ORIGINAL ARTICLE
Female parity, male aggression, and the Challenge Hypothesis
in wild chimpanzees
Marissa E. Sobolewski ? Janine L. Brown ?
John C. Mitani
Received: 22 March 2012 / Accepted: 24 September 2012
 Japan Monkey Centre and Springer Japan 2012
Abstract The Challenge Hypothesis proposes that tes-
tosterone mediates aggression during periods of heightened
conflict between males, especially episodes that have
important fitness consequences. Considerable evidence
from seasonally breeding species provides support for this
hypothesis, but few data exist in animals that mate year-
round. We tested predictions generated by the Challenge
Hypothesis in chimpanzees, a non-seasonally breeding
primate, through a study of individuals living in an
exceptionally large community at Ngogo, Kibale National
Park, Uganda. Results indicated that dominance rank
had no influence on testosterone levels. Instead of rank
influencing testosterone production, additional analyses
revealed an important role for reproductive competition.
Male chimpanzees displayed more aggression when they
were in the same party as parous estrous females than when
reproductively active females were unavailable. Male
chimpanzees competed more intensely for mating oppor-
tunities with parous females than with nulliparas, and as a
consequence, males displayed more aggression around the
former than the latter. When males accompanied parous
estrous females, their urinary testosterone concentrations
were significantly higher than baseline concentrations. In
contrast, urinary testosterone concentrations did not exceed
baseline when males associated with nulliparous estrous
females. These differences in testosterone levels could not
be attributed to mating per se because males copulated
equally often with parous and nulliparous females. Fur-
thermore, variation in testosterone concentrations were not
due to males gathering together in large parties, as their
levels in these situations did not exceed baseline. Taken
together, these findings, derived from a relatively large
sample of males and estrous females, replicate those from a
prior study and furnish additional support for the Challenge
Hypothesis. Our results suggest that the Challenge
Hypothesis is likely to be broadly applicable to chimpan-
zees and increase our understanding of the physiological
costs to males who compete for estrous females.
Keywords Chimpanzee  Testosterone  Aggression 
Challenge Hypothesis
Introduction
Aggression has been a traditional focus of ethological
study (Lorenz 1966), with considerable research devoted to
explain its causation, development, function, and evolution
(Archer 1988; Nelson 2005). Studies regarding the hor-
monal correlates of aggression have featured prominently
in discussions of the causal mechanisms underlying the
aggressive behavior of vertebrates. While testosterone
plays a primary role in reproduction, it has long been
known to mediate aggression, as revealed by early studies
that showed castration limited the manifestation of
aggression in males of several species, including humans
(Baum 2002; Soma 2006). Continued research, however,
has complicated this picture. While differences between
the sexes and between adult and juvenile males furnish
additional support for a putative relationship between tes-
tosterone and aggression (Nelson 2011), in studies where
only intact adult males are considered, a connection
M. E. Sobolewski (&)  J. C. Mitani
Department of Anthropology, University of Michigan,
1085 South University Avenue, Ann Arbor, MI 48109, USA
e-mail: mesobole@umich.edu
J. L. Brown
Smithsonian Conservation Biology Institute,
1500 Remount Road, Front Royal, VA 22630, USA
123
Primates
DOI 10.1007/s10329-012-0332-4
between testosterone and aggression does not always exist
(Hirschenhauser and Oliveira 2006).
The Challenge Hypothesis helps resolve these conflict-
ing findings by proposing that testosterone is associated
with aggression only in specific social and reproductive
contexts (Wingfield et al. 1990). Because chronically ele-
vated testosterone levels can have deleterious effects on
health, its production is limited to aggression that has
particularly important fitness consequences. Aggression
associated with competition for mates is a prime example,
especially in many seasonally breeding primate species.
For example, breeding season testosterone levels of male
rhesus monkeys (Macaca mulatta) vary positively as a
function of how frequently they display aggressive
behavior (Higley et al. 1996). These same rhesus males
exhibit relatively high levels of aggression while compet-
ing for estrous females and display correspondingly high
levels of testosterone compared with male muriquis
(Brachyteles hypoxanthus) who compete less intensely for
mating opportunities (Strier et al. 1999). Additional
observations reveal that males in several species of strep-
sirrhines and haplorrhines predictably increase their pro-
duction of testosterone during the mating season and
female conception cycles (strepsirrhines: Cavigelli and
Pereira 2000; Fichtel et al. 2007; Gould and Ziegler 2007;
Ostner et al. 2008; haplorrhines: Lynch et al. 2002; Bales
et al. 2006; Girard-Buttoz et al. 2009; Ostner et al. 2011).
