Hormones and Behavior 62 (2012) 553–562

Contents lists available at SciVerse ScienceDirect

Hormones and Behavior

journal homepage: www.elsevier.com/locate/yhbeh

Estrogenic plant consumption predicts red colobus monkey (Procolobus rufomitratus)

hormonal state and behavior
Michael D. Wasserman a, b,⁎, Colin A. Chapman b, c, d, Katharine Milton a, Jan F. Gogarten e,
Daniel J. Wittwer f, Toni E. Ziegler f
a
Department of Environmental Science, Policy, and Management, University of California, Berkeley, 130 Mulford Hall #3114, Berkeley, CA 94720–3114, USA
b
Department of Anthropology, McGill University, 855 Sherbrooke St. West, Montreal, QC, Canada H3A 2T7
c
McGill School of Environment, McGill University, 855 Sherbrooke St. West, Montreal, QC, Canada H3A 2T7
d
Wildlife Conservation Society, 185th St. and Southern Blvd., Bronx, NY 10460, USA
e
Department of Biology, McGill University, 1205 Dr. Penfield Ave., Montreal, QC, Canada H3A 1B1
f
Wisconsin National Primate Research Center, University of Wisconsin-Madison, 1220 Capitol Court Madison, WI 53715, USA

a r t i c l e i n f o a b s t r a c t

Available online 23 September 2012 Numerous studies have examined the effects of anthropogenic endocrine disrupting compounds; however, very
little is known about the effects of naturally occurring plant-produced estrogenic compounds (i.e., phytoestrogens)
Keywords: on vertebrates. To examine the seasonal pattern of phytoestrogen consumption and its relationship to hormone
Estradiol levels (407 fecal samples analyzed for estradiol and cortisol) and social behavior (aggression, mating, and
Cortisol grooming) in a primate, we conducted an 11-month field study of red colobus (Procolobus rufomitratus) in Kibale
Environmental endocrinology
National Park, Uganda. The percent of diet from estrogenic plants averaged 10.7% (n=45 weeks; range: 0.7–
Herbivory
Mating
32.4%). Red colobus fed more heavily on estrogenic Millettia dura young leaves during weeks of higher rainfall,
Phytoestrogen and the consumption of this estrogenic item was positively correlated to both their fecal estradiol and cortisol
Aggression levels. Social behaviors were related to estradiol and cortisol levels, as well as the consumption of estrogenic plants
Kibale National Park and rainfall. The more the red colobus consumed estrogenic plants the higher their rates of aggression and copu-
Grooming lation and the lower their time spent grooming. Our results suggest that the consumption of estrogenic plants has
Legume important implications for primate health and fitness through interactions with the endocrine system and changes
in hormone levels and social behaviors.
© 2012 Elsevier Inc. All rights reserved.

Introduction the effects of consuming natural plant-produced estrogenic compounds

(i.e., phytoestrogens; Wynne-Edwards, 2001). Most primates depend
Endocrine disruption is of major concern because many synthetic heavily on a plant-based diet, thus the consumption of estrogenic plants
chemicals pose a threat to humans and wildlife through interference may have important implications for their behavior, ecology, and evolu-
with the vertebrate endocrine system (Hayes, 2005; Propper, 2005). tion. Further, as half of all primate species are at risk of extinction
Much research has focused on the physiological and behavioral effects (Chapman and Peres, 2001), a better understanding of the relationship
of anthropogenic endocrine disrupting compounds, particularly those between primates and naturally occurring phytoestrogens present in
with estrogenic activity (Colborn et al., 1993; Guillette, 2000; Hayes et their foods is needed because such compounds have the potential to
al., 2002; Milnes et al., 2006). Effects include modifying the developing alter fertility and mating behavior (Cederroth et al., 2010a; Simon et
reproductive system (e.g., genital deformation) and altering steroid hor- al., 2004; Whitten and Patisaul, 2001; Wynne-Edwards, 2001).
mone profiles, gamete production, and sex-typical behaviors in exposed Phytoestrogen consumption has been documented to disrupt fertili-
adults (Hayes et al., 2002; Milnes et al., 2006). Much less is known about ty and affects behavior in laboratory and domesticated species, including
rodents, monkeys, sheep, and cattle, as well as humans (Adams, 1990,
1995; Cederroth et al., 2010b; Whitten and Patisaul, 2001). Due to the
conservative nature of the endocrine system across vertebrates, such ef-
⁎ Corresponding author at: Department of Anthropology, McGill University, 855 fects are expected in wild primates feeding on similar estrogenic plant
Sherbrooke St. West, Montreal, QC, Canada H3A 2T7. Fax: + 1 514 398 7476. compounds (Hayes, 2005; Thornton, 2001). Previously we confirmed
E-mail addresses: michael.wasserman@mail.mcgill.ca (M.D. Wasserman),
colin.chapman@mcgill.ca (C.A. Chapman), kmilton@berkeley.edu (K. Milton),
the presence of estrogenic plant foods in the diet of red colobus monkeys
jan.gogarten@gmail.com (J.F. Gogarten), wittwer@primate.wisc.edu (D.J. Wittwer), (Procolobus rufomitratus) of Kibale National Park, Uganda, using tran-
ziegler@primate.wisc.edu (T.E. Ziegler). sient transfection assays (Wasserman et al., 2012). Three staple

0018-506X/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.yhbeh.2012.09.005
554 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562

