Vuren, 1996). Therefore adults are less likely to
fend off potential immigrants while fattening. Dispersal at an early age does not
always preclude young males from encountering aggression by local residents.
Hibemacula are scarce in S. panyii, and adult male and female residents
aggressively defend their hibemacula toward the end of their active season from
other conspecifics, including juvenile immigrants (Carl, 1971). Evolution of
dispersal at a very young age allows males to avoid resident aggression and
increase survival in many but not all of the ground-dwelling sciurids included in
this study.

Delayed dispersal may be adaptive when temporal avoidance of resident
aggression is not possible because adults of one or both sexes defend territories

throughout a species’ active season. This is the case in most prairie dog (C.

85

Iudovicianus; C. gunnisonr) and marmot species (e.g., M. flaviventn’s; hoary
marmot, M. caligata; golden marmot, M. caudata aurea; M. mannota; Olympic
marmot, M. olympus) in which the resident male or all members of a social group
aggressively defend a common territory throughout the species‘ active season
(Armitage and Downhower, 1974; Barash, 1973; Blumstein and Arnold, 1998;
Garrett and Franklin, 1988; Hoogland, 1999; Michener, 1979a; Michener, 1979b).
In many of these species a male's immigration success depends on his ability to
displace a local male or to subsist in low quality areas until a habitat opening
becomes available upon the emigration or death of a local resident male (M.
flaviventn's, Armitage and Downhower, 1974; M. mannota, Arnold, 1990). It is
thus not surprising that males of these species generally disperse after they are
relatively large and therefore presumably better able to counter resident
aggression (Armitage 1981, 1999; Blumstein and Armitage, 1998) facilitating
immigration.

The patterns I detected illustrate the importance of a species’
developmental and social attributes in life history evolution. An important
assumption of all three hypotheses examined in this study is that males are able
to modulate their dispersal activity to maximize their probability of successful
dispersal and immigration. Yet, males may not be able to do so if local residents
evict young males from their natal site. In case of ground-dwelling sciurids, I can
be confident that males of most species are not forced out of their natal areas
since a considerable body of evidence has been gathered to reject this

hypothesis (reviewed in Holekamp, 1984a). Although aggression is known not to

86

 

“152;; '_-_ $2: I:

affect emigration in sciurids it apparently does affect immigration. However,
future studies involving other species will need to examine this hypothesis before
assessing the role of other selection pressures on shaping dispersal behavior.
The possible influence of social resistance on dispersal age is intriguing, and I
hope that future studies will examine the role of resident aggression in shaping

dispersal and immigration patterns.

87

M. mligala

M. olympus

 

M. flavivenln's

 

0.068

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F—w 0'046 M. caudata aurea
0 M. marmots
00“ M. monax
‘M' 0044 rm— 8. beecheyi
[ML S. Iateralis
0‘35 s. cite/[us
0'005 C. gunnisoni
0
C. leucums
C. Iudovicianus
S. tereticaudus
S franklinii
S. In'decemlineatus
'—‘ S. elegans
S. ridrardsonii
S. panyii
0.005 ' ' S. columbianus
0.169
0.041 0.011 I—ML S. armatus
[£36— 8. beldingi
“059 s. moi/is
HM. S. adocetus
]_0£Z5__ S. annulatus

Figure 1: Phylogenetic tree of 22 ground-dwelling sciurid species used in the
current study for comparative analyses. Numbers above branches indicate
branch lengths, which were used as expected units of evolutionary change

between speciation events for the Brownian motion model of character evolution.

88

Table 1. Proposed direction of relationship between predictor variables and

dispersal age based on three different hypotheses.

 

Proposed direction of

Hypothesrs PredIctor vanable relationship

Juvenile maturity
index (M '1)
Age at adult body +

mass
Effective dispersal +
distance
Social aggression +
score

Growth limitation

Risk avoidance

Social aggression

89

Table 2. Results of simple bivariate regression analyses of factors proposed to

influence dispersal age in ground-dwelling sciurids

 

Brownian motion model

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Term in model N b R Test statistics
Juvenile maturity F1,1g=10.30
index (M1,) 20 '2'73 0'36 P=0.005
A99 eggs" may 22 0.24 0.02 F1,21=0.54, P=0.470
Effective dispersal _
distance 13 0.32 0.17 Fug—2.21, P=0.165
S°°'a'sac%$;ess'°" 22 0.45 0.23 F,,2,=5.95, P=0.024
Punctuated equilibrium model
Term in model N b R Test statistics
Juvenile maturity F1,19=25.96,
index (M1,) 20 '3'" 0'59 P<0.001
Age atrj‘a‘gt km” 22 0.45 0.05 F1,21=1.09, P=0.309
Effective dispersal _
distance 13 0.57 0.35 Fug-6.01, P=0.032
S°°'a'sac%$;ess'°" 22 0.53 0.27 F1,21=7.28, P=0.013
Star phylogeny model
Term in model N b R Test statistics
Juvenile maturity F1,1g=28.54,
index (M1,) 20 '3'14 0'61 P<0.001
A99 at 3‘1”" W” 22 1.02 0.21 F1,21=5.28, P=0.030
mass
EffeCt'Ye d'Spe'sa' 13 0.69 0.29 F1,12=4.49, P=0.058
distance
3°95“ aggressm 22 0.94 0.43 F1,21=18.53,P<0.001
SCOI'e

 

N, number of species used to generate regression lines; b, slope of regression
line. Results of binary linear regression analyses using independent contrasts
that were generated using avarying branch lengths or bequal branch lengths, or

°raw species data; dregression was forced through the origin.

90

Table 3. Results of multiple regression analyses of factors proposed to

influence dispersal age in ground-dwelling sciurids.

 

Brownian motion model

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Term in model b SE I3 Test statistics
Juvenile maturity _ t,,16=-3.47
index (M1,) 2'99 0'36 ”0°66 P=0.003
Age at adult body t1,15=-1.37,
mass -0.38 0.28 -0.26 P=0.191
Social aggression t,,15=2.59,
score 0.43 0.17 0.44 P=0.020
Punctuated equilibrium model
Term in model 8 SE 8 Test statistics
Juvenile maturity t1,,5=-4.36,
index (M1,) ‘3'34 0°76 '0'68 P<0.001
Age at adult body t1,,5=-0.56,
mass -0.17 0.31 -0.09 P=0.586
Social aggression t1,15=2.06,
score 0.39 0.19 0.32 P=0.056
Star phylogeny model
Term in model b SE /3 Test statistics
Juvenile maturity t,,,6=-3.69,
index (M1,) '21s 0'59 '0'54 P=0.001
Age at adult body t,,,5=-0.50,
mass 015 0.30 -0.70 P=0.622
Social aggression t1, 161:3.81 ,
score 0.75 0.20 0.52 P=0.002

 

b, slope of regression line; SE, standard error of regression slope; [3, beta

coefficient of regression. Results of multiple regression analyses using

independent contrasts that were generated using avarying branch lengths or

bequal branch lengths, or °raw species data; dregression was forced through the

origin.

91

CHAPTER FIVE
CONCLUSION

To understand why an animal exhibits a specific behavior it is necessary
to examine the behavior from multiple perspectives. In 1961, Mayr proposed that
it is important to examine causal explanations in biology from a proximate and
ultimate level, where proximate factors influence the day-to-day life of an animal
and ultimate factors determine how a trait is passed on across many generations
and even across entire taxa. Expanding this framework, Tinbergen (1963)
proposed that on a proximate level, it is necessary to observe the development of
a behavior during the course of an individual’s ontogeny and to elucidate the
underlying physiological mechanisms that regulate a behavior. Furthermore, he
suggested that on the ultimate level it is not only important to assess the fitness
consequences of a behavior, but also to examine the behavior's evolutionary
origin. An important implication of this framework is that there are at least four
reasons why a specific behavior is observed. In order to fully understand a
behavior it is also necessary to address all four levels of the Mayr - Tinbergen
framework. In this chapter, I apply this general framework to address the
question: “Why do most Belding’s ground squirrel males disperse when they are
relatively young?"