The Challenge Hypothesis was originally developed to
explain the relationship between male aggression and tes-
tosterone secretion in seasonally breeding birds, which
display pronounced temporal variation in their aggressive
behavior (Wingfield et al. 1987). More evidence is needed
to assess its applicability in non-seasonally breeding taxa
that lack large seasonal changes in behavior (Archer 2006;
Beehner et al. 2009). In this regard, chimpanzees represent
a relevant species to test the Challenge Hypothesis as they
mate throughout the year and are non-seasonal breeders.
Despite the absence of breeding seasons, however, they
still provide an opportunity to compare male testosterone
concentrations in situations involving intense male repro-
ductive aggression and other contexts. This opportunity
exists because male competition varies over time and as a
function of the individual identity of females (Muller and
Wrangham 2004; Muller et al. 2006). Individual females
mate during discrete estrous periods where they develop
sexual swellings that last about 12?13 days (Furuichi and
Hashimoto 2002). Moreover, females give birth only once
every 5?6 years (Goodall 1986; Boesch and Boesch-Ach-
ermann 2000; Nishida et al. 2003; Sugiyama 2004), leading
to an operational sex ratio that is skewed heavily toward
males (Emlen and Oring 1977). As a consequence, males
compete vigorously to obtain mating and reproductive
opportunities with estrous females who are available only
rarely (Boesch et al. 2006; Inoue et al. 2008; Wroblewski
et al. 2009; Newton-Fisher et al. 2010). Aggression asso-
ciated with reproductive competition can be extremely
intense, especially for older, parous females, who have
already reproduced successfully (Muller et al. 2006). Males
compete less intensely for young nulliparous females, who
represent less attractive mating partners (ibid.).
One previous study has taken advantage of the temporal
variation in mating behavior by individual females and
differences in male mating preferences to test the Chal-
lenge Hypothesis in chimpanzees (Muller and Wrangham
2004). There it was found that male aggression and tes-
tosterone levels increased in the presence of parous estrous
females. In contrast, aggression and testosterone were not
elevated when males were around nulliparous estrous
females and when estrous females were absent. Because
mating frequencies were the same with both parous and
nulliparous females, increases in male testosterone
appeared to be related to aggression associated with mate
acquisition, rather than the act of mating itself. Finally, a
positive relationship between male dominance rank and
testosterone concentrations existed, with high-ranking
males possessing higher levels than lower ranking indi-
viduals during a period of rank stability. While these
results support the Challenge Hypothesis, it is unclear
whether they can be applied broadly to the behavior of
chimpanzees because a relatively small number of females
and males were observed. Because variation is a charac-
teristic feature of chimpanzee behavior (Wrangham et al.
1994; Boesch et al. 2002), additional data from other
chimpanzee communities, including larger samples of
males and estrous females, are clearly needed to evaluate
the generality of these findings.
One possibility not investigated by Muller and Wrang-
ham (2004) in their previous study was that high male
testosterone levels were a byproduct of large male group
size. Large groups of males gather predictably around
estrous females (Mitani et al. 2002), and in these, elevated
rates of male aggression typically occur (Muller 2002). As
a consequence, party size is a likely confound that requires
examination and possible control. The presence of alpha
males represents another potential confound that warrants
investigation because these males are responsible for a
disproportionate number of charging displays observed
each day (Muller 2002). As a result, male testosterone
concentrations might increase simply due to the heightened
conflict that surrounds alpha males.
Chimpanzees at Ngogo, Kibale National Park, Uganda,
live in an extremely large community containing over 150
individuals and many adult males and females. The
unusually large size of this community creates ample
opportunities to observe male?male aggression and mat-
ings between males and females that differ in their
Primates
123
reproductive states and parity. In this study we extend the
findings of previous work by examining whether associa-
tions exist between male aggression and testosterone dur-
ing interactions with females who differ in parity. Our
observations of male chimpanzees, as they compete for
matings with multiple parous estrous females over several
cycles, provide empirically sufficient samples to test the
Challenge Hypothesis in this non-seasonally breeding pri-
mate. Specifically, we test the prediction that the presence
of parous estrous females leads to heightened rates of
male?male aggression, which elevates male chimpanzee
testosterone levels above their baseline concentrations. In
contrast, we do not expect a similar rise in urinary testos-
terone levels in males when they associate with nulliparous
females. Although males mate nulliparas, these females do
not generate high levels of male competition.
Methods
Study site and subjects
We observed chimpanzees at Ngogo, Kibale National Park,
Uganda. The 30 km2 study area lies at an altitude of about
1,400 m above sea level and consists primarily of mature,
evergreen tropical forest (Struhsaker 1997). The Ngogo
chimpanzees have been under observation since 1995
(Mitani 2009). As a result, they are well habituated to
human presence and can be followed easily and observed
closely. The Ngogo chimpanzee community is the largest
described in the wild thus far and consisted of approxi-
mately 150 individuals at the time of study, including 27
adult males and 19 cycling females. The latter comprised
eight parous females and 11 nulliparas.