estrogenic foods with activity at estrogen receptor beta (ERβ) colobines may be particularly susceptible to the estrogenic activity of
made up 10% of the diet (Millettia dura [Fabaceae] young leaves, plants, as has been documented for foregut-fermenting livestock (e.g.,
Ficus natalensis [Moraceae] young leaves, and bark from non-native “clover disease” of sheep) (Adams, 1990, 1995; Bennetts, 1946). Data
Eucalyptus grandis [Myrtaceae]), while non-staple estrogenic foods show that phytoestrogens are often more active after bacterial metabo-
only made up 0.6% of the diet. Here, our objective was to determine lism (Gultekin and Yildiz, 2006; Setchell and Clerici, 2010). For example,
if the consumption of these estrogenic plant foods by red colobus a number of phytoestrogens (e.g., formononetin, daidzein) are converted
monkeys altered their physiology and behavior via interaction with to the more bioactive compound, equol, via bacterial metabolism in a
their endocrine system. number of species, including foregut fermenters (Setchell and Clerici,
Two hormonal axes are important for understanding the potential for 2010).
phytoestrogens to alter physiology and behavior. The first, the
hypothalamo-pituitary gonadal axis (HPG), plays a central role in regu- Assessment of red colobus diet and social behavior
lating the development and maintenance of the mammalian reproduc-
tive system through the production of sex steroids (i.e., estrogens and We determined the red colobus diet, rates of mating and aggression,
androgens) and their downstream effects on reproductive physiology and time spent grooming by collecting behavioral data from one group
and behavior (Hadley, 1999; Wingfield and Sapolsky, 2003). Interference (~70 individuals) in Kibale from August 13, 2007 to June 27, 2008
with the HPG axis can result in altered fertility in males, due to the role of (258 days of sampling), for a total of 1327 observation hours. Data
testosterone and gonadotropins in sperm production, and females, due were collected six days per week from 0800 to 1300 h using scan sam-
to the role of estrogens and gonadotropins in their estrous cycle ples of five individuals every 30 min. When feeding, plant species and
(Hadley, 1999). The second hormonal axis, the hypothalamo–pituitary the part being consumed were identified. We calculated the percent of
adrenal axis (HPA), plays a central homeostatic role through the produc- diet for each item at the weekly scale by summing the number of obser-
tion of glucocorticoids (Sapolsky, 2005; Wingfield and Sapolsky, 2003). vations of feeding on each plant item, and dividing this by the total num-
There is strong evidence that these two hormone axes interact as the ber of feeding observations that week. Using these values, we calculated
HPA axis can suppress the HPG axis so that stress can suppress reproduc- the percentage of the diet coming from each of the staple estrogenic
tion (Wingfield and Sapolsky, 2003) and estrogens can suppress the neg- foods, as well as a sum of all rare estrogenic plant foods, for each week
ative feedback loop of the HPA axis, thus increasing production of (n=45). A staple food was defined as a plant item that was fed on
glucocorticoids (Weiser and Handa, 2009). >1% of the total time, while a rare food was any plant item that was
We examined patterns of phytoestrogen consumption, fecal hor- fed on b 1% of the total time. Estrogenic activity of plant foods was tested
mone levels, and social behavior of red colobus (P. rufomitratus) in using transient transfection assays (Wasserman et al., 2012). We also
Kibale over 11 months. Specifically, we determined how weekly varia- calculated the weekly percent of diet for each of the 13 most fed on di-
tion in the time spent feeding on estrogenic plants related to fecal estra- etary items as listed in Wasserman et al. (2012). Behavioral data on
diol and cortisol levels in adult males. As both consumption of feeding came from 37 known individuals in the group, including 22
estrogenic plants and red colobus hormone levels may vary seasonally, adult females, 13 adult males (including all ten males examined in the
their relationships to rainfall and temperature were examined. Climate hormonal analyses), and two subadult males. Differences in feeding be-
can mediate the relationship between phytoestrogen consumption and havior between males and females were not apparent, especially for the
primate hormone levels through three main mechanisms that are not three estrogenic staple foods (e.g., M. dura young leaf consumption was
mutually exclusive: (1) its effects on food availability (i.e., if plant 5.2 ± 1.4% for adult males and 5.4± 1.6% for adult females), so we used
phenology is linked to climatic variables) (Chapman et al., 1999; our entire dataset to determine diet for the ten adult males as this likely
Struhsaker, 1997), (2) its effects on phytoestrogen levels (i.e., plants provided the most robust and appropriate statistical sample estimate of
contain more or less of these compounds during certain times of the their weekly diet.
year) (Leopold et al., 1976; Mazur and Adlercreutz, 1998; Morrison et Three types of social behavior were examined: mating, aggression,
al., 2010), and (3) its direct effects on hormone levels (i.e., primates and grooming. Grooming data were collected during the scan samples
may suffer climate related stress) (Gogarten et al., 2012; Wingfield, and calculated as percent of time spent grooming (i.e., providing) or
2005; Wingfield et al., 1983). To determine if any hormonal changes being groomed (i.e., receiving). Copulation rates (#/h) were deter-
related to phytoestrogen consumption translated into altered social be- mined using ad libitum observations. Aggression rates (#/h) (i.e., sum
havior, we examined the influence of climate, phytoestrogen consump- of number of chases and fights) were also collected through ad libitum
tion, and hormonal status on rates of aggression and mating and time observations. All social behavior data used came from only the 13
spent grooming. adult males.

Material and methods Assessment of climate and plant phenology

Study site and species Rainfall and temperature data were collected from an area within
the home range of the group. Trees were monitored monthly for ripe
Kibale National Park (795 km2) is a moist evergreen forest in western fruits, unripe fruits, mature leaves, young leaves, and flowers (abun-
Uganda (0 13′–0 41′ N and 30 19′–30 32′ E) (Chapman et al., 2010). This dance ranked from 0 to 4). We calculated the mean monthly pheno-
forest receives an average of 1696 mm annually (1990–2011; Chapman logical score for the staple estrogenic plant foods to provide an
and Chapman, unpublished data), with most falling during two rainy index of their availability. Young leaves of F. natalensis (n = 2 trees)
seasons (Chapman et al., 2010). The Ugandan red colobus (P. rufomitratus and M. dura (n = 11) were monitored, but Eucalyptus grandis bark
tephrosceles) is considered endangered by the IUCN (Struhsaker, 2008), was not because it does not vary in seasonal availability.
with the only viable population remaining in Kibale (Struhsaker, 2005).
The Kibale population consists of numerous multimale–multifemale Assessment of fecal hormone levels
groups with an average group size of 65 individuals (Snaith et al.,
2008). As a forestomach-fermenting obligate folivore dependent upon Although evidence of effects of phytoestrogen consumption similar
symbiotic gut bacteria for securing nutrients from its leaf based diet to those of laboratory studies are difficult to obtain in field studies, indi-
(Bauchop and Martucci, 1968; Chapman et al., 2002; Lambert, 1998; rect measures can provide evidence of changes in either the HPG or HPA
Milton, 1980), the red colobus is an ideal subject for examining the axes, indicating disruption of reproductive capabilities. These hormonal
effects of phytoestrogens. With their specialized forestomach system, effects can be examined noninvasively using the measurement of
 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562 555