The Belding’s ground squirrel, Spemiophilus beldingi, is a small diurnal
rodent. This species is active four months of the year, between May and August,
and hibemates during the remaining eight months 8. beldingits short active

season can be divided into two parts: breeding and post-breeding periods

92

(Sherman, 1976, Sherman, 1984). During the early part of the breeding period,
mating occurs and males aggressively compete for access to mates. Shortly after
mating, females establish territories, which they aggressively defend from other
conspecifics throughout the breeding season (Nunes et al., 2000). Females give
birth 25 days after having mated, and offspring remain in an underground burrow
for the first 25 - 28 days of their life (Holekamp, 1983). The post-breeding period
is marked by the emergence of juveniles from their natal burrow when
youngsters become nutritionally independent from their mother. Adults begin
fattening for hibernation shortly after juveniles emerge above ground.

As in most mammals, males are the predominant dispersing sex in S.
beldingi, and 70 — 90 % of them leave their natal area as juveniles at 7 -— 10
weeks of age, after having spent only 3 — 6 weeks above-ground (Holekamp,
1984b, Holekamp, 1986). The remaining 10 -— 30 % of males delay dispersal until
after their first hibernation as yearlings around 55 weeks of age. Since dispersal
is a dangerous process (Byrom and Krebs, 1999; Garrett and Franklin, 1988;
Larsen and Boutin, 1994) predominant juvenile dispersal by S. beldingi males
seems to be a paradox since young males should be more vulnerable to the risks
associated with dispersal than males that are older and more experienced. So,
why do most S. beldingi males disperse as juveniles?

Kay Holekamp and her student Scott Nunes have shed light on this
question by examining the proximate factors that regulate dispersal behavior in
S. beldingi. Holekamp (1986)’s work illustrated that a male’s dispersal age is

contingent upon reaching a threshold body mass during a male’s first active

93

season. Males that reached this threshold during their first summer dispersed as
juveniles, whereas males that were still relatively small at the end of their first
active season delayed dispersal until after their first hibernation. Nunes
expanded on Holekamp’s work and revealed that males must not only reach a
threshold body mass but, more specifically, they must acquire a minimum lean
body mass / body fat ratio prior to dispersal (Nunes and Holekamp, 1996, Nunes,
1998). He also found that the magnitude of this body mass / body fat threshold
varied seasonally, with pre-dispersal body fat requirements being low early in the
juvenile”s active season and increasing gradually as the season progressed.
Males able to reach seasonally appropriate body fat levels early in the active
season dispersed as juveniles. Males that reached this threshold towards the
end of the season delayed dispersal until after their first hibernation. Holekamp
(1986) and Nunes et al. (1998) proposed that this body mass / body fat threshold
functions to prevent the dispersal of undersized males who would have little
chance of surviving the energetic demands of dispersal and hibernation. As a
result, dispersal is inhibited at the end of the active season since survival
prospects would be slim for dispersing males, regardless of the magnitude of a
male’s fat stores.

To elucidate the underlying physiological mechanism associated with the
onset of dispersal, Holekamp et al. (1985) and Nunes et al. (1999) turned to
examining the role of hormones in regulating dispersal behavior. Previous work
with other mammals illustrated the importance of pre-natal testosterone in

generating sexual dimorphic behaviors (e.g., Phoenix et al., 1959). Since

94

dispersal is male-biased in S. beldingi, it was therefore reasonable to assume
that testosterone may play a similar role in shaping the dispersal behavior of S.
beldingi. To test this hypothesis, Nunes et al. (1999) treated neonatal females
with testosterone or oil and compared their dispersal behavior to that of untreated
males. Their results showed that testosterone-treated, but not oil-treated females
exhibited the male-typical dispersal behavior. This result supports the hypothesis
that prenatal testosterone is responsible for organizing the neural circuitry
necessary to regulate the dispersal behavior of male 8. beldingi.

In summary, Holekamp and Nunes' work revealed that dispersal in S.
beldingi is associated with a change in a male‘s internal environment, the
acquisition of a threshold lean body mass/body fat ratio, during the course of an
individual’s ontogeny. Arrival at this threshold acts apparently on the neural
circuitry shaped by prenatal exposure to testosterone, which triggers the onset of
a suite of dispersal-related behaviors, including increased locomotor behavior,
reduced fearfulness, and dispersal behavior itself (Holekamp, 1986). Hence, their
work elucidated the factors that determine dispersal age of male 8. beldingi at
both proximate levels of the Mayr - Tinbergen framework. However, in order to
fully explain this behavior it is necessary to reveal the ultimate factors
responsible for shaping the observed dispersal dichotomy as well. The objective
of this dissertation was to examine the ultimate determinants of dispersal age of
male 8. beldingi.

Evidence from the literature suggests that aggression by local residents

towards intruders may prevent the immigration of dispersers, marginalize

95

individuals from resources, and lead to the eviction of recent immigrants (see
references in Chapter 2). Furthermore, immigrants exposed to high levels of
resident aggression may suffer injuries and are forced into marginal habitat
where their survival prospects are slim. Social resistance toward immigrants may
therefore have profound fitness consequences for a disperser. As a result,
selection should favor dispersal at an age when immigrants pose little threat to
local residents, and when the level of aggression directed towards intruders is
low.

In the second chapter of this dissertation, I set out own to determine if
dispersal age influenced the level of social resistance males encounter during the
dispersal process. Differences in the social context of dispersal based on
dispersal age suggested that this might be the case. Juvenile dispersers
emigrate during a time period when females cease to defend territories and when
both males and females are fattening for hibernation. In contrast, yearling
dispersers emigrate during the breeding season, when intra- and inter-specific
aggression is at its peak. Hence, juvenile dispersal in this species may be
adaptive because it reduces the level of aggression males encounter during the
immigration process. Using staged- and naturally—occurring encounters, I tested
the social resistance hypothesis by examining the responses of adult female and
adult male residents towards intruding juvenile and yearling males. I found that
resident females aggressively targeted yearling males, while juvenile intruders
encountered little female aggression. Adult males, however, direct very little

aggression towards juvenile and yearling male intruders. These results indicate

96

that early dispersal is adaptive by allowing juvenile immigrants to avoid
aggression from female residents more effectively than do males who disperse
as yearlings.

The persistent bias toward juvenile dispersal also suggests that early
dispersal imparts significant survival or reproductive benefits to males that are
able to disperse as juveniles. In chapter three, I asked whether dispersal at a
young age enhances fitness in male Belding’s ground squirrels by increasing
male survivorship or mating success. I used live-trapping data to compare the
survivorship among males of known dispersal age, and conducted mating
observations to assess how dispersal age influenced male mating success. I
found that age-specific survivorship differed significantly between early and late
dispersers, which was primarily a result of lower juvenile survivorship among
males that dispersed as juveniles than for males that delayed dispersal until their
yearling summer. My data also showed that juvenile dispersers mated with a
greater number of different females than did yearling dispersers. Overall, these
results indicate that juvenile dispersal may carry significant survival costs early in
life, but ultimately enhances the mating success of surviving males.

To examine the evolutionary origins of dispersal age, I assessed the role
of various selective agents in shaping dispersal age among multiple species of
ground-dwelling sciurids, including the Belding”s ground squirrel (Chapter four).
Three hypotheses have been invoked to explain variation in dispersal age among
species of this group (see references in Chapter four). The growth limitation

hypothesis posits that growth limitations imposed by life history and

97

environmental factors determine dispersal age of ground-dwelling sciurids, and
males that are unable to complete sufficient growth during their first active
season should delay dispersal until they are older and have acquired sufficient
energy resources to support the energetic demands of dispersal. Aggression by
local residents can influence immigration success, and the social resistance
hypothesis suggests that the need to avoid or counter this aggression may have
acted as an important selective agent in shaping dispersal age in sciurids. This
hypothesis predicts that, in species where males are likely to encounter social
resistance during the immigration and settlement process, individuals should
disperse at an older age. Lastly, mortality during transit and early settlement can
be high for dispersers, and the risk avoidance hypothesis suggests that risks
associated with dispersal may affect the age at which individuals emigrate. This
hypothesis predicts that the level of dispersal risk encountered by males of a
particular species during dispersal should influence a male’s dispersal age. In
species where males are exposed to high risks, males should therefore disperse
at older ages than in species where males encounter relatively low levels of risks
during emigration.

