download. - Ol Pejeta Conservancy

THE TRIALS OF MOTHERHOOD:
MATERNAL BEHAVIOR PATTERNS AND ANTIPREDATOR TACTICS
IN THOMSON’S GAZELLE (GAZELLA THOMSONII),
A HIDING UNGULATE
Blair A. Roberts
A DISSERTATION
PRESENTED TO THE FACULTY
OF PRINCETON UNIVERSITY
IN CANDIDACY FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY
RECOMMENDED FOR ACCEPTANCE
BY THE DEPARTMENT OF
ECOLOGY AND EVOLUTIONARY BIOLOGY
Adviser: Daniel I. Rubenstein
January 2014
© Blair Allison Roberts, 2014. All rights reserved.
For my parents,
without whose investment
I would surely have fallen to the jackals.
iii
Abstract
This dissertation examines risk management tactics of female Thomson’s gazelles during the
early stages of motherhood, when fawns experience high predation risk. Fawns’ main defense is
the hiding strategy, in which they spend long periods of time concealed in vegetation apart from
their mothers. Hiding requires maternal cooperation and has the potential to affect maternal risk
and behavioral patterns.
Immediately following birth, the fawn is unable to hide. Avoidance of fawn predation
depends on the social and environmental contexts of parturition. Most mothers either isolate
from conspecifics and give birth in tall grass or remain in social groups and give birth in short
grass. The latter tactic provides greater protection from conventional predators, but pressures
such as conspecific harassment may lead mothers to give birth in isolation. Both tactics improve
neonates’ survival probability compared to behavior that is inconsistent with either tactic.
Once hiding begins, fawns are concealed and relatively safe from detection by predators,
and mother-fawn interactions are limited to brief active periods. This framework affords mothers
the freedom to schedule investment behaviors and minimize the impact of motherhood on their
activity. Peaks in maternal vigilance precede and coincide with active periods, when fawns are at
greatest risk. Outside of active periods, mothers are able to behave identically to non-mothers.
As fawns transition out of the hiding strategy, they increase their exposure to predators.
More frequent activity periods enable fawns to move hiding spots more often, which allows
mothers to track group movements more effectively. Mothers and fawns rely on increased time
spent in social groups and in short grass habitats to mitigate the increase in fawn risk. Thus, as
fawns transition out of hiding, mothers transition back to normal activity, grouping, and habitat
use patterns.
iv
During predator attacks, mothers must respond to actual rather than potential risk. In an
experiment involving simulated predator attacks, we found that mothers do not rely on group
vigilance to detect potential fawn predators, that they exhibit riskier responses than non-mothers,
and that the presence of a mother in a group can affect the group’s response in ways that may
reduce maternal risk.
v
Table of Contents
Abstract
iv
List of Tables
x
List of Figures
xi
Acknowledgements
Chapter 1: Introduction and literature review
xiii
1
Parental investment
1
Ungulates: antipredator behavior and offspring care
3
Risk management
4
Response to immediate risk
7
Preventing offspring predation
Thomson’s gazelles and their enemies
10
17
Natural history
17
Predators
18
Overview of chapters
Chapter 2: Maternal tactics for mitigating neonate predation risk during the 26
28
postpartum period in Thomson’s gazelle Abstract
28
Introduction
29
Methods
32
Study site
32
Observations
33
Analyses
34
vi
Results
Effects of detection by predators on fawn survival
37
39
Prepartum isolation, birth site grass height, and their effect on detection of the fawn by
predators
39
Effect of prepartum isolation on disturbances and postpartum period duration
40
Effect of birth site grass height on fawn visibility and probability of detection by 42
predators
Discussion
44
Acknowledgements
47
Chapter 3: The effects of the hiding strategy on maternal behavioral trade-offs in
Thomson’s gazelle 49
Abstract
49
Introduction
50
Methods
53
Field site
53
Behavioral observations
54
Statistical analyses
55
Results
58
Maternal behavior while fawn was active vs. while fawn was hiding
58
Maternal behavior within hiding periods
62
Discussion
66
Acknowledgements
70
Chapter 4: Coping with transition: fawn risk and maternal behavioral changes at 71
vii
the end of the hiding phase Abstract
71
Introduction
71
Methods
74
Field site
74
Behavioral observations
74
Habitat measurements
76
Analyses
77
Results
80
Risk of predator encounter during hiding bout
80
Variation in fawn risk with age
81
Variation in maternal behavior across fawn age groups and comparisons to 82
non-mothers
Discussion
87
Acknowledgements
91
Chapter 5: Maternal responses to simulated cheetah attacks and their effect on 92
group antipredator responses in Thomson’s gazelle Abstract
92
Introduction
92
Methods
95
Study site
95
Experimental apparatus
95
Group trials
97
viii
Mother-fawn trials Analyses
Results
99
100
101
Group trials
101
Mother-fawn trials, and comparison to group trials
102
Discussion
104
Acknowledgements
106
Chapter 6: Overview and concluding remarks: hiding versus following 107
Appendix I: An attack by a warthog Phacochoerus africanus on a newborn 110
Thomson’s gazelle Gazella thomsonii Introduction
110
Materials and methods
110
Results and discussion
110
Acknowledgements
112
Appendix II: Perinatal behavior of a wild Grevy’s zebra (Equus grevyi) mare and foal 113
Abstract 113
Introduction
113
Study area
114
Materials and methods
115
Results and discussion
115
Acknowledgements
121
References
122
ix
List of Tables
1. Time of parturition, prepartum decisions, postpartum events, and outcomes for the 38
observations discussed in Chapter 2.
2. Results of Wilcoxon tests comparing mean fawn visibility when lying down and 43
standing up across video-recorded observations of parturition and postnatal events.
3. Expected relative levels of fawn predation risk and maternal certainty of predator 53
absence during different stages of the hiding strategy.
4. Ethogram used for instantaneous activity budget sampling
55
5. Results of Wilcoxon tests comparing maternal vigilance levels between time bins. 63
6. Contingency table showing the frequency with which fawns of different ages initiate 81
the end of their own hiding bouts or wait for their mothers to retrieve them.
7. Mean group sizes and percent time spent in groups of various sizes for non-mothers 85
and mothers of old and young fawns during hiding and active periods.
8. Contingency table comparing reaction distances between mother-fawn and group 103
cheetah trials.
9. Contingency table comparing distance gained between mother-fawn and group 103
cheetah trials.
10. Contingency table comparing whether subjects left the area following stimulus 103
presentation between mother-fawn and group cheetah trials.
11. Ethogram of postnatal behaviors of Grevy’s zebra mares and neonates and the 119
time at which the behaviors occurred in this observation.
x
List of Figures
1. Changes in mother-infant contact and distance with offspring age for hiding and 12
following ungulates. 2. The effect of non-predatory disturbances on the time delay from birth to the onset 35
of hiding and the fawn’s first walking bout.
3. Survival curves illustrating the probability of avoiding a predator encounter during 41
postpartum periods of various durations.
4. Mean visibility of fawns while standing and lying down in video-recorded observations 42
of parturition and postpartum periods.
5. Two main behavioral tactics used by Thomson’s gazelle mothers during the postpartum 46
period.
6. Mean percent time mothers spend on each of five behaviors while their fawns are 59
active and hiding.
7. Hiding period and overall maternal activity budgets compared to non-mother activity 60
budgets.
8. Variation in maternal vigilance levels within fawn hiding periods. 61
9. Variation in maternal vigilance levels based on vigilance sample data by 30-minute 63
time block.
10. Analysis of bin means with transformed ranks for vigilance sample data.
11. Activity budgets for mothers while the fawn is active and during 30-minute bins 64
64
during hiding periods.
12. Analysis of bin means with transformed ranks for activity budget data. 65
xi
13. Linear regression comparing grass height measurements from pin drops and 77
pasture disc methods.
14. Activity budgets of non-mothers, mothers of young fawns and mothers of old fawns. 83
Maternal activity budgets are given for all data hiding periods and active periods.
15. Cheetah puppet mounted on RC car. 96
16. Vervet monkey puppet mounted on RC car. 96
17. Differences in group reactions to RC car, vervet monkey and cheetah stimuli. 101
18. Warthog holding gazelle infant in its mouth before shaking it.
111
19. Grevy’s zebra mare and neonate immediately after parturition. 116
20. Grevy’s zebra foal standing with umbilicus still attached.
116
21. Grevy’s zebra foal walking and dragging the placenta by the umbilicus.
117
22. Grevy’s zebra mare attempting to lead the foal toward the departing group members.
117
xii
Acknowledgements
I am lucky to have called Princeton my home for the last five and a half years, and as
beautiful as the place is, it’s the people in it that have made it such an engaging environment.
First, I thank my advisor, Dan Rubenstein, for all of the support he has provided over my
graduate career. Beginning with my first trip to Kenya in the summer before I started at
Princeton and continuing through the stressful writing process, Dan has been instructive and
encouraging while still allowing me a great deal of independence in conducting my research.
I also thank my committee members, Henry Horn, Jeanne Altmann, and Jim Gould, for their
instruction and support throughout my graduate career. Siobhan Condran, Lolly O’Brien, the
late Amy Bordvik, and Richard Smith have all been helpful in managing the logistics of being a
grad student and EEB department member.
I am grateful to my fellow Rubenstein lab members and cohort-mates for their feedback
and support: Corinne Kendall, Cassandra Nuñez, Jenny Schieltz, Caitlin Barale, Ipek Kulahci,
Eili Klein, Stephanie Hauck, Qing Cao, Adrienne Tecza, Nitin Sekar, Molly Schumer, and
Chaitanya Yarlagadda. The graduate student community is one of EEB’s greatest assets, and I
am happy to count many of my colleagues as friends as well. Thanks especially to Sam Rabin,
Ann Tate, Carey Nadell, Caroline Farrior, Andrew Hartnett, Jim Adelman, and Carla Staver for
many good times at trivia, on the softball field, at football days and at the holiday party.
Since 2008 I have made six trips to the field totaling more than sixteen months, and Laikipia
has come to feel a lot like my second home. I am grateful to the Government of Kenya for
granting me permission to work in such a breathtaking environment, to Mpala Research Center
for serving as my home base for my first visits and for providing logistic support thereafter, and
to Ol Pejeta Conservancy for allowing me to live and work on the premises. Special thanks to
xiii
Margaret Kinnard, Mburu Tuni, Josphat Mwangi and John Kitoe at Mpala, and to Ol Pejeta’s
management and Ecological Monitoring department, especially Nathan Gichohi and Martin
Mulama. Through their tireless efforts, the Ol Pejeta Research Center staff – Kosgei, Erick,
Emily, Monica and especially Catherine Mundia – truly made a converted stable feel like home.
I thank them as well as the numerous security guards that assisted with tedious vegetation work
while protecting me from buffalo and other beasts.
The research center got lonely at times, so I am grateful for the friendships and memories
I formed with other residents – especially Kim Vanderwaal, Nicole Sharpe, Jenny Schieltz, and
George Paul. You are the stars in the rotating cast of amazing and interesting people that made
up my field family.
I owe thanks to my family, whose influence began long before graduate school and will
continue long after. My siblings – Tadd, Paige, Brooke and Dane – were the subjects of my first
dabblings in the study of the behavior of wild animals. Thanks to Kathy and Kate for many
fantastic weekend escapes full of wine and food in Long Valley and Old Lyme. And of course,
thanks to my parents. I remember a day when I was very young and had just sorted out that the
“junior high”, “high school”, and “college” mentioned in the TV shows I was watching referred
to various stages of education that still lay ahead of me. After verifying this with my mother, I
asked if after college was when I would finally have a job, to which she replied, “No, after college
you’ll probably go to grad school.” That grad school did in fact follow after college is as much
as a result of my own motivation and hard work as my parents’ unending encouragement. I am
so happy to have been able to travel with them in Kenya and share with them what they have
enabled me to accomplish.
Finally, I thank Mike, my best friend and soon-to-be husband. His patience, love and
encouragement have been unfailing, though not untested, throughout this entire process. Of all
xiv
the amazing experiences I had during my time in Kenya, I will always remember August 20,
2012 as the happiest day.
Financial support for this research came from Princeton University, the Department of
Ecology and Evolutionary Biology, the Philadelphia Chapter of the Explorers Club, the Animal
Behavior Society Student Research Grant program, the American Society of Mammalogists
Student Travel Grant program, and the Dean’s Fund for Scholarly Travel.
xv
Chapter 1: Introduction and literature review
Blair A. Roberts
Parenthood is a study in balance and compromise. Prior to reproduction, an individual’s
mission is to survive long enough to create offspring. Once this is accomplished, the parent is
faced with a series of decisions that carry serious repercussions for its reproductive success:
should the parent care for its young? If so, what proportion of available resources should it
commit to its offspring? How much should it keep in reserve for its own survival and future
reproductive efforts? The parent’s answers to these questions are invariably affected by many
factors, including the current social and environmental contexts, qualities of the offspring, and
characteristics of the parent itself. In this thesis, I examine the interplay between motherhood
and risk in Thomson’s gazelle (Gazella thomsonii), a small African ungulate species. I explore the
tactics mothers use to protect their offspring, variation in maternal strategies, and the effects these
behaviors have on mothers themselves.
Parental investment
Parental care is one means by which individuals may increase their reproductive success. By
spending time and energy provisioning or protecting young, parents can improve their offsprings’
chance of survival and thereby increase the abundance of their own genes in the next generation.
However, such parental investment in one offspring reduces the parent’s ability to invest in its
other offspring (Trivers 1972). For iteroparous parents, the costs of investment take two forms:
fecundity costs occur when the parent’s ability to produce or care for offspring is hindered by a
depletion of bodily resources, whereas survival costs result from the parent perishing before its
1
next reproductive opportunity (Bell 1980). In determining the appropriate amount of investment
for a given offspring, parents must weigh both the likely pay-off and the potential costs.
The desired benefit to the parent of any investment is an increase in inclusive fitness
resulting from the improved chance that the offspring will survive and reproduce. Parental
investment is expected to increase with the offspring’s reproductive value (Clutton-Brock 1991).
Reproductive value is a combination of the offspring’s probability of surviving to reproductive
age and its reproductive potential, or the proportion of the parent’s genes that the offspring is
likely to pass on to subsequent generations. For example, older or larger offspring may be more
likely to survive to reproductive age and thus represent a greater probability of increasing the
parent’s inclusive fitness (Tessier & Consolatti 1989; Svagelj et al. 2012). Likewise, in many
species, reproductive potential varies with offspring sex, resulting in higher investment in
offspring of one sex over the other (Clutton-Brock et al. 1981; Trillmich 1986; Bercovitch et al.
2000). Reproductive value can also vary with other offspring characteristics, such as brood size
(Coleman & Fischer 2010; Teichert et al. 2013), health (Saino et al. 2000), and birth or laying
date (Redmon et al. 2009; Kivanova et al. 2011).
The value of a parent’s investment also depends on the relative contribution the investment
makes to the offspring’s survival and reproductive success. The reward of investing to the parent
is large if the offspring is unlikely to survive or reproduce without the investment, and is small
if the offspring has a high probability of success without parental assistance or if the offspring
is unlikely to survive even with large amounts of investment. These concepts often play out in
parental defense of offspring, where parents are more likely to defend or defend more vigorously
when offspring are more able to capitalize on defense efforts by escaping (Redondo & Carranza
1989), or when the offspring are likely to die without parental intervention (the “harm to
offspring” hypothesis; Koskela et al. 2000; Goubault et al. 2007).
2
As described above, the cost to the parent of an investment in current offspring is a
reduced capacity for investment in future offspring. All else being equal, a parent’s investment
should decrease as costs and risks to their future reproduction rise. Reductions in investment
are commonly seen when resource availability is low or parental body condition is poor (Drent
& Daan 1980; Smith 1987; Scott & Traniello 1990; Smith & Wootton 1995). Under these
circumstances, parental expenditures are more likely to negatively affect future reproductive
opportunities or the survival of the parent. It is often difficult to determine whether observed
reductions in care result from adaptive decisions by parents or are simply the unavoidable
consequences of low resource availability (Clutton-Brock 1991). However, some studies have
successfully separated these two drivers and demonstrated that parents do indeed exercise control
over care levels by adjusting their investment strategies in response to surrounding conditions (e.g.
Festa-Bianchet & Jorgenson 1998).
Parents should weight the benefits and costs described above according to their own relative
reproductive value compared to that of their offspring. Generally, parents’ reproductive value
diminishes with age as the parents’ expected number of future reproductive opportunities
dwindles. Therefore, as parents age they should increase investment in the offspring at hand
rather than saving up for opportunities that may never come (the “terminal investment
hypothesis”; Clutton-Brock 1984). Increases in parental investment with age have been
demonstrated across taxa (e.g. Pugesek 1983; Bercovitch et al. 2009; Hoffman et al. 2010; Heinze
& Schrempf 2012).
Ungulates: antipredator behavior and offspring care
Ungulates are a diverse and widespread group, numbering over 450 species (Groves &
Grubb 2011) and inhabiting most of the world’s terrestrial ecosystems. In these systems they play
3
a variety of ecological roles, including serving as common prey items for many predator species.
Young ungulates are particularly vulnerable due to their smaller size. Infant mortality rates of
50% or more are common in ungulate populations, and where predators are present, predation is
usually the primary cause of infant death (Estes & Estes 1979; Linnell et al. 1995; Janermo 2004;
Lomas & Bender 2007). This pressure has selected for a variety of behaviors designed to mitigate
predation risk in both adults and their offspring.
Risk management
Ungulates employ multiple means of managing predation risk outside of actual predator
encounters, three of which I will discuss here. First, they may utilize habitats or areas that reduce
their risk of encounters with or detection by predators, or that offer protection in the event of
an attack. Second, they engage in vigilance behavior, which helps them detect approaching
or attacking predators so that they may respond rapidly and effectively. Third, they may form
groups to dilute risk and increase the probability of predator detection prior to an attack.
Habitat selection. The “landscape of fear” concept espouses that habitat selection by prey
animals is driven in part by predation risk (Altendorf et al. 2001; Laundré et al. 2001; Laundré
et al. 2010). Many studies have shown ungulate habitat use patterns to align well with predictions
stemming from this hypothesis. For example, ungulates often preferentially utilize habitat types
with low predator densities (Mao et al. 2005; McLoughlin et al. 2005; Rettie & Messier 2008;
Valeix et al. 2009) and avoid localized areas where predators were recently active (Creel et al.
2005; Fischhoff et al. 2007). Many species prefer habitats that feature protective elements: some
select habitats with concealing cover (Hirst 1975; Mduma & Sinclair 1994; Tufto et al. 1996;
4
Creel et al. 2005; Mao et al. 2005), while others require good visibility (Brooks 1961; Hirst 1975;
Fortin et al. 2009; Valeix et al. 2009) or readily-accessible escape terrain (Tilton & Willard 1982;
Namgail et al. 2004; Hamel & Côté 2007). The relative riskiness of habitats often fluctuates
throughout the day and across seasons with predator activity cycles, creating temporal patterns of
ungulate habitat use (Mao et al. 2005; Fischhoff et al. 2007).
The greater vulnerability of ungulate offspring is often reflected in the habitat selection
of their mothers, who tend to be found in safer habitats than their non-maternal counterparts.
For example moose (Alces alces) with calves seek refuge on small islands, which feature low wolf
(Canis lupus) densities and offer water for escape and cover for concealment, or inhabit areas near
human campsites, which wolves avoid (Stephens & Peterson 1984). In many mountain-dwelling
ungulate species, such as Dall’s sheep (Ovis dalli) and Nubian ibex (Capra ibex nubiana) mothers and
their young stay closer to escape terrain than non-mothers (Kohlmann et al. 1996; Rachlow &
Bowyer 1998). Finally, female Thomson’s gazelles with young fawns are often found in taller grass
areas than other adults because these habitats offer better cover for hiding young (Chapter 4;
Fitzgibbon 1988). By selecting safer habitats, mothers often sacrifice access to high-quality forage
patches. For example, for ibex and Dall’s sheep, forage quality increase with distance from escape
terrain, and Thomson’s gazelles have lower foraging efficiency in tall grass compared to short
grass (Gwynne & Bell 1968; Jarman & Sinclair 1979; Wilmshurst et al. 1999).
Vigilance. Vigilance behavior is a common means by which animals detect approaching and
attacking predators and thereby reduce their probability of falling victim to surprise attacks
(Godin & Smith 1988; Quinn & Cresswell 2004). More vigilant individuals detect predators
when they are further away, which increases their survival probability by enabling them to avoid
5
being attacked altogether or by giving them more time to respond defensively (Fitzgibbon 1988,
1990b). More vigilant individuals are also less likely to be targeted by predators in the first place
compared to less vigilant individuals (Fitzgibbon 1988, 1990a; Krause & Godin 1996).
Vigilance behavior exacts an opportunity cost by taking up time that could be spent on other
activities, such as foraging, resting, or social interaction (Lima & Dill 1990; Lima 1998; Toïgo
1999; Hamel & Côté 2008). Animals are expected to manage these trade-offs by becoming more
vigilant when risk of predation is high and engaging in other activities when the risk subsides, a
concept known as the “risk allocation hypothesis” (Lima & Bednekoff 1999). Mismanaging this
trade-off by overinvesting in vigilance can result in low forage intake and reduced fecundity, while
underinvesting in vigilance can result in death from predation.
Again, the greater vulnerability of ungulate offspring is reflected in the higher vigilance
levels of their mothers (Chapters 3 & 4; Burger & Gochfeld 1994; Hunter & Skinner 1998).
Fitzgibbon (1988, 1990b) showed that maternal vigilance is key to increasing the survival
probability of Thomson’s gazelle fawns during cheetah attacks: more vigilant mothers were
able to alert their offspring while the cheetah was further away, enabling the fawn to drop
down without the cheetah being sure of its location. In mothers with hiding young (see below),
heightened maternal vigilance enables mothers to detect predators lying in wait for them to
retrieve their young (Chapter 3; Fitzgibbon 1993; Byers 1997).
Grouping. Group membership dilutes individual predation risk by reducing the probability
that any given individual will be targeted by a predator (Hamilton 1971). Furthermore, although
groups may be easier for predators to detect (Hebblewhite & Pletscher 2002), they may be less
likely to be targeted by predators than solitary animals, as Fitzgibbon (1988) found for cheetahs
6
(Acinonyx jubatus) hunting gazelle. Group membership can also benefit individuals by allowing
them to reduce their vigilance levels without suffering an increase in predation risk (the “many
eyes hypothesis”; Lima & Dill 1990; Lima 1995). Fitzgibbon (1988) found that groups detected
predators at greater distances compared to single gazelle and LaGory (1987) found that larger
groups of white-tailed deer (Odocoileus virginianus) were more likely to detect an approaching
human than were small groups.
In ungulate species with follower young (see below), mothers are usually able to maintain
group membership throughout the infancy stage. However, due to the stationary nature of
hiding young (see below), hider mothers often experience decreases in group size. This results
from females intentionally isolating from conspecifics prior to or following parturition, and
from mothers being left behind as their groups move away from the offspring’s hiding location
(Brooks 1961; Walther 1969; O’Brien 1984; Schwede et al. 1993). Due in part to these changes
in grouping patterns, mothers can be at higher risk of predation than non-mother adults (Estes &
Goddard 1967; Fitzgibbon & Lazarus 1995).
Response to immediate risk
In the event of a predator attack, a prey individual’s behavioral response is critical to the
outcome of the encounter (Gese & Grothe 1995; Kullberg et al. 1998; Wisenden et al. 1999;
Lingle & Pellis 2002; DiRienzo et al. 2013). Among ungulates, common responses include flight,
which is often accompanied by signaling behavior; monitoring or harassment of the predator;
and grouping and defense formations.
Flight and signals. Many ungulate species rely on their speed, agility, and/or endurance to
evade attacking predators (Mloszewski 1983; Fitzgibbon & Lazarus 1995). Flight is an especially
7
attractive option when the attacking predator is large or numerous (Schaller 1972). In addition
to the fast running speeds exhibited by many species (e.g. upwards of 72 kph in pronghorn,
Antilocapra americana; Byers 1997), maneuvers such as rapid zig-zagging can cause predators
to over-run and thereby increase the distance between predator and prey (Howland 1974;
Fitzgibbon 1990b).
In many cases, flight is accompanied by visual signals. For example, white-tailed deer flee
with their tails erected, exposing the bright white fur beneath (Hirth & McCullough 1977; Caro
et al. 1995). Thomson’s gazelles, along with other antelope species, often engage in stotting
during flight. In this peculiar gait the animal springs up and down with its legs held stiff. The
purpose of such signals has been the subject of much debate. Proposed functions include:
alerting conspecifics to the predator’s presence; informing the predator that is has been detected;
distracting, confusing or intimidating the predator; promoting group cohesion; and discouraging
the predator from pursuing by advertising the prey’s vigor and high escape potential (Bildstein
1983; Caro 1986a, b; Fitzgibbon & Fanshawe 1988; Caro et al. 1995). In the above examples,
evidence suggests that tail flagging serves to alert the predator to the deer’s escape ability, but
may also encourage group cohesion (Hirth & McCullough 1977; Bildstein 1983; Caro et al.
1995). Furthermore, once the tail is lowered and the white disappears, the deer may become
suddenly less conspicuous to the predator, causing it to lose sight of its target (Caro et al. 1995).
Meanwhile, stotting is thought to alert the predator that it has been spotted (Caro 1986b), and
potentially advertise the escape ability of the prey (Fitzgibbon & Fanshawe 1988).
Young in following species (see below) rely on flight to a greater degree than those in hiding
species. Wildebeest (Connochaetes taurinus) calves are able to run within an hour of birth and equal
adults in their ability to escape hyenas by two days of age (Estes & Estes 1979). Meanwhile, flight
8
ability in pronghorn and Thomson’s gazelle is still insufficient to escape even small predators
when fawns are between three and four weeks old (Barrett 1978; Fitzgibbon 1990b). Adult flight
behaviors are present even in young ungulates: Thomson’s gazelle fawns and adolescents zigzag while fleeing cheetahs, although the tactic does not successfully increase they prey’s distance
from the predator in these age classes (Fitzgibbon 1990b). Furthermore, stotting is exhibited by
young ungulates more frequently than adults and likely serves to alert the infant’s mother to the
presence of a threat (Caro 1986b).
Monitoring and harassment. Rather than flee, ungulates sometimes remain near or approach
the predator in order to monitor or actively harass it. These behaviors serve to notify the
predator that it has been detected, intimidate the predator, and alert other prey in the area to the
predator’s presence (Berger 1979; Lipetz & Bekoff 1980; Fitzgibbon 1994). All of these factors
can encourage the predator to move on and hunt elsewhere where prey are not yet wary of it. In
Thomson’s gazelles, inspection and monitoring are common in response to stalking predators,
and are typically performed by entire groups (Walther 1969; Kruuk 1972; Fitzgibbon 1994;
Fitzgibbon & Lazarus 1995). Likewise, pronghorn groups sometimes monitor and harass coyotes
(Canis latrans; Berger 1979). These behaviors are risky for individuals, but less dangerous and more
effective when undertaken as cohesive efforts by groups (Kruuk 1972; Berger 1979; Fitzgibbon
1994).
In general, ungulate mothers with dependent young are more likely to harass predators than
non-mothers, and do so in order to prevent predators from successfully encountering or attacking
their offspring (personal observation; Lipetz & Bekoff 1980; Fitzgibbon 1988; Novaro et al. 2009).
Mothers are also more prone to predator monitoring, and remain alert to and in the vicinity of
predators longer than non-mother females (Chapter 5; Fitzgibbon 1993).
9
Grouping. Grouping, in addition to being an effective way to manage potential risk, can also
be useful during actual attacks (Lingle 2001). Predators may be confused or overwhelmed when
faced with a large number of potential prey items (the “embarrassment of riches hypothesis”;
Walther 1969; Pitcher 1986), and, as mentioned above, monitoring and harassing efforts
are more effective when undertaken in groups. Additionally, groups can be used as cover by
vulnerable individuals for concealment from predators, as seen in wildebeest calves (Estes & Estes
1979). Finally, groups allow the use of cooperative defensive formations, such as that seen in
musk oxen (Ovibos moschatus), in which adults create circles with vulnerable young and body parts
inside and their formidable weaponry facing out (Lent 1999).
Preventing offspring predation
We have seen that ungulate infants are particularly vulnerable to predation, and that the
antipredator behaviors described above are often used more frequently or more vigorously
by mothers with dependent young. In addition to using these general methods, mothers and
offspring exhibit behaviors geared specifically toward offspring protection. These include tactics
for controlling the circumstances of parturition, the hiding and following strategies, and active
maternal defense.
The first hours: behavior surrounding parturition. Although ungulates are relatively precocial
among mammals, after birth there is a brief period of time known as the postpartum period
when the offspring is incapable of engaging in defensive strategies (Chapter 2; Lent 1974). The
neonate’s physical inability to stand or walk prevents it from following its mother (Appendix
II; Roberts 2012b) and makes it extremely vulnerable if detected by a predator (Chapter 2).
10
Additionally, the scent of birth fluids and tissues on the offspring’s coat may attract predators
(Appendix I; Leuthold 1977; Roberts 2012a), and the initial inability of the mother and infant
to recognize each other would prevent them from reuniting after accidental or intentional
separation (Lent 1974; Estes & Estes 1979). Therefore, during the postpartum period, the
neonate must develop the ability to stand and walk, the mother must clean it thoroughly, and
the pair must begin to bond and learn to recognize each other through grooming and nursing
(Lent 1974; Estes & Estes 1979; Levy & Poindron 1987). Because mothers are also weakened
and vulnerable during and following parturition, their primary means of mitigating offspring
predation risk during the postpartum period is controlling the environment, social context, and
time of day of parturition.
Females often isolate from conspecifics and seek out special habitats for parturition. Mothers
may select birth sites with concealing vegetation (Chapter 2; Gosling 1969; Jarman 1976;
Leuthold 1977; Lott & Galland 1985), high visibility (Poole et al. 2007; Rearden et al. 2011;
Pinard et al. 2012), or low predator density (Bergerud et al. 1984; Alados & Escos 1988; FestaBianchet 1988) in order to minimize the risk of a predator encountering the neonate. However,
some mothers do not isolate and instead use conspecifics for cover (Chapter 2; Estes & Estes
1979; Lott & Galland 1985). In many species, offspring born within social groups are subject to
harassment and aggression from conspecifics that can cause injury, prevent the mother and young
from bonding appropriately, or delay the neonate’s behavioral development (Chapter 2; Gosling
1969; Lent 1974; Jarman 1976; Guinness et al. 1979; Read & Frueh 1980); however, in other
species harassment does not appear to occur or is easily fended off by the mother (Appendix II;
Estes & Estes 1979; Roberts 2012b).