excreted estradiol and cortisol metabolites in the feces (Heistermann et changes in red colobus hormonal levels were related to changes in phy-
al., 1993; Touma and Palme, 2005; Wasser et al., 1988; Whitten et al., toestrogen consumption and (2) climatic factors influenced hormone
1998; Ziegler and Wittwer, 2005). Thus, fecal samples were collected levels directly and/or indirectly though their effects on the timing of es-
immediately upon defecation from ten known adult males (out of 13 trogenic plant availability and/or consumption. We used fecal estradiol
in the group). Adult males were selected because female reproductive level as an index of reproductive physiology and fecal cortisol level as
state can greatly influence steroid hormone levels (Weingrill et al., an index of stress physiology. To determine the relative importance of
2004). Fecal samples were collected between 0830 and 1230 h to re- phytoestrogen consumption and climatic seasonality to red colobus hor-
duce the contribution of diurnal variation seen in the excretion patterns monal state, we used stepwise regression including six predictor vari-
of fecal steroids (Sousa and Ziegler, 1998; Wasserman, unpublished data ables: (1)% of diet from rare estrogenic foods, (2–4)% of diet from each
for this species). of the three staple estrogenic foods (M. dura young leaves, F. natalensis
We collected samples approximately once a week for each male young leaves, and Eucalyptus grandis bark), (5) weekly total rainfall,
(n = 407). Fecal samples were immediately placed in sterile vials and (6) mean maximum temperature, and two outcome variables:
and stored in a cooler with ice packs until placed in a − 20 °C freezer (1) fecal estradiol and (2) fecal cortisol. We used the median of weekly
later that day. On the day of extraction, samples were taken from the estradiol and cortisol levels in our analyses since both hormone data sets
− 20 °C freezer and thawed. Each sample was homogenized using a were not normally distributed. We also used stepwise regression with
spatula and 0.5 g was weighed into a test tube. We added 5 ml of the red colobus' 13 most fed on dietary items (11 non-estrogenic and
5.0 pH citrate buffer and 5 ml 95% ethanol to each fecal sample, 2 estrogenic, accounting for 66% of total diet; Wasserman et al., 2012)
and this fecal material solution was mixed on a homogenizer for and rainfall as predictor variables and fecal estradiol as the outcome var-
24 h. Steroid hormones were then separated from the fecal pellet iable to test if seasonal changes in diet, other than consumption of
using a centrifuge, and 2 ml of supernatant were passed through a phytoestrogens, influenced male reproductive state and thus hormone
preconditioned solid phase extraction cartridge at a flow rate of levels.
4 ml/min. The steroid hormones were then stored in these cartridges To examine seasonality in diet, relationships between rainfall and
with both ends capped. Capped samples were stored out of direct the red colobus' 13 most fed on dietary items, including the three es-
light until analysis via radioimmunoassay (RIA) for estradiol content trogenic staple foods, were examined using Pearson correlations (n =
and enzyme immunoassay (EIA) for cortisol content by MW at the 45 weeks). To further clarify if seasonality influenced the consump-
Wisconsin National Primate Research Center (WNPRC). tion of estrogenic plants, we examined the relationship between rain-
At WNPRC, the cartridges were washed with 1 ml of 5% methanol and fall and phenology of M. dura young leaves and F. natalensis young
the steroid hormones were collected using 2 ml of 100% methanol passed leaves, as well as the relationship between their availability and the
through the cartridge at a 1 ml/min flow rate. The methanol was then time the red colobus fed on each at the monthly level using Spearman
evaporated off and steroids hormones were reconstituted in 1 ml of rank correlations (n = 11 months).
100% ethanol and stored in a 4 °C refrigerator until analyses. For the estra- To examine the influence of hormonal state, climate, and phytoestro-
diol RIA, 25 μL of sample were used and recovery was 107.44%±2.53%. gen consumption on behavior, we employed a regression tree analysis.
Parallelism was demonstrated with no significant difference between Regression trees are ideally suited for the analysis of such complex data
slopes of the serial dilution of the sample pool and standard curve (De'ath and Fabricius, 2000). Data often display nonlinear relationships
(p>0.05). Inter-assay variation for the high pool was 14.62% and for and complex higher-level interactions between variables (Elith et al.,
the low pool was 9.6%, while intra-assay variation was 4.53% for the 2008). Regression trees explain variation in a single response variable
high pool and 7.51% for the low pool. The estradiol antibody used was an- using one or more explanatory variables and exhibit a number of advan-
tiserum from Holly Hill Biologicals, Inc., Hillsboro, Oregon, diluted so that tages over many commonly used statistics: they are non-parametric,
~50% of the tritiated estradiol was bound in the absence of unlabeled ste- which is often useful when considering behavioral data, well-suited for
roid (French et al., 1983). Cross-reactivity was 0.2% for estriol, 3.12% for examining complex interactions in explanatory variables, and easy to
estrone, and less than 0.01% for all other steroids tested (i.e., testosterone, interpret (De'ath, 2007; De'ath and Fabricius, 2000). Regression trees
DHT, androstenedione, progesterone, cortisone, cortisol, epiandrosterone, use a recursive partitioning approach to split the variable space into
DHEA, and androstenediol). Because the cross-reactivities with other en- smaller regions that minimize a measure of variation within the parti-
dogenous estrogens were low, it is unlikely that our antibody that is spe- tions. We used an algorithm that split the data so as to minimize the
cific to estradiol would have enough cross-reactivity with phytoestrogens sum of squares, where splitting continues until some stopping criterion
in the red colobus fecal samples to influence our results. For the cortisol is achieved; in this case that any additional splits do not improve the
EIA, 50 μL of sample were used and recovery was 125.27%±3.18%. Paral- cost-complexity criterion of the tree (i.e., that increasing complexity
lelism was demonstrated using serial dilution curves, with no significant does not significantly improve the fit of the tree) (De'ath and Fabricius,
difference between the sample pool and standards (p>0.05). Inter-assay 2000).
variation for the high pool was 18.83% and for the low pool was 16.62%, We built regression trees to predict: (1) % of time grooming, (2) cop-
while intra-assay variation was 6.24% for the high pool and 6.26% for ulation rate, and (3) rate of aggression using ‘rpart’ implemented in R
the low pool. The cortisol antibody used was R4866 (anti-cortisol-BSA) version 2.12.2 (R-Development-Core-Team, 2012; Therneau et al.,
developed by Stabenfeldt & Munro at the University of California, Davis. 2012). We predicted red colobus behaviors with median fecal cortisol
It was diluted to 1:22,000 with 50 mM bicarbonate buffer (pH 9.6), as and estradiol levels, rainfall, mean maximum temperature, and the per-
per Ziegler et al. (1995). The final hormone values are given in ng of centage of the diet consisting of estrogenic foods (% diet from all estro-
steroid hormone per g of dry feces. The dry matter content of all genic plants, % diet from staple estrogenic plants, % diet from rare
fecal samples was calculated in the field by drying 0.5 g of each sam- estrogenic plants, and % diet from each of the estrogenic staples). Parti-
ple to a constant weight. Biological validation of our cortisol EIA pre- tions near the top of the regression tree reflect strong relationships be-
viously showed this to be an effective index of stress for the red tween predictor and response variables, with the length of the branch
colobus (Wasserman, unpublished data). corresponding to the proportion of the sum of squares explained by
each split in the tree (De'ath and Fabricius, 2000). Regression trees are
Statistical analyses read from the base of the tree (i.e., the root) down the branches to the
terminal nodes (i.e., leaves), following a number of splits at each
To determine if the consumption of estrogenic plant foods interfered node. The number displayed at the tip is the mean value of the depen-
with the endocrine system and if climate played an important mediating dent variable in that subset of the data and n represents the number
role in this relationship, we tested two hypotheses: (1) temporal of samples in the data that are categorized into that leaf.
556 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562

31
A
30

Mean Maximum Temperature (oC) 29

28

27

26

25

24

23

22
100.0
B

80.0
Rainfall (mm)

60.0

40.0

20.0

0.0
Dry Wet Dry Wet Dry

30.0%
C

25.0%
% of Diet

20.0%

15.0%

10.0%

5.0%

0.0%
11
13
15
17
19
22
24
26
28
30
32
34
36
38
40
42
44
46
1
3
5
7
9

Week of Study
% Millettia dura YL % Ficus natalensis YL % Eucalyptus Bark % Nonstaple Estrogenic Parts

Fig. 1. Weekly (n=45) A) mean maximum temperature (°C) and B) amount of rainfall (mm) at Makerere University Biological Field Station, Kibale National Park, Uganda, with wet and dry
seasons indicated. C) Percent of weekly diet from each of the estrogenic staple plant foods, as well as a sum of the percent of diet from all non-staple estrogenic plant foods, of the Ugandan red
colobus monkey.
 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562 557

Results did not show as much variation for cortisol levels as they did for estra-
diol levels, with a range from 18.63 to 200.12 ng/g (n= 407). The aver-
Estrogenic plants in red colobus diet age SEM was 8.67 ng/g, with a range from 1.86 to 23.12 ng/g.