I reviewed the literature on ground-dwelling sciurids for field studies that
documented dispersal age and used the extracted data from 22 species to test
the predictions of these hypotheses. I found that, relative to adult body mass, the
amount of growth males were able to achieve during their juvenile summer had a
significant negative effect on the age at which males dispersed, which supports

the growth limitation hypothesis. I also observed a positive relationship between

98

the level of social resistance immigrants were likely to encounter and a species’
dispersal age, a result that supports the social resistance hypothesis. However, I
found no significant relationship between mean dispersal distance, my measure
of dispersal risk, and the age at which a particular species generally disperses.
This result suggests that a species’ dispersal age is independent of the level of
risk a male encounters during the dispersal process. However, future work will be
necessary to confirm the latter result.

In summary, work by Holekamp and Nunes and the data presented in this
dissertation helped to elucidate the factors that determine dispersal age in S.
beldingi across all four levels of analysis in the Mayr - Tinbergen framework.

Studies by Holekamp and Nunes unraveled the ontogenetic basis and
physiological mechanisms involved in shaping the observed dispersal dichotomy.
They revealed that three factors: body fat, time of year, and testosterone are
responsible for shaping the observed dispersal pattern of male 8. beldingi on a
proximate level. My work has focused on examining explanations on the ultimate
level by assessing the potential fitness consequences of dispersal age (Chapters
two and three) and by determining its evolutionary origin among species closely
related to S. beldingi (Chapter four). The results presented in chapters two and
three suggest that dispersal as a juvenile carries significant fitness benefits by
reducing the level of social resistance a male experiences during the immigration
process and enhancing male mating success. The comparative study (chapter
four) illustrated that juvenile growth and resident aggression have been important

selection forces in shaping dispersal age among ground-dwelling sciurids.

99

.commmm 328m 9:52... :2: .3395 332:. 8.25ng
8.2: ES mEEsmwm nocgmmm 20>, meow .mmLchE ncm 60mm .8536 693 2:0QO Soc En 205:5L
x 25.. nno>>v $6QO £5 E .3536 5:52... vetoed. 253m wees x 53 6me ..ooomgv 86QO
£5 E .8536 2:52: 52.58. 552m mcoe .3535 mmmE Eon gave E:E_:_E some. mm_mE 52:3

Cm one». ”xva EEBmE 2:52....“ xmxmogv 52m :32. :2: E0: 52536 8.2: Go excomm 5033 um 0W5

 

 

 

 

 

 

 

S2 .222
new seesaw ”B2 .588 68mins
new 59255 ”59. .EmfiEEm v we N3 mmSm mambzmo .E
52 .856: ”$2 88mins
.52 .593 ”22 .593 m -- S; assess S
82
...n as 888 ”22 92:9“.
pen seam ”Sow .82
.82 $2 $2 .2280: SEEKS
”82 .952 can 5:292 e emu 8 2:99:85 .0
meow
.52 .2980: ”t2 {so 85?. is
”com—- ...mcszomr— Dam coammzh _. 1.. ...v «5.5039 .0
moo~.3n
5 92m: ”Bow .58 £2
5558152 .32 $2
.298”. ”E2 $55.53 see SEEKS
29.85:. ”3.2 EEBB m am 9. 2:82:55. .o
1.-.-. C883: --
9on cocfiwfi 35.5 mme :o_mmooom
Coimemmm .mmLonE Eon gape xcmmcowv
woococoaom _m_oow o>=ootm um 39? 52535.... moEm: $6QO
.... 1 1 .mmmzmcm

52:23:50 .8 new: 2322: 258m 9: So: $33.? Go meEsm new moon 59035 .x_ozmdn_<

100

...00000 02.00 0.602... E0... 29.9...

ESEE 602806 00.0.: 6... mEEammm 626.000 0.02. 0000 60.2.06 new dawn 620006 .0008 3.00%
.e: 2e 905%. n. 52 new)... 8.88 2... e. 628% 9.82.. 8:89. .63... 2.9.. u. 52 .22 __eeee>>.
00.00% 0.5 E 60.9.06 0.602...— 60toace 55:0 0.5e A003... 000E >603 ....60 2.3566. .600. 006E
50...; .m 0m<o ”x06... ESEE 0.605..me0605 2.0 .90: E05 E0: 00.0006 006E 6 $8M 20.52. .m 09..m

 

mm? .56.). Home
...mEmm ”mam. .coton. ”Em.

 

.5952 ”52 ..s a 525 $8332.
”vow? £02.05 ”mom. 600.55 F -I no? mwd v 320:. .2
Bow

...m .0 2.000000 ”mam? .665...
see .8528: ”82 ...n
.e 8.92 .32 ..n a 282

 

.33 ..Esz can 66:2. ”mmmr CEOOFE<V
.nommw .mommw .wwmw .U_OE< v —..m v.3. mmd n.3, 305ng .‘<
mam?

...m .0 Emzcom uommw E05>
cm> new? .0mm...E< 6cm c05>
cm> ”mm? .252. 6:0 009E?

...»an .c006:0>m 6an ...m
.0 009E}. ”3.? ..0306260

 

En BEE... ”82 6.8.3.32.
-11 .52 .mnm. ....an .000..E.< m 0mm 3.. 5.0 8.. 05:023.. .3.
C095...
0.000 00:30.6 60.2.. 000... .00.)... 5.000000
8.00230 60.0506 60.. :60 0mm xcmmcmwv
000:0.06m 6.00m 0>..00tw .0 090.0 ..s... 60.0%.? moEmc 00.00....0.

 

835.80 .50sz“?

101

600000 0>=00 0__c0>:._ :05“ £9.95
>026? 000.0006 00.9.: 00.: 90.25000 0009000 0.02, 00mm 00.03% 0:0 .0000 .0000006 6005 @0000
00: 20 0.05:0 .3me 000>>V 000.000 0:: E _00._000_0 0__c0>:._ 0000000.. >030 0:00 AS? 009 =0000>>v
008000 0:: 5 .0000006 0__c0>:_ 00:000.. .0530 0:0.u .Amxzc 000E >000 “.300 EEEEE 5000. 00.9:
4.20.52, 00 095. x005 3:308 0:525 A0x00>>v 0:0 .000: :05 E0: 000006 00.08 00 $3M :05; 0.0 005

 

00910 #0 00:32

”mam? ...0 00 00:32 ”003

...0 00 00:22 Home .QE0x0_0I
0:0 00:32 Home .0E0x0_0I
#02 .3306: #02 50:22

 

 

 

 

0:0 :0E00cw ӣ09 .059 Armmnmi/b

£08.05 ”03? .._0 “0 8:02 A we 00 00.0 0 55300 .m.
002 00520 new £92.80
5me .0052 0:0 9:00.06...
new? .0300 0:0 c0800 H000..

.22 5880 ”$2 .600. 8550;:
0va 00:50.0... 0:0 0c0>m A mN mm 0.0 .0 305000 .0.
002 .5020 ”32 €00 0cm A520E$
000a ”mom? 000.0% 0:0 20.0m 0 -- mm -- 00 050800 .m:

Ammrrrwm¢q
0,02 .5900 ”22 .6900 v -- m: «to 00 03:50 .2 .
C0053:
0.000 0000006 A0033 000E A0035 5000000
5000500 .0000005 >000 .300 000 xc0mc00v
008000001 _0_00m 0200005 00 090.0 £20 .00000050 0080: 00625

 

82:80 .x_0zma..._<

102

:00000 0500 0:52... ..05 0030.5
030:: 000.0000 00_0E 505 050.3000 00:0.000 0.02, 0000 _00.000_0 0:0 .0000 _00.000_0 50000 @0000

.9. 00 0.9.50. ”A .02 00020 000000 25 5 00.00% 2.85.. 09.000. .630 08.. ”A 202 .002 .0085
00.0000 05 :_ _00.000_0 0__:0>:.A 000000. .0530 0:00 .A0xzc 000:. >000 0:00 5255:. 0000. 000:.