Many species also exhibit bias in the time of day at which parturition occurs. In studies
of free-ranging populations, birth peaks tend to coincide with periods of low predator activity
11
(Fraser 1968; Jarman 1976; Estes & Estes 1979), and thus are hypothesized to reduce the risk of
a predator encountering the offspring during the postpartum period. Many captive and domestic
ungulates give birth at night (Dittrich 1967; Campitelli et al. 1982; Manski 1991; Pagan et al.
2009). This tendency may support the antipredator hypothesis, with humans assuming the role
of predators (Rossdale 1968), but some authors have proposed an alternative hypothesis that
birth peaks coincide with the period of least conspecific activity (Lent 1974). The benefit of
giving birth during periods of low conspecific activity is not made clear in the literature, but in
wild populations doing so may reduce maternal risk by allowing mothers to give birth without
being left behind by their social groups during labor. The strategy may also reduce conspecific
disturbances of the mother and young during the postpartum period (Chapter 2).
Hiding and following. After the postpartum period, ungulate mothers and infants engage in one
of two broad strategies of maternal care: hiding or following. The following strategy is used by all
perissodactyls and camelids, most caprids, and several large and highly mobile artiodactyl species
(Lent 1974; Fisher et al. 2002). Follower infants accompany their mothers and their social groups
Figure 1. Changes in motherinfant contact and distance with
offspring age for hiding and
following ungulates. Adapted by
Mike Costelloe from Lent (1974).
12
continuously from birth (Walther 1965; Lent 1974). As the infant ages, the amount of motherinfant contact decreases and mother-infant distances increase until the offspring is fully weaned
(Figure 1).
In contrast, most cervids, all gazelle, and most other antelope species are hiders (Lent 1974;
Fisher et al. 2002). Hider infants spend the majority of their first days or weeks of life hidden in
vegetation apart from their mothers. Hiding begins within a few hours of parturition when the
infant selects a nearby hiding spot and lies down, thus initiating a hiding period. The mother
retrieves the infant periodically throughout the day to initiate active periods, during which she
grooms and feeds the infant and allows it to play (Spinage 1968; Gosling 1969; Lent 1974;
Autenrieth & Fichter 1975; Leuthold 1977; Spinage 1986b). At the conclusion of the active
period, the mother may lead the infant to a different area before allowing it to select its next
hiding spot (White et al. 1972; Autenrieth & Fichter 1975; Ozoga & Verme 1986; Byers 1997).
This alternation of active and hiding periods persists for the duration of the hiding phase, which
lasts anywhere from one day (impala, Aepyceros melampus; Mooring & Hart 1997) to four months
(bushbuck, Tragelaphus scriptus; Kingdon 1982), but in most species lasts between two and four
weeks (Ralls et al. 1987b). Near the end of the hiding phase, the infant undergoes a transition
during which it increases the frequency of active periods by initiating them in the absence of its
mother (Chapter 4; Lent 1974; Leuthold 1977; Fitzgibbon 1990b). This results in shorter average
mother-infant distances and greater amounts of mother-infant contact (Figure 1; Lent 1974).
By the end of the hiding phase, the infant behaves similarly to a follower and accompanies its
mother and her social group constantly.
Follower infants rely on flight and maternal or group defense for protection from predators,
while hiders rely on concealment. Hider infants are often cryptically-colored (Lent 1974). When
13
in hiding they lie nearly motionless with their chins resting on the ground and their ears flattened
against their heads to reduce visibility (Gosling 1969; Spinage 1986a). They also lower their
heart and breathing rates, which may reduce noise and movement (Autenrieth & Fichter 1975;
Jacobsen 1979). Infants have little odor to give them away since their mothers consume their
urine and feces (Leuthold 1977) and the scent glands present in adults are inactive in infants
(Gosling 1969; Walther 1969; Autenrieth & Fichter 1975; Spinage 1986a).
The behavior of mothers is crucial to ensuring that predators do not detect hiding offspring.
Even if an infant is well-hidden, a predator may be able to reduce its search effort by observing
the mother’s behavior, which could give away the infant’s presence or even the location of its
hiding spot. Predators can either reduce the energy required to find the infant by locating a
mother and waiting for her to retrieve her offspring, or decrease the time and energy spent
searching by limiting their search to an area around mothers with hiding young. These strategies
and the methods ungulate mothers use to combat them have been studied in cheetahs (Fitzgibbon
1993) and coyotes (Byers & Byers 1983), predators of Thomson’s gazelle and pronghorn fawns,
respectively. Both strategies require that predators be able to distinguish between mothers with
hiding offspring and other females, which seems to be the case for cheetahs and coyotes, which
identify mothers by their increased vigilance levels and solitude. Pronghorn and gazelle mothers
combat the hide-and-wait strategy by increasing the probability that they will detect predators
through heightened vigilance (Chapter 3) and by delaying retrieval of their fawns if a predator
is present. In both species, mothers that detect predators extend their fawns’ hiding period
significantly, making it more worthwhile for a predator to seek other prey than to wait for the
fawn to emerge. Pronghorn also attempt to foil the “watch and wait” strategy by distributing their
activity evenly over the hiding period such that a coyote cannot tell by the mother’s behavior if
she is about to retrieve the infant (Byers & Byers 1983).
14
Mothers combat the restricted area search strategy by enlarging the area predators must
search to find the fawn. Byers and Byers (1983) found that pronghorn mothers maintain distances
to their hiding fawns that are large enough on average that a coyote searching in a circular
area around the mother would experience a higher energy gain by hunting ground squirrels.
Fitzgibbon (1993) inferred that average mother-infant distances of Thomson’s gazelle were
large enough to discourage the restricted area search strategy since predators were never seen to
systematically search around individual mother gazelle, although they do search around gazelle
herds during peak birthing times. Both species provide behavioral clues that predators could use
to reduce their search area: pronghorns orient and look towards their infants more often than
perpendicular to or away from them and gazelle vary their activity according to their distance
from the fawn such that when they are near their fawns they spend more time lying down and
being vigilant and less time walking and feeding (Fitzgibbon 1993). However, neither author finds
evidence that predators use these clues to narrow their search.
Infant predation events are difficult to observe, so data on the effectiveness of hiding are
scarce, but several studies and observations suggest that hiding is effective at reducing infants’ risk
of detection. Walther (1969) and Fitzgibbon (1990b) observed cheetahs, spotted hyenas (Crocuta
crocuta) and jackals passing within 5 meters of hiding Thomson’s gazelle fawns without detecting
them and Byers (1997) acknowledges that it is nearly impossible to find and capture hiding
pronghorn fawns unless one has witnessed the fawn selecting its hiding spot. Finally, Fitzgibbon
(1990b) found that cheetahs hunted more active Thomson’s gazelle fawns and fewer hiding fawns
than expected given the ratio of active to hiding fawns available and Barrett (1978) found that
pronghorn infant mortality increased with the waning of the hiding phase, both findings that
suggest that hiding is an effective defense against predation.
15
Maternal defense. Maternal defense behavior is documented in a variety of ungulate species,
including pronghorn (Autenrieth & Fichter 1975; Byers 1997), Thomson’s gazelle (Wyman
1967; Walther 1969; Fitzgibbon 1988), white-tailed deer (Smith 1987), mountain goats (Oreamnos
americanus; Côté et al. 1997), musk oxen (Lent 1999), mule deer (Odocoileus hemionus), elk (Cervus
canadensis), and moose (Lent 1974). Many follower ungulates rely heavily on maternal defense to
compensate for their infants’ conspicuousness and helplessness (Lent 1974; Berger 1978; Estes &
Estes 1979). For hiders, however, defense is a secondary strategy for use if hiding fails. Whereas
hiding serves to increase the search effort for predators hunting hidden young, maternal defense
increases the capture effort. Defense usually consists of direct attacks in which the mother butts,
kicks, bites or chases the predator. It can also include distraction behaviors such as darting
between the predator and infant during a chase (Walther 1969; Fitzgibbon 1988). Sometimes the
predator is injured or even killed (Lent 1974), but distraction or exhaustion is often sufficient as it
gives the infant a chance to flee or hide again.
Smith’s (1987) study of maternal defense in white-tailed deer reveals some correlations
that suggest that ungulate mothers tailor their defense behavior according to the expected costs
and benefits. In harsh seasons when the condition of the general population and the probability
of fawn recruitment were low, does defended less often and less aggressively than in bountiful
seasons when adults and fawns were in better condition. Behavioral adjustment is also seen in
Thomson’s gazelle in response to potential costs to the mother. Mother gazelle tailor their defense
behavior according to the species of predator based on the danger that species poses to adult
gazelle (Walther 1969). They readily attack jackals, attempt to distract cheetahs and hyenas,
but rarely interfere with wild dogs (Lycaon pictus; Wyman 1967; Walther 1969; Fitzgibbon 1988).
Gazelle mothers also adjust the frequency of defense according to the age of their offspring with
16
defense being common but not universal for fawns, less common for half-grown individuals,
and nonexistent for adolescents. This pattern suggests that older Thomson’s gazelle are better at
escaping predators on their own, so maternal defense is likely not beneficial enough to offset the
risk to the mother (Daly 1990).
Thomson’s gazelles and their enemies
Natural history
Thomson’s gazelles are small antelope native to East Africa. Adults stand approximately 65
cm at the shoulder, with males weighing 17 to 29 kg and females weighing 13 to 23 kg. Both sexes
have horns, although these are much shorter and thinner and are often broken or deformed in
females. Males’ horns are robust, S-shaped and 25 to 43 cm long (Estes 1991).
Thomson’s gazelles inhabit open grassland habitats, strongly preferring short-grass plains
to long-grass plains (Brooks 1961). Short grass makes predator detection easier and facilitates
efficient feeding by gazelle on young grass parts and forbs, which are more abundant and easier
to access in short grass (Gwynne & Bell 1968; Jarman & Sinclair 1979; Wilmshurst et al. 1999).
Male gazelle are territorial throughout the year and defend territories 100 to 200 meters
in diameter from other males (Walther 1969). Bachelor males and females form separate,
open-membership groups which roam across male territories and whose members can number
in the hundreds (Brooks 1961; Walther 1969). Females are wide-ranging, with day ranges
that encompass multiple male territories and home ranges as large as 9 km2 (Leuthold 1977).
Although males may attempt to prevent females from leaving their territories, there are no
barriers to females leaving or joining groups (Brooks 1961). Thomson’s gazelles breed and fawn
year-round but exhibit two birth peaks, the timing of which varies by location but which are
17
centered around the rains (Brooks 1961; Robinette & Archer 1971). Females can give birth to
two fawns per year: gestation lasts approximately 5.5 months and females can become pregnant
within two or three weeks of giving birth (Brooks 1961; Walther 1969; Hvideberg-Hansen 1970).
Newborn fawns weigh approximately 2.2 to 3 kg (Hvideberg-Hansen 1970). Fawns are weaned
by six months of age and male offspring disperse to bachelor groups two months later (Brooks
1961).
Thomson’s gazelle is a typical hider species. Fawns hide intensively for the first two weeks of
life and decrease hiding behavior thereafter until eight weeks of age when they no longer hide at
all (Walther 1969; Fitzgibbon 1988, 1990b). Females may isolate for parturition (Chapter 2), and
are often found alone and in tall grass during the first weeks of the fawn’s life (Chapter 4; Brooks
1961; Walther 1969; Fitzgibbon 1988).
Predators
Due to their small size, young and adult Thomson’s gazelles are vulnerable to many
predator species including lions (Panthera leo), cheetahs, leopards (Panthera pardus), wild dogs, and
hyenas (Brooks 1961; Estes & Goddard 1967; Walther 1969; Borner et al. 1987). Jackals, birds of
prey, and baboons prey on young gazelles, but pose little threat to adults (Wyman 1967; Walther
1969). Cheetahs, jackals and hyenas are reported to specialize on fawns during birthing peaks
and are likely the most important predators of this age class in most populations (Wyman 1967;
Walther 1969; Fitzgibbon 1990b, 1993). Here we call special attention to black-backed jackals
(Canis mesomelas; the most common threat to fawns at our field site), cheetahs (the subject of most
previous research on gazelle predation), and warthogs (Phacochoerus africanus; an unconventional
predator of fawns).
18
Black-backed jackals. The black-backed jackal is a small-bodied canid species that occupies
two distinct ranges in East and southern Africa (Loveridge & Nel 2004). Jackals are generalist
predators and scavengers, feeding on a mix of food resources including fruits, insects, small
mammals and carrion, but specializing on the hidden young of ungulates during birthing seasons
(Wyman 1967; Klare et al. 2010). Jackals are capable of killing adult ungulates as well (Krofel
2007; Kamler et al. 2010), but this is a rare occurrence and the risk posed to adults by jackals
is low (Estes 1967; Macdonald et al. 2004). In the East African population, Thomson’s gazelle
fawns are the most common prey item, constituting more than half of the animal remains
found in jackal feces. Observations suggest that jackals usually kill fawns themselves rather
than scavenging them (personal observation; Wyman 1967; Klare et al. 2010). In the southern
population where Thomson’s gazelles do not exist, jackals instead specialize on similarly-sized
springbok antelope (Antidorcas marsupialis) fawns and may consume up to a third of the springbok
population in some areas (Klare et al. 2010). Based on their predation of fawns, jackals may be
the most important predator of Thomson’s gazelles overall (Estes 1967) and therefore have the
potential to strongly affect population dynamics in this species.
Jackal hunts of Thomson’s gazelle fawns usually begin with the jackal either coming across
a hidden fawn during general prey searches or spotting an exposed fawn from a distance and
running in to capture it (personal observation; Wyman 1967). Unlike coyotes hunting pronghorn
(Byers 1997), jackals, upon detecting an active fawn, rarely wait for it to hide or for its mother
to leave, and instead rush in and attack immediately (personal observation). This may allow the
jackal to target the visible fawn and avoid the risk of losing site of the fawn’s hiding place during
its approach. Furthermore, although gazelle mothers vigorously defend their fawns from jackals,
this defense rarely results in any apparent injury to the jackals (personal observation; Wyman
1967), so there is little preventing the jackal from attacking while the mother is nearby.
19
Mother gazelle exhibit a range of deceptive and aggressive behaviors aimed at protecting
their fawns against jackals (personal observations). If a mother spots a jackal before it has
detected her active fawn, she immediately rushes away from the fawn as it drops down and
becomes prone on the ground. At a distance of 50 meters or more from her fawn, the mother
monitors the jackal intently. If the jackal has not observed the mother’s response, it often moves
through the area without detecting the fawn; however, if it has noticed the mother’s behavior it
begins to search the area. Until it becomes clear that the jackal has detected or is about to detect
the fawn, the mother does nothing but monitor the jackal, maintaining or increasing her distance
from her young. When the jackal comes very close to the fawn, the fawn will either remain
hidden (very young fawns) or jump up and run away stotting (more common in fawns older than
two weeks; personal observation). In either case the mother begins repeatedly charging the jackal
while butting and kicking at it. This serves to prevent the jackal from finding the hidden fawn, or
to position herself between the jackal and the fleeing fawn (personal observation; Wyman 1967).
If the fleeing fawn gets far enough away from the jackal, it will suddenly drop down in the grass.
Often in this case the jackal has been distracted enough by the mother’s onslaught that is does
not see exactly where the fawn has hidden, and continued aggression by the mother prevents it
from effectively searching for the fawn.
Maternal defensive efforts tend to be more successful when there is sufficient cover for the
fawn to hide in, when there are multiple mothers defending simultaneously, and when only a
single jackal is attacking (personal observation; Wyman 1967). Success also depends heavily on
the fawn’s ability to flee effectively. In three observations of jackal attacks on gazelle neonates
that were not able to walk steadily (Chapter 2), mothers attempted to prevent the jackals from
detecting the young. However, once the fawn was detected all maternal defense stopped, likely
20
because the fawn’s inability to run away would doom the defensive effort. Fitzgibbon (1988)
found that mothers defended in 76% of jackal hunts of fawns and that 50% of fawn hunts were
successful.
Cheetahs. Cheetahs occur in scattered populations across much of Africa, and are present in
most protected areas in Kenya (Ray et al. 2005). In 2012, the cheetah population of Ol Pejeta
Conservancy consisted of approximately 30 individuals (Ol Pejeta Conservancy, unpublished
data). Although I did not observe any cheetah hunts of Thomson’s gazelles, I did observe two
cheetahs consuming recently captured adult prey.
Cheetahs are ambush predators. Typical hunts start with a stalking phase during which
they approach their prey as closely as possible without being detected and select their targeted
prey individual (Schaller 1968; Eaton 1970b; Eaton 1970a). The stalk is followed by an attack
phase consisting of an explosive sprint during which they attempt to trip their prey so that it can
be subdued and suffocated with a bite to the throat (Schaller 1968; Eaton 1970a, b). Cheetahs
are capable of short bursts of speeds up to 103 kph, making them the fastest terrestrial animals
(Sharp 1997). Although they are capable of taking prey ranging in size from hares (2 kg) to adult
zebra (270 kg), they specialize on locally abundant ungulate species weighing between 23 and
56 kg (Hayward et al. 2006a). Where Thomson’s gazelles are available, they are cheetahs’ most
commonly killed prey item and are taken more often than predicted given their abundance
relative to other prey species (Schaller 1968; Hayward et al. 2006a).
When hunting Thomson’s gazelle, cheetahs preferentially target solitary individuals and
small groups (fewer than ten individuals), especially those in tall grass. Within groups, cheetahs
select less vigilant prey and prey that are located on the edge of the group closest to the cheetah
21
(Fitzgibbon 1988, 1990a). The element of surprise is crucial to cheetahs’ hunting success:
cheetahs are successful in only 6% of hunts when the prey group detects them during the stalk
phase compared to 25% when they were not detected. They abandon more than half of hunts
in which they are detected prior to the chase (Fitzgibbon 1988). Cheetahs kill more male gazelle
than female gazelle, likely due to territorial males’ greater tendency to be alone. However, they
are more successful at killing solitary female gazelle than solitary male gazelle, suggesting that
mothers that isolate during the hiding phase are at especially high risk of predation by this species
(Fitzgibbon 1988, 1990a).
In addition to adults, cheetahs regularly hunt gazelle fawns when they are available and are
highly effective at capturing this age class (Eaton 1970b; Fitzgibbon 1988). Whereas only 20.6%
of hunts of adult gazelle are successful, 70.5% of hunts of fawns result in a kill (Fitzgibbon
1988). Mothers never attack cheetahs to defend their fawns, but engage in distraction behavior
– running and stotting alongside the cheetah – in 30% of fawn hunts (Fitzgibbon 1988). Due to
the ineffectiveness of this distraction behavior and the fawn’s lack of running speed and agility, a
fawn’s primary defense when detected by a cheetah is to drop down and hide (Fitzgibbon 1988,
1990b). In Fitzgibbon’s (1990b) study in the Serengeti, 50% fawns that dropped down survived
while 77% of those that fled were captured. Fawns that dropped down and survived dropped
while the cheetah was further away than fawns that were found and killed. The distance from the
cheetah at which a fawn drops depends heavily on its mother: more vigilant mothers were able to
warn their fawns of approaching cheetahs sooner than less vigilant mothers.
Warthogs. Although other suid species, such as wild boar (Sus scrofa; Pavlov & Hone 1982;
Choquenot et al. 1997; Wilcox & Van Vuren 2009), are known to hunt and kill juvenile ungulates,
22
to our knowledge this behavior has never before been documented in warthogs. As with all
suids, animal protein features regularly in warthog diets, usually in the form of small vertebrates,
invertebrates and eggs (Cumming 1975; Leus & Macdonald 1997). Warthogs and other suids are
also known to consume carrion, and large vertebrate matter in suid diets is often assumed to have
been scavenged (Wilcox & Van Vuren 2009).
I observed two instances of warthog predation of gazelle fawns (Chapter 2), one attack
in which the fawn was contacted but not killed (Appendix I; Roberts 2012a), and two events
wherein warthogs chased and knocked down fawns, but the fawns recovered and escaped before
further contact could be made. In three instances the attacker was a single adult male warthog;
in one (wherein the fawn escaped), the attacker was an adult female; and in one (wherein the
fawn was killed) approximately eight male warthogs took part in the attack. Attack sites were
distant from one another, indicating that the same warthogs were unlikely to be involved in
multiple attacks. Based on my observations, warthogs appear poorly equipped to subdue all but
the youngest gazelle fawns: both fawns that were killed were neonates less than an hour old. The
fawn that was attacked but not killed was also less than an hour old; it is unclear why the warthog
did not proceed to eat the fawn after subduing it (Appendix I; Roberts 2012a). Of the two fawns
that escaped, one was several hours old and one was several days old. Similarly to wild boar
(Pavlov & Hone 1982), warthogs catch fawns by chasing them and knocking them to the ground
with their snouts. The warthog then slowly approaches the fawn and sniffs it before holding it
down with a forehoof and tearing at the flesh with its mouth. While newborn gazelles are easy to
catch and incapacitate, fawns that are more than several hours old appear too fast too catch or
are agile enough to recover quickly after being knocked down (personal observation).
Contributing to the vulnerability of gazelle fawns to warthogs is the fact that gazelle mothers
do not appear to regard them as predators. Whereas conventional predators such as jackals and
23
cheetahs are regarded with extreme caution, warthogs often form mixed-species groups with
Thomson’s gazelles. In all attacks, the warthogs were ignored by the mothers until they were close
enough to the fawn to displace the mother. Likewise, mothers did not hesitate to retrieve their
hidden fawns in the presence of warthogs, but never retrieved their fawns if they detected jackals
in the area (Appendix I; Roberts 2012a).
The lack of previous records of this predatory behavior is likely due to the rarity of the
event itself. The window of opportunity for warthogs to kill fawns is short, and the tendency of
some mothers to isolate from social groups for parturition may limit interactions between the
two species during this period (Chapter 2; Appendix I; Roberts 2012a). In 2011 and 2012, when
all observations of warthog predation took place, warthog density was high at my field site (Ol
Pejeta Conservancy, personal communication). This may have increased the rate of interaction
of warthogs and newborn gazelle, leading to my observations.
Other predators. In addition to the above species, gazelle at our field site were vulnerable to
lions, spotted hyenas, wild dogs, leopards, olive baboons (Papio anubis), and various birds of prey
(Chapter 4). Lions tend to prefer large-bodied prey (between 190 and 550 kg) and therefore
are not likely to be major predators of Thomson’s gazelles (Hayward & Kerley 2005). Hyenas
typically take prey between 56 and 102 kg in body weight, but are extremely generalist in their
prey selection. Therefore, Thomson’s gazelles, although smaller than hyenas’ preferred prey,
are still among this species’ most common prey items due to their abundance (Hayward 2006).
Gazelle fall within the leopard’s preferred prey size range of 10 to 40 kg (Hayward et al. 2006c),
and in the Serengeti they constituted approximately a quarter of animals taken by leopards,
although this was less than expected given their high availability (Kruuk & Turner 1967). Adult
24
Thomson’s gazelles are a preferred prey item of wild dogs (Estes & Goddard 1967; Kruuk &
Turner 1967; Hayward et al. 2006b), and this species is an important predator of gazelle where
they co-occur. Baboons typically only prey on gazelle fawns. Kills tend to be opportunistically
carried out by adult males, although intentional searching around gazelle groups does occur
(Harding 1973).
Finally, birds of prey such as martial (Polemaetus bellicosus) and tawny eagles (Aquila rapax) are
thought to be predators of gazelle fawns (Corinne Kendall, personal communication). Martial
eagles as well as other raptors are known to take larger ungulate prey (Wright 1960; Byers 1997;
Hamel & Côté 2009; Kerley 2013). However, to my knowledge no published records of avian
predation specifically on Thomson’s gazelles exist. I observed two unsuccessful attempts by tawny
eagles to take fawns; in one instance the swooping eagle missed the running fawn (which was
soon killed by jackals attracted by the commotion), and in the second instance the mother ran to
stand over her fawn and butted at the swooping eagle, effectively deterring it. The eagle remained
perched nearby and the mother remained lying or grazing within several meters of the hidden
fawn until the eagle departed. The mother’s behavior in the latter observation is similar to that
reported in pronghorn does defending against golden eagles (Aquila chrysaetos; Byers 1997).
Apart from wild dogs, little information exists on gazelle behavior in response to these
predators, and, besides the aforementioned tawny eagles encounters and one half-hearted chase
of a fawn by a male baboon, I observed no predation attempts by these species. Fitzgibbon
(1988) contrasts some aspects of antipredator behavior in response to cheetahs and wild dogs.
The differences in antipredator behavior stem from the disparate hunting styles of ambush
predators (cheetahs), which choose their prey prior to the chase and rely heavily on the element
of surprise, and coursing predators (wild dogs), which evaluate the vulnerability of multiple
25
potential targets during the chase and rely on their superior endurance to run down prey that
may or may not detect them beforehand. Gazelle stot more often in response to wild dogs, and,
aside from mothers and fawns, rarely stot in response to stalking predators such as cheetahs and
lions. This supports the hypothesis that this behavior serves to advertise an individual’s fitness to
choosy predators (Fitzgibbon 1988; Fitzgibbon & Fanshawe 1988). Further differences emerge in
distraction and inspection behavior. Mothers rarely attempt to distract wild dogs hunting their
fawns, but try to distract cheetahs in 30% of hunts (Fitzgibbon 1988). Likewise, less than 5% of
gazelle groups engage in inspection of wild dogs and hyenas, but 39 and 52% of groups inspect
lions and cheetahs, respectively (Fitzgibbon 1994). In the case of cheetahs, the prey’s awareness
of the stalking predator removes the necessary element of surprise and makes harassment or
approach relatively safe. However, given that coursing predators can effectively attack aware prey,
proximity to these species is always very risky for gazelle (Fitzgibbon 1988). Finally, Estes and
Goddard (1967) note that territorial males and mothers with hiding fawns are highly vulnerable
to wild dog predation due to their tendency to eventually turn back toward their territories and
hiding young when chased. This is likely not a determining factor of the outcome of cheetah
chases, which generally are of shorter duration.
Overview of chapters
The chapters of this thesis examine maternal behavior throughout various stages of infancy.
Chapter 2 addresses behavioral tactics surrounding parturition and their effects on the survival
of gazelle neonates. Chapter 3 examines maternal vigilance and activity budgets during the
hiding phase, focusing on the effect of hiding on maternal activity and the use of vigilance to
mitigate fawn risk. Chapter 4 moves to the transitional phase, when fawns are reducing hiding
26
activity and becoming more like followers. This chapter examines the increase in risk to fawns
and the behavioral changes exhibited by their mothers. Chapter 5 discusses maternal behavior in
response to immediate risk, elicited via experimentally-simulated predation events. The behavior
of mothers and non-mother groups are compared and contrasted and the influence of maternal
responses on group responses are investigated. Finally, Chapter 6 offers an overview of the main
findings of the previous chapters along with concluding remarks.
27
Chapter 2: Maternal tactics for mitigating neonate predation risk during the postpartum
period in Thomson’s gazelle1 Blair A. Roberts and Daniel I. Rubenstein
Abstract
The time immediately following birth is a period of high predation risk for ungulate
neonates. Ungulate mothers exhibit perinatal behaviors that appear to mitigate offspring risk
during this time. However, few studies of infant mortality include the postpartum period.
Therefore, the function and effectiveness of these maternal behaviors are untested. We observed
perinatal behavior in eleven Thomson’s gazelle mother-infant pairs in a free-ranging population
under predation pressure. Five of the six fawns that were detected by predators during the
perinatal period were killed. Fawn survival therefore depended on avoiding detection by
predators. Considered individually, neither prepartum isolation from conspecifics nor birth site
selection affected the risk of being detected by a predator. However, analyses revealed two distinct
perinatal tactics: mothers either isolated and gave birth in tall grass or remained in their social
groups and gave birth in short grass. Both of these tactics resulted in lower predator detection
rates compared to behavior that was inconsistent with either tactic. The tactics represent a
maternal trade-off between minimizing the duration of the highly vulnerable postpartum period
and minimizing conspicuousness to predators.
Conference presentations on this work:
Roberts, B.A. and Rubenstein, D.I. 2013. Maternal strategies for reducing calf predation risk in Thomson’s
gazelle. Animal Behavior Society Annual Meeting, Boulder, Colorado. 29 August, 2013.
Roberts, B.A. and Rubenstein, D.I. 2013. The effects of perinatal behavior and maternal decisions on neonate
survival in a hiding ungulate. American Society of Mammalogists Annual Meeting, Philadelphia, Pennsylvania.
17 June, 2013.
1
28
Introduction
Ungulates, like most mammals, experience their highest mortality rates during infancy
(Caughley 1966). Where predators are present predation is the primary threat to offspring
survival (Linnell et al. 1995; Rubenstein & Nuñez 2009). Each species uses one of two main
strategies, following or hiding, to mitigate offspring predation risk. Follower infants accompany
their mothers continuously from birth and rely on flight and maternal defense to escape attacking
predators (Walther 1965; Lent 1974). Hider infants spend the majority of their time lying
motionless in vegetation apart from their mothers, who return several times per day to care for
their young. Both strategies are effective at protecting ungulate offspring (Walther 1965; Lent
1974). However, neither strategy can be implemented during the period immediately after birth,
known as the postpartum period (Lent 1974).
The postpartum period is perhaps the most vulnerable time of an ungulate’s life. Ungulates
are born unable to stand, walk or run and are coated in odiferous birth materials that hinder
their movements and attract predators (Leuthold 1977; Roberts 2012a, b). Furthermore, they
are unable to recognize their mothers, who themselves have not yet bonded to or learned to
recognize their offspring (Lent 1974; Estes & Estes 1979). During the postpartum period, the
mother and infant must efficiently rectify these conditions so that the neonate may engage in and
benefit from the hiding or following strategy. Until hiding or following begins, the neonate’s safety
depends on the behavior of its mother and the conditions under which she has chosen to give
birth.
In many ungulate species, parturient females seem to mitigate neonate risk during the
postpartum period by controlling the ecological and social context in which parturition occurs.