Behavioral observations
MES conducted behavioral observations over 14 months
during three field seasons, May?July 2006, May?Novem-
ber 2007, and February?May 2008. She observed 27 adult
males for 1,378 h, recording their identities and noting
whether they followed estrous females. For purposes of the
following analyses, we defined large groups of males to
form on days that the party included more than half of all
of the adult males in the community, i.e. C14 males.
Although chimpanzees live in fission?fusion communities,
whose members split apart and come together throughout
the day, at Ngogo, large parties of males are fairly cohesive
as they predictably form during periods of high food
availability and whenever several females come into estrus
simultaneously (Mitani et al. 2002).
Male aggression was scored by recording charging dis-
plays directed at specific individuals. These displays occur
frequently and are particularly conspicuous and easy to
observe and record. As a result, they were recorded
ad libitum. In prior research, displays have been used to
assay male aggression as they correlate positively with
male dominance rank, with high-ranking males engaging in
this behavior often (Muller 2002). During charges, males
become piloerect, shake and break trees and foliage, pull
branches in their wake, and run quickly in a straight line
through the forest on the ground. Charges typically elicit
screams from recipients at whom charges are directed and
nearby bystanders; they occasionally result in displaying
individuals physically attacking others. We calculated rates
of aggression for each male by counting the number of
times individuals charged relative to the number of hours
they were observed.
We combined observations of charges, where the reci-
pient was identified, with pant grunts to construct a dom-
inance rank matrix. Pant grunts are vocal signals of
submission given by low-ranking chimpanzees to high-
ranking chimpanzees (Bygott 1979; Hayaki et al. 1989).
We ordered individuals from top to bottom so as to mini-
mize the number of reversals in the matrix. Only 12
reversals occurred in 872 interactions, producing a linearity
index of 98.6 %.
Female chimpanzees display sexual swellings when they
are in estrus. We considered only those females that
exhibited full maximal swellings to be in estrus (Wallis
1992). While observing males during focal sampling ses-
sions, MES recorded male copulations with females
ad libitum. Copulations were recorded when males
mounted estrous females followed by intromission and
pelvic thrusting. We counted the number of times males
mated females per day to determine male copulation rates.
Hormone analyses
Urine collection
Testosterone was assayed using urine collected non-
invasively from male chimpanzee subjects. We analyzed
samples from 26 of the 27 adult males (mean = 108
samples/male, SD = 42). Results of assays from one male
(pi in Table 1) were inadvertently lost while these data
were being transferred to the computer and were thus
unavailable for subsequent analysis. Samples were col-
lected from males while they were in the same party and
associating with estrous females, including those who were
parous (mean = 7.5 samples/male, SD = 3.5) and nullip-
arous (mean = 6.8 samples/male, SD = 4.3). As noted
above, these samples were obtained only when females
possessed full, maximal swellings. The eight parous
females were observed during 12 cycles (mean = 1.5
cycles/female; SD = 0.8, range 1?3), while the 11
Primates
123
nulliparas were observed during 20 cycles (mean = 1.8
cycles/female; SD = 0.8, range 1?3).
We collected 103 samples from 25 males on 25 days
when more than half of all of the males were present in the
daily party (mean = 4.13 samples/male, SD = 1.9). For
these, samples were not collected on days males hunted,
patrolled the boundary of their territory, and followed
estrous females, as these behaviors are known or hypoth-
esized to affect testosterone production (Sobolewski et al.
2012). Finally, the alpha male was not seen on 20 days that
we observed parous females. We collected samples from
males on these days (mean = 2.4 samples/male, SD = 1.1)
to compare with others collected from them when both the
alpha male and parous estrous females were present
(mean = 6.0 samples/male, SD = 3.3). We also compared
the former samples to their baseline levels.
Urine was collected from leaf litter on the ground with a
pipette. Samples were occasionally caught in a plastic bag
when males urinated overhead in trees. Samples cross-
contaminated with feces or blood were discarded. Those
that were retained were placed in 5 ml tubes and labeled
with the male?s name, date, and time. Samples were frozen
within 12 h after collection in the field and shipped to the
United States on ice using certified transport equipment.
All hormone analyses were conducted by MES working
under the supervision of JLB at the Smithsonian Conser-
vation Biology Institute, Front Royal, Virginia.