00.0.5 50 0,00 000:. >530... 0__:0>:._0 ”A3005 0:0 .050: ._05 505 00.0005 00.05 00 0000M 00.02, 50 00.0.m

 

 

 

 

 

 

002

.._0 .0 000:8. .002 8:00 A000 is

.002 00500 .002 020802 2 -- .0 00.0 0 0.00.0.0. .0

:: 800.2040

0.2 0.52 F -- 00 -- 0. 050.20.. .0

- : 200.203

0002 .0002 00.0.0 F 0.0 00 F 0 8000.0 .0
.02 .002 000002, 000
:5 .002 .952 .002 .080
.002 .002 .0000 0:0 000005
.002 005.002, ”N02
.8080 .002 .._0 .0 0000s)
:02 .05. 0:0 0060.00.00“.

202 .952 0:0 080 002 000.203

.952 ”0.2 00.0: 0:0 0:22 F 0.2 00 00.0 00.. 020.0858 .0
00000.00000

.0 .0 50500: 002 .002 00.2.3

.._m .0 00:3 .0000 .._0 .0 .000: F -- 20 0.0 0... 02.0.0 .0

. A.00E::
0.000 00:90.0 A0035 000E A0005 8000000
8002000 000000 >000 2:00 000 0.:0mc0ov
000:0.0000 .0600 0500.5 50 000.0 520 00.0005,m 02.0: 006000

 

00:03:00 .X_Dzmn_0_<

103

.000000 0260 0:002... 005 0000.5

>030_E 000.0006 00_0E 000.. 006.5000 0000600 0.03 0000 620090 000 .0000 60.0006 60x0 36000
000 20 0.0500. .3002 000>>v 006.000 0:: 0_ 000.0006 0:002... 000000. >030 0000 .3000. 0000 =0000>>v
006000 0.7: 0_ 60.806 0.602... 000000. .0500 0000 A005 000:. >000 0:00 05:00:: 0000. 00.20
00:05 00 000.0. .0005 30300.. 0:00;? .A0x0020 000 .0000 005 E0: 00.0000 020:. 00 0000M 007.3 00 00.0.0

 

00%: .0002 0.2030

 

 

 

 

”Q02 £02,900.- 000 00_00>0m F -- 00 F 0 000000.000. .m
0002 000.65.
000 .0000002 ”0000 000060..
000 .0000002 ”NSF 000025
000 00000:). ”:02 0000022 Am F05 ES
000 0000002 ”Q0? 00000002 _. we 00? 5.0 0 000000000... .0
room .x0.0~00_>> 000 >000._
002 .0000. 0:0 5030 002
.00E0m 000 xozm K002 ...0
.0 >800 002 .002 0000.2 £020 is
i.- U0>0F 000.65. ”:02 .000 v 0.0 00 00.0 v .3000 .0,
Bow
.._0 00 0:900. H0002 050:
0:0 800 002 .._0 .0 050: A020 is
-.. - U000— 000000., 000 000.0 F NF 00 F v 0.50:. .m
. : - s.- A.00E:0 .
0.000 0000006 A0003 000E A0003 00600000
00600.000 .00.000_0 >000 0:00 000 0.0200000
3:3- :- 00000.000m_ _06om 0260.5 00 00.0..v 52.. 00.00050 . 000.00 00600w

 

 

00:03:00 .Xazmanz

104

.cOmmmw m>zom m__cm>:_ .65 20:25
3328 nofimamfi mm_mE “m5 mc_E:mmm nmcgmwm 9m; 3% Remamfi new .wmmm .3536 Umxm >20QO
8: 2c 9053. x 53 225 868m $5 5 3982“. 2E9... 8:89 225 m8... x 52 $2 =monm>a
36QO £5 5 $9036 2:52... 8:89 55.6 mcou .Amxzc mmmE >309 228 E:E_:_E comm: meE
:02; um mm<o ”xmnE 3:39: m__cm>:r x9693 mam Em: :9: EC: @9636 3:9: 5 $8M 22;; “m m9?

 

 

 

 

 

 

 

cmfimis
- .......... -- -- -- -- mamsgm .m
AvvamFu¢Q
,. i- .-.;.t- -- -- -- -- 3388 m.
39
umvtmn. ucm .ngmmazow
mm? 6me .8580;
cam 90532an ӣ2 EEm Es
.ncmbmé ”mm? £39m new m2mm§EmombE
-. .mEEmmEom ”82 $2.80: P 3 3 F v i- .m
, c883:
98m mocmumfi A933 89: 3x5 563on
.5685? 73535 >82 gsum mom xcmmcmwv
é... mmocmfifim _m_oow @26th um 35 _=>_p .3835? 8.5: $6QO

 

3:528 .xazmmnz

105

LITERATURE CITED

Adkins-Regan E, 1987. Sexual differentiation in birds. Trends in Neuroscience
10:517-522.

Agrell J, Wolff JO, Ylonen H, 1998. Counter-strategies to infanticide in mammals:
costs and consequences. Oikos 83:507-517.

Alberts SC, Altmann J, 1995. Balancing costs and opportunities: Dispersal in
male baboons. American Naturalist 145:279-306.

Allaine D, Graziani L, Coulon J, 1998. Post-weaning mass gain in juvenile alpine
marmots, Marmota man'nota. Oecologia 113:370-376.

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

Antolin MF, Horne BV, Michael D. Berger J, Holloway AK, Roach JL, Ronald D.
Weeks J, 2001. Effective population size and genetic structure of a Piute
ground squirrel (Spermophilus mollis) population. Canadian Journal of
Zoology 79:26-34.

Armitage KB, 1974. Male behavior and territoriality in yellow-bellied marmots.
Journal of Zoology 172:233-265.

Armitage KB, 1975. Social behavior and population dynamics of marmots. Oikos
26:341-354.

Armitage KB, 1981. Sociality as a life history tactic of ground squirrels. Oecologia
48:36-49.

Armitage KB, 1991. Social and population dynamics of yellow bellied marmots :
Results from long-term research. Annual Review of Ecology and
Systematics 22:379-407.

Armitage KB, 1998. Reproductive strategies of yellow-bellied marmots: Energy
conservation and differences between the sexes. Journal of Mammalogy
79:385-393.

Armitage KB, 1999. Evolution of sociality in marmots. Journal of Mammalogy
8021-10.

Armitage KB, Downhower JF, 1974. Demography of yellow-bellied marmot
populations. Ecology 55:1233-1245.

106

Armitage KB, Downhower JF, Svendsen GE, 1976. Seasonal changes in weights
of marmots. American Midland Naturalist 96:36-51.

Amitage KB, Johns DW, 1982. Kinship, reproductive strategies and social
dynamics of yellow-bellied marmots. Behavioral Ecology and Sociobiology
11:55-63.

Arnold W, 1988. Social thermoregulation during hibernation in alpine marmots
(Marmota marmota). Journal of Comparative Physiology B-Biochemical
Systemic and Environmental Physiology 158:151-156.

Arnold W, 1990a. The Evolution of marmot sociality: 1. Why disperse late?
Behavioral Ecology and Sociobiology 27:229-237.

Arnold W, 1990b. The Evolution of marmot sociality: 2. Costs and benefits of joint
hibernation. Behavioral Ecology and Sociobiology 27:239-246.

Arnold W, 1993. Social evolution in marmots and the adaptive value of joint
hibernation. Verhandlungen der Deutschen Zoologischen Gesellschaft
86:79-93.

Arnold W, Dittami J, 1997. Reproductive suppression in male alpine marmots.
Animal Behaviour 53:53-66.

Arnold W, Klinkicht M, Rassmann K, Tautz D, 1994. Molecular analysis of the
mating system of alpine marmots (Marmota marmota). Verhandlungen der
Deutschen Zoologischen Gesellschaft 87:27.

Baker AJ, Dietz JM, 1996. Immigration in wild groups of golden lion tamarins
(Leontopithecus rosalia). American Journal of Primatology 38:47-56.

Bakker J, Vanophemert J, Eijskoot F, Slob AK, 1994. A Semiautomated test
apparatus for studying partner preference behavior in the rat. Physiology
& Behavior 56:597-601.