Mothers may select birth sites with concealing vegetation (Gosling 1969; Jarman 1976; Leuthold
29
1977), high visibility (Poole et al. 2007; Rearden et al. 2011; Pinard et al. 2012) or low predator
density (Bergerud et al. 1984; Alados & Escos 1988; Festa-Bianchet 1988) in order to minimize
the risk of a predator encountering the neonate. In gregarious species, birth site selection is
often accompanied by isolation from social groups (Fraser 1968; Gosling 1969; Jarman 1976;
Leuthold 1977; Guinness et al. 1979; Byers 1997). Isolation is thought to reduce non-predatory
disturbances such as sexual harassment or aggression by conspecifics (Gosling 1969; Lent 1974;
Estes & Estes 1979; Guinness et al. 1979), which can result in injury or failure of the mother and
infant to bond appropriately and increase predation risk by delaying the onset of following or
hiding (Jarman 1976; Estes & Estes 1979; Read & Frueh 1980).
Within species, females may vary in their tendency to exhibit birth site selection and
prepartum isolation. For example, Lott and Galland (1985) found that female bison (Bison bison)
are more likely to isolate in wooded habitats compared to open plains. The authors suggest
that these behaviors serve to reduce predation risk during parturition: bison females prefer
to isolate and give birth in areas with vegetative cover where they and their fawns will be less
conspicuous to predators, but will use conspecifics for cover when vegetation is unavailable.
This second option is less favorable and only used in the absence of vegetative cover because it
exposes the mother-calf pair to interferences by conspecifics. This study indicates that species
may exhibit multiple conditionally-dependent tactics for reducing neonate predation risk. Major
bison predators were absent in Lott and Galland’s (1985) study and no predation attempts were
observed; therefore, the relative effectiveness of the tactics in preventing predation could not be
determined.
To date, no study has provided an integrated investigation of the effects of habitat cover
and prepartum isolation on offspring survivorship during the perinatal period. The postpartum
period is difficult to predict and observe due to its brevity and the skittishness of parturient
30
females (Fraser 1968). Studies of ungulate infant mortality therefore rely on remote monitoring
of infant survival beginning hours or days after birth, once the offspring has already survived the
vulnerable postpartum period (e.g. Smith & Anderson 1996; Singer et al. 1997; Janermo et al.
2004; Olson et al. 2005; DeVivo et al. 2011; Buuveibaatar et al. 2013). Information regarding
neonate mortality risk and the role of maternal decisions in determining neonate survival during
the postpartum period is thus integral to our understanding of ungulate population dynamics.
For this study we observed perinatal behavior and neonate predation events in free-ranging
Thomson’s gazelles, a hider species. Mothers in the study population exhibited variation in
their use of perinatal tactics, which enabled us to conduct the first integrated analysis of these
behaviors and their effects on neonate survival. Given that gazelle neonates are incapable of
fleeing predators, we expect that detection of the fawn by predators will result in predation.
If this is true, maternal efforts to prevent offspring predation must prevent detection of the
neonate by predators. Although black-backed jackals, the primary fawn predators at our field site,
sometimes detect fawns exclusively through olfaction, our observations of this species’ hunting
behavior suggest that most attacks begin with visual detection of either the fawn itself or its
mother. Mothers can be distinguished from non-mothers by their greater tendency to be in tall
grass and isolated from conspecifics (Fitzgibbon 1993). In most attacks, jackals either run directly
at exposed fawns or identify a likely mother from a distance and then search around her for
hidden offspring (BR, personal observation; Fitzgibbon 1993; Byers 1997).
Based on Lott and Galland’s (1985) findings in bison, we hypothesize that prepartum
isolation and birth site selection do not vary independently, but rather that some combinations of
habitat type and social context will be more effective at preventing fawn predation than others,
and that these combinations will be favored by parturient females. After determining overall
patterns of perinatal tactic use, we examine possible mechanisms by which these behaviors serve
31
to reduce fawn predation risk. Specifically, we test whether isolation from conspecifics prior to
birth reduces the occurrence of non-predatory disturbances, thereby minimizing the duration of
the postpartum period and reducing the fawn’s risk of being detected by predators before hiding
commences. We also investigate whether vegetative concealment of the fawn at the birth site
affects its risk of detection by predators.
Methods
Study site
We conducted all fieldwork at Ol Pejeta Conservancy (OPC) in Laikipia, Kenya. OPC
is a fenced, 360 km2 conservancy consisting of discrete grassy plains less than 10 km2 in
area separated by Acacia drepanolobium and Euclea divinorum woodlands. Animals on OPC are
well-habituated to vehicles; therefore, we were able to approach individuals to distances of
approximately 100 to 150 meters. At this distance we were able to use binoculars to observe
behavior of parturient females and their fawns while causing minimal disturbance to our subjects.
We searched for parturient females and observed births and perinatal behavior opportunistically
from approximately 7:30 AM to 6:30 PM during four field trips from March to June 2010, March
to June 2011, August to November 2011, and June to September 2012. Thomson’s gazelle give
birth year-round but exhibit birth peaks roughly coinciding with the semiannual rains (Brooks
1961; Hvideberg-Hansen 1970; Robinette & Archer 1971), which at OPC occur from March
to April and October to November. In 2012, OPC had a population of approximately 1300
Thomson’s gazelles. Other herbivore species are abundant, as are predators, including lions,
cheetahs, black-backed jackals, spotted hyenas, olive baboons, and birds of prey.
32
Observations
We observed perinatal behavior by locating parturient females or females with neonates
that were still wet and unable to walk steadily. Parturient females were easiest to spot when birth
membranes were already protruding from the vulva or adhering to the hind legs, but we were
also able to identify females in earlier stages of labor by their raised tails, enlarged abdomens,
and swollen udders. After finding a female, we observed her for the remainder of the postpartum
period, until her fawn was killed or began hiding. We recorded seven observations using a video
camera and three using a voice recorder and still photos. One further birth was witnessed and
voice-recorded by a colleague.
In addition to narrating the behavior of the female and her fawn, during each observation
we noted the time of birth, whether the female was alone or in a social group, and the height
of the grass at the birth site. For two of the three observations in which we did not witness
parturition, we estimated time of birth based on the fawn’s level of development when we
encountered it; for observation 10, we could only confidently estimate time of birth to within
a half hour. Females were considered to be in a group if they were less than 50 m from their
nearest neighbor. We classified the grass at the birth site as either short or tall (below and above
the hock, respectively).
From the recorded observations, we quantified disturbances to the mother-fawn pair and
determined the fawn’s conspicuousness. Any separation of the mother and fawn by more than
one adult body length resulting from an outside influence or a fear reaction by the mother,
but which did not end with fawn predation, was considered a disturbance. Disturbances were
caused by male Thomson’s gazelles attempting to engage in sexual activity with the mother or
neighboring females (three observations); the neonate eliciting fear and flight reactions from the
33
mother (one observation); female Thomson’s gazelles enticing the fawn away from its mother
or aggressively interacting with the fawn or mother (two observations); Grant’s gazelles (Gazella
granti; one observation), plains zebras (Equus burchelli; one observation), and female Thomson’s
gazelles investigating the fawn or birth site (one observation); humans passing near the mother
on foot (one observation); and an attack of a fawn by a warthog (one observation; Appendix I;
Roberts 2012a). We counted the disturbances for each observation and measured their durations.
We could measure fawn conspicuousness for the video-recorded observations only. We
quantified the conspicuousness of the neonate using still frames taken every 60 seconds beginning
one minute after birth. For each image we noted whether the fawn was lying down or standing up
and estimated the percentage of the fawn that was visible to the nearest 10%. We then compared
mean visibility for each fawn position across observations.
Analyses
We ran all statistical analyses in JMP v10.0. We first used Fisher’s exact test, which is valid
for small sample sizes, to determine if detection by predators resulted in fawn mortality. We
then tested an overall nominal logistic model to determine whether maternal decisions affected
the fawns’ risk of being detected by predators. We included birth site grass height (short or tall),
social context (isolated or not isolated), and the interaction between these two variables as model
effects, and used detected (yes or no) as the response variable. We evaluated the model using effect
likelihood ratio tests.
After establishing the overall effect of maternal behavior on fawn detection by predators,
we went on to examine the underlying mechanisms mediating this effect. We investigated
whether prepartum isolation reduced the number of non-predatory disturbances of the mother
34
Time delay to fawn behavior (min)
Time to hiding
Time to walking
200
150
y=47.1+1.0x
p=0.0002
r2=0.98
100
50
0
y=20.6+0.9x
p=0.014
r2=0.82
0
50
100
150
Disturbance duration (min)
Figure 2. The effect of non-predatory disturbances on the time delay from birth to the onset of hiding
(open circles, grey line) and the fawn’s first walking bout (closed circles, black line). Fawns that spend more
time separated from their mothers due to disturbances during the postpartum period take longer to walk
and hide.
and fawn, and whether these disturbances prolonged the duration of the postpartum period
and thereby increased the fawn’s risk of being encountered by a predator. We used Fisher’s
exact test to determine if prepartum isolation reduced disturbances. To determine the effect of
disturbances on the duration of the postpartum period, we first used linear regression to examine
the relationship between total disturbance duration and the time it took the fawn to hide (Figure
2). This relationship, although strong (r2=0.98, p=0.0002), is based on six observations and driven
in part by an outlying data point. Furthermore, in two of these observations, the mother-fawn
pairs experienced no disturbances. The dearth of intermediate data points is due to the fact
that fawns that are killed never hide; therefore, we have no data on the time delay from birth to
hiding for these fawns. In order to include fawns that were killed, we also examined the effect
of disturbances on the time it took the fawn to walk. This landmark falls late enough in the
fawn’s behavioral development that disturbances were likely to happen beforehand, but early
enough that most fawns attained the landmark before being killed. Although there were only four
35
observations in which the time delays to both walking and hiding were recorded, these points had
a high correlation coefficient (r=0.83). Given this apparent relationship and because walking is a
prerequisite for hiding, we proceeded under the assumption that fawns that take longer to walk
will also take longer to hide.
We constructed two Kaplan-Meier survival curves to examine the effect of postpartum
period duration on risk of a predator encounter. One curve modeled the risk of encounter
by conventional predators and the other modeled the risk of encounter by all predators. The
first curve was constructed using encounter times for fawns that were encountered by jackals,
with encounter times by other predators and hiding times included as censored values. In the
second curve, all encounter times were used regardless of predator, with hiding times included
as censored values. The hiding time for observation 1 was not included in either curve since
this fawn was encountered by a predator before hiding and its encounter time is included in
the model. We fit log-normal curves to the Kaplan-Meier curves and plotted these on a semilog plot. The use of the logarithimic scale on the y-axis allows better visualization of encounter
rates: in this format, lines with constant slope represent constant rates of predator encounter
(Deevey 1947). We used the log-normal curves to determine the survival probability for observed
postpartum period durations and to predict the postpartum period durations necessary to yield a
variety of survival probabilities.
We also examined whether birth site grass height affected fawn visibility, and whether more
visually conspicuous fawns were more likely to be encountered by predators. The Shapiro-Wilk
W Test showed our data to be non-normal. We therefore used the Mann-Whitney U-Test, which
is appropriate for nonparametric data, to compare fawn visibility between different grass heights
across video-recorded observations. We used Fisher’s exact test to determine whether the visibility
of the fawn at the birth site affected the risk of predator encounter.
36
All Fisher’s exact tests were two-tailed, and significance in all tests was accepted at p<0.05.
Because samples sizes are necessarily small, we identify ‘strong trends’ when p<0.1. We do so
hoping that such trends will stimulate further studies. In JMP, the Mann-Whitney U-Test is
performed using the Wilcoxon test command, and the Z-statistic rather than the U-statistic is
reported. Therefore, we report our results using the Z-statistics and Wilcoxon test format.
Results
In more than 13 months of studying maternal behavior in Thomson’s gazelles, we observed
eleven instances of perinatal behavior. Eight observations included parturition and prepartum
behavior, and three began very shortly after parturition when the fawn was still wet and unable to
walk steadily. All fawns exhibited similar levels of movement and vigor after birth and there was
no indication that any were born debilitated. Table 1 provides data from each of our observations
including time of birth, use of antipredator tactics, and the fate of the neonates.
Mothers in our study did appear to actively choose the circumstances of parturition rather
than simply giving birth wherever they happened to be when labor began. We detected signs
of labor as early as 90 minutes prior to parturition and females did not settle at the birth site
until between 8 and 43 minutes prior to parturition. In the interim, we observed some subjects
leave their groups and travel (in one case more than a kilometer) from their original position,
sometimes moving through multiple grass heights and joining and leaving multiple groups. All
females had access to both tall and short grass habitats on the plains on which they gave birth.
Females that sought isolation invested considerable effort in shedding herding males and females
that followed them away from groups. In two cases the female’s decision to isolate or remain
in the group was negated when the group rejoined or moved away from her after she became
immobile during labor. In both of these instances, the fawn was killed by jackals.
37
38
1:43 PM
10:33 AM
7:46 AM
10:16 AM
10:50 AM
11:05 AM
3:04 PM
9:55 AM
3:06 PM
10:30 11:30 AM
2:15 PM
1
2
3
4
5
6
7
8
9
10
11
No
No
Yes
Yes
No
Yes
No
No
Attempt
to isolate?
No
No
Yes
No
Yes
Yes
Yes
No
No
No
No
Isolated
at birth?
Short
Tall
Tall
Short
Tall
Tall
Short
Short
Tall
Short
Short
Birth site
grass height
49.8
NR
NR
19.6
21.5
NR
30.4
63.9
11.8
NR
Fawn
walks
NR
23.5
36.0
68.4
13.7
Fawn
killed
70.0
Yes
53.0
Yes
Yes
No
No
No
No
No
No
Yes
No
Consp.
No
No
No
No
Yes
No
No
No
Yes
Yes
Yes
Heterosp.
Disturbed?
46.6
33.6
63.9
220.0
Fawn
hides
Postpartum events
(minutes after birth)
22.3
NR
9.0
2.0
6.5
0
0
0
52.8
2.8
166.0
Total
disturb.
duration
(minutes)
Killed
(warthogs)
Survived
to hiding
Killed
(jackals)
Survived
to hiding
Survived
to hiding
Survived
to hiding
Killed
(jackals)
Survived
to hiding
Killed
(jackals)
Killed
(warthog)
Fawn fate
Survived
to hiding
Table
Time ofof
parturition,
prepartum
decisions, postpartum
events,
and outcomesevents,
for the observations
discussedfor
in this
study.
NR indicates that
the event occurred,
the datum
not recorded.
Table
1.1.Time
parturition,
prepartum
decisions,
postpartum
and outcomes
the
observations
discussed
in this but
study.
NR was
indicates
that
Blank cells indicate that the event did not occur during the observation (e.g. the fawn was killed before the event occurred). Parturition was not observed for observations 5, 8 and 10; parturition times
the for
event
but the are
datum
wasbased
notonrecorded.
Blank
cells indicate
thatwhen
theweevent
did not
occur during
the ofobservation
(e.g.(italicized)
the fawn
was
these occurred,
observations (italicized)
estimated
the fawn’s level
of behavioral
development
encountered
the mother-fawn
pair. Times
postpartum events
for these
observations
are based
off of
estimated birth
times. Because
timenot
of birth
for observation
could only be broadly
estimated,
times for postpartum
events
notobservations
be calculated for this
observation.are
killed
before the
event
occurred).
Parturition
was
observed
for 10
observations
5, 8, and
10; parturition
times
for could
these
(italicized)
estimated based on the fawn’s level of behavioral development when we encountered the mother-fawn pair. Times of postpartum events (italicized)
for these observations are based off of estimated birth times. Because time of birth for observation 10 could only be broadly estimated, times for
postpartum events could not be calculated for this observation.
Time of
birth
Obs.
Prepartum decisions
Fawns were attacked by black-backed jackals and warthogs. Females did not appear to
regard warthogs as predators. Mothers reacted to jackals and other conventional fawn predators
with high levels of vigilance, whereas warthogs were often interspersed in gazelle groups and
were ignored until they displaced the mother. During jackal attacks, mothers became alert and
fled the birth site as soon as they detected the predator. In contrast, in warthog attacks, the
mother continued caring for the fawn until the warthog approached to within a body length
(Roberts 2012a). The differences in maternal reactions to warthogs and jackals suggest that
warthogs are an emerging predator at our field site.
Effects of detection by predators on fawn survival
Six (55%) fawns were detected by predators during the postpartum period. Three of these
were found by black-backed jackals and killed. Three were found by warthogs. Of these, two
were killed and eaten and one was attacked but survived (see Appendix I; Roberts, 2012a). A
predator encounter during the postpartum period thus drastically decreased a fawn’s chance of
surviving (Fisher’s exact test; p=0.015). Therefore, in order to improve fawn survival probability
maternal antipredator tactics needed to reduce the risk of neonate detection by predators.
Prepartum isolation, birth site grass height, and their effect on detection of the fawn by predators
Four (36%) females were isolated at birth while seven gave birth within 50 m of at least
one conspecific. Of the four females that were isolated at parturition, three (75%) gave birth in
tall grass, while five (71%) of the seven females that remained in their groups gave birth in short
grass.
Our nominal logistic model reveals that neither prepartum isolation (effect likelihood
ratio test; χ2=1.81e-9, p=1.00) nor grass height (χ2=1.93e-9, p=1.00) affected whether a fawn
39
was detected by predators. We found the same results when we only considered detection by
jackals (isolation: χ2=3.02, p=0.08; grass height: χ2=5.88e-9, p=1.00). However, the interaction
between these two factors had a significant effect on fawn detection by all predators (χ2=4.32,
p=0.038) and by jackals (χ2=5.57, p=0.033). Fawns born either in isolation and in tall grass or
in groups and in short grass suffered lower rates of predator detection than fawns born under
other conditions. Predators detected only one out of three (33%) fawns born in isolation and in
tall grass and only two out of five (40%) fawns born in groups and in short grass. In contrast,
all fawns born either in isolation and in short grass (one fawn) or in groups and in tall grass (two
fawns) were detected by predators.
Effect of prepartum isolation on disturbances and postpartum period duration
Non-predatory disturbances of the mother-fawn pair by conspecifics and heterospecifics
during the postpartum period were common and occurred in seven of our eleven observations
(Table 1). Non-predators disturbed only one (25%) of the females that isolated for birth, but
six (86%) that did not isolate. Although the difference was not significant, there was a trend
for mothers that isolated to experience fewer non-predatory disturbances than non-isolated
mothers (Fisher’s exact test; p=0.088). At first glance, isolation appeared to prevent conspecific
disturbances in our observations: no isolated female was disturbed by conspecifics, while four
(57%) females that did not isolate were disturbed by conspecifics. However, this difference was not
significant (Fisher’s exact test p=0.19).
Figure 2 shows the effect of the total duration of disturbances on the time delay from birth
until the fawn hid and time delay from birth until the fawn walked. The more time the mother
spent displaced from the fawn during the postpartum period, the longer it took for the fawn to
40
walk and hide. Thus, non-predatory disturbances prolonged the postpartum period.
Our survival analysis revealed that risk of a predator encounter was relatively constant
during the postpartum period (Figure 3). Therefore, longer postpartum periods incurred greater
risk of encounter by predators. The shortest postpartum periods in our observations were
exhibited by fawns born in isolation (Table 1). These postpartum period durations of 33.6 and
46.6 minutes corresponded to survival probabilities of 0.84 and 0.71, respectively, when only
jackals were considered and 0.69 and 0.55, respectively, when all predators were considered. In
contrast, postpartum period durations for fawns born in groups ranged from 53 to 70 minutes,
with survival probabilities of between 0.50 and 0.65 when only jackals were considered and
between 0.37 and 0.49 when all predators were considered.
Proportion of fawns not encountered
1.0
0.8
0.6
0.4
0.2
0
30
60
Time since birth (min)
90
Figure 3. Probability of avoiding a predator encounter during postpartum periods of various durations.
We generated the lognormal curves by fitting them to Kaplain-Meier survival curves based on our data.
The dashed curve includes only jackal encounters and the solid curve includes all predators. Vertical
lines represent hiding times (dotted line) for our observations and the mean hiding time (solid line).
The logarithmic scale on the y-axis allows for clearer visualization of encounter rates, with straight lines
representing constant risk of encounter (Deevey 1947)
41
100
Short Grass
Tall Grass
Percent of Fawn Visible
90
80
70
60
50
40
30
20
10
0
2
4
8
11
7
9
10
Observation
Fawn standing
Fawn lying down
Figure 4. Mean visibility of fawns while standing and lying down in video-recorded observations. Except for
observation 4, fawns in short grass are significantly more visible while lying down than fawns in tall grass. All
fawns born in short grass are significantly more visible while standing than the fawn in observation 7, which
was born in tall grass. Our analysis yielded no data on the visibility while standing for the fawn in observation
9. Error bars indicate standard error.
Effect of birth site grass height on fawn visibility and probability of detection by predators
The mothers and fawns remained at or near their birth sites for the majority of the
postpartum period; therefore, fawn visibility was largely determined by grass height at the birth
site. Figure 4 shows the mean percentage of the fawn’s body that was visible when it was standing
up and lying down for all video-taped observations. Birth occurred in short grass in observations
2, 4, 8 and 11 and in tall grass in observations 7, 9 and 10. Except for the fawn in observation 4,
all fawns born in short grass were significantly more visible when lying down than the fawns born
in tall grass. All fawns born in short grass were also significantly more visible when standing than
the fawns in observations 7 and 10. Our analysis yielded no standing data for observation 9. The
Z-scores and p-values for the Wilcoxon tests comparing these means are given in Table 2.
42
Table 2. Results of Wilcoxon tests comparing mean fawn visibility when lying down and standing up across
video-recorded observations. Comparisons for standing fawns are in the upper right portion of the table
and comparisons for lying fawns are in the lower left portion. Results cells contain the Z-score followed by
the p-value. Cells comparing significantly different pairs are shaded.
Three (60%) concealed fawns (born in tall grass) and three (50%) exposed fawns (born in
short grass) were detected by predators. Therefore, the fawn’s visual conspicuousness had no
detectable effect on whether or not the fawn was encountered by a predator (Fisher’s exact test,
p=1.00).
43
Discussion
The postpartum period is a time of potentially high predation risk for ungulate neonates:
more than half of Thomson’s gazelle fawns in our study were encountered by predators during
this time, and encountered fawns were almost universally killed. The powerful effect of infant
mortality rates on ungulate population dynamics is well-documented (Gaillard et al. 1998;
Gaillard et al. 2000; Raithel et al. 2007). However, the postpartum period is not usually included
in studies of infant mortality (Smith & Anderson 1996; Singer et al. 1997; Janermo et al. 2004;
Olson et al. 2005; DeVivo et al. 2011; Buuveibaatar et al. 2013), suggesting that most studies
underestimate infant mortality and predation risk. Just as in Lott and Galland’s (1985) study on
bison, Thomson’s gazelle mothers in our study exhibited two primary perinatal tactics, either
isolating and giving birth in areas of vegetative cover or remaining in their social groups and
giving birth in open habitat. These combinations of social context and habitat type resulted
in fawn survival rates of approximately 65% and were thus more successful at mitigating fawn
predation risk than other combinations, in which all observed fawns were killed. Prepartum
isolation and birth site selection thus combine to form behavioral suites rather than functioning
as independent tactics.
Although it is clear that females actively choose the social and environmental contexts under
which they give birth, the factors leading to the choice of one tactic over another are unknown
and require further investigation. Unlike in Lott and Galland’s (1985) study, our females were not
subject to obvious differences in habitat availability. The decision may be influenced by habitat
variables that we did not detect or on other unknown factors, such as the female’s parity or social
rank, the composition of available groups, or recent predator activity in the area. In choosing the
context of parturition, females may also consider factors beyond offspring survival such as their
44
own level of predation risk. Due to the energetic expenditure and temporary immobility inherent
to labor, mothers are also at heightened risk during the postpartum period and may be safer
in groups where risk is diluted or in short grass habitats where predators can be spotted more
easily (Fitzgibbon 1988, 1990a). Unlike for some species (e.g. elk, Rearden et al. 2011), we do
not expect forage availability to play a large role in birth habitat selection. Gazelle in our system
chose birth habitats at a small scale, selecting patches of tall or short grass within their resident
plain rather than leaving their normal home range altogether. At this scale, after the postpartum
period mothers could leave the microhabitat in which they gave birth and select optimal foraging
locations nearby while still maintaining typical distances of 50 to 400 meters (Chapter 4) from
their hiding fawns. Furthermore, mothers essentially ceased foraging at the onset of heavy labor
and resumed feeding when they left the birth site after the postpartum period.
The two tactics seem to represent a trade-off between reducing the duration of the
postpartum period and reducing conspicuousness to predators (Figure 5). Due to our small
sample size, we were unable to demonstrate statistically that isolation reduces non-predatory
disturbances; however, trends in our data suggest that future studies may establish such an effect.
Disturbances prolong the duration of the hiding period, thereby increasing the fawn’s risk of
attack during this time. However, fawns born in tall grass appear to be vulnerable to detection
by jackals, comprising two of the three jackal kills we observed. We attribute this vulnerability to
maternal conspicuousness: when neonates were in tall grass, they were difficult for us to observe
even when they were standing; however, the mothers were always easily visible.
Mothers that give birth in groups may be less conspicuous to jackals than mothers in tall
grass due to social concealment. Indeed, the two fawns born in groups and in short grass that
were detected by predators were detected by warthogs. Antipredator behavior is not expected to
45
Stra
teg
yB
gy A
ate
r
t
S
tion
isola
m
tu
s
par ll gras
e
r
a
P
T
No pr
epar
Sho tum is
rt g
ras olatio
s
n
e of
anc pair
b
r
n
tu
Dis er-faw
h
t
mo
of
tion um
a
r
t
Du stpar
po period
M
consp aterna
icuo l
usn
ess
f predator enc
ility o
oun
b
a
ter
ob
Pr
n predation
Faw
Figure 5. Two main behavioral tactics used by Thomson’s gazelle mothers. Solid and dashed arrows
represent positive and negative correlations, respectively. Giving birth in isolation and in tall grass serves
to minimize non-predatory disturbances and thereby shorten the postpartum period. However, mothers
using this tactic are conspicuous to predators. Giving birth within social groups and in short grass serves
to minimize maternal conspicuousness to predators, but these mother-fawn pairs are at higher risk of
disturbance from conspecifics.
have evolved to counter warthog predation since gazelle do not regard this species as a predator.
In fact, to our knowledge this study is the first recorded instance of warthogs preying on ungulate
fawns. Warthogs are known to consume carrion and small vertebrates (Cumming 1975), and
other suids, such wild boar (Pavlov & Hone 1982; Choquenot et al. 1997; Wilcox & Van Vuren
2009), hunt and eat juvenile ungulates with some regularity. Similarly to wild boar, warthogs
catch fawns by knocking them to the ground with their snouts and then eat them while holding
them down with a forehoof. Fawn agility increases rapidly in the first hours of life; therefore the
window available to warthogs for fawn predation may be limited to the postpartum period.
Given that fawns born in groups and in short grass suffered fewer jackal attacks than fawns
born in tall grass and in isolation, and given that we have already ruled out the possibility of
46
mothers accounting for risk of warthog predation, why do any mothers choose to give birth
under the latter conditions? In addition to minimizing the duration of the postpartum period,
avoiding non-predatory disturbances may reduce the risk of fawn injury or accidental adoption
(Lent 1974; Jarman 1976; Estes & Estes 1979; Guinness et al. 1979). Fawns born in groups
were often knocked over by males attempting to court the mother, or temporarily “stolen” by
conspecific females. If other neonates are present in the group, the mother may also be at risk of
adopting an unrelated fawn at the expense of her actual offspring. Separating from conspecifics
and using tall grass as a visual barrier may therefore be an advantageous option during
peak fawning periods or for subordinate individuals that may not be able to effectively deter
conspecifics.
This study suggests that further investigation of the understudied postpartum period may
yield interesting information regarding maternal risk management and decision-making. Further
research involving more individuals is necessary to fully understand factors driving maternal
behavior patterns during the postpartum period. Future studies should attempt to include
observations of nocturnal and crepuscular births as well as diurnal births. Our observation
schedule did not allow us to investigate the effects of time of birth on maternal behavior, but
differences in predator activity level and the visibility of mothers and fawns at night may affect
maternal behavior and offspring survival.
Acknowledgements
We thank Ol Pejeta Conservancy for permitting us to work on the premises and the
Conservancy’s Research Department and security staff for their assistance in the field. We are
grateful to Jennifer Schieltz for the use of her recorded observation and to Ipek Kulahci, Molly
47
Schumer, Adrienne Tecza, and Chaitanya Yarlagadda for their feedback during manuscript
preparation. We thank Mike Costelloe for assistance in figure preparation, especially for the
design of Figure 5. This research was conducted with permission of the Government of Kenya
(permit NCST/5/002/R/648). Funding was provided by Princeton University’s Department
of Ecology and Evolutionary Biology, the Philadelphia chapter of the Explorers Club, and the
Animal Behavior Society’s student research grant program. DIR was supported by NSF grants
IBN-9874523, CNS-025214, and IOB-9874523.
48
Chapter 3: The effects of the hiding strategy on maternal behavioral trade-offs in Thomson’s
gazelle1 Blair A. Roberts and Daniel I. Rubenstein
Abstract
In many species, vigilance is an effective means of avoiding predator attacks, but is mutually
exclusive to other activities such as foraging and resting. Balancing the costs and benefits of
this trade-off can be especially difficult for parents with dependent offspring due to the high
vulnerability and energetic demands of infants. Here we examine whether hiding, a maternal
care strategy used by many ungulate species, helps to reduce the behavioral changes required
for motherhood by enabling mothers to reduce risk without experiencing significant loss of
foraging time when compared to non-mothers. Through behavioral observation of free-ranging
Thomson’s gazelles in Kenya, we find that maternal behavior varies with fawn activity: mothers
spend more time vigilant and less time resting and foraging while their fawns are active compared
to while they are hiding. During hiding periods, mothers exhibit their highest vigilance rates
prior to retrieving the fawn from hiding, and their lowest vigilance rates immediately after the
fawn resumes hiding. While their fawns are hidden, mothers are able to achieve activity budgets
equivalent to those of non-mother females. Overall, mothers exhibit no loss in foraging time
compared to non-mother females despite being significantly more vigilant. Our findings suggest
that the hiding strategy, by protecting the offspring and providing information regarding predator
absence, helps mothers to reduce risk while suffering minimal loss of time for other activities.