Testosterone analysis
We analyzed testosterone in unprocessed urine using a
single antibody enzyme immunoassay (EIA) provided by
Coralie Munro from the University of California, Davis
(Kersey et al. 2010). Microtitre plates (96 well; Nunc-
Immuno, Maxisorp) were coated with a polyclonal testos-
terone antiserum (R 156/157; 50 ll per well; diluted
1:7,500 in coating buffer, 0.05 M NaHCO3, pH 9.6) and
allowed to set for 12?18 h at 4 C. Unabsorbed antiserum
was removed with wash solution (0.149 M NaCl, 0.5 %
Tween 20). Testosterone standards (50 ll, range 2.3?600
pg/well, diluted in assay buffer, 0.1 M NaPO4, 0.149 M
NaCl, 0.1 % bovine serum albumin, pH 7.0) in triplicate
and samples (50 ll) in duplicate were then added to the
wells, followed immediately with testosterone-horseradish
peroxidase (50 ll, 1:80,000 dilution in assay buffer). Fol-
lowing incubation at room temperature for 2 h, plates were
washed five times before 100 ll substrate buffer (0.4 mM
2, 20-azino-di-(3-ethylbenzthiazoline sulfonic acid) diam-
monium salt, 1.6 mM H2O2, 0.05 M citrate, pH 4.0) was
added to each well. After incubation for 30?60 min, the
absorbance was measured at 405 nm (540 reference filter)
when the optical density in the total binding wells reached
*1.0. Intra-assay and inter-assay coefficients of variation
(CV) for the internal controls (n = 124 assays) were below
10 and 15 %, with 9.34 % (mean binding, 23.6 %) and
11.89 % (mean binding, 69.5 %) for the high and low
samples, respectively, while the CV for the 50 % binding
point of the standard curve was 6.36 %. The assay was
validated for chimpanzee urine by demonstrating that serial
dilutions of pooled urine samples produced displacement
curves parallel to those of the testosterone standard curve
and that there was significant recovery ([90 %) of exog-
enous testosterone added to urine before analysis. We
achieved high levels of recovery and accuracy of mea-
surement and maintained strict controls for individual
variation, sample quality, and assay variance.
Creatinine analysis
We indexed all urine samples for creatinine (Cr) to account
for variations in water content (Taussky 1954). Creatinine
is a by-product of muscle breakdown and under normal
Table 1 Male dominance rank, age, aggression, and testosterone
levels
Male Rank Age class Aggression
(charges/h)
Baseline testosterone
(ng/mg Cr)
bt 1 Prime 1.04 69
ho 2 Prime 0.83 73
mi 3 Prime 0.97 113
bs 4 Prime 0.55 80
cr 5 Prime 0.58 71
mt 6 Prime 0.28 80
wb 7 Young 0.61 104
lo 8 Prime/old 0.36 99
br 9 Prime 0.38 85
mo 10 Prime/old 0.26 94
hr 11 Prime/old 0.11 88
mw 12 Old 0.13 89
bg 13 Prime 0.14 109
do 14 Old 0.05 77
mg 15 Prime/old 0.18 132
ga 16 Prime 0.57 123
pk 17 Old 0.11 95
or 18 Prime 0.24 119
ro 19 Young 0.29 107
pi 20 Old 0.20 No data
bf 21 Old 0.06 77
rh 22 Young 0.29 112
dx 23 Prime 0.04 82
ri 24 Young 0.10 112
ta 25 Young 0.23 111
gz 26 Young 0.12 62
di 27 Prime 0.13 112
Primates
123
conditions is excreted at a constant rate per individual. The
creatinine concentrations in urine were determined using a
Jaffe reaction. Samples with creatinine concentrations
below 0.01 ng/ml were considered too dilute and excluded
from analysis; this involved less than 5 % of all samples.
Hormone concentrations were divided by creatinine con-
centrations and expressed as the concentration of testos-
terone (ng)/per mg Cr.
Statistical analyses
We determined baseline concentrations of testosterone for
each male through an iterative process described by
Moreira et al. (2001), with minor modifications. This
technique has been used successfully in the past to identify
baseline values from biologically relevant peaks (Moreira
et al. 2001). We began by computing the mean value of all
samples for each male. We then removed values that were
above 2 standard deviations from the mean. We iterated
this procedure until no values outside 2 standard deviations
of the mean remained. We employed the resulting mean as
the statistical baseline value for that male. We excluded
urine samples collected before 0900 h from analyses to
control for diurnal variation in testosterone production
(Muller and Wrangham 2004).
Testosterone varies considerably among individual
males (Kempenaers et al. 2008). Therefore, we used each
male as his own control for statistical purposes. Specifi-
cally, we employed a non-parametric Wilcoxon matched-
pairs signed-ranks test to determine if male testosterone
concentrations assayed on days when estrous females were
present differed from their baseline levels. We also used
Wilcoxon tests to investigate whether male aggression and
copulation rates varied as a function of female parity. For
all of these tests, we separated estrous females into two
categories, those who had given birth, i.e. parous females,
and those who had not, i.e. nulliparous females. In two
additional Wilcoxon tests we examined whether urinary
testosterone levels were elevated on days males gathered in
large parties compared to their baseline levels and whether
male testosterone concentrations differed on days they
followed parous estrous females in the presence and
absence of the alpha male. Finally, we utilized the mat-
ched-pairs design to examine whether male testosterone
levels were high on days they followed parous estrous
females in the absence of the alpha male. We did so by
comparing these to their baseline concentrations.