Balph DF, Stokes AW, 1963. On the ethology of a population of Uinta ground
squirrels. American Midland Naturalist 69:106-126.

Barash DP, 1973. The social biology of the Olympic marmot. Animal Behavior
Monograph 6:173-245.

Barash DP, 1974. The Evolution of marmot societies: A general theory. Science
185:415-420.

107

 

Barash DP, 1975. Ecology of paternal behavior in the hoary marmot (Marmota
caligata): An evolutionary interpretation. Journal of Mammalogy 56:613-
618.

Barash DP, 1981. Mate guarding and gallivanting by male hoary marmots
(Marmota caligata). Behavioral Ecology and Sociobiology 92187-193.

Barash DP, 1989. Marrnots: Social behavior and ecology. Stanford: Stanford
University Press.

Barbraud C, Johnson AR, Bertault G, 2003. Phenotypic correlates of post-
fledging dispersal in a population of greater flamingos: The importance of
body condition. Journal of Animal Ecology 72:246-257.

Barnes BM, 1984. Influence of energy stores on activation of reproductive
function in male golden-mantel ground squirrels. Journal of Comparative
Physiology B-Biochemical Systemic and Environmental Physiology
154:421-425.

Blumstein DT, 1997. lnfanticide among golden marmots (Marmota caudata
aurea). Ethology, Ecology, and Evolution 9:169-173.

Blumstein DT, Armitage KB, 1998. Life history consequences of social
complexity: a comparative study of ground-dwelling sciurids. Behavioral
Ecology 9:8-19.

Blumstein DT, Armitage KB, 1999. Cooperative breeding in marmots. Oikos
84:369-382.

Blumstein DT, Arnold W, 1998. Ecology and social behavior of golden marmots
(Marmota caudata aurea). Journal of Mammalogy 79:873-886.

Blumstein DT, Foggin JM, 1997. Effects of vegetative variation on weaning
success, overwinter survival, and social group density in golden marmots
(Marmota caudata aurea). Journal of Zoology 243:57-69.

Boag DA, Murie JO, 1981. Weight in relation to sex, age, and season in
Columbian ground squirrels (Sciuridae, Rodentia). Canadian Journal of
Zoology 59:999-1004.

Boellstorff DE, Owings DH, 1995. Home range, population structure, and spatial-

organization of California ground-squirrels. Journal of Mammalogy 76:551-
561.

108

Boellstorff DE, Owings DH, Penedo MCT, Hersek MJ, 1994. Reproductive-
behavior and multiple paternity of California ground squirrels. Animal
Behaviour 47: 1 057-1 064.

Boydston EE, Kapheim KM, Horn RCV, Smale L, Holekamp KE, in press. Sexual
dimorphic patterns of space use throughout ontogeny in the spotted
hyaena (Crocuta crocuta). Journal of ZoologY. London.

Brandt CA, 1992. Social factors in immigration and emigration. In: Animal
dispersal (Stenseth NC, Lidicker WZ, eds): Chapman and Hall; 96-141.

Brockelman WY, Reichard U, Treesucon U, Raemaekers JJ, 1998. Dispersal,
pair formation and social structure in gibbons (Hylobates Iar). Behavioral
Ecology and Sociobiology 42:329-339.

Bronson FH, 1963. Some correlates of interaction rate in natural populations of
woodchucks. Ecology 44:637-643.

Bronson FH, 1964. Agonistic behaviour in woodchucks. Animal Behaviour
12:470-478.

Buck CL, Barnes BM, 19993. Annual cycle of body composition and hibernation
in free-living arctic ground squirrels. Journal of Mammalogy 80:430-442.

Buck CL, Barnes BM, 1999b. Temperatures of hibemacula and changes in body
composition of arctic ground squirrels over winter. Journal of Mammalogy
80:1264-1276.

Byrom AE, Krebs CJ, 1999. Natal dispersal of juvenile arctic ground squirrels in
the boreal forest. Canadian Journal of Zoology 77:1048-1059.

Carl EA, 1971. Population control in Arctic ground squirrels. Ecology 52:395-413.

Caro JF, Sinha MK, Kolazynski JW, Zhang PL, Considine RV, 1996. Leptin: the
tale of an obesity gene. Diabetes 45.

Cheney DL, 1983. Proximate and ultimate factors related to the distribution of
male migration. In: Primate social relationships: an integrated approach
(Hinde RA, ed). Sunderland, Mass: Sinauer Associates; 241-294.

Clark TW, 1977. Ecology and ethology of the white-tailed prairie dog: Milwaukee
Public Museum.

Cockbum A, Scott MP, Scotts DJ, 1985. Inbreeding avoidance and male-biased

natal dispersal in Antechinus spp. (Marsupialia: Dasyuridae). Animal
Behaviour 33:908-915.

109

Colbert J, Danchin E, Dondt A, Nichols J, 2001. Dispersal. Oxfrod, United
Kingdom: Oxford University Press.

Crockett CM, Sekulic R, 1984. lnfanticide in red howler monkeys (Alouatta
seniculus). In: lnfanticide: Comparative and evolutionary perspectives
(Hausfater G, Hrdy SB, eds). New York: Hawthome.

Deevey ES, Jr., 1947. Life tables for natural populations of animals. Quarterly
Review of Biology 22:283-314.

Diaz-Uriate R, Garland T, Jr., 1996. Testing hypotheses of correlated evolution
using phylogenetically independent contrasts: Sensitivity to deviations
from Brownian motion. Systematic Biology 45:27-47.

Dobson FS, 1979. An experimental study of dispersal in the California ground
squirrel. Ecology 60:1103-1109.

Dobson FS, 1982. Competition for mates and predominant juvenile male
dispersal in mammals. Animal Behaviour 30:1183-1 192.

Dobson FS, 1983. Agonism and territoriality in the California ground squirrel.
Journal of Mammalogy 64:218-225.

Dobson FS, 1992. Body mass, structural size, and life history patterns of the
Columbian ground squirrel. American Naturalist 140:109-125.

Dobson FS, Chesser RK, Hoogland JL, Sugg DW, Foltz DW, 1998. Breeding
groups and gene dynamics in a socially structured population of prairie
dogs. Journal of Mammalogy 79:671-680.

Dobson FS, Davis DE, 1986. Hibernation and solicality in the California ground
squirrel. Journal of Mammalogy 67:416-421.

Doolan SP, Macdonald DW, 1996. Dispersal and extra-territorial prospecting by
slender-tailed meerkats (Sun'cata sun'catfa) in the south-westem Kalahari.
Journal of Zoology,_London 240259-73.

Drews C, 1996. Contexts and patterns of injuries in free-ranging male baboons
(Papio cynocephalus). Behaviour 133:443-474.

Dufty AM, Jr., Belthoff JR, 2001. Proximate mechanisms of natal dispersal: the
role of body condition and hormones. In: Dispersal (Colbert J, Danchin E,
Dhondt AA, Nichols JD, eds). Oxford: Oxford University Press; 217-229.

Dunford C, 1977a. Behavioral limitation of round-tailed ground squirrel density.
Ecology 58:1254-1268.

110

Dunford C, 1977b. Social System of round-tailed ground squirrels. Animal
Behaviour 25:885-906.

Ellsworth EA, Belthoff JR, 1999. Effects of social status on the dispersal
behaviour of juvenile western screech-owls. Animal Behaviour 57:883-
892.

Emlen ST, 1991. Evolution of cooperative breeding in birds and mammals. In:
Behavioral Ecology: An evolutionary approach (Krebs JR, Davies NB,
eds). Oxford: Blackwell Scientific Publications; 301-337.

Evans FC, Holdenried R, 1943. A population study of the Beechey ground
squirrel in central California. Journal of Mammalogy 24:231-260.

Felsenstein J, 1985. Phylogenies and the comparative method. American
Naturalist 12521-15.

Festa-Bianchet M, King WJ, 1984. Behavior and dispersal of yearling Columbian
ground squirrels. Canadian Journal of Zoology 62:161-167.

Fitch HS, 1948. Ecology of the California ground squirrel on grazing land.
American Midland Naturalist 39:513-596.

Fitzgerald JP, Lechleitner RR, 1974. Observations on the biology of Gunnison's
prairie dog in central Colorado. American Midland Naturalist 92:146-163.