Conference presentations on this work:
Roberts, B.A. and Rubenstein, D.I. 2013. Maternal strategies for reducing calf predation risk in Thomson’s
gazelle. Animal Behavior Society Annual Meeting, Boulder, Colorado. 29 August, 2013.
1
49
Introduction
For prey species, predation is a major threat to reproductive success and therefore exerts a
powerful selective force on behavior (Lima & Dill 1990). Vigilance behavior is a common means
by which animals preempt predator attacks and thereby mitigate predation risk: more vigilant
individuals are less likely to fall victim to predation than less vigilant individuals (Fitzgibbon
1988; Godin & Smith 1988; Fitzgibbon 1990a; Krause & Godin 1996; Quinn & Cresswell
2004). However, for many species vigilance and other behaviors, such as foraging and resting,
are mutually exclusive (Lima & Dill 1990; Lima 1998; Toïgo 1999; Hamel & Côté 2008). These
species face a trade-off between mitigating risk through vigilance and engaging in activities that
increase forage intake or conserve energetic resources. Mismanaging this trade-off can be costly:
individuals that do not maintain sufficient body condition suffer reduced fecundity (Hester 1964;
Smith 1988; Peckarsky et al. 1993; Scrimgeour & Culp 1994; Benton et al. 1995; Cook et al.
2004; Lomas & Bender 2007), whereas those that are not vigilant enough or not vigilant at the
right time increase their risk of death from predation.
According to the risk allocation hypothesis, in order to best manage the trade-off between
vigilance and other activities animals should be more vigilant when risk of predation is high
and engage in other activities when risk of predation subsides (Lima & Bednekoff 1999). The
components determining an individual’s risk level can be divided roughly into the probability of
encountering a predator and the probability of surviving an encounter, both factors that can be
assessed by prey individuals (Lima & Dill 1990). However, individuals estimate these probabilities
using information of varying accuracy (Sih 1992). In particular, the probability of encountering
a predator depends in large part on whether or not a predator is nearby, information that many
predator species have evolved to conceal. Uncertainty therefore can also affect vigilance levels.
The immediate cost of underestimating risk is higher than overestimating; therefore, greater
50
uncertainty should cause an increase in vigilance behavior beyond that predicted simply by risk
(Sih 1992).
Parents with dependent offspring face a complex iteration of the vigilance trade-off. In
determining their appropriate vigilance levels they must account not only for their own risk, but
for that of their young as well. The opportunity cost of vigilance is also higher for parents than
for other adults: parents often must meet their offspring’s energetic needs in addition to their
own, thus increasing the value of foraging time. Parenting strategies that reduce the amount of
additional vigilance required to protect the young or that enable parents to accurately gauge risk
levels will increase the amount of foraging time available. Hiding, a cryptic strategy used by many
ungulate species to minimize risk of detection of offspring by predators, may offer mothers these
benefits.
While many ungulate species use a following strategy of maternal care in which the
offspring accompanies the mother constantly from birth, most artiodactyls use a hiding strategy
instead. Hiding is a cooperative behavioral strategy characterized by long periods of separation
of the mother and offspring (Walther 1965; Lent 1974). Beginning shortly after birth, the
offspring spends the majority of its time hidden in vegetation. The mother retrieves the infant
several times per day, thus initiating active periods. During these brief periods, the mother
grooms the infant and the infant nurses and plays. At the conclusion of the active period, the
infant selects a hiding spot, lies down, and thus begins a hiding period, which usually lasts several
hours. This alternation of short active and long hiding periods persists for the duration of the
hiding phase, which can last several weeks or months (Fisher et al. 2002). Hiding mitigates
offspring predation risk by reducing the probability of a predator finding the infant (Lent 1974;
Barrett 1978; Fisher et al. 2002) and may also free mothers to focus on foraging.
51
In this study we examine vigilance behavior in free-ranging Thomson’s gazelle females in
order to determine if the hiding strategy benefits mothers by enabling them to structure their
behavioral schedules to effectively mitigate risk with minimal loss of time for other activities. We
hypothesize that this may be accomplished by two mechanisms: a reduction of the total risk that
must be mitigated through maternal vigilance during hiding periods, and periodic increases in
certainty regarding predator absence. The concealment of the offspring during hiding periods
minimizes the role of maternal vigilance in ensuring offspring survival, enabling mothers to invest
in other activities while the fawn is hiding and engage more fully in vigilance behavior while the
fawn is active. The necessary exposure of the highly vulnerable offspring during active periods
yields as a byproduct reliable evidence regarding the absence of predators. Thomson’s gazelle
fawns are easy prey for many predators and maternal defense behaviors do not pose great risk
of injury to fawn predators. Therefore, predators tend to attack immediately upon detecting the
fawn (BR, personal observation). The non-occurrence of an attack during an active period is thus
a reliable indicator of the absence of predators.
These hypotheses suggest that maternal behavior should vary temporally between hiding
and active periods and within hiding periods to reflect changing levels of risk and certainty of
predator absence, with vigilance varying directly with risk and inversely with certainty. Other
activities, such as foraging and resting, are expected to vary inversely with vigilance. Maternal risk
is not expected to vary with fawn activity, but does vary with habitat type and group membership
(Fitzgibbon 1988, 1990a). All else being equal, maternal risk should be comparable to the risk of
a female without dependent offspring. Meanwhile, fawns are at higher risk during active periods
compared to while they are hiding (Table 3; Fitzgibbon 1988; Fitzgibbon 1990b). Within hiding
periods, fawns are at the most risk at the end of the hiding period when their mothers approach
52
Fawn predation risk
Time period
Active period
Hiding period
Immediately prior to retrieval of fawn
Immediately after fawn resumes hiding
High
Low
High
Low
Certainty of
predator absence
Low
Low
Low
High
Predicted maternal
vigilance
High
Low
High
Very low
Table 3. The expected relative levels of fawn predation risk and maternal certainty of predator absence
during different stages of the hiding strategy. Maternal predation risk is not expected to vary between
different time periods within the hiding phase. Expected levels of fawn predation risk and certainty of
predator absence are used to predict relative levels of maternal vigilance.
the hiding spot to retrieve them, potentially exposing their location to nearby predators. Maternal
certainty as to predator absence should be highest immediately after the fawn resumes hiding; as
the hiding period wears on, the information regarding predator absence grows more obsolete. If
the mother is not constantly monitoring for predators entering the area, her certainty regarding
their absence should decrease with time. Drawing on these expected patterns, we make the
following predictions:
(1) Mothers will be less vigilant and engage more in other activities while their offspring are
hiding compared to when they are active;
(2) Maternal vigilance levels and activity budgets during hiding periods will be equivalent to
those of females without dependent young;
(3) Within hiding periods, maternal vigilance rates will be highest prior to fawn retrieval;
(4) Within hiding periods, maternal vigilance rates will be lowest immediately following the
fawn’s resumption of hiding.
Methods
Field site
We conducted all fieldwork in three field trips from March to June 2011, August to
November 2011, and June to September 2012 at Ol Pejeta Conservancy (OPC) in Laikipia,
53
Kenya. OPC is a fenced, 360 km2 conservancy consisting of discrete grassy plains divided
by Acacia drepanalobium and Euclea divinorum woodlands. In 2012, OPC had a population of
approximately 1300 Thomson’s gazelles. The conservancy also has high densities of gazelle
predators including lions, cheetahs, black-backed jackals, spotted hyenas, olive baboons and birds
of prey.
Behavioral observations
Animals at OPC are well-habituated to vehicles, allowing slow approaches to within 100
m. At this distance we were able to observe gazelle through binoculars without disturbing them
or eliciting vigilance reactions. In order to further minimize disturbance of the subject during
observations, we waited to begin observations until the vehicle had been stationary for five
minutes, and thereafter moved the vehicle only as necessary to keep the subject in sight.
We observed both non-mother females and mothers. We identified mothers by their swollen
udders or the presence of a hiding-aged fawn. Hiding-aged fawns can be identified by their size
relative to their mothers and their coloration (Walther 1973b; Fitzgibbon 1990b). We observed
each female for a minimum of two and a maximum of four hours. Observations ended after
two hours unless the subject was a mother whose fawn had not yet emerged from hiding. In this
case, we observed the subject for an additional two hours or until the fawn emerged, whichever
occurred first. If the fawn did not emerge after four hours, we excluded the observation from
this study. We photographed females head-on with a 500 mm lens and used unique, natural horn
shapes and facial markings to differentiate individuals after observation (Walther 1973b).
Every 30 minutes during each observation, we collected two-minute continuous focal
samples of the subject’s vigilance behavior, noting when she began and ended vigilance bouts
54
Behavior
Vigilant
Foraging
Resting
Self-grooming
Grooming
Out of sight
Other
Description
Subject is standing with her head raised above shoulder level
Subject is grazing with her head lowered to ground level
Subject is lying down
Subject is standing and nibbling, licking and/or scratching one part of her body with another
Subject is nibbling or licking her fawn
Observer’s view of the subject is obstructed
Subject’s behavior is not described by above categories
Table 4. Ethogram used for instantaneous activity budget sampling
(Altmann 1974). Following Fitzgibbon (1990b), we considered a subject vigilant if her head
was above shoulder level and not vigilant if her head was below shoulder level. We did not take
samples when the subject was lying down, and terminated samples if the subject lay down before
two minutes were finished. We discarded terminated samples shorter than 60 s in duration. To
construct activity budgets for each subject, we instantaneously sampled the subject’s behavior
every five minutes following the ethogram in Table 4 (Altmann 1974). We also noted when the
fawn emerged from or began hiding, and whether the fawn emerged by waiting for its mother to
retrieve it or by standing up on its own. We noted any predators sighted during observations and
any time the subject was startled by a non-predator (i.e. a vehicle). Finally, every 15 minutes we
noted the subject’s group size and surrounding grass height relative to the subject as short or tall
(above or below the subject’s hock, respectively).
Statistical analyses
This paper deals with management of potential rather than immediate risk, and predator
presence strongly affect vigilance rates in gazelle (Walther 1969). Therefore, we excluded from
analysis all vigilance samples from observations during which a predator was spotted, as well as
any individual samples during which the female was startled by a non-predator. We performed
55
all statistical analyses in JMP 10.0. Our data were not normally distributed; therefore, we used
nonparametric methods in our analyses.
Vigilance samples. For each sample we divided the number of seconds that the subject was
vigilant by the total number of seconds in the sample and used this proportion as the measure of
maternal vigilance. We then classified each sample as having occurred while the fawn was active
or while it was hiding, and assigned each sample a time value corresponding to the number of
minutes (rounded to the nearest five) until or since an active period. We assigned zero values to
samples taken while the fawn was active, and negative and positive values to samples taken before
the fawn emerged and after the fawn resumed hiding, respectively. If a sample fell between two
observed active periods, it was assigned the number with the lower absolute value. For example,
if the sample was taken 30 minutes after one active period and five minutes before the next active
period, we considered the second active period to be more relevant and assigned the sample a
value of −5. An exception to this rule occurred when the latter active period was initiated by the
fawn rather than by the mother. In these cases, samples falling between two active periods were
assigned values relative to the first active period. This is because the mother’s behavior cannot be
expected to anticipate an event that she does not initiate.
We used nonlinear regression to analyze vigilance samples occurring before and after active
periods separately. We fit two-parameter sigmoid curves to each set of data. We chose twoparameter sigmoid curves because we expected mothers to exhibit relatively swift transitions
between states of higher and lower vigilance. Solitary adult gazelle and gazelle in tall grass
habitats are at an increased risk of predation (Fitzgibbon 1988, 1990a). We therefore expected
mothers under these conditions to exhibit higher overall rates of vigilance behavior, which may
56
weaken the relationship between maternal vigilance behavior and fawn activity. In order to
disentangle the effects of maternal and fawn risk, we performed the curve-fitting analysis on all
data as well as the following subsets of data: samples in which mothers were solitary, samples in
which mothers were in groups, and observations in which the mother spent the majority of her
time in short grass. We did not have enough observations in which the female spent the majority
of her time in tall grass to test this subset independently.
We grouped vigilance samples into seven bins for means comparison: 1 to 30 minutes before
fawn is retrieved, 31 to 60 minutes before fawn is retrieved, >60 minutes before fawn is retrieved,
1 to 30 minutes after fawn resumes hiding, 31 to 60 minutes after fawn resumes hiding, >60
minutes after fawn resumes hiding, and fawn active. We used Wilcoxon tests to compare means
between each bin and analysis of means with transformed ranks to compare vigilance rates in
each bin to the overall mean. Finally, for each observation we averaged the proportion of time
spent vigilant from samples taken while the fawn was active and while the fawn was hiding and
compared these means using a paired t-test.
Activity budget data. We used the 5-minute instantaneous behavior samples to create activity
budgets, excluding samples in which the female was out of sight and the single sample in which
the subject’s behavior was classified as “other”. To create activity budgets for non-mothers, we
divided the number of instantaneous behavior samples in the vigilant, foraging, resting, grooming
and self-grooming categories by the total number of samples for each observation to yield the
percent time each subject spent on each activity. For mothers, we divided samples within each
observation according to whether the fawn was active or hiding and calculated separate activity
budgets for each fawn state. To create overall activity budgets for mothers, we recombined the
57
active and hiding period budgets after weighting them according to Fitzgibbon’s (1990b) finding
that infants spend 75% of their time hiding and 25% of their time active. For mothers we also
assigned each instantaneous sample a time value and binned the data as described above. We
calculated activity budgets for each observation within each bin, excluding activity budgets
based on two or fewer instantaneous samples. We used Wilcoxon tests to compare mean time
spent on each behavior between activity budgets and paired t-tests to compare maternal activity
budgets between hiding and active periods. We used analysis of means with transformed ranks to
compare means of binned data to the overall mean.
Results
We conducted 51 observations of 47 mother gazelle and 38 observations of 37 nonmothers. We excluded from all analyses 16 observations during which predators were sighted,
and eight vigilance samples during which the subject was startled by nearby vehicles or humans
on foot. This left for analysis 37 observations of 36 mothers and 36 observations of 35 nonmothers. These yielded 80.3 hours of activity budget data and 115 vigilance samples for mothers,
and 69.2 hours of activity budget data and 88 vigilance samples for non-mothers.
Maternal behavior while fawn was active vs. while fawn was hiding
Mothers were significantly more vigilant while their fawns were active compared to while
their fawns were hiding. Our two-minute vigilance samples indicate that mothers spent 50.2%
(SE=6.7%, n=35) of their time vigilant while their fawns were active compared to 29.8%
(SE=3.8, n=80) while their fawns were hiding (Wilcoxon test, Z=3.10, p=0.0019). A paired t-test
for mothers whose vigilance we sampled both while the fawn was active and while the fawn was
58
Percent time spent performing behavior
70
60
Fawn active
Fawn hiding
50
40
30
20
10
0
Vigilant
Foraging
Resting
Selfgrooming
Grooming
Figure 6. Mean percent time
mothers spend on each of
five behaviors while their
fawns are active and hiding.
Mothers spend significantly
more time vigilant and less
time foraging and lying down
while their fawns are active
compared to while they are
hiding. Error bars indicate
standard error.
Behavior
hiding showed mothers to be significantly more vigilant while their fawns were active compared
to while they were hiding (paired t-test, t(21)=-3.74, p=0.0012). Our activity budget data (Figure
6) yielded similar results, showing that mothers spent 56.3% (SE=4.5%, n=34) and 20.1%
(SE=4.3%, n=37) of their time vigilant while their fawns were active and hiding, respectively
(Wilcoxon test, Z=4.63, p<0.0001). The 36.2% decrease in vigilance while the fawn was hiding
was accounted for by increases in foraging and resting behavior. Mothers spent 32.9% (SE=4.8)
of their time foraging while their fawns were active compared to 55.7% (SE=4.6) while their
fawns were hiding (Wilcoxon test Z=-3.02, p=0.0025). Mothers rarely lay down while their fawns
were active (0.8%, SE=3.0), but spent 22.7% (SE=2.9) of their time lying down while their fawns
were hiding (Wilcoxon test Z=-4.67, p<0.0001). There was no significant difference in time spent
self-grooming between active (1.48%, SE=0.8) and hiding periods (1.64%, SE=0.8; Wilcoxon
test Z=-1.93, p=0.051). Mothers spent 6.5% (SE=1.5) of their time grooming their fawns during
active periods.
59
Percent time spent performing behavior
70
Mother (fawn hiding)
Mother (overall)
60
Non-mother
50
40
30
20
10
0
Vigilant
Foraging
Resting
Behavior
Selfgrooming
Figure 7. Hiding period
and overall maternal
activity budgets compared
to non-mother activity
budgets. While their fawns
are hiding, mothers exhibit
activity budgets equivalent
to those of non-mothers.
Overall, mothers spend
significantly more time
vigilant and less time resting
and self-grooming than nonmothers. Overall time spent
foraging is equivalent for
mothers and non-mothers.
Error bars indicate standard
error.
Activity budgets of non-mothers and mothers whose fawns were hiding were not
significantly different (Figure 7). Non-mothers spent 15.4% (SE=3.0, n=36) of their time vigilant,
53.6% (SE=3.8) of their time foraging, 28.0% (SE=3.6) of their time resting, and 3.0% (SE=0.5)
of their time self-grooming compared to 20.5% (SE=2.9, n=37; Wilcoxon test, Z=-1.43, p=0.15)
vigilant, 55.0% (SE=3.7; Z=-0.22, p=0.83) foraging, 22.9% (SE=3.6; Z=1.13, p=0.26) resting,
and 1.6% (SE=0.5; Z=1.71, p=0.09) self-grooming for mothers with hiding fawns. When both
active and hiding period activity budgets were considered, mothers spent significantly more time
vigilant (29.6%, SE=3.1, n=34; Wilcoxon test, Z=3.72, p=0.0002) and less time resting (17.3%,
SE=3.2; Z=-2.22, p=0.03) and self-grooming (1.5%, SE=0.51; Z=-2.41, p=0.02) than did nonmothers (Figure 7). Overall maternal foraging time was not significantly different from that of
non-mothers (50.0%, SE=3.8; Z=-0.80, p=0.42).
60
100
80
1
1+Exp(-0.02*(x+94.15))
r2= 0.30
1
1+Exp(-0.03*(x-91.09))
r2= 0.32
y=
y=
Fawn active
A.
60
40
20
0
100
80
1
1+Exp(-0.06*(x-69.47))
r2= 0.66
y=
1
y=
1+Exp(-0.03*(x+80.70))
r2= 0.57
Fawn active
B.
40
20
0
100
80
1
1+Exp(-0.01*(x+100.34))
r = 0.20
1
1+Exp(-0.02*(x-133.59))
r = 0.15
y=
y=
2
2
Fawn active
C.
Percent of time mother is vigilant
60
60
40
20
0
100
80
y=
1
1+Exp(-0.07*(x+77.17))
y=
r2= 0.62
1
1+Exp(-0.03*(x-89.72))
r2= 0.34
Fawn active
D.
60
40
20
0
-240 -210 -180 -150 -120
-90
Minutes until
fawn retrieval
-60
-30
0
30
60
90
Minutes since
fawn resumed hiding
Figure 8. Variation in maternal vigilance levels within fawn hiding periods. Horizontal dashed lines represent
overall mean maternal vigilance during fawn hiding periods. Curves are two parameter sigmoid curves fitted
to vigilance sample data, equations with parameter estimates given in grey. Curves are fitted using a) all data,
b) samples in which mother was solitary, c) samples in which mother was in a group, and d) observations in
which the mother spent most of her time in short grass.
61
Maternal behavior within hiding periods
Within hiding periods, maternal vigilance varied according to the amount of time until or
since an active period. Maternal vigilance increased leading up to fawn retrieval, with 30% of
the variance in maternal vigilance explained by the fitted two-parameter sigmoid curve (Figure
8). We found the same pattern when performing the analysis on subsets of data for solitary
mothers, mothers in groups, and mothers in short grass. The pattern was stronger for short
grass and solitary data subsets, in which the fitted curves explained 62 and 57% of the variance,
respectively. After fawns resumed hiding, maternal vigilance levels were low, but increased
over time. The fitted two-parameter sigmoid curve explained 32% of the variance in maternal
vigilance levels with time until retrieval. Again, this pattern held for all data subsets.
Binning the data revealed that maternal vigilance was significantly higher in the hour
preceding fawn retrieval than in the hour after the fawn resumed hiding (Figure 9; Table 5).
Vigilance in the hour prior to retrieval was significantly higher than average while vigilance in
the 30 min after the fawn resumes hiding was significantly lower than average (Figure 10). Our
activity budget data did not capture the pre-retrieval increase in vigilance or any corresponding
decreases in other behaviors (Figure 10). This is likely because the instantaneous scan samples we
used to construct our activity budgets were biased towarddocumenting long behavioral bouts and
therefore missed many of the short vigilance bouts detected by our vigilance sampling method.
However, the decrease in vigilance after the fawn resumes hiding was reflected in our binned
activity budgets and was accompanied by an increase in foraging behavior (Figure 11). The
amount of time spent foraging in the half hour after the fawn resumes hiding was significantly
higher than average (Figure 12).
62
80
60
40
Calf Active
Percent of time mother spends vigilant
100
20
0
>60
31–60
1–30
0
Minutes before
mother retrieves fawn
1–30
31–60
>60
Minutes after
fawn resumes hiding
Figure 9. Variation in maternal vigilance levels based on vigilance sample data by 30-minute time block.
Error bars indicate standard error. Results of Wilcoxon tests comparing means of each column pair are given
in Table 5.
Minutes Before Fawn
Retrieval
Minutes After
Fawn Restumes
Hiding
Fawn
Active
Minutes Before Fawn
Retrieval
Time
N
Mean±
>60
14
37.7±7.6
31 to 60
5
72.8±12.7
1 to 30
9
73.4±9.5
0
35
50.2±4.8
1 to 30
32
10.8±5.0
31 to 60
13
17.1±7.9
31 to 60
5
72.8±12.7
1.36
0.18
Fawn Active
Minutes After Fawn Resumes Hiding
1 to 30
9
73.4±9.5
1.81
0.07
0
35
50.2±4.8
1.09
0.28
1 to 30
32
10.8±5.0
-2.51
0.01*
31 to 60
13
17.1±7.9
-1.44
0.15
>60
7
38.0±10.8
-0.26
0.79
-0.07
0.95
-1.41
0.16
-2.77
0.006*
-2.33
0.02*
-1.63
0.10
-2.02
0.04*
-4.12
<0.0001*
-3.08
0.002*
-1.91
0.06
-5.04
<0.0001*
-3.14
0.002*
-0.95
0.34
0.59
0.55
1.57
0.12
1.04
0.30
Table 5. Results of Wilcoxon tests comparing maternal vigilance levels between time bins. Results cells
contain the Z-score followed by the p-value. Cells comparing significantly different pairs are shaded.
63
Mean transformed rank
(maternal vigilance proportion)
1.6
1.4
UDL
1.2
1.0
0.8
Mean=0.793
0.6
0.4
LDL
0.2
0.0
>60
31–60
1–30
Minutes before
mother retrieves fawn
0
1–30
Fawn
active
31–60
>60
Minutes after
fawn resumes hiding
Figure 10. Analysis of bin means with transformed ranks for vigilance sample data. The dotted line
represents the overall mean rank score. Bin ranks that exceed the upper decision limit (UDL) or lower
decision limit (LDL) are significantly higher or lower, respectively, than the overall mean. Decision limits were
set using α=0.05.
Percent of time spent performing behavior
80
70
Vigilant
Foraging
Resting
Self-grooming
Grooming
60
50
40
30
20
10
0
>60
31–60
Minutes before
mother retrieves fawn
1–30
0
Fawn
active
1–30
31–60
Minutes after
fawn resumes hiding
>60
Figure 11. Activity budgets for mothers while the fawn is active and during 30-minute bins during hiding
periods. Error bars indicate standard error.
64
Mean transformed rank
(% time spent vigilant)
A.
1.2
UDL
1.0
0.8
Mean=0.794
0.6
LDL
0.4
B.
Mean transformed rank
(% time spent foraging)
>60
C.
1–30
0
1–30
31–60
>60
1.2
UDL
1.0
0.8
Mean=0.793
0.6
0.4
0.2
LDL
>60
31–60
1–30
0
1–30
31–60
>60
Mean transformed rank
(% time spent resting)
1.4
Mean transformed rank
(% time spent self-grooming)
D.
31–60
1.2
UDL
1.0
0.8
Mean=0.787
0.6
0.4
0.2
LDL
>60
31–60
1–30
0
1–30
31–60
>60
1.2
UDL
1.0
0.8
Mean=0.755
0.6
0.4
0.2
LDL
>60
31–60
1–30
Minutes before
mother retrieves fawn
0
Fawn
active
1–30
31–60
>60
Minutes after
fawn resumes hiding
Figure 12. Analysis of bin means with transformed ranks for activity budget data. The dotted line represents
the overall mean rank score for each behavior. Bin ranks that exceed the upper decision limit (UDL) or
lower decision limit (LDL) are significantly higher or lower, respectively, than the grand mean. Decision limits
were set using α=0.05.
65
Discussion
Ungulate mothers face high stakes in the vigilance-activity trade-off due to the vulnerability
of their offspring to predation and the heavy energetic burden of lactation. Predation is the
main cause of the high infant mortality rates observed in most ungulate populations (Linnell et
al. 1995), and maternal vigilance is the primary line of defense when offspring are out of hiding
(Fitzgibbon 1990b). However, vigilance behavior is mutually exclusive with grazing and resting,
and thus limits the amount of time mothers can devote to meeting their high energetic demands
or conserving energetic resources. Maternal energetic costs are highest for ungulates during early
lactation due to rapid infant growth rates, which must be fueled almost entirely by maternal
resources (Case 1978; Robbins et al. 1981; Loudon 1985; Clutton-Brock et al. 1989). In addition
to meeting their offspring’s needs, mothers must maintain adequate body condition to capitalize
on their next reproductive opportunity (Benton et al. 1995; Lomas & Bender 2007). Thomson’s
gazelle females can reproduce twice per year and enter estrus within two weeks of parturition
(Brooks 1961; Hvideberg-Hansen 1970; Robinette & Archer 1971; Leuthold 1972); therefore, loss
of foraging or resting time can directly affect the survival of current and future offspring. This
study suggests that the hiding strategy benefits ungulate mothers by enabling them to effectively
manage the behavioral trade-off between vigilance and other activities, thus improving their longterm reproductive potential.
Fawn predation risk is reduced for long periods of time through concealment during hiding
periods. This reduces the vigilance required by mothers to ensure the safety of themselves and
their fawns: as we predicted, mothers were significantly less vigilant while their fawns were hiding
compared to while they are active. Similar results have been reported for Thomson’s gazelle
in the Serengeti (Fitzgibbon 1988) and other hider species (Clutton-Brock et al. 1982; White
66
& Berger 2001). The increase in time spent vigilant during the active period compared to the
hiding period was accounted for by decreases in time spent foraging and resting. There was no
difference in time spent on self-grooming behavior between the active and hiding periods.
When their fawns were hiding, mother gazelle in our study were able to achieve activity
budgets equivalent to those of females without dependent young. This is in contrast to
Fitzgibbon’s (1988) and Byers’ (1997) findings, which indicate that mother ungulates are more
vigilant and forage less than non-mothers even when their fawns are hidden. The cause for
this discrepancy is unclear. However, our finding suggests that hiding can potentially benefit
ungulate mothers by limiting the behavioral costs of motherhood to discrete brief periods during
which the offspring is active. This benefit of the hiding strategy is further supported by White
and Berger’s (2001) finding that moose cows reduce their vigilance and increase their foraging
behavior during hiding periods to levels below and above, respectively, those of non-mothers.
This compensates for the behavioral changes required during active periods and results in moose
mothers having overall activity budgets similar to non-mothers when both active and hiding
period behavior is considered.
When both hiding and active period activity budgets were considered, Thomson’s gazelle
mothers did not exhibit a loss in foraging time compared to non-mothers, despite being
significantly more vigilant. Rather, the increase in vigilance was afforded by a decrease in time
spent resting. In contrast to our findings, Fitzgibbon (1988, 1993) reported that Thomson’s
gazelle mothers in the Serengeti spent significantly less time foraging overall than non-mothers.
Fitzgibbon did not report percent time spent on resting or self-grooming; therefore, it is unclear
whether the maternal increase in vigilance was subsidized entirely by a decrease in foraging or
also by decreases in resting or other activities. Studies of other ungulate species tend to match
67
our findings more closely than those of Fitzgibbon, reporting that mothers exceed non-mother
vigilance rates at the expense of time spent resting rather than time spent foraging (Toïgo
1999; Hamel & Côté 2008). Although we did not measure rumination behavior specifically,
our observations suggest that mothers increase their time spent vigilant by continuing to stand
while they ruminate, while non-mothers tend to conduct most of their rumination while lying
down. Ungulates exert between 21 and 37% more energy while standing compared to lying
down (Parker et al. 1984). Loss of resting time therefore represents an energetic cost to mothers,
albeit presumably lower than if mothers’ increase in vigilance behavior was subsidized by a loss
of foraging time. Interestingly, alpine ibex (Capra ibex ibex), a follower species of similar body
size to Thomson’s gazelle, exhibits similar differences in activity budgets between mothers and
non-mothers as Thomson’s gazelles (Toïgo 1999). This suggests that hiding and following may
be equal alternative strategies for infant care with regards to their effects on maternal activity
budgets.
A byproduct of the exposure of the fawn during active periods is reliable evidence of
absence of predators. Mothers appear to use this information by taking the opportunity to
reduce their vigilance levels in favor of foraging immediately after the fawn resumes hiding. Of
course, exposing a vulnerable offspring to heightened predation risk in order to gain information
regarding predator presence would not be an evolutionarily prudent strategy and we do not
propose that such information gathering was a driving factor in the evolution of the hiding
strategy. However, the ability to utilize the information that results from hiding periods may
benefit mothers by helping to subsidize the observed increase in maternal vigilance prior to
fawn retrieval. Pre-retrieval peaks also have been described in pronghorn (Byers 1997) and red
deer (Clutton-Brock et al. 1982). Such peaks serve to mitigate the impending increase in fawn
68
vulnerability by increasing the probability that the mother will detect lurking predators prior to
revealing the fawn’s location (Fitzgibbon 1988; Byers 1997). In several instances, mothers in our
observations assumed a heightened state of vigilance while approaching their hiding fawns only
to halt their advances and retreat upon detecting jackals or humans approaching on foot. In two
instances, the mother’s heightened vigilance failed to detect nearby jackals, resulting in the fawn’s
death only seconds after it emerged from hiding.