Results
Table 1 shows individual males, their dominance ranks,
relative ages, aggression rates, and baseline testosterone
concentrations. In accord with previous studies (e.g. Muller
2002), high-ranking males at Ngogo displayed higher rates
of aggression than did lower ranking males (Spearman
r = 0.70, p \ 0.001, N = 27). High-ranking males, how-
ever, did not possess high levels of testosterone. There was
no relationship between male dominance rank and baseline
testosterone levels (Spearman r = 0.34, p [ 0.05, N = 26).
To test the Challenge Hypothesis, we examined the
relationship between testosterone and aggression as males
sought reproductive opportunities with females. Male
chimpanzees compete for estrous females, as manifest by
their aggressive behavior. Competition, however, varied as
a function of female parity. Males displayed more
aggression around parous estrous females than they did
when no estrous females were present (Wilcoxon test:
Z = 4.55, p \ 0.001, N = 27 males; Fig. 1). In contrast,
male aggression was not elevated in the presence of nul-
liparous estrous females compared to when estrous females
were absent (Wilcoxon test: Z = 1.33, p [ 0.15, N = 27
males; Fig. 1). As a consequence, male chimpanzees
showed more aggression in the presence of parous estrous
females than they did in the presence of nulliparous estrous
females (Wilcoxon test: Z = 4.13, p \ 0.001, N = 27
males; Fig. 1). Despite these heightened rates of aggression
around parous females, male chimpanzees mated these
females as frequently as they did nulliparous females
(Wilcoxon test: Z = 1.37, p [ 0.10, N = 27; Fig. 2).
The higher rates of aggression stimulated by the pres-
ence of parous females in estrus were related to male tes-
tosterone levels in predictable ways. Male testosterone was
Fig. 1 Male chimpanzee aggression varies as a function of the
presence and absence of estrous females. Rates of aggression, assayed
by charging displays, when estrous females, either nulliparous or
parous individuals, were present are compared with those when
estrous females were absent. Displayed are means of individual male
means ? 1 SE. N = 27 males. *p \ 0.001 for comparisons between
parous estrous females and estrous females absent and between
parous estrous females and nulliparous estrous females
Primates
123
elevated above baseline levels on days when parous
females were present (Wilcoxon test: Z = 4.32, p \ 0.01,
N = 26; Fig. 3). In contrast, testosterone concentrations
did not differ from baseline on days that males accompa-
nied nulliparous females in estrus (Wilcoxon test:
Z = 0.70, p [ 0.45, N = 26; Fig. 3). As a result, male
urinary testosterone concentrations were also higher when
they followed parous estrous females compared to when
they accompanied nulliparous estrous females (Wilcoxon
test: Z = 3.98, p \ 0.001, N = 26; Fig. 3).
The rise in testosterone experienced by males while
following parous estrous females could not be attributed to
the fact that they formed large parties on these days; male
testosterone levels did not exceed baseline on days that
they formed large parties in the absence of estrous females
(Wilcoxon test: Z = 1.01, p [ 0.30, N = 25). In addition,
urinary testosterone concentrations of males who followed
parous estrous females did not vary as a function of the
presence or absence of the alpha male (Wilcoxon test:
Z = 0.03, p [ 0.95, N = 23). Male testosterone levels
were also high on days that they followed parous estrous
females in the absence of the alpha male; their values on
these days consistently exceeded baseline (Wilcoxon test:
Z = 2.03, p \ 0.05, N = 24).
Discussion
Prior research has revealed that dominance rank affects
testosterone in male chimpanzees. High-ranking males
display relatively high testosterone levels compared with
lower ranking individuals (Muehlenbein et al. 2004; Muller
and Wrangham 2004). In contrast, there was no relation-
ship between these two variables among males in this
study. Age is a likely factor that might account for the
failure to document a relationship between male rank and
testosterone at Ngogo (Table 1). Testosterone decreases
with age in human males (Bribiescas 2001), and some high
and middle ranking male chimpanzees in our sample were
old. Further research will be necessary to investigate this
possibility.
Additional analyses indicate that male chimpanzees at
Ngogo compete for mating opportunities for females, but
that they do so primarily for parous females. Male
aggression increased on days these females were present
compared to days when males followed nulliparous
females and on days when estrous females were absent.
Testosterone is associated with the heightened aggression
over more attractive parous females, as concentrations in
the presence of these females were higher than baseline
levels and those when males accompanied nulliparous
females. The elevation in male testosterone is directly
attributable to reproductive competition for females and
not associated with increased mating activity per se
because males copulated with parous females as often as
they did with nulliparous females. Large group size cannot
explain these differences because urinary testosterone
concentrations did not vary from baseline on days when
males formed large parties in the absence of estrous
females. In addition, the elevated testosterone levels that
we documented in males as they competed for parous
estrous females could not be attributed to the presence of
the alpha male; male testosterone concentrations did not
Fig. 2 Female parity does not affect male chimpanzee copulation
rates. Male copulation rates with nulliparous and parous females are
shown. Displayed are means of individual male means ? 1 SE.