Foltz DW, Schwagmeyer PL, 1989. Sperm competition in the thirteen-lined
ground squirrel: differential fertilization success under field conditions.
American Naturalist 133:257-265.

French JA, Schaffner CM, Shepherd RE, Miller ME, 1995. Familiarity with
intruders modulates agonism towards outgroup conspecifics in Wieds
black-tufted-ear marrnoset (Callithrix kuhli, Primates, Callitrichidae).
Ethology 99:24-38.

Garland T, Jr., Harvey PH, lves AR, 1992. Procedures for the analysis of
comparative data using phylogenetically independent contrasts.
Systematic Biology 41:18-32.

Garrett MG, Franklin WL, 1988. Behavioral ecology of dispersal in the black-
tailed prairie dog. Journal of Mammalogy 69:236-250.

Gese EM, 2001. Territorial defense by coyotes (Canis Iatrans) in Yellowstone

National Park, Wyoming: Who, how, where, when, and why. Canadian
Journal of Zoology 79:980-987.

11]

 

Goossens B, Chikhi L, Taberlet P, Waits LP, Allaine D, 2001. Microsatellite
analysis of genetic variation among and within Alpine marmot populations
in the French Alps. Molecular Ecology 10:41-52.

Greenwood PJ, 1980. Mating systems, philopatry and dispersal in birds and
mammals. Animal Behaviour 28:1140-1162.

Hacklander K, Arnold W, 1999. Male-caused failure of female reproduction and
its adaptive value in alpine marmots (Marmota marmota). Behavioral
Ecology 10:592-597.

Hagman M, Forsman A, 2003. Correlated evolution of conspicuous coloration
and body size in poison frogs (Dentrobatidae). Evolution 57:2904-2910.

Hanby JP, Bygott JD, 1987. Emigration of subadult lions. Animal Behaviour
35:161-169.

Hanski I, 1999. Metapopulation ecology. Oxford: Oxford University Press.

Harvey PH, Page MD, 1991. The Comparative Method in Evolutionary Biology.
Oxford: Oxford University Press.

Haynie ML, Van den Bussche RA, Hoogland JL, Gilbert DA, 2003. Parentage,
multiple paternity, and breeding success in Gunnison's and Utah prairie
dogs. Journal of Mammalogy 84:1244-1253.

Herron MD, Castoe TA, Parkinson CL, 2004. Sciurid phylogeny and the
paraphyly of Holarctic ground squirrels (Spermophilus). Molecular
Phylogenetics and Evolution 31 :1015-1030.

Hoffmann IE, Millesi E, Huber S, Everts LG, Dittami JP, 2003a. Population
dynamics of European ground squirrels (Spermophilus cite/Ius) in a
suburban area. Journal of Mammalogy 84:615-626.

Hoffmann IE, Millesi E, Pieta K, Dittami JP, 2003b. Anthropogenic effects on the
population ecology of European ground squirrels (Spermophilus citellus)
at the periphery of their geographic range. Mammalian Biology 68:205-
213.

Holekamp KE, 1983. Proximal mechanisms of natal dispersal in Belding's ground
squirrels (Spermophilus beldingi beldingi) (Ph.D.). Berkeley: University of
California, Berkeley.

Holekamp KE, 1984a. Dispersal in ground-dwelling sciurids. In: Biology of

ground-dwelling squirrels (Murie JO, Michener GR, eds). Lincoln,
Nebraska: University of Nebraska Press; 297-320.

112

Holekamp KE, 1984b. Natal dispersal in Belding's ground squirrels
(Spermophilus beldingi). Behavioral Ecology and Sociobiology 16:21 -30.

Holekamp KE, 1986. Proximal causes of natal dispersal in Belding's ground
squirrels (Spermophilus beldingi). Ecological Monographs 56:365-391.

Holekamp KE, Nunes S, 1989. Seasonal variation in body weight, fat, and
behavior of free-living California ground squirrels (Spermophilus
beecheyl). Canadian Journal of Zoology 67:1425-1433.

Holekamp KE, Sherman PW, 1989. Why male ground squirrels disperse?
American Scientist 77:41-48.

Holekamp KE, Simpson HB, Smale L, 1985. Endocrine influences on natal
dispersal in Belding's ground squirrels (Spermophilus beldingi). In:
Migration: Mechanisms and adaptive significance contributions in marine
sciences (Supplement) (Rankin MA, ed); 397-408.

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

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

Holmes WG, 1984. The ecological basis of monogamy in alaskan horay
marmots. In: The biology of ground-dwelling squirrels (Murie JO, Michener
GR, eds). Lincoln: University of Nebraska Press; 250-274.

Hoogland JL, 1981. The evolution of coloniality in white-tailed and black-tailed
prairie dogs (Sciuridae: Cynomys Ieucurus and C. Iudovicianus). Ecology
62:252-272.

Hoogland JL, 1985. lnfanticide in prairie dogs: Lactating females kill offspring of
close kin. Science 230:1037-1040.

Hoogland JL, 1986. Nepotism in prairie dogs (Cynomys Iudovicianus) varies with
competition but not with kinship. Animal Behaviour 34:263-270.

Hoogland JL, 1992. Levels of inbreeding among prairie dogs. American
Naturalist 139:591-602.

Hoogland JL, 1995. The black-tailed prairie dog. Chicago: University of Chicago
Press.

113

Hoogland JL, 1999. Philopatry, dispersal, and social organization of Gunnison's
prairie dogs. Journal of Mammalogy 80:243-251.

Hoogland JL, 2001. Black-tailed, Gunnison's, and Utah prairie dogs reproduce
slowly. Journal of Mammalogy 82:917-927.

Hoogland JL, 2003. Sexual dimorphism of prairie dogs. Journal of Mammalogy
84:1254-1266.

Hoogland JL, Foltz DW, 1982. Variance in male and female reproductive success
in a harem-polygynous mammal, the black-tailed prairie dog (Sciuridae,
cynomysludovicianus). Behavioral Ecology and Sociobiology 112155-63.

Horne BV, Olson GS, Schooley RL, Corn JG, Bumham KP, 1997. Effects of
drought and prolonged winter on Townsend's ground squirrel demography
in shrubsteppe habitats. Ecological Monographs 672295-315.

Huber S, Millesi E, Dittami JP, 2002. Paternal effort and its relation to mating
success in the European ground squirrel. Animal Behaviour 63:157-164.

Ims RA, 1989. Kinship and origin effects on dispersal and space sharing in
Clethrionomys rufocanus. Ecology 702607-616.

lsbelI LA, Van Vuren D, 1996. Differential costs of locational and social dispersal
and their consequences for female group-living primates. Behaviour
13321-36.

Jones KE, Purvis A, 1997. An optimum body size for mammals? Comparative
evidence from bats. Functional Ecology 11:751-756.

Kenagy GJ, Sharbaugh SM, Nagy KA, 1989. Annual cycle of energy and time
expenditure in a golden-manteled ground squirrel population. Oecologia
78:269-282.

Koenig WD, Van Vuren D, Hooge PN, 1996. Detectability, philopatry, and the
distribution of dispersal distances in vertebrates. Trends in Ecology and
Evolution 112514-517.

Lacey EA, Wieczorek JR, 2001. Territoriality and male reproductive success in
arctic ground squirrels. Behavioral Ecology 12:626-632.

Lacey EA, Wieczorek JR, Tucker PK, 1997. Male mating behaviour and patterns

of sperm precedence in Arctic ground squirrels. Animal Behaviour 53:767-
779.

114

Larsen KW, Boutin S, 1994. Movements, survival, and settlement of red squirrel
(Tamiasciurus hudsonicus) offspring. Ecology 75:214-223.

Lehmann L, Perrin N, 2003. Inbreeding avoidance through kin recognition:
Choosy females boost male dispersal. American Naturalist 162:638-652.

Lenihan C, Van Vuren D, 1996. Growth and survival ofjuvenile yellow-bellied
marmots (Marmota flaviventn's). Canadian Journal of Zoology 74:297-302.