It is possible that the drop in vigilance at the onset of the hiding period is an attempt by
mothers to compensate for lost foraging time and does not indicate any insight into predator
absence. Perhaps low investment in foraging during and prior to the active period causes the
mother’s need to forage to outweigh the potential risk of reduced vigilance. Although our data
are insufficient to disentangle these two hypotheses, the fact that maternal feeding rates prior
to and during active periods were not significantly below average suggests that mothers were
not suffering a large loss of feeding time to enable increased vigilance. Additionally, the drop
in vigilance was strongly exhibited even by solitary mothers, suggesting that mothers perceive a
lower level of risk to themselves during this time.
Motherhood exacts costs on females, often by necessitating behavioral changes to meet
offspring needs or investment demands that negatively impact future reproductive opportunities
(Trivers 1972, 1974). Our study suggests that the hiding strategy serves to alleviate these costs
in two ways. First, it conceals the offspring from predators for long periods of time and thus
minimizes the maternal behavioral investment required to ensure the survival of the current
offspring. Second, it provides reliable information regarding maternal and fawn predation
risk, enabling mothers to optimally structure their activity schedules so as to avoid predation at
minimal opportunity cost for other activities. These findings may be widely applicable to hiding
69
ungulate species, as well as to other taxa in which young are stationary and concealed, including
some macropods, primates, and denning species.
Acknowledgements
We thank Ol Pejeta Conservancy for permitting us to work on the premises and the
Conservancy’s Research Department and security staff for their assistance in the field.
This research was conducted with permission of the Government of Kenya (permit
NCST/5/002/R/648). Funding was provided by Princeton University’s Department of Ecology
and Evolutionary Biology, the Philadelphia chapter of the Explorers Club, and the Animal
Behavior Society’s student research grant program. DIR was supported by National Science
Foundation grants IBN-9874523, CNS-025214, and IOB-987452.
70
Chapter 4: Coping with transition: fawn risk and maternal behavioral changes at the end of
the hiding phase Blair A. Roberts and Daniel I. Rubenstein
Abstract
The transition out of hiding is a risky time for hiding ungulate young due to their increased
exposure to predators. In this study, we examine the behavioral patterns of Thomson’s gazelle
mothers with hiding and transitioning fawns to determine the effect of fawn behavioral changes
and heightened risk on maternal behavior. We find that during the transition, mothers change
their grouping behavior, habitat use, and activity budgets to more closely resemble those of nonmothers. Mothers and transitioning fawns rely on group membership and collective vigilance
rather than heightened maternal vigilance to mitigate increases in fawn risk. Group membership
is made possible by the shorter hiding bouts of transitioning fawns relative to younger fawns;
more frequent active bouts enable the fawns to relocate more frequently, helping the mother and
fawn to keep up with group movements.
Introduction
Ungulate species use one of two maternal care strategies to mitigate offspring predation
risk (Walther 1965; Lent 1974). In the following strategy, the infant accompanies its mother and
her social group immediately after the postpartum period. The pair maintains small motherinfant distances and high rates of interaction. In hiding species, the infant hides in vegetation
shortly after birth. The offspring spends long periods separated from its mother, who returns and
retrieves the infant several times per day to feed and care for it. These short active periods are
71
characterized by small mother-infant distances and high rates of interaction, while the longer
hiding periods feature greater separation and no direct interaction between mother and offspring
(Ralls et al. 1987a, b). The alternation of active and hiding periods continues for the duration of
the hiding phase, which varies among species from several days to several months (Lent 1974).
Near the end of the hiding phase, the hider infant begins a transition out of the strategy.
This entails emerging from hiding more often, resulting in fewer, shorter hiding bouts and more
time spent out of hiding (Fitzgibbon 1990b). Additionally, the infant initiates the end of hiding
bouts independently by standing up on its own more frequently, without waiting for its mother
to retrieve it (Alados & Escos 1988; Fitzgibbon 1990b). It appears that this transition typically
occurs before the infant has developed sufficient speed and agility to escape predators (Fitzgibbon
1990b; Byers 1997); therefore, this period of transition is particularly risky for ungulate infants.
Hider infants that survive the transition behave essentially as followers: they no longer conceal
themselves but instead remain constantly active and in the same social group as their mothers.
While it is clear that the transition out of hiding is risky for the infant, the effect of this
period on maternal behavior is not known. During the hiding phase, mothers typically spend
more time alone or in smaller groups than females without dependent offspring, either because
they intentionally isolate from conspecifics or because they are left behind when groups depart
the area in which the mother’s offspring is hidden (Brooks 1961; Walther 1969; O’Brien 1984;
Schwede et al. 1993). Solitary individuals and those in smaller groups are at greater risk of
predation than individuals in large groups (Hamilton 1971; Fitzgibbon 1988, 1990a). The shorter
hiding bouts and more frequent active periods of transitioning infants may allow mothers to
relocate their offspring more often and track group movements more effectively, thereby reducing
maternal risk by increasing the mother’s time spent in social groups. The hiding strategy can also
72
cause mothers to leave their normally preferred habitats in favor of areas with greater cover for
their hiding offspring. These habitats are often riskier for mothers (Fitzgibbon 1990a) or poorer
for foraging (White & Berger 2001; Ciuti et al. 2009). As infants rely less on hiding for protection,
mothers may be able to return to preferred habitats.
The transition out of hiding may have negative effects on mothers as well. By protecting
infants from predators and isolating them from their mothers, hiding may reduce the impact of
motherhood on maternal activity budgets. In some cases, mothers are able to behave identically
to non-mothers while their offspring are hidden (Chapter 2). During the transition, mothers
may need to increase their vigilance rates to compensate for their infants’ higher risk levels.
Transitioning infants may also be able to solicit more care from their mothers, further disrupting
their activity patterns.
In this study, we examine fawn risk and maternal behavior in Thomson’s gazelle and
compare behavior of mothers with fawns of different ages to that of non-mothers. Because
non-mothers do not need to compensate for offspring risk, we assume their behavioral patterns
represent the optimal balance of risk management and resource acquisition for adult female
gazelle and that departures from these patterns by mothers represent investments in offspring. In
Thomson’s gazelle, fawns hide intensively for the first month of life, and then begin the transition
out of hiding. The transition is complete by the time the fawn is two months old (Fitzgibbon
1990b). We investigate differences in fawn risk between hiding (young) and transitioning (old)
fawns. We then compare the behavior of mothers with old fawns to that of mothers with young
fawns and non-mothers to determine the effects of the hiding transition on maternal grouping
patterns, habitat use, and activity budgets. We hypothesize that mothers of old fawns will be
intermediate between mothers of young fawns and non-mothers in their grouping behavior
and habitat selection, but that transitioning mothers will be less similar to non-mothers in their
73
activity budgets than will mothers of younger fawns. Specifically, we expect that mothers of
transitioning fawns will be more vigilant than mothers of young fawns due to the decreasing
reliance on crypsis for defense with fawn age, and that mothers of transitioning fawns will spend
more time caring for their fawns than mothers of young fawns simply because older fawns spend
more time active and therefore can solicit more maternal care (Trivers 1974).
Methods
Field site
We conducted all fieldwork from March to June 2011, August to November 2011, and
June to September 2012 at Ol Pejeta Conservancy (OPC) in Laikipia, Kenya. OPC is a fenced,
360 km2 conservancy consisting of discrete grassy plains divided by Acacia drepanalobium and
Euclea divinorum woodlands. In 2012, OPC had a population of approximately 1300 Thomson’s
gazelles. The conservancy also has high densities of gazelle predators including lions, cheetahs,
black-backed jackals, spotted hyenas and olive baboons. OPC is also frequented by several
raptor species that have either been observed attacking Thomson’s gazelle fawns or are thought
to be capable of preying on fawns, including tawny eagles, martial eagles, lappet-faced vultures
(Torgos tracheliotos), spotted eagle-owls (Bubo africanus), Verreaux’s eagle-owl (Bubo lacteus), longcrested eagles (Lophaetus occipitalis), and bateleurs (Terathopius ecaudatus) (Corinne Kendall, personal
communication).
Behavioral observations
Animals at OPC are well-habituated to vehicles, allowing slow approaches to within 100
meters. At this distance, we were able to observe gazelle through binoculars without disturbing
74
them or eliciting vigilance reactions. In order to further minimize disturbance of the subject
during observations, we waited to begin observations until the vehicle had been stationary for five
minutes, and thereafter moved the vehicle only as necessary to keep the subject in sight.
We observed both non-mothers and mothers. We identified mothers by their swollen udders
or the presence of a hiding-aged fawn. We grouped hiding-aged fawns into two age groups
according to their size and coloration (Walther 1973a; Fitzgibbon 1990b): young fawns are one
to four weeks old and old fawns are five to eight weeks old. Observations lasted two hours unless
the subject was a mother whose fawn had not yet emerged from hiding. In this case, we observed
the subject for an additional two hours or until the fawn emerged, whichever occurred first.
If the fawn did not emerge after four hours, we excluded the observation from this study. We
photographed females head-on with a 500 mm lens and used unique, natural horn shapes and
facial markings to differentiate individuals after observation (Walther 1973a).
During each observation, we instantaneously sampled the subject’s behavior every five
minutes, recording her behavior as either vigilant, feeding, lying down, self-grooming, or
grooming the fawn (Altmann 1974). Every fifteen minutes we used a laser rangefinder and
compass to measure the distance and bearing from our vehicle to the subject, the fawn and
the subject’s nearest adult conspecific neighbor. With these measurements we used the law of
cosines to calculate distances between the subject and her fawn, and the subject and her nearest
neighbor. Every fifteen minutes we also noted the subject’s group size, and for subjects in groups
of five or more individuals, we noted whether the subject was in the center or on the edge of her
group (Fitzgibbon 1990a). Following Fitzgibbon (1988, 1990b), subjects were considered to be in
a group if they were less than 50 meters from their nearest neighbor.
We recorded all predator sightings and the subject’s reaction to predators during each
observation. We defined predator sightings as a predator or group of predators being within sight
75
of the fawn’s hiding spot. We did not include warthogs as they are not regarded as predators by
gazelle mothers (Chapter 2). The subject reactions we recorded consisted of vigilance, moving
away from the fawn (in the case of raptors), butting the predator, approaching the predator,
retrieving the fawn and fleeing, and chasing and butting at a jackal that chased the fawn. We
considered vigilance to be a passive response to the predator and all other actions to be an active
response. We also noted when fawn hiding bouts began and ended, as well as whether each
hiding bout was terminated by the mother retrieving the fawn or by the fawn standing up on its
own.
Habitat measurements
After each observation we returned to sites recorded for each subject and measured
habitat openness and grass characteristics. For mothers, we measured habitat characteristics at
subject locations recorded during both active and hiding periods. In 2011 we measured grass
characteristics along a 25 meter transect. The transect began at the subject’s location and
proceeded along a randomly determined compass bearing. Every meter along the transect we
inserted a thin metal pin vertically into the ground. We recorded the number and color (green
or brown) of each grass leaf and stem that contacted the pin, as well as the height of the tallest
point of contact. For each transect we calculated the mean height and the percent of all contacts
that were green plant material (percent green) and grass leaves (percent leaf). We also recorded
habitat openness (Bitterlich 1948). In 2012 we conducted five pin measurements at each site: one
at the subject’s location and one five meters away in each of the cardinal directions. We also used
a pasture disc to measure biomass at each of these five points. The pasture disc was constructed
of a 16-inch square of 0.08-inch thick acrylic with a pole inserted through a hole cut in the
76
Mean Pin Height (cm)
Figure 13. Linear
regression comparing
grass height
measurements from
pin drops and pasture
disc methods.
y = -0.571902 + 1.7445261x
p < 0.0001
r2 = 0.71
50
40
30
20
10
0
0
5
10
15
Mean Pasture Disc Height (cm)
20
center. The square was raised and lowered with string and was weighted with approximately 500
grams of weight. To measure biomass, we raised the square and placed the pole in the location to
be measured. We then gently lowered the square until it rested on the grass. We then recorded the
height at which the square rested using centimeter measurements marked onto the pole.
To compare vegetation height measurements made using the different 2011 and 2012
methods, we conducted both types of measurements at 20 randomly chosen sites within known
gazelle habitat. We used linear regression to compare the mean pin height and pasture disc
measurements for site. The two measurements were strongly correlated (Figure 13). We used the
equation for the line of best fit to convert the pasture disc measurements to equivalent pin heights:
1.745d – 0.572 = p
where d is the mean pasture disc measurement and p is the equivalent pin height.
Analyses
We performed all statistical analyses in JMP 10.0. We tested all data for normalcy using the
Shapiro-Wilk W test. Except where noted, all data were non-normal and therefore subjected to
nonparametric tests in our analyses. In all tests, significance was accepted at α=0.05.
77
Fawn risk and hiding behavior. To calculate predation risk during hiding we determined the
probability of a predator being sighted during a given ten-minute period of observation.
Excluding fawn active periods, we divided each mother observation into 10-minute blocks. We
then tallied the predator sightings that occurred within each block. Single predator sightings
could span multiple blocks if the predator stayed in the area for an extended period. We
disregarded blocks shorter than 10 minutes in length, which resulted from fawns emerging from
hiding or the observation ending before the end of a block. Disregarding incomplete blocks
excluded 499 minutes and one predator sighting from analysis. We calculated the proportion of
blocks with predator sightings.
From our observation notes we counted the number of times fawns in each observation
emerged from hiding and whether the fawn emerged on its own or waited for its mother
to retrieve it. For each observation we calculated the rate of emergence events per hour of
observation, and compared the mean rates across fawn age groups using a Wilcoxon test, which is
valid for non-parametric data. We used Fisher’s exact test, which is valid for small sample sizes, to
measure the difference in relative frequency with which fawns of different ages initiate the end of
hiding bouts. For each fawn emergence event, we calculated the proportion of the two preceding
maternal behavior samples that were scored as vigilance, and compared the mean proportions
across fawn age groups using a Wilcoxon test.
Maternal activity budgets. In calculating maternal activity budgets we excluded observations in
which predator sightings occurred to remove any effect of the predator on maternal behavior.
Maternal behavior patterns differ when the fawn is hidden compared to when it is active, and our
sampling method did not ensure that hiding and active periods were sampled in the proportions
78
in which they actually occur. To compensate for this, we divided maternal behavior samples into
those taken while the fawn was active and while the fawn was hiding, and from these datasets
calculated two activity budgets for each mother. We excluded from our analyses any activity
budgets based on fewer than four behavior samples. To create overall activity budgets, we
recombined the hiding and active budgets after weighting them according to Fitzgibbon’s (1990b)
data on the proportion of time spent active and hiding by fawns in each age group. Across all
activity budgets, percent time spent feeding was normally distributed but all other activity data
were non-normal. We used Student’s T-test to compare the mean percentage of time nonmothers and mothers in each fawn age class spent feeding, and Wilcoxon tests to compare mean
time spent on all other activities. We conducted comparisons using both overall maternal activity
budgets (active and hiding periods) and hiding period activity budgets.
Habitat measurements. We used Wilcoxon tests to compare mean grass heights, percent leaf and
percent green between mothers in different fawn age groups and non-mothers.
Grouping patterns. We calculated mean group size for each subject and compared means of
mothers of old and young fawns and non-mothers using a Wilcoxon test. We also calculated the
percent of the observation that each subject spent alone, with one other adult, in small groups
(3 to 4 animals), medium groups (5 to 19 animals) and large groups (more than 20 animals). We
compared the mean percent time spent in each group by non-mothers and mothers of old and
young fawns using Wilcoxon tests.
We used Wilcoxon tests to compare the mean distances between subjects and their fawns
and nearest neighbors. In order to detect differences in the amount of variation in the distances
79
subjects maintain from their fawns and nearest neighbors we compared mean standard deviations
in distances as well. Finally, we calculated the percent of time that each subject spent on the edge
of social groups and compared means using a Wilcoxon test.
Results
We conducted 57 observations on mothers and 47 observations on non-mothers. The
observation time totals for mothers and non-mothers were 129.8 hours and 93.9 hours,
respectively. Fawns were hiding for at least part of every mother observation, a total of 103.7
hours. Fawns were active during 47 of the mother observations, a total of 26.1 hours.
Risk of predator encounter during hiding bout
We documented a total of 24 predator sightings during our observations. Sightings
occurred in 18 (17.3%) out of 104 observations. Sixteen sightings were of jackals, six sightings
were of raptors, and two sightings were of baboons. Four sightings (three jackal, one baboon)
occurred during four (8.5%) out of 47 observations of non-mothers. Sixteen sightings (nine
jackal, six raptor, one baboon) occurred during 11 (19.3%) out of 57 observations of mothers
while the fawn was hiding, and four sightings (all jackals) occurred during four (8.5%) out of 47
observations of mothers while the fawn was active.
There were 572 10-minute blocks of observation time of mothers with hiding fawns;
predator sightings occurred in 20 blocks. Risk of a predator encounter during a given 10-minute
period during which the fawn was hiding was therefore 3.5%. Furthermore, although all
predators were clearly visible to the subject mothers, mothers responded to predators in only
11 out of 16 (69%) sightings. Only seven sightings (44%) warranted an active response by the
80
mother and only two (12.5%) sightings resulted in an actual attack on the fawn. In a given
10-minute period, there was only a 1.9% (11/572 blocks) chance of an active maternal response
to a predator being required. Maternal responses may have been important to preventing attacks
after a predator sighting occurs; nevertheless, the risk of a predator encounter that required an
active maternal response in any 10-minute period was less than 2%. In contrast, four predator
sightings occurred while fawns were active. All four elicited active responses from mothers, and
three resulted in attacks on the fawn, one of which was successful.
Variation in fawn risk with age
As in previous studies (Fitzgibbon 1990b), old fawns in our observations emerged from
hiding more frequently than young fawns. Young fawns (n=43) stood up 0.53±0.10 times per
hour on average. This is approximately half as frequently as old fawns (n=13), which stood up
0.98±0.18 times per hour (Wilcoxon test, Z=1.99, p=0.046).
Old fawns also terminated hiding bouts on their own more often, while young fawns were
more likely to wait for the mothers to retrieve them (Table 6; Fisher’s exact test, p<0.0001). In
instances where the mother terminated hiding bouts, the fawn’s emergence was significantly
more likely to be preceded by maternal vigilance behavior. On average, when they retrieved
fawns (n=30), mothers were vigilant in 1.03 out of the two behavior samples preceding fawn
emergence; when fawns emerged on their own (n=30), mothers were only vigilant in 0.47 of the
two previous behavior samples (Wilcoxon test, Z=2.76, p=0.006).
Initiator
Fawn age
Young
Mother
30
Fawn
12
Total
42
Old
3
18
21
Total
33
30
63
Table 6. Contingency table showing the
frequency with which fawns of different
ages initiate the end of their own hiding
bouts or wait for their mothers to
retrieve them.
81
Variation in maternal behavior across fawn age groups and comparisons to non-mothers
Activity budgets. When both hiding and active periods were considered, mothers of
transitioning fawns seemed to represent an intermediate step between mothers of young fawns
and non-mothers (Figure 14A). However, perhaps due to the gradual nature of the transition,
this step did not yield significant differences between mothers of young and transitioning fawns
in the amount of time spent vigilant, feeding, lying down, self-grooming or grooming their fawns.
However, there were differences between maternal activity budgets and those of non-mothers.
Mothers of both young (n=19) and old fawns (n=10) spent more time vigilant than non-mothers
(n=47; Wilcoxon tests, non-mother vs young: Z=4.17, p<0.0001; non-mother vs old: Z=2.37,
p=0.018). Compared to non-mothers, mothers of young fawns also spent less time lying down
(Z=-3.20, p=0.0014) and self-grooming (Z=-2.43, p=0.015).
When fawns were hidden, there were no significant differences between mothers of old and
young fawns in time spent feeding, lying down, or self-grooming (Figure 14B). Neither mother
category differed from non-mothers in time spent feeding or self-grooming. However, mothers of
young fawns (n=33) spent more time vigilant than mothers of old fawns (n=10; Wilcoxon test,
Z=-3.49, p=0.0005). Mothers of young fawns were also more vigilant than non-mothers (n=47;
Z=3.02, p=0.0025) while mothers of old fawns were less vigilant than non-mothers (Z=-2.10
p=0.036). Finally, mothers of young fawns spent a lower percentage of their time lying down
than non-mothers (Z= -2.62, p=0.009).
When fawns were active, mothers of young fawns (n=21) spent less time feeding than
mothers of transitioning fawns (n=10; Students t-test, t(73)=-2.35, p=0.02). Otherwise there were
no significant differences between mothers of old and young fawns in the time spent on each
activity (Figure 14C). Mothers of both old (n=10) and young (n=21) fawns were significantly
82
A. Overall
80
Percent time
60
40
20
0
B. Fawn
hidden
80
Percent time
60
40
20
0
C. Fawn
active
80
Percent time
60
40
20
0
Vigilant
Mothers
(young fawns)
Feeding
Lying
down
Mothers
(transitioning fawns)
Selfgrooming
Grooming
fawn
Non-mothers
Figure 14. Activity budgets of non-mothers, mothers of young fawns and mothers of old fawns. Maternal
activity budgets are given for all data (A), hiding periods (B), and active periods (C). Error bars indicate
standard error.
83
more vigilant than non-mothers (Wilcoxon tests, young vs non-mother: Z=5.04, p<0.0001; old vs
non-mother: Z=3.22, p=0.0013) and spent less time lying down (young vs non-mother: Z=5.20,
p<0.0001; old vs non-mother: Z=3.53, p=0.0004) and more time grooming fawns (young vs nonmother: Z=4.95, p<0.0001; old vs non-mother: Z=2.99, p=0.003). Mothers of young fawns also
spent less time feeding (Students t-test, t(73)=-3.54, p=0.0007) and self-grooming (Wilcoxon test,
Z=-2.71, p=0.007) than non-mothers.
Habitat type. Our vegetation analyses revealed that when fawns were hidden, mothers of
young fawns (n=35) selected habitats with taller grass than non-mothers (n=41; x̄young=8.2±1.0
cm, x̄non-mother=4.9±1.0 cm; Wilcoxon test, Z=2.27, p=0.02), and mothers of old fawns (n=8)
were found in areas with greener grass than mothers of young fawns (n=31; x̄young=52.0±3.8%,
x̄old=76.0±7.6%; Z=-2.71, p=0.007) and non-mothers (n=40; x̄=53.7±3.4%; Z=-2.66, p=0.008).
There were no significant differences between groups in the mean percent leaf of subject
locations, with means for all groups ranging from 74.7 to 79.0%. There were also no differences
in habitat openness, with all groups selecting very open habitats (mean Bitterlich <1 for all
groups). Finally, mothers of old fawns, mothers of young fawns, and non-mothers did not differ
in any habitat characteristics during active periods.
Grouping patterns. On average, mothers of young fawns (n=42) stayed further away from
their hiding offspring than did mothers of old fawns (n=12; x̄young=102.6±8.9 m, x̄old=59.8±16.7
m; Wilcoxon test, Z=-2.23, p=0.026). Mothers of young fawns also maintained higher
variance in their mother-fawn distances, as indicated by standard deviation (nyoung=42, nold=11,
x̄young=45.2±4.4 m, x̄old=24.3±8.6 m; Z=-2.75, p=0.006). However, when fawns were active,
84
Fawn
Hiding
% Time in
Medium
Groups (5
to 19)
26.7±7.11
% Time
in Large
Groups
(>20)
5.4±6.31
1.5±6.81
27.8±9.31
15.4±9.01
33.3±11.01
22.1±9.72
17.2±2.22A
5.7±3.71A
9.9±4.62A
17.6±4.01A
35.1±4.71A
30.3±4.92A
5.9±2.3B
26.2±3.8B
31.6±4.7B
16.7±4.1A
17.9±4.9A
7.0±5.1B
14.0±4.2A
9.9±7.0A
36.4±8.7B
6.6±7.5A
17.8±9.0A
29.4±9.3A
% Time
Alone
18.1±4.51
13.2±4.22
Non-mother 47
Young 43
Old 13
Fawn Age
Group
Fawn
Active
% Time in
Pairs
32.6±6.01
% Time
in Small
Groups (3
to 4)
17.2±5.81
Mean
Group
Size
5.7±2.71
n
Young 31
Old 13
Table 7. Mean group sizes and percent time spent in groups of various sizes for non-mothers and mothers
of old and young fawns during hiding and active periods. Values given are group means and standard error.
Values in each column connected by the same superscript letter (hiding period) or number (active period)
are not significantly different.
mothers of young fawns (n=32) stayed closer to their offspring than di mothers of old fawns
(n=12; x̄young=6.5±3.5 m; x̄old=13.4±5.7 m; Z=2.67, p=0.008) and vary less in their mother-fawn
distances (nyoung=15, nold=11; x̄young=5.2±3.2 m; x̄old=14.7±3.8 m; Z=2.27, p=0.024).
There were no significant differences in average group size between mothers of old and
young fawns during hiding periods but, mothers of young fawns associated with smaller groups
than non-mothers (Wilcoxon test, Z=-4.27, p<0.0001; Table 7). When fawns were active,
mothers of young fawns associated with smaller groups than both non-mothers (Z=-3.97,
p<0.0001) and mothers of old fawns (Z=2.30, p=0.02).
When we examined the percentage of time that subjects spent alone, in pairs, in small
groups (3-4 animals), in medium groups (5-19 animals), and in large groups (>20 animals;
Table 7), we found that while their fawns were hiding, mothers of young fawns spent more than
a quarter of their time alone, which was significantly more time than mothers of old fawns
(Wilcoxon test, Z=2.19, p=0.03) and non-mothers (x̄=5.7±3.7; Z=4.73, p<0.0001). Mothers of
young fawns also spent less than ten percent of their time in large groups, which was significantly
less than mothers of old fawns (Z=-2.23, p=0.026) and non-mothers (Z=-2.97, p=0.003).
Mothers of old and young fawns did not differ in the time they spent in pairs or in small or
85
medium groups, but both groups spent more time in pairs than non-mothers (young vs nonmother: Z=3.50, p=0.0005; old vs non-mother: Z=2.41, p=0.016). The composition of pairs did
not differ across female groups, with approximately two-thirds (62.5% to 71.4%) of pairs being
composed of the subject and a territorial male. Likewise, females that were alone did not differ
in the availability of groups to join: across fawn age groups and non-mothers, there was a group
within sight in 94.1% to 100% of instances where the subject was alone.
When fawns were active, there were no significant differences across subject groups in time
spent alone or in small or medium groups (Table 7). However, mothers of young fawns spent
significantly less time in large groups than mothers of old fawns (Wilcoxon test, Z=-2.07, p=0.04)
and non-mothers (Z=-3.46, p=0.0005), and mothers in both fawn age groups spent more time
in pairs than non-mothers (young vs non-mother: Z=1.99, p=0.046; old vs non-mother: Z=2.08,
p=0.037).
When fawns were hidden, mothers of young fawns (n=24) in medium and large groups
spent significantly more time on the edge of the group compared to non-mothers (n=37;
x̄young=59.0±7.4%, x̄non-mother=34.0±5.9%; Wilcoxon test, Z=2.42, p=0.016). When groups
smaller than five were included (and all animals in these groups were considered to be on the
edge), mothers of young fawns (n=43) spent more than 85% of their time on the edge, again
significantly more time than non-mothers (n=47; x̄young=87.9±5.1%; x̄non-mother=54.1±4.9%;
Z=4.27, p<0.0001). When fawns were active, there were no differences between subject groups in
percent time spent on the edge of medium and large groups. However, when smaller groups were
included, mothers of young fawns (n=31) spent 89.6±6.2% of their time on the edge on average,
which was significantly more than mothers of old fawns (n=13; x̄old=66.3±9.6%; Z=2.76,
p=0.006) and non-mothers (n=47; x̄non-mother=54.1±5.1%; Z=4.46, p<0.0001).
86
When fawns were hidden, mothers of young fawns (n=42) were on average significantly
farther from their nearest neighbor than were non-mothers (n=47; x̄young=58.1±7.9 m, x̄nonmother
=32.2±7.5 m; Wilcoxon test, Z=3.83, p=0.0001), and varied more in their nearest neighbor
distances (standard deviations: x̄young=35.7±4.5 m, x̄non-mother=26.1±4.3 m; Z=2.81, p=0.005).
When fawns were active, there were no significant differences in average nearest neighbor
distance between subject groups.
Discussion
Older fawns in our system were at greater risk of predation than their younger counterparts
due to their tendency to stand up more frequently and terminate more hiding bouts on their
own. Fitzgibbon (1990b) described the same behavioral patterns in 3 to 8 week old gazelle fawns
in the Serengeti, and demonstrated that these behaviors resulted in a steady increase in the
amount of time older fawns spent out of hiding and therefore exposed to predators. Indeed,
in our study, more than half of the predator sightings that occurred while fawns were hiding
required active maternal intervention, and fewer than 15% resulted in an attack on the fawn;
however, all predator sightings that occurred while fawns were active elicited defensive action by
mothers and 75% of them resulted in an attack on the fawn. This reflects the effectiveness of the
hiding strategy at concealing fawns and the risk transitioning fawns assume when they reduce
their hiding behavior (Barrett 1978; Fitzgibbon 1990b). Furthermore, because mothers cannot
anticipate the end of hiding bouts when fawns emerge on their own, they are less vigilant prior to
fawn-initiated emergences than mother-initiated emergences. Mothers returning to retrieve their
offspring increase their vigilance behavior in order to augment their chances of detecting lurking
predators before revealing their fawns (Chapter 3; Byers 1997). Thus, fawns that emerge on their
87
own forfeit the benefit of maternal vigilance and thereby put themselves at higher risk than those
that wait for their mothers to retrieve them.