N = 27 males
Fig. 3 Male chimpanzee testosterone levels vary as a function of
female parity. Male baseline testosterone concentrations are com-
pared with their concentrations when they accompanied nulliparous
estrous females and parous estrous females. Displayed are means of
individual male means ? 1 SE. N = 26 males. *p \ 0.001 for the
comparison of male baseline concentrations and those when they
followed parous estrous females. **p \ 0.01 for the comparison
between nulliparous estrous females and parous estrous females
Primates
123
differ in the presence or absence of the alpha, and they
continued to exceed baseline when parous estrous females
were present and he was not there. Taken together, our
results are consistent with the Challenge Hypothesis, which
proposes that the production of testosterone will be asso-
ciated with aggression directly related to reproduction.
The fact that testosterone did not increase in the pres-
ence of estrous females, who have not yet given birth, may
be surprising. Nulliparas, however, are invariably recent
immigrants who experience a prolonged period of infer-
tility after moving into their new communities. As a result,
these females cycle consistently, sometimes for several
years, before giving birth for the first time (Goodall 1986;
Boesch and Boesch-Achermann 2000; Nishida et al. 2003;
Sugiyama 2004). Because many cycles do not represent
legitimate reproductive opportunities, males do not com-
pete for nulliparas as vigorously as they do for older,
parous females (Muller et al. 2006).
Our findings are consistent with the only previous study
that has tested the Challenge Hypothesis in wild chim-
panzees. In the Kanyawara chimpanzee community, also
located in the Kibale National Park, Muller and Wrangham
(2004) found that males there selectively increased their
rates of aggression in the presence of parous estrous
females. Elevated testosterone levels were also displayed
by males who accompanied parous estrous females com-
pared with those who followed nulliparous estrous females
or those who were alone in the absence of any estrous
females. Because males failed to mate parous females more
frequently than they did nulliparas, Muller and Wrangham
(2004) concluded, like us, that the increase in testosterone
could be attributed to heightened aggression around parous
estrous females in general, rather than mating activity
alone. One limitation of this prior study was that data were
collected from a relatively small number of males
(N = 8?11), nulliparous females (N = 2), and parous
females (N = 3), with observations focused on a single
parous estrous female during a single estrous cycle. These
small samples raise the possibility that the findings of this
previous study cannot be generalized to chimpanzees in
other communities and as a whole. The results presented
here were based on much larger samples of males
(N = 26), parous females (N = 8), nulliparous females
(N = 11), and reproductive cycles of females (parous
females: N = 12; nulliparous: N = 20). Our findings sup-
port those derived from studies of other vertebrates
(Wingfield 2005) and suggest that the Challenge Hypoth-
esis is likely to be broadly applicable to chimpanzees.
This study also adds to our understanding of how social
interactions influence male testosterone in a species that
does not typically display pronounced seasonal variation in
aggressive behavior. Most previous tests of the Challenge
Hypothesis have been conducted with seasonally breeding
taxa that show large fluctuations in aggressive behavior
over time (reviews in Oliveira 2004; Hirschenhauser and
Oliveira 2006). Few data exist regarding non-seasonally
breeding species because such changes in aggression are
less obvious, making it difficult to relate changes in tes-
tosterone to changes in behavior. The Challenge Hypoth-
esis, however, proposes that social interactions that have
particularly significant fitness consequences will lead to
transient changes in androgen levels irrespective of the
time of year, and as a consequence, it should apply to
seasonally-breeding and non-seasonally breeding species
alike. Because many primates mate year-round, they pro-
vide a model taxon to investigate in this regard. Obtaining
additional data from these animals represents an important
area for future research.
Acknowledgments Our fieldwork was sponsored by the Uganda
Wildlife Authority, Uganda National Council for Science and Tech-
nology, and Makerere University Biological Field Station. We thank
A. Magoba, G. Mbabazi, L. Ndangizi, and A. Tumusiime for field
assistance, J. Lwanga for logistic aid, and J. Beehner, K. Hosaka, and
an anonymous reviewer for comments on the manuscript. N. Presley
and S. Putnam provided help in the laboratory with hormone analyses.
This research was supported by grants from the L.S.B. Leakey
Foundation and the U.S. National Science Foundation (IOB-0516644
and BCS-0752637). This research was conducted in compliance with
all legal requirements of Uganda and was approved by the University
Committee on the Use and Care of Animals (UCUCA) at the Uni-
versity of Michigan (Protocol #1092).