On average 10.7% of the diet was from estrogenic plants (range: Influence of phytoestrogen consumption, non-estrogenic dietary items, and
0.7%–32.4% across weeks; Fig. 1C). This was mainly due to the consump- climate on fecal hormone levels
tion of three staple estrogenic plants and alone their contribution to the
diet averaged 10.1% (range: 0.7%–31.0%). The red colobus fed on at least In the stepwise regressions predicting the median fecal estradiol
one of the three staple estrogenic plants during each week, but never levels with estrogenic dietary items and climatic variables, the best pre-
fed on all three during the same week. Of these three estrogenic staples, dictor was simply the percent of diet from M. dura young leaves (R2 =
M. dura young leaves were fed on most, with an average of 5.1% of the 0.362, pb 0.001). The relationship between fecal estradiol and M. dura
diet and a range from none during 8 weeks to 15.5%. Eucalyptus grandis consumption was also significant when only examining weeks in which
bark was fed on second most with an average of 3.4% of the diet and a it was fed on (r=0.574, pb 0.001, n=37; Fig. 3A). The median fecal cor-
range from none during 33 weeks to 31.0%. F. natalensis young leaves tisol levels were also best predicted by percent of diet from M. dura young
were fed on least of these staple estrogenic plants with an average of leaves (R2 =0.128, p=0.016) in the stepwise regressions. The relation-
1.5% of the diet and a range from none fed on during 28 weeks to ship between fecal cortisol and M. dura consumption was also significant
11.8%. As for the rare (non-staple) estrogenic plant foods (Erythrina when only examining weeks in which it was fed on (r=0.461, p=0.004,
abyssinica young leaves and flowers, Ficus sansibarica unripe fruit and n=37; Fig. 3B).
young leaves, F. natalensis unripe fruit), their contribution to the diet av- In the stepwise regression predicting median fecal estradiol level
eraged 0.6%, with a range from none fed on during 33 weeks to 6.1%. that included the red colobus' 13 most fed on dietary items (11 non-
estrogenic and 2 estrogenic) and rainfall, two significant models were
Climatic variation and plant phenology produced: 1) just M. dura young leaves (R2 =0.355, pb 0.001) and 2)
M. dura young leaves and Prunus africana young leaves (R2 = 0.424,
Rainfall was highly variable temporally, both between weeks during pb 0.001). Although P. africana consumption was not related to rainfall
the dry seasons and wet seasons and across weeks within a particular (r=−0.028, p= 0.855), it was negatively related to median estradiol
season (Fig. 1B). Total rainfall during the 45 weeks was 1304.5 mm, levels (r=−0.395, p=0.007). It is important to note that it also showed
with an average of 28.99 mm falling per week (range=0–100.33 mm). a marginally significant negative relationship with M. dura consumption
Weekly mean maximum temperature averaged 25.7 °C (range=23.1– (r=−0.267, p= 0.076).
30.3 °C) (Fig. 1A).
Young leaves of both M. dura and F. natalensis were available through- Influence of phytoestrogen consumption, hormonal state, and climate on
out the study. The availability of M. dura young leaves was related to rain- social behavior
fall with a one-month time lag (i.e., rainfall in month 1 resulted in an
increase in availability of M. dura young leaves in month 2; rs =0.778, In the regression tree predicting aggression rate, we found median
p=0.008). The availability of F. natalensis young leaves was unrelated fecal estradiol to be the primary explanatory factor, followed by the per-
to rainfall (rs =−0.075, p=0.836). Even though some young leaves of centage of diet from estrogenic staple foods (Fig. 4A; R2 = 0.437), with
F. natalensis were available year-round, there were months when the the highest rates of aggression associated with intermediate levels of
red colobus did not feed on them, while M. dura young leaves were fed fecal estradiol and lowest rates when estradiol was low. Estrogenic
on in every month. However, there was no relationship between propor- plant consumption was related to aggression rate when estradiol was
tion of their feeding time spent eating M. dura young leaves and its avail- greater than 34.35 ng / g, with more consumption relating to higher
ability (rs =0.232, p=0.492), nor did they feed on F. natalensis young rates of aggression.
leaves more when they were more available (rs =−0.041, p=0.905). With copulation rate the primary split was based on median fecal
There was a positive relationship between rainfall and the time cortisol levels, with more copulations when cortisol was lower, and sec-
spent feeding on M. dura young leaves (r= 0.414, p = 0.005), while ondary splits based on rainfall and M. dura young leaf consumption
there were no significant relationships between the time spent feeding (Fig. 4B; R 2 = 0.388). Copulations were highest when median fecal cor-
on any of the 11 most fed on non-estrogenic plant items, or the other tisol was low and consumption of M. dura young leaves was high, while
two estrogenic foods, and rainfall (all p > 0.107). they were lowest when cortisol and rainfall were high (Fig. 4B).
With grooming time, we found median fecal cortisol levels to be the
Adult male red colobus fecal estradiol and cortisol levels primary explanatory factor followed by the percent of all estrogenic
foods in the diet and then median estradiol (Fig. 4C, R2 =0.352). Red
Fecal estradiol levels averaged 80.03 ng/g, with a range from 27.32 colobus males groomed most when cortisol was low, less estrogenic
to 297.69 ng/g (n= 45; Fig. 2A). However, individual samples had foods were consumed, and estradiol was high, while they groomed the
much greater variation in estradiol levels, with a range from 19.62 to least when cortisol was high (Fig. 4C).
2237.81 ng/g (n= 407). Due to one or two individuals having extreme-
ly high levels during a few weeks (e.g., 2237.81 ng/g), certain weeks Discussion
had very high levels of variation, with a range in the standard error
of the mean (SEM) from 1.07 to 216.95 ng/g and an average of Our results suggest that plant chemistry could act as a selective pres-
38.66 ng/g. Five of the weeks had one sample >1000 ng/g, while sure on primates, and other herbivores, in ways that have yet to be fully
14 weeks had one or two samples > 500 ng/g. These one or two ex- appreciated because we found that adult male red colobus fecal estradiol
tremely high samples were responsible for the high SEM during those and cortisol levels were related to the consumption of estrogenic plant
weeks. There was no apparent reason to suspect that these high values foods, while social behaviors were related to both endogenous hormonal
were a result of methodological error, as the cortisol values for these state and consumption of phytoestrogens. By altering steroid hormone
same samples were not outliers. If the samples had been contaminated levels that are important to reproductive physiology and the tendency
or incorrectly processed, it is expected that their cortisol levels would to engage in aggressive, grooming, and mating behaviors, phytoestrogens
have also reflected this error. could directly alter the fitness of the individuals ingesting them. The
Mean fecal cortisol levels averaged 71.58 ng/g dry feces, with a strongest relationship we documented was the positive association be-
range from 33.17 to 107.59 ng/g (n= 45; Fig. 2B). Individual samples tween median estradiol level and the consumption of M. dura young
558 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562

500 A

Mean Fecal Estradiol Level (ng / g)

400

300

200

100

0
120
B

100
Mean Fecal Cortisol Level (ng / g)

80

60

40

20

0
1
3
5
7
9
11
13
15
17
19
22
24
26
28
30
32
34
36
38
40
42
44
46
Week of Study

Fig. 2. Weekly mean (+/−SEM) fecal A) estradiol level and B) cortisol level (ng/g dry feces) from ten adult male red colobus living in one group at Kibale National Park, Uganda.