Mothers of transitioning fawns did not appear to compensate for this higher fawn risk
by increasing their vigilance behavior relative to mothers of young fawns. Nor did they invest
more time in fawn care. In fact, contrary to our expectations, mothers of transitioning fawns
had overall activity budgets that were more similar to those of non-mothers than did mothers
of young fawns. Mothers of young fawns sacrificed time spent self-grooming and resting in
favor of maintaining time spent feeding and increasing vigilance rates. Meanwhile, mothers of
transitioning fawns also exhibited heightened overall vigilance rates compared to non-mothers,
but to a lesser degree than mothers of young fawns. Mothers of transitioning fawns were able
to accomplish this lesser increase in vigilance without significant loss of foraging, resting, or
self-grooming time. Mothers of old fawns further violated our expectations by exhibiting lower
vigilance rates than non-mothers while their fawns were hiding. This suggests that mothers do
not attempt to compensate for their inability to predict their fawn’s emergence from hiding by
increasing overall vigilance rates. The only major differences in activity budgets between mothers
of old fawns and non-mothers occured during active periods when mothers decreased their
time spent lying down in favor of increasing vigilance levels. High vigilance during this time
likely has a high pay-off for mothers since infant survival of a predator attack depends heavily
on the mother being able to spot the predator and warn the fawn while the predator is still far
away (Fitzgibbon 1990b). Even though mothers of transitioning fawns were more vigilant than
non-mothers due to these high-risk active periods, we found that they were still significantly less
vigilant overall than mothers of younger fawns. This finding agrees with previous results from
gazelles in the Serengeti (Fitzgibbon 1988).
88
Rather than increased vigilance, gregarious grouping behavior combined with the lower
mother-fawn distances maintained by mothers of old fawns during hiding periods appear to
be the key to mitigating fawn risk during the hiding transition. Whereas mothers of young
fawns differed greatly from non-mothers in their grouping patterns, mothers of old fawns were
similar to non-mothers in nearly every social aspect we measured. Mothers of old fawns and
non-mothers did not differ in their mean group sizes, nearest neighbor distances, or in the
proportions of time they spent alone and in small, medium and large groups. When old fawns
were hidden, their mothers were less than 60 meters away on average. At this distance, if the
mother is in a group and not on the edge (both usually the case for mothers of old fawns), then
the hidden fawn is either within or on the edge of the group. If the fawn emerges from hiding
independently, it does so in the immediate vicinity of a group. Furthermore, Fitzgibbon (1988)
found that, compared to mothers of young fawns, mothers with older fawns were less likely to
detect approaching cheetahs before a control non-mother female in the same social group. This
suggests that mothers of older fawns rely more heavily on the collective vigilance of the group for
offspring protection compared to mothers of younger fawns.
It therefore appears that gazelle mothers’ primary strategy for mitigating the increased levels
of fawn risk during the transition out of hiding is to rejoin social groups. The fawn’s shorter
hiding bouts and more frequent active periods make this possible: rather than being tied to a
hidden infant that is stationary for hours, fawns emerge from hiding and can therefore change
locations much more frequently, which allows mothers to track group movements. Spending
more time in groups results in mothers of transitioning fawns spending more time in the short
grass habitats preferred by non-mothers (Brooks 1961) rather than the taller grass areas used by
mothers of young fawns. Shorter grass may make it harder for fawns to hide effectively, but Byers
89
(1997) suggests that hiding may already be less effective for older fawns due to their larger body
size and that body size increases may be a driving factor behind the timing of the transition out
of hiding.
Spending less time alone and more time in larger groups not only mitigates fawn risk, but
also decreases maternal risk and may facilitate fawn socialization. Although older fawns spend
more time active, their mothers do not spend more time grooming them compared to mothers
of younger fawns. Furthermore, older fawns are further from their mothers during active periods
than are younger fawns. This suggests that older fawns are spending their active periods apart
from their mothers but within her social group, possibly socializing with other group members.
The tendency for mothers to rejoin groups during their offsprings’ transition out of hiding
is evident in other hider species as well. White-tailed deer does are aggressive and defend small
territories around fawn hiding spots (Schwede et al. 1993). However, this behavior wanes as
hiding decreases, resulting in more frequent maternal associations with conspecifics as the fawn
ages. In pronghorn, does isolate for parturition but begin forming groups with other does and
their fawns starting approximately three weeks after birth, when the fawn begins its transition out
of hiding (Autenrieth & Fichter 1975). Interactions between fawns and group members during
this time are critical to fawn socialization.
Although the transition from hiding to following is a risky time for fawns, mothers do not
appear to bear the cost of this risk. Rather than mitigating fawn risk through vigilance, they rely
on group concealment, risk dilution and collective vigilance to protect the fawn. Mothers in fact
lower their own risk during the transition by spending more time in the safety of groups. Mothers
also benefit from the transition by occupying better habitats and exhibiting activity budgets more
similar to those of non-mothers than they do at the beginning of the hiding phase.
90
Acknowledgements
We are grateful to Ol Pejeta Conservancy for permitting us to work on the premises and
to the Conservancy’s Ecological Monitoring department and security staff for their assistance in
the field. This research was conducted with permission from the Government of Kenya (permit
NCST/5/002/R/648). Funding was provided by Princeton University’s Department of Ecology
and Evolutionary Biology, the Philadelphia chapter of the Explorers Club, and the Animal
Behavior Society’s student research grant program. DIR was supported by National Science
Foundation grants IBN-9874523, CNS-025214, and IOB-987452.
91
Chapter 5: Maternal responses to simulated cheetah attacks and their effect on group
antipredator responses in Thomson’s gazelle Blair A. Roberts and Daniel I. Rubenstein
Abstract
Group membership during predator attacks benefits prey individuals by increasing the
likelihood of predator detection prior to the attack and by allowing for cohesive group responses.
We examine whether mother ungulates additionally benefit from group membership by 1)
relying on group vigilance to detect potential fawn predators and 2) recruiting group members
to participate in defense responses. We observe the responses of Thomson’s gazelle groups
without mothers and mothers in groups to simulated cheetah attacks. We find that mothers
detect predators sooner than groups and therefore do not rely on group vigilance to detect fawn
predators. Mothers’ responses are riskier than group responses in that mothers approach the
predator more often than do groups. Finally, maternal responses influence groups to stay in the
vicinity of the predator, which may reduce the risk to defending mothers. This study offers novel
findings regarding the antipredator behavior of mothers and its effects on other group members,
as well as a new approach to experimentally simulating predator attacks.
Introduction
An individual’s behavioral response to a predator attack can strongly affect whether the
individual survives the attack or not (Gese & Grothe 1995; Kullberg et al. 1998; Wisenden et al.
1999; Lingle & Pellis 2002; DiRienzo et al. 2013). Among non-cryptic ungulate species flight is
perhaps the most common reaction to attacking predators, but other responses include visual and
92
auditory displays such as stotting, a stiff-legged leaping gait performed in response to predators
(Bildstein 1983; Caro 1986b, 1994; Stankowich & Coss 2008; Pipia et al. 2009); approach,
monitoring, and/or harassment of the predator (Leuthold 1977; Berger 1979; Novaro et al.
2009); and grouping with or without the use of cooperative defense formations (Sikes 1971; Gray
1974; Lingle 2001). The risk associated with each type of behavioral response (and the likelihood
of a response being employed) depends on various circumstances of the attack, including the
species and number of predators, characteristics of the habitat in which the attack occurs, and
the prey individual’s group size, age and condition (Leuthold 1977; Fitzgibbon 1988; Caro &
Fitzgibbon 1992; Fitzgibbon 1994; Bleich 1999; Lingle & Pellis 2002). However, in general those
responses that distance the prey from the predator tend to confer lower risk of injury or death
than those that increase proximity to the predator (Kruuk 1972; Sordahl 1990; Dugatkin 1991;
Fitzgibbon 1994).
In determining their response to a predator, individuals may be influenced by the actions of
conspecifics. For example, Fitzgibbon (1994) found that it was rare for only a portion of a gazelle
group to engage in predator inspection behavior; rather, either all members or no members
of a given group displayed such behavior. Group cohesiveness and similarity of response often
increases the effectiveness of an antipredator behavior through greater intimidation, confusion,
or distraction of the predator (Landeau & Terborgh 1986; Pitcher 1986; Lingle 2001; Novaro et
al. 2009). Furthermore, group membership dilutes individual risk (Hamilton 1971), and similarity
to other group members reduces the chance that an individual will be singled out and targeted by
a predator (Landeau & Terborgh 1986; Theodorakis 1989; Caro & Fitzgibbon 1992). Therefore,
an individual may actually be safer when participating in an approach or attack behavior as part
of a group than it would be if it abandoned the group and fled on its own.
93
Ungulate infants are at particularly high risk due to their smaller size, which renders
them vulnerable to a greater number of predator species (Sinclair et al. 2003), and their
underdeveloped flight ability, which makes them more likely to be caught once detected (Barrett
1978; Fitzgibbon 1990b). In order to protect their offspring, mothers with dependent young
engage in approach and attack responses more frequently than other classes of ungulate (Walther
1969; Lipetz & Bekoff 1980; Gese 1999) and increase their vigilance behavior to ensure early
detection of approaching predators (Chapter 3; Chapter 4; Fitzgibbon 1990b,1993). Defensive
behavior often prevents offspring from being killed (Lingle et al. 2005; Novaro et al. 2009), but
also increases risk of maternal injury or predation (BR, personal observation; Estes & Estes 1979;
Smith 1987). Likewise, vigilance behaviors take time away from other activities, such as resting
and foraging (Chapter 3; Chapter 4), and thus may negatively affect maternal fecundity. The
costs of these investments may be reduced through the use of group members. If a mother’s
reaction to a predator influences the reaction of the group, she may benefit from group cohesion
or risk dilution as described above. Furthermore, mothers in groups may be able to rely more on
group vigilance for predator detection and thereby increase their foraging rates.
Here we investigate the antipredator behavior of Thomson’s gazelle groups and mothers
within social groups in response to simulated predator attacks. We first examine whether mothers
detect predators sooner than non-mothers to determine if mothers use group vigilance to
subsidize their own vigilance behavior. If mothers do rely on group vigilance, we expect them
to detect approaching predators later than groups without mothers. We then examine whether
individual mothers with hidden fawns take on more risk in their response compared to groups of
non-mothers. Riskier responses may include approaching the stimulus rather than fleeing, fleeing
shorter distances from the stimulus, or remaining in the vicinity of the stimulus following the trial.
94
We expect mothers to accept more risk in order to actively defend their fawns or maintain visual
contact with the “predator” and their hidden offspring. Finally, we examine whether groups that
include a mother exhibit riskier responses than groups with no mothers. Because all groups have
only one mother and group defense of unrelated offspring has never been reported in Thomson’s
gazelles, we do not expect group members to engage in active defense of the fawn. However, we
predict that mothers’ riskier responses will influence group behavior, resulting in riskier responses
by groups with mothers compared to those without mothers.
Methods
Study site
We conducted all fieldwork at Ol Pejeta Conservancy (OPC) in Laikipia, Kenya between
June and September, 2012. OPC is a fenced, 360 km2 conservancy consisting of discrete grassy
plains separated by Acacia drepanolobium and Euclea divinorum woodlands. Animals on OPC are wellhabituated to vehicles; therefore, we were able to approach individuals and groups to distances
of approximately 100 to 150 meters without disturbing our subjects. In 2012, OPC had a
population of approximately 1300 Thomson’s gazelles. Other herbivore species are abundant, as
are predators, including lions, cheetahs, black-backed jackals, spotted hyenas, olive baboons, and
birds of prey.
Experimental apparatus
In each trial we presented gazelle with one of three stimuli: a cheetah puppet, a vervet
monkey (Chlorocebus pygerythrus) puppet, or no puppet. Puppets were crafted out of baling wire and
fabric and mounted on a battery-powered remote control (RC) car (Traxxas E-Maxx) with an
95
Figure 15. Cheetah puppet mounted on RC car. Photo courtesy of John Herzfeld.
Figure 16. Vervet monkey puppet mounted on RC car. Photo courtesy of John Herzfeld.
96
operating range of approximately 100 meters. The cheetah puppet featured yellow fabric with
black spots, a face constructed out of a child’s cheetah mask, and a long tail topped with black
fur stripes and a black tip (Figure 15). Due to the flexibility of the baling wire, the tail bounced
up and down as the remote control car moved, simulating the balancing motion of a running
cheetah’s tail. The vervet monkey featured a buff grey body, black face and long tail (Figure 16).
The puppets were similarly sized. In trials with no puppet, the RC car was presented with no
puppet affixed. Cheetahs are effective predators of both adult and juvenile Thomson’s gazelles
(Fitzgibbon & Fanshawe 1989) while vervet monkeys pose no risk to any age class. Both cheetahs
and vervet monkeys are common at OPC and frequently seen in gazelle habitat.
The RC car was capable of maneuvering over or through small tufts of grass but tended
to become stuck on large grass tufts, in dense tall grass, and in large holes; therefore, we
limited trials to short grass areas and maneuvered around obstacles when necessary. Princeton
University’s Institutional Animal Care and Use Committee reviewed and approved all
experimental procedures used in this study (protocol #1842).
Group trials
To determine the validity of our cheetah stimulus, we tested whether gazelle groups
reacted differently to our experimental (cheetah) and control (vervet monkey, RC car) stimuli.
We conducted 30 trials on eleven groups. Nine groups were subjected to three trials each, one
trial with each stimulus. One group was subjected to one trial, using the vervet stimulus, and
one group was subjected to a vervet trial and a cheetah trial. We varied the order in which the
stimuli were presented so that no more than three groups experienced the stimuli in the same
order, and allowed at least five days between trials on any given group. Because gazelle groups
97
are fluid in their composition, it is highly unlikely that consecutive trials on any given group were
performed on the same group of individuals. However, plains are far enough apart and separated
by dense enough woodland to suggest that gazelle rarely switch plains. Furthermore, individuallyrecognized females could be found on the same plain from week to week. Therefore, where
possible we only conducted trials on one group per plain. However, due to the limited number of
plains available, we had to conduct trials on multiple groups on three large plains. In these cases,
we conducted trials in distinct areas of the plain where gazelle were normally found. Within these
plains, the order of trials was held constant and trials with the same stimulus were performed on
the same day to prevent individual gazelle from being subjected to the same stimulus twice. Trials
conducted on the same day were conducted far enough apart that individuals from the first trial
were not affected by the second trial and vice versa.
Groups ranged in size from six to 49 individuals. After locating a group for trial, we parked
our vehicle approximately 120 meters from the nearest group member; as noted above, OPC
gazelles do not flee from vehicles at this distance. After observing the group’s behavior for 15
minutes and ensuring that the animals were no longer alert to us, we recorded the location of the
group using a laser rangefinder and a compass, and drove the stimulus slowly toward the group.
We drove the stimulus in as straight a trajectory as possible; however due to obstacles, equipment
malfunction or operator error, in ten trials (three RC car, three cheetah, and four vervet) the
stimulus followed an oblique trajectory aimed at the edge of the group. We noted when the group
became alert to the stimulus (detection distance) and stopped moving the stimulus when the
group fled (reaction distance). In three vervet trials, two cheetah trials and one car trial the RC
car became stuck on grass tufts or in holes after the group detected it but prior to the reaction
distance. In these trials, we recorded the detection distance but were unable to collect data for
98
other distances. We recorded the locations of the stimulus when the group became alert and fled,
and recorded the location to which the group fled. We continued observing the group’s behavior
for 15 minutes or until the group left the area and went out of sight.
Mother-fawn trials
We performed mother-fawn trials opportunistically upon locating a mother with a hidingaged fawn and waited for fawns to hide before conducting the trial. Based on their coloration
and relative size to their mothers, all fawns were less than one month old (Fitzgibbon 1990b).
Each mother-fawn pair was only subjected to one trial, and all trials were performed using the
cheetah puppet. The shortest time between trials on the same plain was two days. In every trial,
the mother was in a group with other adult gazelle and was the only apparent mother in the
group. Group size ranged from two individuals (trial 4) to 41 individuals (trial 2). In each trial, we
observed baseline behavior as in the group trials and recorded the location of the fawn and the
mother prior to the trial. We then drove the stimulus slowly toward the hidden fawn, stopping it
when the mother reacted. In one trial (trial 3), the mother was less than 5 meters from the fawn
when the stimulus was presented and therefore the stimulus was also driven toward the mother
and her group. In the other four trials, the mother and her group were more than 100 meters
away from the fawn and therefore not targeted by the stimulus. We described the mother’s
reaction in ad libitum notes, and recorded her location after her initial reaction as well as the
location of the stimulus. We observed mothers for at least 15 minutes following the presentation
of the stimulus.
99
Analyses
We performed all statistical analyses in JMP 10.0. In addition to calculating detection
and reaction distances, for each group and mother-fawn trial, we calculated the distance
between the subject and the stimulus when the subject stopped after fleeing (safety distance).
We also calculated the distance between the subject’s starting point and where it stopped after
fleeing (displacement distance), and subtracted the reaction distance from the safety distance
to determine the amount of separation from the predator that the subject gained by fleeing
(distance gained). Because the subject’s detection distances may have been limited by the starting
distance of the stimulus from the group (that is, detection distances cannot be greater than the
maximum distance of the stimulus from the subject), we normalized the detection distance
by dividing it by the maximum distance of the stimulus from the group. We then compared
the means of these measurements for group trials across treatments. We tested the data for
normalcy using the Shapiro-Wilk W test. The normalized detection distance (W=0.91, p=0.06),
safety distance (W=0.96, p=0.53) and distance gained (W=0.92, p=0.09) data were normally
distributed, but reaction distance (W=0.88, p=0.008) and displacement distance (W=0.88,
p=0.01) were not. We used Student’s t-tests to compare means for the parametric data and
Wilcoxon tests to compare nonparametric data. We also used two-tailed Fisher’s exact tests to test
whether groups’ tendency to leave sight of the stimulus varied across treatments.
In comparing mother-fawn trials to group trials, we used non-parametric tests due to
the skewed and bimodal distributions resulting from our small sample sizes. Our sample size
also prevented effective means comparisons for some variables; in these cases we pooled data
into binary groups based on relevant threshold values. We used a Wilcoxon test to compare
normalized detection distances between mothers and groups. We pooled reaction distances into
100
categories of greater and less than 100 meters, which is the median detection distance of the
cheetah for mothers and groups. We lumped distances gained into categories of greater and
less than 20 meters. We chose 20 meters as a threshold because Caro (1986b) reports that this
is the median distance from prey at which cheetahs began their chases in successful hunts of
Thomson’s gazelle. Therefore, gazelle that increase their distance from the predator by more than
20 meters will likely be far enough from the predator to avoid being caught. We used two-tailed
Fisher’s exact tests to compare these pooled variables and the frequency with which the subject
left sight of the stimulus following the trial across trial types.
Results
Group trials
Groups reacted more strongly to the mobile cheetah puppet compared to the other two
Mean distance (m)
stimuli (Figure 17). They detected the cheetah stimulus (n=7) at greater distances than the vervet
120
RC car
100
Cheetah
Vervet
80
60
40
20
0
Detection
distance
Reaction
distance
Displacement
distance
Safety
distance
Distance
gained
Figure 17. Differences in group reactions to RC car, vervet monkey and cheetah stimuli. Detection and
reaction distances are the distances between the group and stimulus when the group noticed and fled from
the stimulus, respectively. Displacement distance is the distance the group fled from its original location.
Safety distance is the distance between the group after it fled and the stimulus (which stopped moving once
the group fled). The distance gained is the difference between the safety distance and the reaction distance.
Error bars indicate standard error.
101
(n=11; normalized detection distance, Student’s t-Test, t(18)=-3.0, p=0.007), and reacted to the
cheetah stimulus (n=8) at greater distances than the vervet (n=8; Wilcoxon test, Z=-2.21, p=0.03)
and the RC car (n=7; Z=2.03, p=0.04). Groups ran further when presented with the cheetah
(n=8) compared to the vervet (n=8; Wilcoxon test, Z=-2.48, p=0.01) and the RC car (n=6;
Z=2.39, p=0.02), resulting in greater safety distances for the cheetah (n=8) than for the vervet
(n=7; Student’s t-test, t(18)=-2.40, p=0.03) and the RC car (n=6; t(18)=3.65, p=0.002). There
were no significant differences between treatments in the distance the groups gained by running
from the stimulus or in the group’s tendency to leave sight of the stimulus. Means for vervet
and RC car trials did not differ in any metric. We also found no effect of the order in which the
stimuli were presented on the normalized detection, reaction or flight distances of groups. We
found no significant differences in means for any response measure between straight and oblique
stimulus presentations within each stimulus group, suggesting that the presence of the moving
stimulus near the group was more important than its trajectory relative to the group.
Mother-fawn trials, and comparison to group trials
Mothers’ normalized detection distances averaged 0.98, indicating that they detected the
stimulus almost as soon as it was exposed. This is significantly greater than the group’s mean
normalized detection distance (x̄=0.89; Wilcoxon test, Z=2.17, p=0.03). Compared to groups,
mothers were also significantly more likely to react when the stimulus was more than 100 meters
away (Table 8; Fisher’s exact test, p=0.04). Mothers’ reactions resulted in less distance gained
between themselves and the cheetah stimulus: all groups but only two out of five mothers
increased their distance from the stimulus by 20 meters or more (Table 9; Fisher’s exact, p=0.04).
This result is partially driven by two trials (1 and 4) in which the mother approached the stimulus
102
Reaction
distance
<100 m
>100 m
Group
trials
8
0
8
Motherfawn trials
2
10
3
3
5
13
Distance
gained
<20 m
>20 m
Group
trials
8
0
8
Motherfawn trials
3
11
2
2
5
13
Leave
sight?
Yes
No
Group
trials
6
4
10
Motherfawn trials
0
6
5
9
5
15
Table 8. Contingency table comparing reaction distances
between mother-fawn and group cheetah trials.
Table 9. Contingency table comparing distance gained
between mother-fawn and group cheetah trials.
Table 10. Contingency table comparing whether subjects
left the area following stimulus presentation between
mother-fawn and group cheetah trials.
rather than fleeing from it. In trial 4, the mother left her group and ran toward the cheetah and
fawn while stotting. She then appeared to try to entice the cheetah to chase her by rushing to
within five meters of it before sharply turning and dashing away to a distance of approximately
15 meters. In trial 1, the mother was on the far edge of her social group from her hidden fawn
when the stimulus was presented. She became alert to the cheetah and then ran through the
social group to the edge closest to the cheetah. The tendency for mothers to approach the
cheetah was also evident in trial 3, where, after her initial flight away from the stimulus, the
mother then moved to the edge of the group closest to the cheetah.
Mothers never left sight of the cheetah stimulus following the trial. This is a significant
difference compared to group trials, in which six out of ten groups left sight of the cheetah within
15 minutes of its presentation (Table 10; Fisher’s exact, p=0.04). Furthermore, in every motherfawn trial the mother’s social group also remained in sight of the stimulus.
103
Discussion
Mothers detected the cheetah stimulus sooner after its presentation than did groups without
mothers, suggesting that mothers do not rely on group vigilance for predator detection. This is
a somewhat surprising result given that foraging and resting time are at a premium for mothers,
who face high energetic costs from lactation (Loudon 1985). Relying on group vigilance may not
be effective for the defense of fawns hidden outside the group since group members’ vigilance
may be focused on their own surroundings rather than the fawn’s hiding spot. Furthermore,
adult group members may not respond as readily as mothers to predator species, such as jackals,
that are effective fawn predators but pose little risk to adults (Estes 1967; Macdonald et al. 2004).
Group members’ behavior is geared toward managing their own risk factors, which differ from
those of hidden fawns; therefore mothers must exhibit increased vigilance even in group settings
in order to ensure detection of fawn predators.
Compared to groups, mothers reacted to the predator when it was further away. However,
this does not necessarily result in lower risk to mothers compared to groups since mothers gained
less distance through their reaction. This is because in many cases mothers even approached
the predator suggesting that the benefit of heightened reaction is to reduce the fawn’s risk, not
her own. In one such case (trial 4), the mother’s response – stotting and apparently attempting
to entice the cheetah to follow her – closely matches descriptions of maternal defense behavior
in response to real cheetahs (Fitzgibbon 1988). Fitzgibbon reports that mothers performed this
behavior in approximately 30% of cheetah hunts and that the behavior is indeed risky: in 19
observed instances of this behavior, one mother was caught by the cheetah and killed.
No group engaged in active defense of the fawn, as expected. However, group behavior
did appear to be affected by the presence of the mother in that groups in mother-fawn trials
104
remained within sight of the cheetah for at least fifteen minutes while 60% of groups without
mothers left within fifteen minutes of the trial. It is unlikely that this subtle behavioral change
would affect fawn survival in the event of an actual cheetah attack. However, it may decrease
maternal risk during defense by providing a refuge for her to flee to if the cheetah did begin
to chase her. Remaining in the vicinity of predators, and even approaching them, is a reaction
commonly exhibited by gazelles that may reduce risk of predation by encouraging the predator
to hunt elsewhere or by removing the element of surprise in the event of a subsequent attack
(Fitzgibbon 1994). This may suggest that by staying in the area gazelle groups were decreasing
rather than increasing their risk levels. However, individuals that stay and monitor the predator
do increase their risk of predation if the predator is not deterred (Fitzgibbon 1994). Furthermore,
in our experiment the “predator” never leaves the area. Groups seeking to monitor the predator
are then faced with a choice of either leaving the area themselves or staying and watching
indefinitely. Monitoring reduces foraging activity and thus incurs an opportunity cost (Fitzgibbon
1994); furthermore, predators that do not leave are more likely to hunt again in the same area.
Therefore, after an extended period of monitoring, leaving the area likely becomes the less costly
choice, both in terms of predation risk and opportunity cost.
Further study is necessary to determine the extent to which mothers influence the behavior
of their social groups, the benefits of this influence to mothers and its costs to other group
members. Our novel experimental approach using mobile predator stimuli offers one viable
means to pursue such study. Stationary model predators have been used in previous studies of
ungulate antipredator behavior (Stankowich & Coss 2008; Eby & Ritchie 2012), but the mobility
afforded by the use of the RC car allows for targeting of individual gazelles and using their
responses, especially detection and reaction distances and behaviors, to assess risk. In the future
105
such controlled predator mobility experiments could be used to alter risk by varying different
angles of stimulus presentation.
Acknowledgements
We are grateful to Youngin Lim, Jennifer Schieltz, and Michael Costelloe for their assistance
in conducting trials in the field and to John Herzfeld for the use of his photos. We thank Ol
Pejeta Conservancy for permitting us to conduct our experiment on the premises. This research
was conducted with permission of the Government of Kenya (permit NCST/5/002/R/648).
Funding was provided by Princeton University’s Department of Ecology and Evolutionary
Biology, the Philadelphia chapter of the Explorers Club, and the Animal Behavior Society’s
student research grant program. DIR was supported by National Science Foundation grants
IBN-9874523, CNS-025214, and IOB-987452.
106
Chapter 6: Overview and concluding remarks: hiding versus following Blair A. Roberts
Beginning before parturition, mother gazelle bear nearly all the costs of fawn risk reduction.
Mothers take on increased risk by adjusting their grouping patterns (Chapter 2; Chapter 4)
and habitat use (Chapter 4), and forgo resting time in order to increase their vigilance behavior
(Chapter 3; Chapter 4). Meanwhile, they also must forage enough to fuel the bulk of their fawns’
rapid body growth through lactation (Carl & Robbins 1988).
The hiding strategy helps to ease these burdens by concealing fawns and reducing risk for
long periods of time, enabling mothers to approximate non-mothers in their activity budgets
(Chapter 3). Mother-fawn interactions and the high vigilance rates required for fawn risk
reduction are largely confined to brief bouts of fawn activity, although mothers still rely on
heightened vigilance to detect predators, especially prior to fawn retrieval (Chapter 3, Chapter 5).
As the hiding strategy wanes and fawn risk increases again, maternal costs drop as mothers and
fawns rely more on group membership and collective vigilance for fawn protection (Chapter 4).
In parsing out the costs and benefits of the hiding strategy to mothers, the obvious
inclination is to compare this strategy to its alternative, the following strategy. However such
comparisons are difficult due to differences in body size, habitat type, offspring number,
gregariousness, and predation pressure, among other factors, between potentially comparable
species. Existing comparative studies, including two intraspecific comparisons, are few and all
were conducted on captive populations or under conditions of extremely low infant predation risk
(O’Brien 1984; Carl & Robbins 1988; Kohlmann et al. 1996). Perhaps the best comparison is that
of Kohlmann et al. (1996) on a Nubian ibex (Capra ibex nubiana) herd in which several follower
107
young became accidentally trapped for several days in natural, rock-walled depression. For the
period of their entrapment, these kids were functional hiders and their mothers exhibited fewer
restrictions in habitat use and spent more time foraging than mothers with typical follower kids.
This suggests that hiding does in fact liberate mothers from some constraints of motherhood,
but since predation risk was apparently low or absent in this system and the “nursery” may have
secluded trapped young from predators anyway, offspring survival costs could not be compared
between the two strategies.
We see in Chapter 4 that maternal costs in Thomson’s gazelle appear to decline during the
transition out of hiding as fawns begin acting more like followers. Mothers are able to act more
like follower mothers as well by joining social groups, returning to favored habitats and reducing
the effects of the fawn on their activity budgets. However, as we have also seen, this is a time of
high mortality for fawns, which are yet unable to escape predators effectively. The seemingly
early timing of the transition and the resulting high mortality cost is explained by Byers (1997) in
pronghorn as an unavoidable situation given the theoretical decline in the effectiveness of hiding
with increasing fawn body size. It is fallacious to conclude, based on the findings in Chapter 4,
that the following strategy would incur lower costs on gazelle mothers if employed instead of
the hiding strategy: given the vulnerability of exposed fawns, the vigilance behavior necessary
to bring fawn survival probabilities to acceptable levels would likely cause significant negative
impacts on maternal fecundity.
Although the effects of species characteristics such as body size, habitat structure, and
mobility obviously favor one strategy over another in some species, in general the factors
determining the distribution of the hiding and following strategies across ungulate species have
yet to be elucidated (Lent 1974; Fisher et al. 2002). The small body size of Thomson’s gazelle
108
likely favors hiding in this species: small offspring are less able to escape larger predators through
flight and maternal defense by small-bodied mothers can be ineffective, making following a poor
strategy. Small body size also contributes to the effectiveness of hiding and enables the use of
this strategy even in the open plains habitat preferred by this species. However, as we have seen,
negative effects include a reduction in maternal mobility resulting in changes in microhabitat use
and grouping patterns, and a lack of opportunities for fawns to develop speed and motor skills
resulting in high mortality when hiding is no longer a viable option (Fitzgibbon 1990b; Byers
1997).