References
Archer J (1988) The behavioural biology of aggression. Cambridge
University Press, Cambridge, p 257
Archer J (2006) Testosterone and human aggression: an evaluation of
the challenge hypothesis. Neurosci Biobehav Rev 30:319?345
Bales KL, French JA, McWilliams J, Lake RA, Dietz JM (2006)
Effects of social status, age, and season on androgen and cortisol
levels in wild male golden lion tamarins (Leontopithecus
rosalia). Horm Behav 49:88?95
Baum M (2002) Sex steroids and testosterone. In: Becker J, Breedlove
M, Crews D (eds) Behavioral endocrinology, 2nd edn. MIT
Press, Cambridge, pp 120?159
Beehner JC, Gesquiere L, Seyfarth RM, Cheney DL, Alberts SC,
Altmann J (2009) Testosterone related to age and life-history
stages in male baboons and geladas. Horm Behav 56:472?480
Boesch C, Boesch-Achermann H (2000) The chimpanzees of the Ta??
forest. Oxford University Press, Oxford, p 316
Boesch C, Hohmann G, Marchant L (2002) Behavioural diversity in
chimpanzees and bonobos. Cambridge University Press, Cam-
bridge, p 300
Boesch C, Kohou G, Nene H, Vigilant L (2006) Male competition and
paternity in wild chimpanzees of the Ta?? forest. Am J Phys
Anthropol 130:103?115
Bribiescas RG (2001) Reproductive ecology and life history of the
human male. Yearb Phys Anthropol 44:148?176
Bygott D (1979) Agonistic behavior, dominance, and social structure
in wild chimpanzees of the Gombe National Park. In: Hamburg
D, McCown E (eds) The great apes. Menlo Park, CA, Benjamin-
Cummings, pp 405?427
Primates
123
Cavigelli SA, Pereira ME (2000) Mating season aggression and fecal
testosterone levels in male ring-tailed lemurs (Lemur catta).
Horm Behav 37:246?255
Emlen ST, Oring LW (1977) Ecology, sexual selection, and evolution
of mating systems. Science 197:215?223
Fichtel C, Kraus C, Ganswindt A, Heistermann M (2007) Influence of
reproductive season and rank on fecal glucocorticoid levels in
free-ranging male Verreaux?s sifakas (Propithecus verreauxi).
Horm Behav 51:640?648
Furuichi T, Hashimoto C (2002) Why female bonobos have a lower
copulation rate during estrus than chimpanzees. In: Boesch C,
Hohmann G, Marchant L (eds) Behavioural diversity in chim-
panzees and bonobos. Cambridge University Press, Cambridge,
pp 156?167
Girard-Buttoz C, Heistermann M, Krummel S, Engelhardt A (2009)
Seasonal and social influences on fecal androgen and glucocor-
ticoid excretion in wild male long-tailed macaques (Macaca
fascicularis). Physiol Behav 98:168?175
Goodall J (1986) The chimpanzees of Gombe. Belknap Press,
Cambridge, p 673
Gould L, Ziegler TE (2007) Variation in fecal testosterone levels,
inter-male aggression, dominance rank and age during mating
and post-mating periods in wild adult male ring-tailed lemurs
(Lemur catta). Am J Primatol 69:1325?1339
Hayaki H, Huffman M, Nishida T (1989) Dominance among male
chimpanzees in the Mahale Mountains National Park, Tanzania.
Primates 30:187?197
Higley JD, King ST, Hasert MF, Champoux M, Suormi SJ, Linnoila
M (1996) Stability of inter-individual differences in serotonin
function and its relationship to severe aggression and competent
social behavior in rhesus macaque females. Neuropsychophar-
macology 14:67?76
Hirschenhauser K, Oliveira RF (2006) Social modulation of andro-
gens in male vertebrates: meta-analyses of the challenge
hypothesis. Anim Behav 71:265?277
Inoue E, Inoue-Murayama M, Vigilant L, Takenaka O, Nishida T
(2008) Relatedness in wild chimpanzees: influence of paternity,
male philopatry, and demographic factors. Am J Phys Anthropol
137:256?262
Kempenaers B, Peters A, Foerster K (2008) Sources of individual
variation in plasma testosterone levels. Proc R Soc Lond B Biol
Sci 363:1711?1723
Kersey DC, Wildt DE, Brown JL, Huang Y, Snyder RJ, Monfort SL
(2010) Parallel and seasonal changes in gonadal and adrenal
hormones in male giant pandas (Ailuropoda melanoleuca).
J Mammal 91:1496?1507
Lorenz K (1966) On aggression. Harcourt, Brace and World, New
York, pp 306
Lynch JW, Ziegler TE, Strier KB (2002) Individual and seasonal
variation in fecal testosterone and cortisol levels of wild male
tufted capuchin monkeys, Cebus apella nigritus. Horm Behav
41:275?287
Mitani JC (2009) Cooperation and competition in chimpanzees:
current understanding and future challenges. Evol Anthropol
18:215?227
Mitani J, Watts D, Lwanga J (2002) Ecological and social correlates
of chimpanzee party size. In: Boesch C, Hohmann G, Marchant
L (eds) Behavioural diversity in chimpanzees and bonobos.