leaves. M. dura is a legume of the Papilionoideae subfamily, which is supports the conclusion that the consumption of estrogenic plant
known to contain estrogenic isoflavonoids (Reynaud et al., 2005). Studies foods affected red colobus physiology. The relationship with fecal estra-
demonstrate that the ingestion of papilionoids by livestock can dra- diol may simply have been due to the RIA antibody binding to metabo-
matically impair reproduction (Adams, 1990, 1995; Bennetts and lites of the phytoestrogens passing through the monkeys, but this is
Underwood, 1951), and similar detrimental effects on fertility may unlikely due to our antibody's cross-reactivities. However, the addition-
occur in humans (Cederroth et al., 2010a). However, it is also suggested al relationship with cortisol provided strong evidence that the
that phytoestrogens may provide health benefits such as cancer preven- phytoestrogens were absorbed and affected the endocrine system
tion or alleviation of menopausal disorders (Dixon, 2004; Leitman et al., through the HPG and HPA axes, since metabolites of phytoestrogens
2010; Ososki and Kennelly, 2003; Setchell and Cassidy, 1999). Based on would not bind to the cortisol antibody. Increased cortisol production
these studies and the relationships found here, it is possible that from phytoestrogen consumption may result in a synergistic threat to
M. dura exerted an important influence on the red colobus endocrine sys- endangered primates which live in environments with unusually high
tem. However, as red colobus have had a long evolutionary relationship proportions of estrogenic plants, both through the suppression of their
with this native species, it is possible that they have evolved adaptations immune and reproductive systems by elevated cortisol (Sapolsky,
to protect against any potential endocrine disruption caused by ingesting 2005) and by altered fertility through phytoestrogen interaction with
leaves of M. dura (Wynne-Edwards, 2001). estrogen receptors (Cederroth et al., 2010a).
Feeding on M. dura young leaves also had a positive relationship The relationships found here between the consumption of estrogenic
with fecal cortisol, and it is likely that the phytoestrogens interacted plants and red colobus hormone levels were correlative; therefore, as
not only with the HPG axis, but also with the HPA axis. Studies on am- with any non-experimental approach, there is the possibility that these
phibians have shown that anthropogenic endocrine disruptors in pesti- relationships were due to spurious correlations. However, by including
cides can lead to an increase in stress hormone production (Hayes et al., rainfall and temperature in the multiple regressions, we have attempted
2006). Further, estrogens are able to alter the negative feedback loop of to control for other factors related to seasonality that could influence hor-
the HPA axis, thus changing production of glucocorticoids (Weiser and mone levels. In the tropics, rainfall is the most important factor influenc-
Handa, 2009). Therefore, it is possible that the phytoestrogens of ing seasonality. Because the consumption of M. dura young leaves was the
M. dura young leaves influenced cortisol levels in red colobus. This only significant variable for predicting both median estradiol and cortisol
 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562 559

A 55

50

Median Fecal Estradiol Level (ng / g)

45

40

35

30

25
y = 92.546x + 32.718
R² = 0.33
20
0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 16.0%
Percent of Diet from Millettia duraYoung Leaves

B 120
110
Median Fecal Cortisol Level (ng / g)

100

90

80

70

60

50

40

30
y = 190.52x + 56.608
R² = 0.2115
20
0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 16.0%
Percent of Diet from Millettia dura Young Leaves

Fig. 3. Relationships between percent of weekly diet from Millettia dura young leaves and A) median fecal estradiol level and B) median fecal cortisol level for weeks when the red colobus fed on
that particular estrogenic plant item (n=37). Hormone levels are weekly calculations from ten adult male red colobus living in the same group at Kibale National Park, Uganda. Correlations are
significant (pb 0.05).

in the multiple regressions, we think this strongly supports that this es- estradiol levels. In the multiple regression model incorporating non-
trogenic plant influenced red colobus hormone levels and that this rela- estrogenic dietary items, estrogenic M. dura young leaves explained
tionship was not driven by another seasonal process. most of the variation in red colobus estradiol levels. Of the other 12
Furthermore, red colobus do not have a breeding season, as births most fed on items, only non-estrogenic P. africana young leaves slightly
occur in all months of the year, and Kibale is known for being relatively improved the model. However, their consumption was negatively corre-
aseasonal in terms of plant phenology and availability of primate plant lated with consumption of M. dura and negatively correlated with estradi-
foods (Chapman et al., 2005; Struhsaker, 1997). Supporting the ol. Thus, this relationship was likely due to the red colobus feeding less on
aseasonality of Kibale, we found none of the other 13 most fed on dietary M. dura in weeks they fed more on P. africana. Alternatively, there is pub-
items correlated with rainfall, besides M. dura young leaves. Thus, large lished evidence that P. africana interacts with the endocrine system as it is
changes in red colobus male reproductive state related to dietary quality used to prevent and treat prostate disorders and its extract displaces both
or climatic variables were not expected. Variation in the consumption of estrogens and androgens (Shenouda et al., 2007). Although our previous-
estrogenic plants remains as the most plausible explanation for the ly reported transfection assay did not find this species to have estrogenic
changes in hormonal state we found, and these changes could translate activity at either estrogen receptor (Wasserman et al., 2012), steroidal ac-
into variation in fertility. tivity through other mechanisms may be responsible for the negative re-
Additionally, if components of the diet other than phytoestrogens lationship between its consumption and estradiol levels found here.
were responsible for the relationships found here it would be expected Therefore, it does not appear that the relationships between consumption
that consumption of other staple dietary items would relate to rainfall of estrogenic plants and red colobus hormone levels found here were due
(none were) and non-estrogenic items would explain variation in to seasonal changes in non-estrogenic components of diet. This was
560 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562

A) Aggression rate [N acts/hour] B) Copulation rate [N acts/hour]

| |
0.103 0.398
Median n=45 Median n=45
estradiol estradiol
< 33.35 >=33.35

Median Median
0.0114 cortisol cortisol
0.133 >=67.43 < 67.43
n=11 n=34

Median Median
estradiol estradiol
>=34.35 < 34.35
% of diet % of diet
Rainfall 0.270 Rainfall consisting 0.511 consisting
% of diet consisting n=21 of Millettia n=24 of Millettia
% of diet consisting >=17 < 17
of estrogenic of estrogenic dura dura
0.207 0.372 young leaves young leaves
staple foods staple foods 0.287
0.0931 n=13 n=8 < 1.3 >=1.3
< 6.75 >=6.75 n=7
n=27 0.354 0.575
0.0436 0.118
n=7 n=17
n=9 n=18

C) Percentage of time spent grooming [%]

|
5.38
n=45
Median Median
cortisol cortisol
>=79.91 < 79.91

3.975 5.891
n=12 n=33
% of diet consisting % of diet consisting
of estrogenic foods of estrogenic foods
>=12.05 < 12.05

4.609 Median 6.532 Median

n=11 estradiol n=22 estradiol
< 33.7 >=33.7
5.386 7.067
n=7 n=15

Fig. 4. Regression trees explaining variance in A) number of aggressive interactions per hour, B) number of copulations per hour, and C) % of time spent grooming for red colobus males using
median fecal cortisol and estradiol, phytoestrogen consumption, rainfall, and temperature as predictor variables. Values in the ovals represent the mean value of the dependent variable at that
node and the number of samples falling along that branch. Values in the rectangles represent the leaves of the regression tree and correspond to the mean value of the dependent variable and
the number of samples that fall along that particular branch. For example, in Fig. 4C; when median cortisol was lower than 79.91 and the consumption of estrogenic plants was greater than
12.05% of the diet, the mean percent of time spent grooming was 4.609% and 11 of our samples fell into this category. For a detailed description of the mechanics and interpretation of regression
trees see De'ath and Fabricius (2000); in brief the tree is created by repeatedly splitting the data on a single explanatory variable such that the variance in the dependent variable is minimized.
The splitting is continued trying all explanatory variables, until the added variance explained is outweighted by the complexity of adding more splits.