Lidicker WZ, 1975. The role of dispersal in the demography of small mammals.
In: Small Mammals: Their productivity and population dynamics (Golley
FB, Petrusewicz K, Ryszkowski L, eds). London: Cambridge University
Press; 103-128.

Longhurst W, 1944. Observations on the ecology of the Gunnison prairie dog in
Colorado. Journal of Mammalogy 25:24-36.

Maddison W, Maddison J, 2000. MacClade 4: Analysis of Phylogeny and
Character Evolution: Sinauer, Sunderland, MA.

Martins EP, 2000. Adaptation and the comparative method. Trends in Ecology
and Evolution 15:296-299.

Mayr E, 1961. Cause and effect in biology. Science 131 :1 501-1 506.

McCarley H, 1966. Annual cycle, population dynamics, and adaptive behavior of
Citellus tridecemlineatus. Journal of Mammalogy 47:294-316.

McKeever S, 1964. The biology of the golden-manteled ground squirrel, Cite/[us
Iateralis. Ecological Monographs 34:383-401.

McLean IG, 1979. Paternal behavior and intraspecific killing by the arctic ground
squirrel. American Zoologist 19:936-936.

McLean IG, 1980. Kin association in arctic ground squirrels. American Zoologist
20:828-828.

McLean IG, 1982. The association of female kin in the Arctic ground squirrel
Sperrnophilus parryii. Behavioral Ecology and Sociobiology 10:91-99.

McNutt JW, 1996. Sex-biased dispersal in African wild dogs, Lycaon pictus.
Animal Behaviour 5221067-1077.

Mennoz ME, Omelas JF, 2004. Phylogenetic analysis of life-history adaptations
in parasitic cowbirds. Behavioral Ecology 152109-1 19.

115

Metzgar LH, 1967. An experimental comparison of screech owl predation on
resident and transient white-footed mice (Peromyscus Ieucopus). Journal
of Mammalogy 48:387-391.

Michener DR, Michener GR, 1971. Sex Ratio and inter-year residence in a
population of Spennophilus richardsonii. Journal of Mammalogy 52:853

Michener GR, 1973. Intraspecific aggression and social organization in ground
squirrels. Journal of Mammalogy 5421001-1003.

Michener GR, 1979a. Spatial relationships and social organization of adult
Richardson's ground squirrels. Canadian Journal of Zoology 57:125-139.

Michener GR, 1979b. Yearly Variations in the Population-Dynamics of
Richardsons Ground-Squirrels. Canadian Field-Naturalist 93:363-370.

Michener GR, 1983. Kin identification, matriarchies, and the evolution of sociality
in ground-dwelling sciurids. Advances in the Study of Mammalian
Behavior 72528-572.

Michener GR, Locklear L, 1990. Differential costs of reproductive effort for male
and female Richardsons ground squirrels. Ecology 71:855-868.

Michener GR, McLean IG, 1996. Reproductive behaviour and operational sex
ratio in Richardson's ground squirrels. Animal Behaviour 52:743-758.

Michener GR, Michener DR, 1977. Population structure and dispersal in
Richardson's ground squirrels. Ecology 58:359-368.

Michener GR, Murie JO, 1983. Black-tailed prairie dog coteries: Are they
cooperatively breeding units? American Naturalist 121:266-274.

Michener GR, Sheppard DH, 1972. Social behavior between adult female
Richardson's ground squirrels (Spermophilus richardsonii) and their own
and alien young. Canadian Journal of Zoology 50:1343-1349.

Millesi E, Huber S, Dittami J, Hoffmann I, Daan S, 1998. Parameters of mating
effort and success in male European ground squirrels, Spennophilus
cite/Ius. Ethology 104:298-313.

Millesi E, Strijkstra AM, Hoffman IE, Dittami JP, Daan S, 1999. Sex and age
differences in mass, morphology, and annual cycle in European ground
squirrels, Sperm0philus cite/Ius. Journal of Mammalogy 80:218-231.

Moore J, Ali R, 1984. Are dispersal and inbreeding avoidance related? Animal
Behaviour 32:94-112.

116

Morton ML, Maxwell CS, Wade CE, 1974. Body size, body composition, and

behavior of juvenile Belding ground squirrels. Great Basin Naturalist
34:121-134.

Murie JO, 1973. Population characteristics and phenology of a Franklin ground
squirrel (Spermophilus franklinii) colony in Alberta, Canada. American
Midland Naturalist 90:334-340.

Murie JO, 1980. Social Interactions and dominance relationships between female
and male Columbian ground squirrels. American Zoologist 20:827-827.

Murie JO, Harris MA, 1978. Territoriality and dominance in male Columbian
ground squirrels (Spermophilus columbianus). Can. J. Zool 56:2402-2412.

Murie JO, Michener GR, 1984. The biology of ground-dwelling sciurids. Lincoln:
University of Nebraska Press.

Nilsson JA, Smith HG, 1985. Early fledgling mortality and the timing ofjuvenile
dispersal in the marsh tit, Parus palustn‘s. Omis Scandinavica 16:293-
298.

Noordwijk MAV, Schaik CPV, 2001. Career Moves: Transfer and rank challenge
decisions by male long-tailed macaques. Behaviour 1382359-395.

Nowak RM, 1999. Walker's mammals of the world, 6 ed. Baltimore: Johns
Hopkins University Press.

Nunes S, Duniec TR, Schweppe SA, Holekamp KE, 1999. Energetic and
endocrine mediation of natal dispersal behavior in Belding's ground
squirrels. Hormones and Behavior 35:113-124.

Nunes S, Ha CDT, Garrett PJ, Muecke EM, Smale L, Holekamp KE, 1998. Body
fat and time of year interact to mediate dispersal behaviour in ground
squirrels. Animal Behaviour 55:605-614.

Nunes S, Holekamp KE, 1996. Mass and fat influence the timing of natal
dispersal in Belding's ground squirrels (Spermophilus beldingi). Journal of
Mammalogy 772807-817.

Nunes S, Muecke E-M, Ross HE, Bartholomew PA, Holekamp KE, 2000. Food
availability affects behavior but not circulating gonadal hormones in
maternal Belding's ground squirrels. Physiology and Behavior 71:447-455.

Olson GS, Horne BV, 1998. Dispersal patterns of juvenile Townsend's ground

squirrels in southwestern Idaho. Canadian Journal of Zoology 76:2084-
2089.

117

h

Packer C, 1979. Inter-troop transfer and inbreeding avoidance in Papio anubis.
Animal Behaviour 2721-36.

Pfeifer S, 19823. Disappearance and dispersal of Sperm0philus elegans
juveniles in relation to behavior. Behavioral Ecology and Sociobiology
10:237-243.

Pfeifer SR, 1982b. Variability in reproductive output and success of
Spennophilus elegans ground squirrels. Journal of Mammalogy 63:284-
289.

Phoenix CH, Goy RW, Gerall AA, 1959. Organizing action of prenatally
administered testosterone propionate on the tissues mediating mating
behavior in the female guinea pig. Endocrinology 652369-82.

Purvis A, Rambaut A, 1995. Comparative analysis of independent contrasts
(CAIC): an Apple Macintosh application for analyzing comparative data.
Comparative Applied Bioscience 11:247-251.

Pusey AE, 1987. Sex-biased dispersal and inbreeding avoidance in birds and
mammals. Trends in Ecology and Evolution 22295-299.

Randall JA, Hekkala ER, Cooper LD, Barfield J, 2002. Familiarity and flexible
mating strategies of a solitary rodent, Dipodomys ingens. Animal
Behaviour 6421 1-21.

Rayor LS, 1984. Yearling behavior, reproduction, and dispersal in Gunnison
prairie dogs. American Zoologist 242A2-A2.

Rayor LS, 1985. Effects of Habitat Quality on growth, age of first reproduction,
and dispersal in Gunnison prairie dogs (Cynomys gunnisom). Canadian
Journal of Zoology 63:2835-2840.

Rayor LS, 1988. Social organization and space-use in Gunnison prairie dog.
Behavioral Ecology and Sociobiology 22:69-78.

Reynolds HG, Turkowski F, 1972. Reproductive variations in the round-tailed
ground squirrel as related to winter rainfall. Journal of Mammalogy 53:893-
898.