Cambridge University Press, Cambridge, pp 102?111
Moreira N, Monteiro-Filho ELA, Moraes W, Swanson WF, Graham
LH, Pasquali OL, Gomes MLF, Morais RN, Wildt DE, Brown JL
(2001) Reproductive steroid hormones and ovarian activity in
felids of the Leopardus genus. Zoo Biol 20:103?116
Muehlenbein MP, Watts DP, Whitten PL (2004) Dominance rank and
fecal testosterone levels in adult male chimpanzees (Pan
troglodytes schweinfurthii) at Ngogo, Kibale National Park,
Uganda. Am J Primatol 64:71?82
Muller M (2002) Agonistic relations among Kanyawara chimpanzees.
In: Boesch C, Hohmann G, Marchant L (eds) Behavioural
diversity in chimpanzees and bonobos. Cambridge University
Press, Cambridge, pp 112?124
Muller MN, Wrangham RW (2004) Dominance, aggression and
testosterone in wild chimpanzees: a test of the ?challenge
hypothesis?. Anim Behav 67:113?123
Muller MN, Thompson ME, Wrangham RW (2006) Male chimpan-
zees prefer mating with old females. Curr Biol 16:2234?2238
Nelson R (2005) Biology of aggression. Oxford University Press,
Cary, p 529
Nelson R (2011) An introduction to behavioral endocrinology, 4th
edn. Sinauer Associates, Sunderland, p 712
Newton-Fisher NE, Thompson ME, Reynolds V, Boesch C, Vigilant
L (2010) Paternity and social rank in wild chimpanzees (Pan
troglodytes) from the Budongo Forest, Uganda. Am J Phys
Anthropol 142:417?428
Nishida T, Corp N, Hamai M, Hasegawa T, Hiraiwa-Hasegawa M,
Hosaka K, Hunt KD, Itoh N, Kawanaka K, Matsumoto-Oda A,
Mitani JC, Nakamura M, Norikoshi K, Sakamaki T, Turner L,
Uehara S, Zamma K (2003) Demography, female life history,
and reproductive profiles among the chimpanzees of Mahale.
Am J Primatol 59:99?121
Oliveira RF (2004) Social modulation of androgens in vertebrates:
mechanisms and function. Adv Stud Behav 34:165?239
Ostner J, Kappeler PM, Heistermann M (2008) Androgen and
glucocorticoid levels reflect seasonally occurring social chal-
lenges in male redfronted lemurs (Eulemur fulvus rufus). Behav
Ecol Sociobiol 62:627?638
Ostner J, Heistermann M, Schulke O (2011) Male competition and its
hormonal correlates in Assamese macaques (Macaca assamen-
sis). Horm Behav 59:105?113
Sobolewski ME, Brown JL, Mitani JC (2012) Territoriality, tolerance
and testosterone in wild chimpanzees. Anim Behav. http://
dx.doi.org/10.1016/j.anbehav.2012.09.018
Soma KK (2006) Testosterone and aggression: Berthold, birds and
beyond. J Neuroendocrinol 18:543?551
Strier K, Ziegler T, Wittwer D (1999) Seasonal and social correlates
of fecal testosterone and cortisol levels in wild male muriquis
(Brachyteles arachnoides). Horm Behav 35:125?134
Struhsaker TT (1997) Ecology of an African rainforest. University
Press of Florida, Gainesville, p 432
Sugiyama Y (2004) Demographic parameters and life history of
chimpanzees at Bossou, Guinea. Am J Phys Anthropol 124:
154?165
Taussky HH (1954) A microcolorimetric determination of creatinine
in urine by the Jaffe reaction. J Biol Chem 208:853?861
Wallis J (1992) Chimpanzee genital swelling and its role in the
patterns of sociosexual behavior. Am J Primatol 28:101?113
Wingfield JC (2005) A continuing saga: the role of testosterone in
aggression. Horm Behav 48:253?255
Wingfield JC, Ball GF, Dufty AM, Hegner RE, Ramenofsky M
(1987) Testosterone and aggression in birds. Am Sci 75:602?608
Wingfield JC, Hegner RE, Dufty AM, Ball GF (1990) The challenge
hypothesis?theoretical implications for patterns of testosterone
secretion, mating systems, and breeding strategies. Am Nat
136:829?846
Wrangham RW, McGrew WC, de Waal FBM, Heltne P (1994)
Chimpanzee cultures. Harvard University Press, Cambridge, p 448
Wroblewski EE, Murray CM, Keele BF, Schumacher-Stankey JC,
Hahn BH, Pusey AE (2009) Male dominance rank and repro-
ductive success in chimpanzees, Pan troglodytes schweinfurthii.
Anim Behav 77:873?885
Primates
123