further supported by the fact that P. africana did not show seasonality in (Simon et al., 2004). Male rats fed a high-isoflavone diet spent less
consumption, while M. dura and estradiol did. time in social interactions and had higher corticosterone response to
Our results showing behavioral relationships with the consumption stress than rats fed an isoflavone-free diet (Hartley et al., 2003). Similar-
of estrogenic plants provided further support that the correlations ly, the ERβ agonist equol increased aggression and anxiety in male rats
found here between phytoestrogens and red colobus estradiol and corti- that were exposed neonatally (Patisaul and Bateman, 2008).
sol levels had biological significance. Although estradiol levels were most In contrast, our finding that the more the red colobus fed on M. dura
important to predicting rates of aggression and cortisol levels were most young leaves the more they copulated does not agree with other studies
important to predicting rates of copulations and percent of time spent that document fewer copulations with consumption of phytoestrogens
grooming, the consumption of estrogenic plants was secondarily impor- (Jaroenporn et al., 2006; Wisniewski et al., 2003). It may be that the nu-
tant to predicting these social behaviors. We found that the greater the tritional benefits obtained from this plant item outweigh any reproduc-
consumption of estrogenic plants, the more the red colobus males dis- tive costs through increased cortisol and estradiol or the red colobus
played aggressive behaviors and copulated and the less they groomed. may use this plant as a chemical cue for reproductive behavior, where
These results are supported by captive studies that documented changes males are confused by the altered physiology of females due to feeding
in behavior due to phytoestrogens. In captive adult male cynomolgus on the same plant (Berger et al., 1981; Higham et al., 2007; Leopold et
monkeys (Macaca fascicularis), a 15-month soy-based high-isoflavone al., 1976). In future studies we will examine tradeoffs between nutrient
diet resulted in increased aggressive behaviors (i.e., 67% more frequent content and phytoestrogen presence in the foods of the red colobus, as
compared to individuals fed an isoflavone-free diet) and a decrease in well as the physiological and behavioral effects of these estrogenic
affiliative behaviors (i.e., 68% less time in body contact and 30% more plants on female red colobus.
time alone) (Simon et al., 2004). It was postulated that these effects Finally, although red colobus did not feed more heavily on M. dura
were likely due to the weaker action of isoflavones on ERβ than estradiol, young leaves when they were more available, they did feed more heavi-
thus reducing the inhibition of the aggression‐promoting action of ERα ly on them during weeks with more rain. This might mean that variation
 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562 561

in the phytoestrogen content of these leaves in response to increased of the National Institutes of Health (NIH), to the Wisconsin National Pri-
rainfall is more important than availability in determining when and mate Research Center, University of Wisconsin-Madison. This research
how much red colobus will feed on them. This is further supported by was conducted in part at a facility constructed with support from Re-
the fact that consumption of the other 13 most fed on dietary items search Facilities Improvement Program grant numbers RR15459-01
was not related to rainfall. Increased phytoestrogen content may either and RR020141-01. This publication's contents are solely the responsibil-
be attractive as a form of self-medication (Forbey et al., 2009; Glander, ity of the authors and do not necessarily represent the official views of
1980; Huffman, 1997; Strier, 1993) or act as a feeding deterrent if phy- NCRR or NIH. These funding sources had no involvement with our
toestrogen consumption reduces fertility (Harborne, 1993; Hughes, research other than providing financial support. Research conducted
1988; Wynne-Edwards, 2001). Increases in isoflavone concentration complied with all regulations regarding the study of field animals and
(i.e., phytoestrogens) with increased precipitation have been docu- with Ugandan and U.S. laws. We would like to thank Musunguzi Hillary,
mented in soybeans (Glycine max), which are members of the same sub- Baguma Charles, Katusabe Swaibu, Mutegeki Richard, Patrick Omeja,
family (Papilionoideae) of legumes as M. dura (Morrison et al., 2010). Dennis Twinomugisha, Kaganzi Clovis, Lauren Chapman, Jessica
Variation in the nutritional content of plant foods is important to pri- Rothman, and Julie Wasserman for providing assistance in the field at
mate diet selection (Milton, 1979; Rothman et al., 2006, 2011) and can Kibale. Tyrone Hayes, Isao Kubo, Paul Falso, Alexandra Taylor-Gutt,
affect reproduction and steroid production (Lu et al., 2011), so tradeoffs Mariah Hopkins, Tarek Milleron, Karen Strier, and Andrew Ritchie pro-
between phytoestrogen consumption and nutrient intake may be criti- vided useful comments throughout the project. We also thank Kim
cal to a primate's decision to feed on a specific plant and warrant future Wallen and the anonymous reviewers for their insightful comments.
study.
References
Conclusions
Adams, N.R., 1990. Permanent infertility in ewes exposed to plant oestrogens. Aust.
Our results show that the ingestion of estrogenic plants related to Vet. J. 67, 197–201.
Adams, N.R., 1995. Detection of the effects of phytoestrogens on sheep and cattle. J. Anim.
adult male red colobus fecal estradiol and cortisol levels, as well as ag- Sci. 73, 1509–1515.
gressive, mating, and grooming behaviors. Further, rainfall was important Bauchop, T., Martucci, R.W., 1968. Ruminant-like digestion of the langur monkey. Sci-
for understanding seasonality in the consumption of certain estrogenic ence 161, 698–700.
Bennetts, H.W., 1946. Metaplasia in the sex organs of castrated male sheep maintained
plant foods, but not to variation in estradiol and cortisol levels or social
on early subterranean clover pastures. Aust. Vet. J. 22, 70–78.
behaviors. Climate can affect red colobus phytoestrogen consumption Bennetts, H.W., Underwood, E.J., 1951. The oestrogenic effects of subterranean clover