109
Appendix I: An attack by a warthog Phacochoerus africanus on a newborn Thomson’s gazelle
Gazella thomsonii1 Blair A. Roberts
Introduction
This note reports a previously undescribed behaviour of an attack by a warthog on a
newborn Thomson’s gazelle. Most instances of interspecific aggression in wild animals occur
in the contexts of predation (Polis et al. 1989; Kamler et al. 2007) or competition (e.g. Moore
1978; Berger 1985; Loveridge & Macdonald 2002; Schradin 2005). However, warthogs are
omnivores that are not known to prey on gazelle and only rarely include animal protein in their
diets (Cumming 1975). Also, the two species typically associate closely without overt signs of
aggression and exhibit subtle differences in diet, which minimize competition for forage (Mwangi
& Western 1998). Examination of unusual interspecific interactions such as this challenges the
scientific community to consider and evaluate alternative drivers for the behaviour, which may
then be useful in accounting for other seemingly aberrant occurrences.
Materials and methods
The incident occurred on 2 October 2011 at Ol Pejeta Conservancy (OPC) in Laikipia,
Kenya. I observed the events from a vehicle parked between 70 and 150 m from the subjects.
Results and discussion
At 1:12 PM, I spotted a female Thomson’s gazelle in labour. She delivered a fawn at 1:43
PM. Twenty-four minutes later, while the fawn was standing unsteadily after suckling and after
Work published:
Roberts, B. A. 2012. An attack by a warthog Phacochoerus africanus on a newborn Thomson’s gazelle
Gazella thomsonii. African Journal of Ecology 50, 507-508.
1
110
Figure 18.
Warthog holding
gazelle infant in
its mouth before
shaking it.
the mother had consumed all visible birth materials from the neonate and the birth site, an adult
male warthog approached the pair. When it came within several meters of the gazelles, the
mother turned to face it, leaving the fawn between her and the warthog. The warthog rushed at
the fawn, hooked it with its tusk and tossed it approximately 3 m in the air. The warthog then
turned to the mother, who first lowered her horns but quickly retreated. The warthog approached
the fawn, which had not moved since landing on the ground. It sniffed the fawn, nudging it
with its snout. It then grasped the fawn’s hindquarters in its mouth (Figure 18) and vigorously
shook the neonate from side to side for several seconds before dropping it. The fawn fell to the
ground and the warthog walked away. As the warthog left, the mother approached the fawn. The
warthog turned towards her, but then continued to walk away. The mother licked the infant once
before leaving in the opposite direction from which she had come.
At 4:32 PM, the mother returned to a distance of approximately 100 m from the fawn. At
4:53 PM, she approached and groomed the infant. The fawn stood and suckled. At 5:20 PM, the
pair moved away from the birth site. Shortly thereafter, the fawn hid in a tussock of grass and the
mother left.
There is no clear explanation for the attack. It is possible the warthog was attempting to
eat the neonate. Warthogs are reported to include animal food in their diet on rare occasions
111
(Cumming 1975), and other suids, including wild boar and bush pigs (Potamochoerus porcus), are
known to kill and eat small mammals, including ungulates (Kingdon 1979; Estes 1991). However,
the fact that the warthog did not kill or consume the fawn and its rapid loss of interest in the
neonate suggest that predation of the fawn was not the purpose of the attack.
The warthog may have sought to scavenge the birth materials. Warthogs sometimes feed on
carcasses at OPC (personal observation; N. Sharpe & K. Vanderwaal, personal communication)
and once were observed nibbling the amniotic sac of a Grant’s gazelle Sharpe & Vanderwaal,
personal communication). Although the fawn was clean and dry by the time of the attack, the
scent of birth tissues may have persisted and attracted the warthog. However, this explanation
does not account for the extreme aggression towards the fawn.
Perhaps the most plausible possibility is that the incident was a displacement activity or
a display of object aggression. Warthogs display by thrashing vegetation during aggressive
interactions with conspecifics (Kingdon 1979; Estes 1991). I did not notice the warthog before its
approach, but did note that there were other warthogs in the direction from which it had come. It
is possible that an intraspecific conflict prompted an aggressive display and that the gazelle fawn
was simply a convenient target.
Acknowledgements
I am grateful to Kim Vanderwaal and Nicole Sharpe for their observation notes. I thank the
OPC Research Department for their support and the Kenyan government for their permission
to conduct research in Kenya. This paper was made possible through the support of my advisor,
Daniel Rubenstein.
112
Appendix II: Perinatal behavior of a wild Grevy’s zebra (Equus grevyi) mare and foal1 Blair A. Roberts
Abstract
This paper describes the postpartum behavior of a wild Grevy’s zebra (Equus grevyi)
mare and foal. The foal first attempted to stand within five minutes of birth and succeeded
at approximately ten minutes after birth. Within forty minutes of birth, the infant could walk
steadily. The foal did not suckle during the observation time. An ethogram of observed behaviors
is provided.
Introduction
For many mammalian species, the behavior of mother and young immediately after birth
can strongly affect the survival of both individuals and the development of the neonate (Le Bouef
et al. 1972; Townshend & Bailey 1975; Ozoga et al. 1982; Nowak et al. 2000). It is during this
time that the first socialization of the young takes place and the mother-infant bond is formed
(Collias 1956). Rapid attainment of mobility is often crucial to the ability of the neonate to feed
or avoid predation. Risk of predation and injury tends to be elevated during the early postpartum
period due to the young’s helplessness and lack of coordination and the mother’s exhaustion from
labor (Estes & Estes 1979).
Maternal care strategy and social structure, among other factors, can affect the challenges
mothers and neonates face after parturition and the strategies they employ to overcome such
Work published:
Roberts, B. A. 2012. Perinatal behavior of a wild Grevy’s zebra (Equus grevyi) mare and foal. Journal of
Ethology 30, 205-209.
1
113
obstacles. Grevy’s zebra is an African equid species. Like all equids, Grevy’s zebra exhibits a
follower-type maternal care strategy in which the young accompanies the mother after birth
rather than concealing itself in vegetation (Walther 1965; Lent 1974). Socially, Grevy’s zebra is
similar to wild and domestic ass species in that males typically defend territories through which
females roam rather than forming harems likes those exhibited by horses and other zebra species
(Klingel 1974, 1998; Kimura 2000).
Grevy’s zebra is an endangered species with a large portion (~20%) of the remaining
individuals held in captivity in zoos and sanctuaries (Williams 2002). No account of the perinatal
behavior of this species exists from wild or captive populations. Accounts of such behavior in
other wild equid species are scarce. This is due to the rare and unpredictable nature of births and
to the secretive behavior of parturient females (Fraser 1968). Additionally, equids may be able
to delay parturition by several hours if they are disturbed in the early phases of labor (Rossdale
1968; Klingel 1969; Campitelli et al. 1982). As such, current knowledge of parturition and
perinatal behavior in equids is based largely on observations of captive individuals and domestic
species (Lickliter 1984). Descriptions of these behaviors in wild equids are therefore crucial to the
understanding of natural parturient behavior.
I present here observations of Grevy’s zebra perinatal behavior collected during an
opportunistic encounter with a parturient mother.
Study area
The observation took place between 4:00 and 5:00 pm on June 24, 2009 on Mpala Ranch
(36º52’E, 0º17’N) in Laikipia, Kenya. Parturition occurred in an area of black cotton soil and
Acacia drepanolobium woodland.
114
Materials and methods
I observed all events from a vehicle. Observations began when I spotted the mother in the
second stage of labor and ended when the adults and neonate moved out of sight into dense
bush. Observations were interrupted once for fewer than 10 minutes after the initial sighting
when the mother moved deeper into the trees, requiring that I reposition the vehicle. I took care
to ensure that the vehicle remained as far from the birth site as possible while still affording a
clear view in order to minimize disturbance to the mother and foal. The distance between the
vehicle and the zebras varied between approximately 60 and 100 meters. Times of behaviors
were obtained from timestamps of photographs taken during the observation.
Results and discussion
I spotted one adult male and two adult female Grevy’s zebras at approximately 4:15 pm.
One female was lying down when the vehicle approached but stood when the car stopped. She
was in the second phase of labor (Fraser 1968) with the hooves and legs of a foal protruding
approximately 30 cm from her vagina. The zebras moved out of sight into the acacias and I
repositioned the vehicle. At 4:24 pm I resighted the mare. She had completed parturition and
was sternally recumbent next to the foal, whose back and hindquarters were still encased in
the amnion (Figure 19). The other two group members stood some 50 m away, grazing and
periodically looking toward the mother. I estimate the birth to have occurred at approximately
4:20 pm.
The foal first attempted to stand up at 4:25 pm, but was unsuccessful and fell back to the
ground. The mare stood up at 4:28 pm and expelled the afterbirth. She did not attempt to help
the foal stand. Nor did she lick the foal or consume the afterbirth or amnion. The foal continued
115
Figure 19. Mare and neonate immediately after parturition. The amnion (indicated by arrow) is visible on
the foal’s back.
Figure 20. Foal standing successfully. The umbilicus (indicated by arrow) is still attached and visible, wrapped
around the foal’s right hind leg just above the hock.
116
Figure 21. Foal walking and dragging the placenta by the umbilicus (indicated by arrow).
Figure 22. Mare attempting to lead the foal toward the departing group members.
117
its attempts to stand and was successful at 4:34 pm (Figure 20). At 4:35 it attempted to walk but
was hindered by the attached umbilical cord and placenta. By 4:36 the foal managed to take
several steps, dragging the placenta by the umbilicus (Figure 21). At this point the male and other
female began to move away from the mare and my vehicle. The mare attempted to lead the foal
after the rest of the group by walking away from and then looking back at the infant (Figure
22). The foal followed for a few steps and then stopped. The mother periodically snorted, which
seemed to call the infant’s attention to her. The foal attempted to follow her, but was still unable
to walk steadily enough. When the mare moved beyond a certain distance from the infant, the
foal stopped moving toward its mother and began moving towards my vehicle. At this point, the
mother moved toward the foal and snorted to call it back. By 4:48, the infant’s movements had
severed the umbilical cord. The mother continued her attempts to lead the foal off and at 5:07
pm, when the infant could walk steadily, the entire group moved out of sight together. The foal
did not suckle or attempt to suckle from the mother during my observations.
Table 11 presents an ethogram of behaviors noted during this observation. All of these
behaviors except “induce following” have been reported in parturition accounts for other equid
species (Wackernagel 1965; Rossdale 1967,1968; Klingel 1969; Jeffcott 1972; Tyler 1972;
Blakeslee 1974; Boyd 1980, 1988; McCort 1980; Campitelli et al. 1982; Waring 1982). The
timing of these behaviors in Grevy’s zebra is comparable to that reported for other equid species.
The mare and foal stood, the foal first attempted to stand, and the placenta was expelled within
or near the range of times reported for other equid species. However, the umbilical cord broke
later than in other equid species, and the foal attempted to walk and succeeded earlier.
The umbilicus was severed 28 minutes postpartum (mpp) compared to 8-15 mpp in other
equid species (Rossdale 1967, 1968; Klingel 1969; Blakeslee 1974; Boyd 1988). The cord broke
118
Behavior
Mare behaviors
Stand
Expel placenta
Induce following
Foal behaviors
Attempt to stand
Stand
Attempt to walk
Walk
Follow
Other behaviors
Sever umbilicus
Description
Time (min
after birth)
Mother stands after parturition
The placenta is expelled from the mother’s vagina
Mother encourages foal to follow as she walks away. May include snorting,
glancing back at foal, and returning to foal if it ceases to follow
8
8
16
Foal attempts to rise from a recumbent position but falls before standing
Foal rises from a recumbent position and stands still on all four legs for at
least one second
Foal attempts to walk but falls before taking one step with each foot
Foal takes one step with each foot without falling down
Foal walks towards the mare as she walks away from the foal
5
14
Umbilical cord ruptures (usually due to the mare standing up or the foal’s
movements)
28
15
16
Not
recorded
Table 11. Ethogram of postnatal behaviors of Grevy’s zebra mares and neonates and the time at which the
behaviors occurred in this observation.
later because it remained intact after the mare stood up whereas this action broke the cord in
most other accounts. The apparent precocity of the Grevy’s zebra foal in attempting to walk (at
15 mpp compared to 20-40 mpp in other equids; Boyd 1980, 1988) and succeeding (at 47 mpp,
compared to 68-74 mpp in other equids; Waring 1982) is likely a product of a lack of data. These
behaviors are rarely reported for other equids, thus the timing recorded here may prove to be
within the normal range of equids if more data are collected.
In addition to these disparities, it is atypical in equids for the mare to not isolate for birth
and to not lick the foal. Mothers of most equid species isolate before birth in order to avoid
harassment of the neonate by conspecifics, which can cause injury to the foal or the disruption
of the mother-infant bond necessary for parental care (Tyler 1972; Blakeslee 1974; Boyd
1980, 1988; McCort 1980; Sundaresan et al. 2007; Fischhoff et al. 2010). However, neither
accompanying zebra approached the mare or her foal during my observation. Possibly Grevy’s
zebras do not isolate because conspecifics do not tend to interfere during the perinatal period.
119
Although equid mothers usually do not consume the placenta and birth tissues, most do
engage in intense and prolonged licking of the foal (Jeffcott 1972; Tyler 1972; Blakeslee 1974;
Boyd 1980, 1988; McCort 1980). However, the Grevy’s zebra mare did not lick her foal and
plains zebras lick only superficially if at all (Wackernagel 1965; Klingel 1969). Licking is proposed
to stimulate the neonate to move and encourage maternal behavior, bonding, and recognition
of the foal through olfactory and gustatory stimuli present in the birth tissues (Lent 1974; Levy
& Poindron 1987). Why licking should be prominent in some equid species but not others is not
clear.
Finally, the timing of the birth in the late afternoon appears maladaptive (Tyler 1972;
Estes & Estes 1979). Morning is thought to be the ideal time for ungulates to give birth as this
allows maximum time for the mother and young to bond and for the neonate to become steady
on its feet before nightfall (Tyler 1972). Morning parturition seems especially advantageous for
free-ranging African ungulates given that most major African predators are nocturnal (Estes &
Estes 1979). Many of these species, including lions, hyenas, wild dogs, and leopards are present
at Mpala Ranch and pose a threat to Grevy’s zebra foals. It seems plausible, therefore, that the
timing of the observed birth may put the foal at an increased risk of predation.
Births are rare, rapid events that are difficult to predict and observe in wild animals.
Therefore, bodies of data on parturition and perinatal behavior in natural settings must be
obtained piecemeal through fortuitous observations such as the one reported here (Fraser 1968).
Additional instances of Grevy’s zebra perinatal behavior must be documented before any
conclusions can be drawn about the behavioral patterns of parturient mothers and neonates of
this species. It is hoped that the ethogram provided here can be expanded and used in future
observations to build a consistent database of Grevy’s zebra and equid perinatal behavior.
120
Acknowledgements
I thank the Ministry of Education, Government of Kenya for allowing me to work in
Kenya, and Mpala Ranch and its administration and staff for permitting me to work on the
premises and for supporting me in the field. I also wish to thank John Kitoe for his assistance
and guidance in the field, Elaine Bigelow for sharing her photos for reference for this paper,
and Cassandra Nuñez, Siva Sundaresan, and Michael Costelloe for their advice and edits. This
paper was made possible by funding from Princeton University’s Department of Ecology and
Evolutionary Biology and from my graduate advisor, Dr. Daniel Rubenstein.
121
References
Alados, C. L. & Escos, J. (1988). Parturition dates and mother-kid behavior in Spanish ibex (Capra
pyrenaica) in Spain. Journal of Mammalogy 69, 172-175.
Altendorf, K. B., Laundré, J. W., Lopez González, C. A. & Brown, J. S. (2001). Assessing effects
of predation risk on foraging behavior of mule deer. Journal of Mammalogy 82, 430-439.
Altmann, J. (1974). Observational study of behavior: sampling methods. Behaviour 49, 227-267.
Autenrieth, R. E. & Fichter, E. (1975). On the behavior and socialization of pronghorn fawns.
Wildlife Monographs 42, 3-111.
Barrett, M. W. (1978). Pronghorn fawn mortality in Alberta. Proceedings of the Pronghorn
Antelope Workshop 8, 429-444.
Bell, G. (1980). The costs of reproduction and their consequences. The American Naturalist 116,
45-76.
Benton, T. G., Grant, A. & Clutton-Brock, T. H. (1995). Does environmental stochasticity matter?
Analysis of red deer life-histories on Rum. Evolutionary Ecology 9, 559-574.
Bercovitch, F. B., Loomis, C. P. & Rieches, R. G. (2009). Age-specific changes in reproductive
effort and terminal investment in female Nile lechwe. Journal of Mammalogy 90, 40-46.
Bercovitch, F. B., Widdig, A. & Nürnberg, P. (2000). Maternal investment in rhesus macaques
(Macaca mulatta): reproductive costs and consequences of raising sons. Behavioral Ecology and
Sociobiology 48, 1-11.
Berger, J. (1978). Maternal defensive behavior in bighorn sheep. Journal of Mammalogy 59, 620621.
122
Berger, J. (1979). “Predator harassment” as a defensive strategy in ungulates. American Midland
Naturalist 102, 197-199.
Berger, J. (1985). Interspecific interactions and dominance among wild Great Basin ungulates.
Journal of Mammalogy 66, 571-573.
Bergerud, A. T., Butler, H. E. & Miller, D. R. (1984). Antipredator tactics of calving caribou:
dispersion in mountains. Canadian Journal of Zoology 62, 1566-1575.
Bildstein, K. L. (1983). Why white-tailed deer flag their tails. The American Naturalist 121, 709715.
Bitterlich, W. (1948). Die Winkelzahlprobe. Allgemeine Forst-und-Holzwirtschaftliche Zeitung 59,
4-5.
Blakeslee, J. (1974). Mother-young relationships and related behavior among free-ranging
Appaloosa horses. M.S. Thesis, Idaho State University.
Bleich, V. C. (1999). Mountain sheep and coyotes: patterns of predator evasion in a mountain
ungulate. Journal of Mammalogy 80, 283-289.
Borner, M., Fitzgibbon, C. D., Borner, M., Caro, T. M., Lindsay, W. K., Collins, D. A. & Holt, M.
E. (1987). The decline of the Serengeti Thomson’s gazelle populations. Oecologia 73, 32-40.
Boyd, L. E. (1980). The natality, foal survivorship, and mare-foal behavior of feral horses in
Wyoming’s Red Desert. M.S. Thesis, University of Wyoming.
Boyd, L. E. (1988). Ontogeny of behavior in Przewalski horses. Applied Animal Behaviour
Science 21, 41-69.
Brooks, A. C. (1961). A study of the Thomson’s gazelle (Gazella thomsoni gunther) in Tanganyika.
Colonial Research Publications 25.
123
Burger, J. & Gochfeld, M. (1994). Vigilance in African mammals: differences among mothers,
other females, and males. Behaviour 131, 153-169.
Buuveibaatar, B., Young, J. K., Berger, J., Fine, A. E., Lkhagvasuren, B., Zahler, P. & Fuller, T.
K. (2013). Factors affecting survival and cause-specific mortality of saiga calves in Mongolia.
Journal of Mammalogy, 127-136.
Byers, J. A. (1997). American pronghorn: social adaptations and the ghosts of predators past.
University of Chicago Press, Chicago.
Byers, J. A. & Byers, K. Z. (1983). Do pronghorn mothers reveal the locations of their hidden
fawns? Behavioral Ecology and Sociobiology 13, 147-156.
Campitelli, S., Carenzi, C. & Verga, M. (1982). Factors which influence parturition in the mare
and development of the foal. Applied Animal Ethology 9, 7-14.
Carl, G. R. & Robbins, C. T. (1988). The energetic cost of predator avoidance in neonate
ungulates: hiding versus following. Canadian Journal of Zoology 66, 239-246.
Caro, T. M. (1986a). The functions of stotting in Thomson’s gazelles: some tests of the
predictions. Animal Behaviour 34, 663-684.
Caro, T. M. (1986b). The functions of stotting: a review of the hypotheses. Animal Behaviour 34,
649-662.
Caro, T. M. (1994). Ungulate antipredator behaviour: preliminary and comparative data from
African bovids. Behaviour 128, 189-228.
Caro, T. M. & Fitzgibbon, C. D. (1992). Large carnivores and their prey: the quick and the dead.
In: Natural Enemies: The Population Biology of Predators, Parasites and Diseases (M. J.
Crawley, ed). Blackwell Scientific Publications, Oxford, p. 117-142.
124
Caro, T. M., Lombardo, L., Goldizen, A. W. & Kelly, M. (1995). Tail-flagging and other
antipredator signals in white-tailed deer: new data and synthesis. Behavioral Ecology 6, 442450.
Case, T. J. (1978). On the evolution and adaptive significance of postnatal growth rates in the
terrestrial vertebrates. Quarterly Review of Biology 53, 243-282.
Caughley, G. (1966). Mortality patterns in mammals. Ecology 47, 906-918.
Choquenot, D., Lukins, B. & Curran, G. (1997). Assessing lamb predation by feral pigs in
Australia’s semi-arid rangelands. Journal of Applied Ecology 34, 1445-1454.
Ciuti, S., Pipia, A., Grignolio, S., Ghiandai, F. & Apollonio, M. (2009). Space use, habitat
selection and activity patterns of female Sardinian mouflon (Ovis orientalis musimon) during the
lambing season. European Journal of Wildlife Research 55, 589-595.
Clutton-Brock, T. H. (1984). Reproductive effort and terminal investment in iteroparous animals.
The American Naturalist 123, 212-229.
Clutton-Brock, T. H. (1991). The Evolution of Parental Care. Princeton University Press,
Princeton.
Clutton-Brock, T. H., Albon, S. D. & Guinness, F. E. (1981). Parental investment in male and
female offspring in polygnous mammals. Nature 289, 487-489.
Clutton-Brock, T. H., Albon, S. D. & Guinness, F. E. (1989). Fitness costs of gestation and
lactation in wild mammals. Nature 337, 260-262.
Clutton-Brock, T. H., Guinness, F. E. & Albon, S. D. (1982). Red Deer: Behavior and Ecology of
Two Sexes. University of Chicago Press, Chicago.
125
Coleman, R. M. & Fischer, R. U. (2010). Brood size, male fanning effort and the energetics
of a nonshareable parental investment in bluegill sunfish, Lepomis macrochirus (Teleostei:
Centrarchidae). Ethology 87, 177-188.
Collias, N. E. (1956). The analysis of socialization in sheep and goats. Ecology 37, 228-239.
Cook, J. G., Johnson, B. K., Cook, R. C., Riggs, R. A., Delcurto, T., Bryant, L. D. & Irwin, L. L.
(2004). Effects of summer-autumn nutrition and parturition date on reproduction and survival
of elk. Wildlife Monographs 155, 1-61.
Côté, S. D., Peracino, A. & Simard, G. (1997). Wolf, Canis lupus, predation and maternal defensive
behavior in mountain goats, Oreamnos americanus. Canadian Field-Naturalist 111, 389-392.
Creel, S., Winnie Jr., J., Maxwell, B., Hamlin, K. & Creel, M. (2005). Elk alter habitat selection as
an antipredator response to wolves. Ecology 86, 3387-3397.
Cumming, D. H. M. (1975). A field study of the ecology and behaviour of warthog. Museum
Memoirs, National Museums and Monuments of Rhodesia 7, 1-179.
Daly, M. (1990). Evolutionary theory and parental motives. In: Mammalian Parenting (N. A.
Krasnegor & R. S. Bridges, eds). Oxford University Press, Oxford, p. 25-39.
Deevey, E. S. (1947). Life tables for natural populations of animals. The Quarterly Review of
Biology 22, 283-314.
Devivo, M. T., Cottrell, W. O., Deberti, J. M., Duchamp, J. E., Heffernan, L. M., Kougher, J. D.
& Larkin, J. L. (2011). Survival and cause-specific mortality of elk Cervus canadensis calves in a
predator rich environment. Wildlife Biology 17, 156-165.
Dirienzo, N., Pruitt, J. N. & Hedrick, A. V. (2013). The combined behavioural tendencies of
predator and prey mediate the outcome of their interaction. Animal Behaviour 86, 317-322.
126
Dittrich, L. (1967). Breeding Kirk’s dik-dik Madoqua kirki thomasi at Hanover Zoo. International
Zoo Yearbook 7, 171-173.
Drent, R. H. & Daan, S. (1980). The prudent parent: energetic adjustments in avian breeding.
Ardea 68, 225-252.
Dugatkin, L. A. (1991). Tendency to inspect predators predicts mortality risk in the guppy (Poecilia
reticulata). Behavioral Ecology 3, 124-127.
Eaton, R. L. (1970a). Hunting behavior of the cheetah. The Journal of Wildlife Management 34,
56-67.
Eaton, R. L. (1970b). The predator sequence, with emphasis on killing behavior and its ontogeny
in the cheetah (Acinonyx jubatus Schreber). Zeitschrift für Tierpsychologie 27, 492-504.
Eby, S. & Ritchie, M. E. (2012). The impacts of burning on Thomson’s gazelles’, Gazella thomsonii,
vigilance in Serengeti National Park, Tanzania. African Journal of Ecology 51, 337-342.
Estes, R. D. (1967). The comparative behavior of Grant’s and Thomson’s gazelles. Journal of
Mammalogy 48, 189-209.
Estes, R. D. (1991). The Behavior Guide to African Mammals. University of California Press,
Berkeley.
Estes, R. D. & Estes, R. K. (1979). The birth and survival of wildebeest calves. Zeitschrift für
Tierpsychologie 50, 45-95.
Estes, R. D. & Goddard, J. (1967). Prey selection and hunting behavior of the African wild dog.
The Journal of Wildlife Management 31, 52-70.
Festa-Bianchet, M. (1988). Seasonal range selection in bighorn sheep: conflicts between forage
quality, forage quantity, and predator avoidance. Oecologia 75, 580-586.
127
Festa-Bianchet, M. & Jorgenson, J. T. (1998). Selfish mothers: reproductive expenditure and
resource availability in bighorn ewes. Behavioral Ecology 9, 144-150.
Fischhoff, I. R., Sundaresan, S. R., Cordingley, J. & Rubenstein, D. I. (2007). Habitat use
and movements of plains zebra (Equus burchelli) in response to predation danger from lions.
Behavioral Ecology 18, 725-729.
Fischhoff, I. R., Sundaresan, S. R., Larkin, H. M., Sellier, M.-J., Cordingley, J. E. & Rubenstein,
D. I. (2010). A rare fight in female plans zebra. Journal of Ethology 28, 201-205.
Fisher, D. O., Blomberg, S. P. & Owens, I. P. F. (2002). Convergent maternal care strategies in
ungulates and macropods. Evolution 56, 167-176.
Fitzgibbon, C. D. (1988). The antipredator behaviour of Thomson’s gazelles. Ph.D. Thesis,
University of Cambridge.
Fitzgibbon, C. D. (1990a). Anti-predator strategies of immature Thomson’s gazelles: hiding and
the prone response. Animal Behaviour 40, 846-855.
Fitzgibbon, C. D. (1990b). Why do hunting cheetahs prefer male gazelles? Animal Behaviour 40,
837-845.
Fitzgibbon, C. D. (1993). Antipredator strategies of female Thomson’s gazelles with hidden
fawns. Journal of Mammalogy 74, 758-762.
Fitzgibbon, C. D. (1994). The costs and benefits of predator inspection behaviour in Thomson’s
gazelles. Behavioral Ecology and Sociobiology 34, 139-148.
Fitzgibbon, C. D. & Fanshawe, J. H. (1988). Stotting in Thomson’s gazelles: an honest signal of
condition. Behavioral Ecology and Sociobiology 23, 69-74.
128
Fitzgibbon, C. D. & Fanshawe, J. H. (1989). The condition and age of Thomson’s gazelles killed
by cheetahs and wild dogs. Journal of the Zoological Society of London 218, 99-107.
Fitzgibbon, C. D. & Lazarus, J. (1995). Antipredator behavior of Serengeti ungulates: individual
differences and population consequences. In: Serenget II: Dynamics, Management and
Conservation of an Ecosystem (A. R. E. Sinclair & P. Arcese, eds). University of Chicago Press,
Chicago, p. 274-296.
Fortin, D., Fortin, M.-E., Beyer, H. L., Duchesne, T., Courant, S. & Dancose, K. (2009). Groupsize-mediated habitat selection and group fusion-fission dynamics of bison under predation
risk. Ecology 90.
Fraser, A. F. (1968). Reproductive behaviour in ungulates. Academic Press, London.
Gaillard, J.-M., Festa-Bianchet, M. & Yoccoz, N. G. (1998). Population dynamics of large
herbivores: variable recruitment with constant adult survival. Trends in Ecology and Evolution
13, 58-63.
Gaillard, J.-M., Festa-Bianchet, M., Yoccoz, N. G., Loison, A. & Toïgo, C. (2000). Temporal
variation in fitness components and population dynamics of large herbivores. Annual Review
of Ecological Systems 31, 267-393.
Gese, E. M. (1999). Threat of predation: do ungulates behave aggressively towards different
members of a coyote pack? Canadian Journal of Zoology 77, 499-503.
Gese, E. M. & Grothe, S. (1995). Analysis of coyote predation on deer and elk during winter in
Yellowstone National Park, Wyoming. American Midland Naturalist 133, 36-43.
Godin, J.-G. J. & Smith, S. A. (1988). A fitness cost of foraging in the guppy. Nature 333, 69-71.
129
Gosling, L. M. (1969). Parturition and related behaviour in Coke’s hartebeest, Alcelaphus buselaphus
cokei Günther. Journal of Reproduction and Fertility 6, 265-286.
Goubault, M., Scott, D. & Hardy, I. C. W. (2007). The importance of offspring value: maternal
defence in parasitoid contests. Animal Behaviour 74, 437-446.
Gray, D. R. (1974). The defence formation of the musk-ox. Musk-Ox 14, 25-29.
Groves, C. & Grubb, P. (2011). Ungulate Taxonomy. The Johns Hopkins University Press,
Baltimore, Maryland.