Ricklefs RE, 2000. The economy of nature, 5 ed. New York: W.H. Freeman and
Company.

Rieger JF, 1996. Body size, litter size, timing of reproduction, and juvenile

survival in the Uinta ground squirrel, Spennophilus armatus. Oecologia
107:463-468.

118

Roach JL, Stapp P, Van Horne B, Antolin MF, 2001. Genetic structure of a

metapopulation of black-tailed prairie dogs. Journal of Mammalogy
82:946-959.

Rood JP, 1987. Dispersal and intergroup transfer in the dwarf mongoose. In:
Mammalian dispersal patterns (Chepko-Sade BD, Halpin ZT, eds).
Chicago: University of Chicago Press; 85-103.

Samo RJ, Bank MS, Stern HS, Franklin WL, 2003. Forced dispersal of juvenile
guanacos (Lama guanicoe): Causes, variation, and fates of individuals

dispersing at different times. Behavioral Ecology and Sociobiology 54:22-
29.

Schwagmeyer PL, Brown CH, 1983. Factors affecting male-male competition in
thirteen-lined ground squirrels. Behavioral Ecology and Sociobiology 13:1-
6.

Schwagmeyer PL, Parker GA, 1987. Queuing for mates in thirteen-lined ground
squirrels. Animal Behaviour 35:1015-1025.

Schwagmeyer PL, Woontner SJ, 1985. Mating competition in an asocial ground
squirrel. Behavioral Ecology and Sociobiology 17:291-296.

Schwagmeyer PL, Woontner SJ, 1986. Scramble competition polygyny in
thirteen-lined ground squirrels: the relative contributions of overt conflict
and competitive mate searching. Behavioral Ecology and Sociobiology
19:359-364.

Schwartz OA, Armitage KB, Van Vuren D, 1998. A 32-year demography of
yellow-bellied marmots (Marmota flaviventris). Journal of Zoology
246:337-346.

Sherman PW, 1976. Natural selection among group-living organisms (Ph.D.).
Ann Arbor: University of Michigan.

Sherman PW, 1977. Nepotism and the evolution of alarm calls. Science
197:1246-1253.

Sherman PW, 1981a. Kinship, demography, and Belding ground squirrel
Nepotism. Behavioral Ecology and Sociobiology 82251-259.

Sherman PW, 1981b. Reproductive competition and infanticide in Belding's
ground squirrels and other animals. In: Natural selection and social
behavior: Recent research and new theory (Alexander RD, Tinkle DW,
eds). New York: Chiron; 311-331.

119

Sherman PW, 1989. Mate guarding as paternity insurance in Idaho ground
squirrels. Nature 338:418-420.

Sherman PW, Morton ML, 1984. Demography of Belding ground squirrels.
Ecology 6521617-1628.

Shields WM, 1982. Philopatry, inbreeding, and the evolution of sex. Albany: State
University of New York Press.

Slade NA, Balph DF, 1974. Population ecology of Uinta ground squirrels.
Ecology 552989-1003.

Smale L, Nunes S, Holekamp KE, 1997. Sexually dimorphic dispersal in
mammals: patterns, causes, and consequences. Advances in the Study of
Behavior 262181-250.

Smith GW, Johnson DR, 1985. Demography of a Townsend ground squirrel
population in Southwestern Idaho. Ecology 66:171-178.

Snyder RL, Davis DE, Christian JJ, 1961. Seasonal changes in the weights of
woodchucks. Journal of Mammalogy 422297-312.

Sokal RR, Rohlf FJ, 1995. Biometry: the principles and practice of statistics in
biological research, Third ed. New York: W. H. Freeman and Company.

Spinks AC, Bennett NC, Jarvis JUM, 2000. A comparison of the ecology of two
populations of the common mole-rat, Cryptomys hottentotus hottentotus:
the effect of aridity on food, foraging and body mass. Oecologia 1252341-
349.

Sprague DS, 1992. Life history and male intertroop mobility among Japanese
macaques (Macaca fuscata). International Journal of Primatology 13:437-
454.

Summers K, Clough ME, 2001. The evolution of coloration and toxicity in the
poison frog family (Dendrobatidae). PANS 98:6227-6232.

Sun LX, Muller-Schwarze D, Schulte BA, 2000. Dispersal pattern and effective
population size of the beaver. Canadian Journal of Zoology 78:393-398.

Svendsen GE, 1974. Behavioral and environmental factors in the spatial
distribution and population dynamics of a yellow-bellied marmot
population. Ecology 552760-771.

Switzer PV, Stamps JA, Mangel M, 2001. When should a territory resident
attack? Animal Behaviour 622749-759.

120

Tang-Martinez Z, Mueller LL, Taylor GT, 1993. Individual odors and mating
success in the golden hamster, Mesocricetus auratus. Animal Behaviour
45:1141-1151.

Tileston JV, Lechleitner RR, 1966. Some comparisons of the black-tailed and
white-tailed prairie dogs in north-central Colorado. American Midland
Naturalist 75:292-316.

Tinbergen N, 1963. On aims and methods of Ethology. Zeitschrift fur
Tierpsychologie 202410-433.

Travis SE, Slobodchikoff CN, Keim P, 1996. Social assemblages and mating
relationships in prairie dogs: A DNA fingerprint analysis. Behavioral
Ecology 7195-100.

Van Ballenberghe V, 1983. Extraterritorial movements and dispersal of wolves in
south-central Alaska. Journal of Mammalogy 642168-171.

van Noordwijk MA, van Schaik CP, 1985. Male migration and rank acquisition in
wild long-tailed macaques (Macaca fascicularis). Animal Behaviour
33:849-861.

Van Vuren D, 1990. Dispersal of yellow-bellied marmots (Ph.D.). Lawrence:
University of Kansas.

Van Vuren D, Armitage KB, 1994. Survival of dispersing and philopatric yellow-
bellied marmots: What is the cost of dispersal? Oikos 692179-181.

Van Vuren D, 1990. Dispersal of yellow-bellied marmots (Ph.D.). Lawrence:
University of Kansas.

Waser PM, Jones WT, 1983a. Natal philopatry among solitary mammals.
Quarterly Review of Biology 58:355-390.

Waser PM, Jones WT, 1983b. Natal philopatry among solitary mammals.
Quarterly Review of Biology 58:355-390.

Waterman J, 1992. The Use of Space by yearling Columbian ground squirrels
before male dispersal. Canadian Journal of Zoology 70:2490-2493.

Webb DR, 1981. Macro-habitat patch structure, environmental harshness, and
Marmota flaviventn‘s. Behavioral Ecology and Sociobiology 8:175-182.

Weddell BJ, 1989. Dispersion of Columbian Ground-Squirrels (Spermophilus

columbianus) in Meadow Steppe and Coniferous Forest. Journal of
Mammalogy 702842-845.

121

 

Weddell BJ, 1991. Distribution of movements of Columbian ground squirrels
(Spermophilus columbianus (Ord))2 are habitat patches like islands?
Journal of Biogeography 18:385-394.

Wiggett DR, Boag DA, 1991. Natal dispersal in Columbian ground squirrels: Is
body mass the proximate stimulus? Canadian Journal of Zoology 70:649-
653.

Wiggett DR, Boag DA, 1992. Natal dispersal in Columbian ground squirrels: ls
body mass the proximate stimulus. Canadian Journal of Zoology 70:649-
653.

Wiggett DR, Boag DA, 1993. The Proximate causes of male-biased natal
emigration in Columbian ground squirrels. Canadian Journal of Zoology
71:204-218.

Wiggett DR, Boag DA, Wiggett ADR, 1989. Movements of intercolony natal
dispersers in the Columbian ground squirrel. Canadian Journal of Zoology
67:1447-1452.

Wistrand H, 1974. Individual, social, and seasonal behavior of the thirteen-lined
ground squirrel (Spermophilus tridecemlineatus). Journal of Mammalogy
552329-347.

Wolff JO, Peterson JA, 1998. An offspring-defense hypothesis for territoriality in
female mammals. Ethology, Ecology, and Evolution 10:227-239.

122

 

   

lllllllllllllllllllIll-Ill