and hormone levels through three main mechanisms: (1) effects on (Trifolium subterraneum); uterine maintenance in the ovariectomised ewe on clo-
ver grazing. Aust. J. Exp. Biol. Med. Sci. 29, 249–253.
food availability/nutritional content, (2) effects on phytoestrogen levels,
Berger, P.J., Negus, N.C., Sanders, E.H., Gardner, P.D., 1981. Chemical triggering of repro-
and (3) effects on hormone levels. Although these mechanisms are not duction in Microtus montanus. Science 214, 69–70.
mutually exclusive, our results suggest that the effect of rainfall on phyto- Cederroth, C.R., Auger, J., Zimmermann, C., Eustache, F., Nef, S., 2010a. Soy, phyto-
estrogen levels of M. dura young leaves is the most important factor af- oestrogens and male reproductive function: a review. Int. J. Androl. 33, 304–316.
Cederroth, C.R., Zimmermann, C., Beny, J.L., Schaad, O., Combepine, C., Descombes, P.,
fecting the timing and magnitude of red colobus feeding on estrogenic Doerge, D.R., Pralong, F.P., Vassalli, J.D., Nef, S., 2010b. Potential detrimental effects
plants. Subsequently, the amount of M. dura young leaves in the diet of a phytoestrogen-rich diet on male fertility in mice. Mol. Cell. Endocrinol. 321,
related to changes in red colobus estradiol and cortisol levels, which 152–160.
Chapman, C.A., Peres, C.A., 2001. Primate conservation in the new millennium: the role
influenced social behaviors. We suggest that the consumption of estro- of scientists. Evol. Anthropol. 10, 16–33.
genic plant foods has important implications for their health and fitness Chapman, C.A., Wrangham, R.W., Chapman, L.J., Kennard, D.K., Zanne, A.E., 1999. Fruit
through changes in physiology and behavior. In addition, this study raises and flower phenology at two sites in Kibale National Park, Uganda. J. Trop. Ecol.
15, 189–211.
the possibility of additive effects of natural plant-based endocrine dis- Chapman, C.A., Chapman, L.J., Gillespie, T.R., 2002. Scale issues in the study of primate
ruptors and climate change on primates, which might threaten their foraging: red colobus of Kibale National Park. Am. J. Phys. Anthropol. 117, 349–363.
long-term survival. Three findings warrant this: (1) climate change in Chapman, C.A., Chapman, L.J., Struhsaker, T.T., Zanne, A.E., Clark, C.J., Poulsen, J.R., 2005.
A long-term evaluation of fruiting phenology: importance of climate change.
the form of increased rainfall (c. 300 mm more per year now as com-
J. Trop. Ecol. 21, 31–45.
pared to the early 1900's) has been documented in Kibale (Chapman et Chapman, C.A., Chapman, L.J., Jacob, A.L., Rothman, J.M., Omeja, P., Reyna-Hurtado, R.,
al., 2005), (2) red colobus were found to feed more heavily on M. dura Hartter, J., Lawes, M.J., 2010. Tropical tree community shifts: implications for wild-
life conservation. Biol. Conserv. 143, 366–374.
young leaves during wetter weeks, and (3) agricultural studies have doc-
Colborn, T., vom Saal, F.S., Soto, A.M., 1993. Developmental effects of endocrine-disrupting
umented increased levels of phytoestrogens with increased rainfall in soy chemicals in wildlife and humans. Environ. Health Perspect. 101, 378–384.
(Morrison et al., 2010). Further study of interactions between estrogenic De'ath, G., 2007. Boosted trees for ecological modeling and prediction. Ecology 88,
plants and primates in a natural setting should provide many exciting 243–251.
De'ath, G., Fabricius, K.E., 2000. Classification and regression trees: a powerful yet sim-
new discoveries relevant to a broad range of fields, including medicine, ple technique for ecological data analysis. Ecology 81, 3178–3192.
conservation, ecology, and evolution. Dixon, R.A., 2004. Phytoestrogens. Annu. Rev. Plant Biol. 55, 225–261.
Elith, J., Leathwick, J.R., Hastie, T., 2008. A working guide to boosted regression trees. J. Anim.
Ecol. 77, 802–813.
Acknowledgments Forbey, J.S., Harvey, A.L., Huffman, M.A., Provenza, F.D., Sullivan, R., Tasdemir, D., 2009.
Exploitation of secondary metabolites by animals: a response to homeostatic chal-
We would like to thank the Uganda Wildlife Authority and Uganda lenges. Integr. Comp. Biol. 49, 314–328.
French, J.A., Abbott, D.H., Scheffler, G., Robinson, J.A., Goy, R.W., 1983. Cyclic excretion of uri-
National Council for Science and Technology for providing permission nary estrogens in female tamarins (Saguinus oedipus). J. Reprod. Fertil. 68, 177–184.
to conduct this research. MW received funding from the National Sci- Glander, K.E., 1980. Reproduction and population growth in free-ranging mantled
ence Foundation (DDIG #0823651 and Graduate Research Fellowship howling monkeys. Am. J. Phys. Anthropol. 53, 25–36.
Gogarten, J.F., Brown, L.M., Chapman, C.A., Cords, M., Doran-Sheehy, D., Fedigan, L.M.,
Program), the International Primatological Society, the University of Grine, F.E., Perry, S., Pusey, A.E., Sterck, E.H.M., Wich, S.A., Wright, P.C., 2012. Seasonal
California, Berkeley (UCB) Department of Environmental Science, Poli- mortality patterns in non-human primates: Implications for variation in selection
cy, and Management, UCB Center for African Studies, the UCB Chang- pressures across environments. Evolution 66 (10), 3252–3266. http://dx.doi.org/
10.1111/j.1558-5646.2012.01668.x.
Lin Tien Scholars Program (funded by the Philomathia Foundation),
Guillette Jr., L.J., 2000. Contaminant-induced endocrine disruption in wildlife. Growth
and McGill University's Tomlinson Postdoctoral Fellowship to support Horm. IGF Res. 10 (Suppl. B), S45–S50.
this research. JFG was supported by the National Science Foundation Gultekin, E., Yildiz, F., 2006. Introduction to phytoestrogens. In: Yildiz, F. (Ed.),
Graduate Research Fellowship under Grant No. (NSF 11–582). This pub- Phytoestrogens in Functional Foods. CRC Press, Boca Raton, Florida, pp. 3–18.
Hadley, M., 1999. Endocrinology, 5th ed. Prentice Hall, Upper Saddle River, New Jersey.
lication was made possible in part by Grant Number P51 RR000167 Harborne, J., 1993. Introduction to Ecological Biochemistry, 4th ed. Elsevier Academic
from the National Center for Research Resources (NCRR), a component Press, San Francisco, California.
562 M.D. Wasserman et al. / Hormones and Behavior 62 (2012) 553–562