Guinness, F. E., Hall, M. J. & Cockerill, R. A. (1979). Mother-offspring association in red deer
(Cervus elaphus L.) on Rhum. Animal Behaviour 27, 536-544.
Gwynne, M. D. & Bell, R. H. V. (1968). Selection of vegetation components by grazing ungulates
in the Serengeti National Park. Nature 220, 390-393.
Hamel, S. & Côté, S. D. (2007). Habitat use patterns in relation to escape terrain: are alpine
ungulate females trading off better foraging sites for safety? Canadian Journal of Zoology 85,
933-943.
Hamel, S. & Côté, S. D. (2008). Trade-offs in activity budget in an alpine ungulate: contrasting
lactating and nonlactating females. Animal Behaviour 75, 217-227.
Hamel, S. & Côté, S. D. (2009). Maternal defensive behavior of mountain goats against
predation by golden eagles. Western North American Naturalist 69, 115-118.
Hamilton, W. D. (1971). Geometry for the selfish herd. Journal of Theoretical Biology 31, 295311.
Harding, R. S. O. (1973). Predation by a troop of olive baboons (Papio anubis). American Journal
of Physical Anthropology 38, 587-591.
130
Hayward, M. W. (2006). Prey preferences of the spotted hyaena (Crocuta crocuta) and degree of
dietary overlap with the lion (Panthera leo). Journal of Zoology 270, 606-614.
Hayward, M. W., Henschel, P., O’Brien, J., Hofmeyr, M., Balme, G. & Kerley, G. I. H. (2006a).
Prey preferences of the leopard (Panthera pardus). Journal of Zoology 270, 298-313.
Hayward, M. W., Hofmeyr, M., O’Brien, J. & Kerley, G. I. H. (2006b). Prey preferences of the
cheetah (Acinonyx jubatus) (Felidae: Carnivora): morphological limitations or the need to capture
rapidly consumable prey before kleptoparasites arrive? Journal of Zoology 270, 615-627.
Hayward, M. W. & Kerley, G. I. H. (2005). Prey preferences of the lion (Panthera leo). Journal of
the Zoological Society of London 267, 309-322.
Hayward, M. W., O’Brien, J., Hofmeyr, M. & Kerley, G. I. H. (2006c). Prey preferences of the
African wild dog (Lycaon pictus) (Canidae: Carnivora): ecological requirements for conservation.
Journal of Mammalogy 87, 1122-1131.
Hebblewhite, M. & Pletscher, D. H. (2002). Effects of elk group size on predation by wolves.
Canadian Journal of Zoology 80, 800-809.
Heinze, J. & Schrempf, A. (2012). Terminal investment: individual reproduction of ant queens
increases with age. PLOS ONE 7, e35201.
Hester, F. J. (1964). Effects of food supply on fecundity in the female guppy, Lebistes reticulatus
(Peters). Journal of the Fisheries Research Board of Canada 21, 757-764.
Hirst, S. M. (1975). Ungulate-habitat relationships in a South African woodland/savanna
ecosystem. Wildlife Monographs 44, 3-60.
Hirth, D. H. & Mccullough, D. R. (1977). Evolution of alarm signals in ungulates with special
reference to white-tailed deer. The American Naturalist 111, 31-42.
131
Hoffman, C. L., Higham, J. P., Mas-Rivera, A., Ayala, J. E. & Maestripieri, D. (2010). Terminal
investment and senscence in rhesus macaques (Macaca mulatta) on Cayo Santiago. Behavioral
Ecology 21, 972-978.
Howland, H. C. (1974). Optimal strategies for predator avoidance: the relative importance of
speed and manoeuverability. Journal of Theoretical Biology 47, 333-350.
Hunter, L. T. B. & Skinner, J. D. (1998). Vigilance behaviour in African ungulates: the role of
predation pressure. Behaviour 135, 195-211.
Hvideberg-Hansen, H. (1970). Contribution to the knowledge of the reproductive physiology of
the Thomson’s gazelle (Gazella thomsoni Günther). Mammalia 34, 551-563.
Jacobsen, N. K. (1979). Alarm bradycardia in white-tailed deer fawns (Odocoileus virginianus).
Journal of Mammalogy 60, 343-349.
Janermo, A. (2004). Neonatal mortality in roe deer. Ph.D. Thesis, Swedish University of
Agricultural Sciences.
Janermo, A., Liberg, O., Lockowandt, S., Olsson, A. & Wahlström, K. (2004). Predation by red
fox on European roe deer fawns in relation to age, sex and birthdate. Canadian Journal of
Zoology 82, 416-422.
Jarman, M. V. (1976). Impala social behaviour: birth behaviour. East African Wildlife Journal 14,
153-167.
Jarman, P. J. & Sinclair, A. R. E. (1979). Feeding strategy and the pattern of resource partitioning
in ungulates. In: Serengeti: Dynamics of an Ecosystem (A. R. E. Sinclair & M. NortonGriffiths, eds). University of Chicago Press, Chicago.
132
Jeffcott, L. B. (1972). Observations on parturition in crossbred pony mares. Equine Veterinary
Journal 4, 209-214.
Kamler, J. F., Davis-Mostert, H. T., Hunter, L. & Macdonald, D. W. (2007). Predation on blackbacked jackals (Canis mesomelas) by African wild dogs (Lycaon pictus). African Journal of Ecology
45, 667-668.
Kamler, J. F., Fogt, J. L. & Collins, K. (2010). Single black-backed jackal (Canis mesomelas) kills
adult impala (Aepyceros melampus). African Journal of Ecology 48, 847-848.
Kerley, L. L. (2013). First documented predation of a sika deer (Cervus nippon) by golden eagle
(Aquila chrysaetos) in Russian Far East. Journal of Raptor Research 47, 328-330.
Kimura, R. (2000). Relationship of the type of social organization to scent-marking and mutualgrooming behaviour in Grevy’s (Equus grevyi) and Grant’s zebras (Equus burchelli bohmi). Journal
of Equine Science 11, 91-98.
Kingdon, J. (1979). East African Mammals: An Atlas of Evolution in Africa. Vol 3B (Large
Mammals). Academic Press, London.
Kingdon, J. (1982). East African Mammals: An Atlas of Evolution in Africa. Vol 3C and 3D
(Bovids). University of Chicago Press, Chicago.
Kivanova, A., Horakova, D. & Exnerova, A. (2011). Nest defence intensity in house sparrows
Passer domesticus in relation to parental quality and brood value. Acta Ornithologica.
Klare, U., Kamler, J. F., Stenkewitz, U. & Macdonald, D. W. (2010). Diet, prey selection and
predation impact of black-backed jackals in South Africa. Journal of Wildlife Management 74,
1030-1042.
133
Klingel, H. (1969). Reproduction in the plains zebra, Equus burchelli boehmi: behaviour and
ecological factors. Journal of Reproduction and Fertility Supplement 6, 339-345.
Klingel, H. (1974). A comparison of the social behaviour of the Equidae. In: The Behaviour
of Ungulates and its Relation to Management (V. Geist & F. Walther, eds). IUCN, Morges,
Switzerland, p. 124-132.
Klingel, H. (1998). Observations on social organization and behaviour of African and Asiatic
wild asses (Equus africanus and Equus hemionus). Applied Animal Behaviour Science 60, 103-113.
Kohlmann, S. G., Müller, D. M. & Alkon, P. U. (1996). Antipredator constraints on lactating
Nubian ibexes. Journal of Mammalogy 77, 1122-1131.
Koskela, E., Juutistenaho, P., Mappes, T. & Oksanen, T. A. (2000). Offspring defence in relation
to litter size and age: experiment in the bank vole Clethrionomys glareolus. Evolutionary Ecology
14, 99-109.
Krause, J. & Godin, J.-G. J. (1996). Influence of prey foraging posture on flight behavior and
predation risk: predators take advantage of unwary prey. Behavioral Ecology 7, 264-271.
Krofel, M. (2007). Opportunistic hunting behaviour of black-backed jackals in Namibia. African
Journal of Ecology 46, 220-222.
Kruuk, H. (1972). The Spotted Hyena: A Study of Predation and Social Behavior. University of
Chicago Press, Chicago.
Kruuk, H. & Turner, M. (1967). Comparative notes on predation by lion, leopard, cheetah and
wild dog in the Serengeti area, East Africa. Mammalia 31, 1-27.
Kullberg, C., Jakobsson, S. & Fransson, T. (1998). Predator-induced take-off strategy in great tits
(Parus major). Proceedings of the Royal Society of London B 265, 1659-1664.
134
Lagory, K. E. (1987). The influence of habitat and group characteristics on the alarm and flight
response of white-tailed deer. Animal Behaviour 35, 20-25.
Landeau, L. & Terborgh, J. (1986). Oddity and the ‘confusion effect’ in predation. Animal
Behaviour 34, 1372-1380.
Laundré, J. W., Hernández, L. & Altendorf, K. B. (2001). Wolves, elk and bison: reestablishing
the “landscape of fear” in Yellowstone National Park, USA. Canadian Journal of Zoology 79,
1401-1409.
Laundré, J. W., Hernández, L. & Ripple, W. J. (2010). The landscape of fear: ecological
implications of being afraid. The Open Ecology Journal 3, 1-7.
Le Bouef, B. J., Whiting, R. J. & Gantt, R. F. (1972). Perinatal behavior of Northern elephant seal
females and their young. Behaviour 43, 121-156.
Lent, P. C. (1974). Mother-infant relationships in ungulates. In: The Behaviour of Ungulates and
its Relation to Management (V. Geist & F. Walther, eds). IUCN, Morges, p. 14-55.
Lent, P. C. (1999). Muskoxen and Their Hunters: a History. University of Oklahoma Press,
Norman, Oklahoma.
Leus, K. & Macdonald, A. A. (1997). From babirusa (Babyrousa babyrussa) to domestic pig: the
nutrition of swine. Proceedings of the Nutrition Society 56, 1001-1012.
Leuthold, W. (1972). Gestation period in Thomson’s gazelle. East African Wildlife Journal 1972,
309-310.
Leuthold, W. (1977). African Ungulates: a Comparative Review of their Ethology and Behavioral
Ecology. Springer-Verlag, Berlin.
135
Levy, F. & Poindron, P. (1987). The importance of amniotic fluids for the establishment of
maternal behaviour in experienced and inexperienced ewes. Animal Behaviour 35, 1188-1192.
Lickliter, R. E. (1984). Behavior associated with parturition in the domestic goat. Applied Animal
Behaviour Science 13, 335-345.
Lima, S. L. (1995). Back to the basics of anti-predatory vigilance: the group size effect. Animal
Behaviour 49, 11-20.
Lima, S. L. (1998). Nonlethal effects in the ecology of predator-prey interactions. BioScience 48,
25-34.
Lima, S. L. & Bednekoff, P. A. (1999). Temporal variation in danger drives antipredator behavior:
the predation risk allocation hypothesis. The American Naturalist 6, 649-659.
Lima, S. L. & Dill, L. M. (1990). Behavioral decisions made under the risk of predation: a review
and prospectus. Canadian Journal of Zoology 68.
Lingle, S. (2001). Anti-predator strategies and grouping patterns in white-tailed deer and mule
deer. Ethology 107, 295-314.
Lingle, S. & Pellis, S. M. (2002). Fight or flight? Antipredator behavior and the escalation of
coyote encounters with deer. Oecologia 131, 154-164.
Lingle, S., Pellis, S. M. & Wilson, W. F. (2005). Interspecific variation in antipredator behaviour
leads to differential vulnerability of mule deer and white-tailed deer fawns early in life. Journal
of Animal Ecology 74, 1140-1149.
Linnell, J. D. C., Aanes, R. & Andersen, R. (1995). Who killed Bambi? The role of predation in
the neonatal mortality of temperate ungulates. Wildlife Biology 1, 209-223.
136
Lipetz, V. E. & Bekoff, M. (1980). Possible functions of predator harassment in pronghorn
antelopes. Journal of Mammalogy 61, 741-743.
Lomas, L. A. & Bender, L. C. (2007). Survival and cause-specific mortality of neonatal mule deer
fawns, North-Central New Mexico. The Journal of Wildlife Management 71, 884-894.
Lott, D. F. & Galland, J. C. (1985). Parturition in American bison: precocity and systematic
variation in cow isolation. Zeitschrift für Tierpsychologie 69, 66-71.
Loudon, A. S. I. (1985). Lactation and neonatal survival of mammals. Symposia of the
Zoological Society of London 54, 183-207.
Loveridge, A. J. & Macdonald, D. W. (2002). Habitat ecology of two sympatric species of jackal
in Zimbabwe. Journal of Mammalogy 83, 599-607.
Loveridge, A. J. & Nel, J. A. (2004). Black-backed jackal. In: Canids: Foxes, Wolves, Jackals and
Dogs. Status Survey and Conservation Action Plan (C. Sillero-Zubiri, M. Hoffman & D.
W. Macdonald, eds). International Union for the Conservation of Nature/Species Survival
Commission, Canid Specialist Group, Cambridge, United Kingdom, p. 161-166.
Macdonald, D. W., Loveridge, A. H. & Atkinson, R. P. D. (2004). A comparative study of sidestriped jackals in Zimbabwe: the influence of habitat and congeners. In: The Biology and
Conservation of Wild Canids (D. W. Macdonald & C. Sillero-Zubiri, eds). Oxford University
Press, Oxford, United Kingdom.
Manski, D. A. (1991). Reproductive behavior of addax antelope. Applied Animal Behaviour
Science 29, 39-66.
137
Mao, J. S., Boyce, M. S., Smith, D. W., Singer, F. J., Vales, D. J., Vore, J. M. & Merrill, E. H.
(2005). Habitat selection by elk before and after wolf reintroduction in Yellowstone National
Park. Journal of Wildlife Management 69, 1691-1707.
McCort, W. D. (1980). The behavior and social organization of feral asses (Equus asinus) on
Ossabaw Island, Georgia. Ph.D. Thesis. Pennsylvania State University.
McLoughlin, P. D., Dunford, J. S. & Boutin, S. (2005). Relating predation mortality to broad-scale
habitat selection. Journal of Animal Ecology 74, 701-707.
Mduma, S. a. R. & Sinclair, A. R. E. (1994). The function of habitat selection by oribi in
Serengeti, Tanzania. African Journal of Ecology 32, 16-29.
Mloszewski, M. J. (1983). The behavior and ecology of the African buffalo. Cambridge
University Press, Cambridge, UK.
Moore, F. R. (1978). Interspecific aggression: toward whom should a mockingbird be aggressive?
Behavioral Ecology and Sociobiology 3, 173-176.
Mooring, M. S. & Hart, B. L. (1997). Reciprocal allogrooming in wild impala lambs. Ethology
103, 665-680.
Mwangi, E. M. & Western, D. (1998). Habitat selection by large herbivores in lake Nakuru
National Park, Kenya. Biodiversity Conservation 7, 1-8.
Namgail, T., Fox, J. L. & Bhatnagar, Y. V. (2004). Habitat segregation between sympatric Tibetan
argali Ovis ammon hodgsoni and blue sheep Pseudois nayaur in the Indian Trans-Himalaya. Journal
of Zoology 262, 57-63.
138
Novaro, A. J., Moraga, C. A., Briceño, C., Funes, M. C. & Marino, A. (2009). First records
of culpeo (Lycalopex culpaeus) attacks and cooperative defense by guanacos (Lama guanicoe).
Mammalia 73, 148-150.
Nowak, R., Porter, R. H., Lévy, F. & Orgeur, P. (2000). Role of mother-young interactions in the
survival of offspring in domestic mammals. Reviews of Reproduction 5, 153-163.
O’Brien, P. H. (1984). Leavers and stayers: maternal postpartum strategies in feral goats. Applied
Animal Behaviour Science 12, 233-243.
Olson, K. A., Fuller, T. K., Schaller, G. B., Lhagvasuren, B. & Odonkhuu, D. (2005).
Reproduction, neonatal weights, and first-year survival of Mongolian gazelles (Procapra
gutturosa). Journal of Zoology 265, 227-233.
Ozoga, J. J. & Verme, L. J. (1986). Relation of maternal age to fawn-rearing success in whitetailed deer. The Journal of Wildlife Management 50, 480-486.
Ozoga, J. J., Verme, L. J. & Bienz, C. S. (1982). Parturition behavior and territoriality in whitetailed deer: impact on neonatal mortality. The Journal of Wildlife Management 46, 1-11.
Pagan, O., Von Houwald, F., Wenker, C. & Steck, B. L. (2009). Husbandry and breeding of
Somali wild ass Equus africanus somalicus at Basel Zoo, Switzerland. International Zoo Yearbook
43, 198-211.
Parker, K. L., Robbins, C. T. & Hanley, T. A. (1984). Energy expenditures for locomotion by mule
deer and elk. The Journal of Wildlife Management 48, 474-488.
Pavlov, P. M. & Hone, J. (1982). The behaviour of feral pigs, Sus scrofa, in flocks of lambing ewes.
Australian Wildlife Research 9, 101-109.
139
Peckarsky, B. L., Cowan, C. A., Penton, M. A. & Anderson, C. (1993). Sublethal consequences of
stream-dwelling predatory stoneflies on mayfly growth and fecundity. Ecology 74, 1836-1846.
Pinard, V., Dussault, C., Ouellet, J.-P., Fortin, D. & Courtois, R. (2012). Calving rate, calf survival
rate, and habitat selection of forest-dwelling caribou in a highly managed landscape. The
Journal of Wildlife Management 76, 189-199.
Pipia, A., Ciuti, S., Grignolio, S., Luchetti, S., Madau, R. & Apollonio, M. (2009). Effect of
predation risk on grouping pattern and whistling behaviour in a wild mouflon Ovis aries
population. Acta Theriologica 54, 77-86.
Pitcher, T. J. (1986). Functions of shoaling behaviour in teleosts. In: The Behaviour of Teleost
Fishes (T. J. Pitcher, ed). Croon Helm, Beckenham, p. 294-337.
Polis, G. A., Myers, C. A. & Holt, R. D. (1989). The ecology and evolution of intraguild
predation: potential competitors that eat each other. Annual Review of Ecology and
Systematics 20, 297-330.
Poole, K. G., Serrouya, R. & Stuart-Smith, K. (2007). Moose calving strategies in interior
montane ecosystems. Journal of Mammalogy 88, 139-150.
Pugesek, B. H. (1983). The relationship between parental age and reproductive effort in the
California gull (Larus californicus). Behavioral Ecology 13, 161-171.
Quinn, J. L. & Cresswell, W. (2004). Predator hunting behaviour and prey vulnerability. Journal
of Animal Ecology 73, 143-154.
Rachlow, J. L. & Bowyer, R. T. (1998). Habitat selection by Dall’s sheep (Ovis dalli): maternal
trade-offs. Journal of Zoology 245, 457-465.
140
Raithel, J. D., Kauffman, M. J. & Pletscher, D. H. (2007). Impact of spatial and temporal
variation in calf survival on the growth of elk populations. Journal of Wildlife Management
71, 795-803.
Ralls, K., Lundigran, B. & Kranz, K. (1987a). Mother-young relationships in captive ungulates:
behavioral changes over time. Ethology 75, 1-14.
Ralls, K., Lundigran, B. & Kranz, K. (1987b). Mother-young relationships in captive ungulates:
spatial and temporal patterns. Zoo Biology 6, 11-20.
Ray, J. C., Hunter, L. & Zigouris, J. (2005). Setting conservation and research priorities for larger
African carnivores. Wildlife Conservation Society, New York.
Read, B. & Frueh, R. J. (1980). Management and breeding of Speke’s gazelle, Gazella spekei, at the
St. Louis Zoo, with a note on artificial insemination. International Zoo Yearbook 20, 99-106.
Rearden, S. N., Anthony, R. G. & Johnson, B. K. (2011). Birth-site selection and predation risk of
Rocky Mountain elk. Journal of Mammalogy 92, 1118-1126.
Redmon, L. J., Murphy, M. T., Dolan, A. C. & Sexton, K. (2009). Parental investment theory and
nest defense by eastern kingbirds. Wilson Journal of Ornithology 121, 1-11.
Redondo, T. & Carranza, J. (1989). Offspring reproductive value and nest defense in the magpie
(Pica pica). Behavioral Ecology and Sociobiology 25, 369-378.
Rettie, W. J. & Messier, F. (2008). Hierarchical habitat selection by woodland caribou: its
relationship to limiting factors. Ecography 23, 466-478.
Robbins, C. T., Podbielancik-Norman, R. S., Wilson, D. L. & Mould, E. D. (1981). Growth and
nutrient consumption of elk calves compared to other ungulate species. Journal of Wildlife
Management 45, 172-186.
141
Roberts, B. A. (2012a). An attack by a warthog Phacochoerus africanus on a newborn Thomson’s
gazelle Gazella thomsonii. African Journal of Ecology 50, 507-508.
Roberts, B. A. (2012b). Perinatal behavior of a wild Grevy’s zebra (Equus grevyi) mare and foal.
Journal of Ethology 30, 205-209.
Robinette, W. L. & Archer, A. L. (1971). Notes on ageing criteria and reproduction of Thomson’s
gazelle. East African Wildlife Journal 9, 83-98.
Rossdale, P. D. (1967). Clinical studies on the newborn thoroughbred foal. British Veterinary
Journal 123, 470-481.
Rossdale, P. D. (1968). Perinatal behavior in the thoroughbred horse. In: Abnormal Behavior in
Mammals (M. W. Fox, ed). Saunders, Philadelphia.
Rubenstein, D. I. & Nuñez, C. (2009). Sociality and reproductive skew in horses and zebras. In:
Reproductive Skew in Vertebrates: Proximate and Ultimate Causes (R. Hager & C. B. Jones,
eds). Cambridge University Press, Cambridge, p. 196-226.
Saino, N., Ninni, P., Incagli, M., Calza, S., Sacchi, R. & Moller, A. P. (2000). Begging and
parental care in relation to offspring need and condition in the barn swallow (Hirundo rustica).
The American Naturalist 156, 637-639.
Schaller, G. B. (1968). The hunting behaviour of the cheetah in the Serengeti National Park,
Tanzania. East African Wildlife Journal 6, 95-100.
Schaller, G. B. (1972). The Serengeti Lion. University of Chicago Press, Chicago.
Schradin, C. (2005). Nest-site competition in two diurnal rodents from the succulent karoo of
South Africa. Journal of Mammalogy 86, 757-762.
142
Schwede, G., Hendrichs, H. & Mcshea, W. (1993). Social and spatial organization of female
white-tailed deer, Odocoileus virginianus, during the fawning season. Animal Behaviour 45, 10071017.
Scott, M. P. & Traniello, J. F. A. (1990). Behavioural and ecological correlates of male and female
parental care and reproductive success in burying beetles (Nicrophorus spp.). Animal Behaviour
39, 274-283.
Scrimgeour, G. J. & Culp, J. M. (1994). Feeding while evading predators by a lotic mayfly: linking
short-term foraging behaviours to long-term fitness consequences. Oecologia 100, 128-134.
Sharp, N. C. C. (1997). Timed running speed of a cheetah (Acinonyx jubatus). Journal of the
Zoological Society of London 241, 493-494.
Sih, A. (1992). Prey uncertainty and the balancing of antipredator and feeding needs. The
American Naturalist 139, 1052-1069.
Sikes, S. K. (1971). The Natural History of the African Elephant. Weidenfeld and Nicolson,
London.
Sinclair, A. R. E., Mduma, S. & Brashares, J. S. (2003). Patterns of predation in a diverse
predator-prey system. Nature 425, 288-290.
Singer, F. J., Harting, A., Symonds, K. K. & Coughenour, M. B. (1997). Density dependence,
compensation, and environmental effects on elk calf mortality in Yellowstone National Park.
The Journal of Wildlife Management 61, 12-25.
Smith, A. T. (1988). Patterns of pika (Genus Ochotona) life history variation. In: Evolution of Life
Histories (M. S. Boyce, ed). Yale University Press, New Haven, Connecticut, p. 233-256.
143
Smith, B. L. & Anderson, S. H. (1996). Patterns of neonatal mortality of elk in northwest
Wyoming. Canadian Journal of Zoology 74, 1229-1237.
Smith, C. & Wootton, R. J. (1995). The costs of parental care in teleost fishes. Reviews in Fish
Biology and Fisheries 5, 7-22.
Smith, W. P. (1987). Maternal defense in Columbian white-tailed deer: when is it worth it? The
American Naturalist 130, 310-316.
Sordahl, T. A. (1990). The risks of avian mobbing and distraction behavior: an anecdotal review.
The Wilson Bulletin 102, 349-352.
Spinage, C. A. (1968). Naturalistic observations on the reproductive and maternal behaviour of
the Uganda defassa waterbuck Kobus defassa ugandae Neumann. Zeitschrift für Tierpsychologie
26, 39-47.
Spinage, C. A. (1986a). Maternal reproduction and health in the Grant’s gazelle (Gazella granti).
Journal of Zoology 1, 461-520.
Spinage, C. A. (1986b). The Natural History of Antelopes. Facts on File Publications, New York.
Stankowich, T. & Coss, R. G. (2008). Alarm walking in Columbian black-tailed deer: its
characterization and possible antipredatory signaling functions. Journal of Mammalogy 89,
636-645.
Stephens, P. W. & Peterson, R. O. (1984). Wolf-avoidance strategies of moose. Holarctic Ecology
7, 239-244.
Sundaresan, S. R., Fischhoff, I. R. & Rubenstein, D. I. (2007). Male harassment influences female
movements and associations in Grevy’s zebra (Equus grevyi). Behavioral Ecology 18, 860-865.
144
Svagelj, W. S., Trivellini, M. M. & Quintana, F. (2012). Parental investment theory and nest
defence by imperial shags: effects of offspring number, offspring age, laying date and parent
sex. Ethology 118, 251-259.
Teichert, N., Keith, P., Valade, P., Richarson, M., Metzger, M. & Gaudin, P. (2013). Breeding
pattern and nest guarding in Sicyopterus lagocephalus, a widespread amphidromous Gobiidae.
Journal of Ethology 31, 239-247.
Tessier, A. J. & Consolatti, N. L. (1989). Variation in offspring size in Daphnia and consequences
for individual fitness. Oikos 56, 269-276.
Theodorakis, C. W. (1989). Size segregation and the effects of oddity on predation risk in
minnow schools. Animal Behaviour 38, 496-502.
Tilton, M. E. & Willard, E. E. (1982). Winter habitat selection by mountain sheep. Journal of
Wildlife Management 46, 359-366.
Toïgo, C. (1999). Vigilance behavior in lactating female Alpine ibex. Canadian Journal of
Zoology 77, 1060-1063.
Townshend, T. W. & Bailey, E. D. (1975). Parturitional, early maternal, and neonatal behavior in
penned white-tailed deer. Journal of Mammalogy 56, 347-362.
Trillmich, F. (1986). Maternal investment and sex-allocation in the Galapagos fur seal,
Arctocephalus galapagoensis. Behavioral Ecology and Sociobiology 19, 157-164.
Trivers, R. L. (1972). Parental investment and sexual selection. In: Sexual Selection and the
Descent of Man 1871-1971 (B. Campbell, ed). Aldine Publishing Company, Chicago.
Trivers, R. L. (1974). Parent-offspring conflict. American Zoologist 14, 249-264.
145
Tufto, J., Andersen, R. & Linnell, J. (1996). Habitat use and ecological correlates of home range
size in a small cervid: the roe deer. Journal of Animal Ecology 65, 715-724.
Tyler, S. J. (1972). The behaviour and social organization of the New Forest ponies. Animal
Behaviour Monographs 5, 87-196.
Valeix, M., Loveridge, A. J., Chamaillé-Jammes, S., Davidson, Z., Murindagomo, F., Fritz, H. &
Macdonald, D. W. (2009). Behavioral adjustments of African herbivores to predation risk by
lions: spatiotemporal variations influence habitat use. Ecology 90, 23-30.
Wackernagel, H. (1965). Grant’s zebra, Equus burchelli boehmi, at Basle Zoo - a contribution to
breeding ecology. International Zoo Yearbook 5, 38-41.
Walther, F. (1965). Verhaltensstudien an der Gattung Tragelaphus de Blainville (1816) in
Gefangenschaft unter besonderer Berücksichtigung des Sozialverhaltens. Zeitschrift für
Tierpsychologie 21, 393-467.
Walther, F. (1973a). On age class recognition and individual recognition of Thomson’s gazelle in
the field. Journal of South Africa Wildlife Management Association 2, 9-15.
Walther, F. (1973b). Round-the-clock activity of Thomson’s gazelle (Gazella thomsoni Günther
1884) in the Serengeti National Park. Zeitschrift für Tierpsychologie 32, 75-105.
Walther, F. R. (1969). Flight behaviour and avoidance of predators in Thomson’s gazelle (Gazella
thomsoni Guenther 1884). Behaviour 34, 184-221.
Waring, G. H. (1982). Onset of behaviour patterns in the newborn foal. Equine Practice 4, 2831, 34.
White, K. S. & Berger, J. (2001). Antipredator strategies of Alaskan moose: are maternal tradeoffs influenced by offspring activity? Canadian Journal of Zoology 79, 2055-2062.
146
White, M., Knowlton, F. F. & Glazener, W. C. (1972). Effects of dam-newborn fawn behavior on
capture and mortality. The Journal of Wildlife Management 36, 897-906.
Wilcox, J. T. & Van Vuren, D. H. (2009). Wild pigs as predators in oak woodlands of California.
Journal of Mammalogy 90, 114-118.
Williams, S. D. (2002). Status and action plan for Grevy’s zebra (Equus grevyi). In: Equids: Zebras,
Asses, and Horses: Status Survey and Conservation Action Plan (P. D. Moehlman, ed). IUCN/
SSC Equid Specialist Group, Gland, Switzerland, p. 11-27.
Wilmshurst, J. F., Fryxell, J. M. & Colucci, P. E. (1999). What constrains daily intake in
Thomson’s gazelles? Ecology 80, 2338-2347.
Wisenden, B. D., Cline, A. & Sparkes, T. C. (1999). Survival benefit to antipredator behavior in
the amphipod Gammarus minus (Crustacea: Amphipoda) in response to injury-released chemical
cues from conspecifics and heterospecifics. Ethology 105, 407-414.
Wright, B. S. (1960). Predation on big game in East Africa. The Journal of Wildlife Management
24, 1-15.
Wyman, J. (1967). The jackals of the Serengeti. Animals 10, 79-83.
147