Hartley, D.E., Edwards, J.E., Spiller, C.E., Alom, N., Tucci, S., Seth, P., Forsling, M.L., File, Setchell, K.D., Cassidy, A., 1999. Dietary isoflavones: biological effects and relevance to
S.E., 2003. The soya isoflavone content of rat diet can increase anxiety and stress human health. J. Nutr. 129, 758S–767S.
hormone release in the male rat. Psychopharmacology 167, 46–53. Setchell, K.D., Clerici, C., 2010. Equol: history, chemistry, and formation. J. Nutr. 140,
Hayes, T.B., 2005. Welcome to the revolution: integrative biology and assessing the im- 1355S–1362S.
pact of endocrine disruptors on environmental and public health. Integr. Comp. Shenouda, N.S., Sakla, M.S., Newton, L.G., Besch-Williford, C., Greenberg, N.M.,
Biol. 45, 321–329. MacDonald, R.S., Lubahn, D.B., 2007. Phytosterol Pygeum africanum regulates pros-
Hayes, T.B., Collins, A., Lee, M., Mendoza, M., Noriega, N., Stuart, A.A., Vonk, A., 2002. tate cancer in vitro and in vivo. Endocrine 31, 72–81.
Hermaphroditic, demasculinized frogs after exposure to the herbicide atrazine at Simon, N.G., Kaplan, J.R., Hu, S., Register, T.C., Adams, M.R., 2004. Increased aggressive be-
low ecologically relevant doses. Proc. Natl. Acad. Sci. U. S. A. 99, 5476–5480. havior and decreased affiliative behavior in adult male monkeys after long-term con-
Hayes, T.B., Case, P., Chui, S., Chung, D., Haeffele, C., Haston, K., Lee, M., Mai, V.P., Marjuoa, sumption of diets rich in soy protein and isoflavones. Horm. Behav. 45, 278–284.
Y., Parker, J., Tsui, M., 2006. Pesticide mixtures, endocrine disruption, and amphibian Snaith, T.V., Chapman, C.A., Rothman, J.M., Wasserman, M.D., 2008. Bigger groups have
declines: are we underestimating the impact? Environ. Health Perspect. 114, 40–50. fewer parasites and similar cortisol levels: a multi-group analysis in red colobus
Heistermann, M., Tari, S., Hodges, J.K., 1993. Measurement of faecal steroids for moni- monkeys. Am. J. Primatol. 70, 1072–1080.
toring ovarian function in New World primates, Callitrichidae. J. Reprod. Fertil. 99, Sousa, M.B., Ziegler, T.E., 1998. Diurnal variation on the excretion patterns of fecal steroids
243–251. in common marmoset (Callithrix jacchus) females. Am. J. Primatol. 46, 105–117.
Higham, J.P., Ross, C., Warren, Y., Heistermann, M., MacLarnon, A.M., 2007. Reduced re- Strier, K.B., 1993. Menu for a monkey. Nat. Hist. 102, 34–43.
productive function in wild baboons (Papio hamadryas anubis) related to natural Struhsaker, T., 1997. Ecology of an African Rain Forest: Logging in Kibale and the Con-
consumption of the African black plum (Vitex doniana). Horm. Behav. 52, 384–390. flict between Conservation and Exploitation. University Press of Florida, Gaines-
Huffman, M.A., 1997. Current evidence for self-medication in primates: a multidisciplinary ville, Florida.
perspective. Yearb. Phys. Anthropol. 40, 171–200. Struhsaker, T.T., 2005. Conservation of red colobus and their habitats. Int. J. Primatol.
Hughes Jr., C.L., 1988. Phytochemical mimicry of reproductive hormones and modulation of 26, 525–538.
herbivore fertility by phytoestrogens. Environ. Health Perspect. 78, 171–174. Struhsaker, T.T., 2008. Procolobus rufomitratus ssp. tephrosceles, IUCN Red List of
Jaroenporn, S., Malaivijitnond, S., Wattanasirmkit, K., Trisomboon, H., Watanabe, G., Threatened Species, Version 20112 ed. IUCN . 2011.
Taya, K., Cherdshewasart, W., 2006. Effects of Pueraria mirifica, an herb containing Therneau, T.M., Atkinson, B., Ripley, M.B., 2012. rpart: Recursive Partitioning, R package ver-
phytoestrogens, on reproductive organs and fertility of adult male mice. Endocrine sion 3.1-51. http://mayoresearch.mayo.edu/mayo/research/biostat/splusfunctions.cfm.
30, 93–101. Thornton, J.W., 2001. Evolution of vertebrate steroid receptors from an ancestral estro-
Lambert, J.E., 1998. Primate digestion: interactions among anatomy, physiology, and gen receptor by ligand exploitation and serial genome expansions. Proc. Natl. Acad.
feeding ecology. Evol. Anthropol. 7, 8–20. Sci. U. S. A. 98, 5671–5676.
Leitman, D.C., Paruthiyil, S., Vivar, O.I., Saunier, E.F., Herber, C.B., Cohen, I., Tagliaferri, Touma, C., Palme, R., 2005. Measuring fecal glucocorticoid metabolites in mammals and
M., Speed, T.P., 2010. Regulation of specific target genes and biological responses birds: the importance of validation. Ann. N. Y. Acad. Sci. 1046, 54–74.
by estrogen receptor subtype agonists. Curr. Opin. Pharmacol. 10, 629–636. Wasser, S.K., Risler, L., Steiner, R.A., 1988. Excreted steroids in primate feces over the
Leopold, A.S., Erwin, M., Oh, J., Browning, B., 1976. Phytoestrogens: adverse effects on menstrual cycle and pregnancy. Biol. Reprod. 39, 862–872.
reproduction in California quail. Science 191, 98–100. Wasserman, M.D., Taylor-Gutt, A., Rothman, J.M., Chapman, C.A., Milton, K., Leitman,
Lu, A., Beehner, J.C., Czekala, N.M., Koenig, A., Larney, E., Borries, C., 2011. Phytochemicals and D.C., 2012. Estrogenic plant foods of red colobus monkeys and mountain gorillas
reproductive function in wild female Phayre's leaf monkeys (Trachypithecus phayrei in uganda. Am. J. Phys. Anthropol. 148, 88–97.
crepusculus). Horm. Behav. 59, 28–36. Weingrill, T., Gray, D.A., Barrett, L., Henzi, S.P., 2004. Fecal cortisol levels in free-ranging
Mazur, W., Adlercreutz, H., 1998. Naturally occurring oestrogens in food. Pure Appl. Chem. female chacma baboons: relationship to dominance, reproductive state and envi-
70, 1759–1776. ronmental factors. Horm. Behav. 45, 259–269.
Milnes, M.R., Bermudez, D.S., Bryan, T.A., Edwards, T.M., Gunderson, M.P., Larkin, I.L., Moore, Weiser, M.J., Handa, R.J., 2009. Estrogen impairs glucocorticoid dependent negative
B.C., Guillette Jr., L.J., 2006. Contaminant-induced feminization and demasculinization of feedback on the hypothalamic–pituitary–adrenal axis via estrogen receptor alpha
nonmammalian vertebrate males in aquatic environments. Environ. Res. 100, 3–17. within the hypothalamus. Neuroscience 159, 883–895.
Milton, K., 1979. Factors influencing leaf choice by howler monkeys: a test of some hy- Whitten, P.L., Patisaul, H.B., 2001. Cross-species and interassay comparisons of phyto-
potheses of food selection by generalist herbivores. Am. Nat. 114, 362–378. estrogen action. Environ. Health Perspect. 109 (Suppl. 1), 5–20.
Milton, K., 1980. The Foraging Strategy of Howler Monkeys: A Study in Primate Eco- Whitten, P.L., Brockman, D.K., Stavisky, R.C., 1998. Recent advances in noninvasive
nomics. Columbia University Press, New York. techniques to monitor hormone–behavior interactions. Am. J. Phys. Anthropol.
Morrison, M.J., Cober, E.R., Saleem, M.F., McLaughlin, N.B., Fregeau-Reid, J., Ma, B.L., (Suppl. 27), 1–23.
Woodrow, L., 2010. Seasonal changes in temperature and precipitation influence Wingfield, J.C., 2005. The concept of allostasis: coping with a capricious environment.
isoflavone concentration in short-season soybean. Field Crop Res. 117, 113–121. J. Mammal. 86, 248–254.
Ososki, A.L., Kennelly, E.J., 2003. Phytoestrogens: a review of the present state of re- Wingfield, J.C., Sapolsky, R.M., 2003. Reproduction and resistance to stress: when and
search. Phytother. Res. 17, 845–869. how. J. Neuroendocrinol. 15, 711–724.
Patisaul, H.B., Bateman, H.L., 2008. Neonatal exposure to endocrine active compounds Wingfield, J.C., Moore, M.C., Farner, D.S., 1983. Endocrine responses to inclement weather in
or an ERbeta agonist increases adult anxiety and aggression in gonadally intact naturally breeding populations of white-crowned sparrows (Zonotrichia leucophrys
male rats. Horm. Behav. 53, 580–588. pugetensis). Auk 100, 56–62.
Propper, C.R., 2005. The study of endocrine-disrupting compounds: past approaches Wisniewski, A.B., Klein, S.L., Lakshmanan, Y., Gearhart, J.P., 2003. Exposure to genistein dur-
and new directions. Integr. Comp. Biol. 45, 194–200. ing gestation and lactation demasculinizes the reproductive system in rats. J. Urol. 169,
R-Development-Core-Team, 2012. R: A Language and Environment for Statistical Com- 1582–1586.
puting. R Foundation for Statistical Computing, Vienna, Austria. Wynne-Edwards, K.E., 2001. Evolutionary biology of plant defenses against herbivory
Reynaud, J., Guilet, D., Terreux, R., Lussignol, M., Walchshofer, N., 2005. Isoflavonoids in and their predictive implications for endocrine disruptor susceptibility in verte-
non-leguminous families: an update. Nat. Prod. Rep. 22, 504–515. brates. Environ. Health Perspect. 109, 443–448.
Rothman, J.M., Van Soest, P.J., Pell, A.N., 2006. Decaying wood is a sodium source for Ziegler, T.E., Wittwer, D.J., 2005. Fecal steroid research in the field and laboratory: im-
mountain gorillas. Biol. Lett. 2, 321–324. proved methods for storage, transport, processing, and analysis. Am. J. Primatol.
Rothman, J.M., Raubenheimer, D., Chapman, C.A., 2011. Nutritional geometry: gorillas pri- 67, 159–174.
oritize non-protein energy while consuming surplus protein. Biol. Lett. 7, 847–849. Ziegler, T.E., Scheffler, G., Snowdon, C.T., 1995. The relationship of cortisol levels to so-
Sapolsky, R.M., 2005. The influence of social hierarchy on primate health. Science 308, cial environment and reproductive functioning in female cotton-top tamarins,
648–652. Saguinus oedipus. Horm. Behav. 29, 407–424.