hormonal correlates of dominance in meerkats (suricata suricatta)
TRANSCRIPT
www.elsevier.com/locate/yhbeh
Hormones and Behavior 46 (2004) 141–150
Hormonal correlates of dominance in meerkats (Suricata suricatta)
Anne A. Carlson,a,* Andrew J. Young,a Andrew F. Russell,a Nigel C. Bennett,b
Alan S. McNeilly,c and Tim Clutton-Brocka
aLarge Animal Research Group, Department of Zoology, University of Cambridge, Cambridge CB2 3EJ, UKbMammal Research Institute, Department of Zoology and Entomology, University of Pretoria 0002, South Africa
cMRC HRSU, Centre for Reproductive Biology, The University of Edinburgh, Academic Centre, Edinburgh EH16 4SB, UK
Received 27 January 2003; revised 22 April 2003; accepted 19 January 2004
Available online 18 May 2004
Abstract
In cooperatively breeding meerkats (Suricata suricatta), individuals typically live in extended family groups in which the dominant male
and female are the primary reproductives, while their offspring delay dispersal, seldom breed, and contribute to the care of subsequent litters.
Here we investigate hormonal differences between dominants and subordinates by comparing plasma levels of luteinizing hormone (LH),
estradiol and cortisol in females, and testosterone and cortisol in males, while controlling for potential confounding factors. In both sexes,
hormone levels are correlated with age. In females, levels of sex hormone also vary with body weight and access to unrelated breeding
partners in the same group: subordinates in groups containing unrelated males have higher levels of LH and estradiol than those in groups
containing related males only. When these effects are controlled, there are no rank-related differences in circulating levels of LH among
females or testosterone among males. However, dominant females show higher levels of circulating estradiol than subordinates. Dominant
males and females also have significantly higher cortisol levels than subordinates. Hence, we found no evidence that the lower levels of
plasma estradiol in subordinate females were associated with high levels of glucocorticoids. These results indicate that future studies need to
control for the potentially confounding effects of age, body weight, and access to unrelated breeding partners before concluding that there are
fundamental physiological differences between dominant and subordinate group members.
D 2004 Elsevier Inc. All rights reserved.
Keywords: Suricata suricatta; Luteinizing hormone; Reproductive suppression; Steroid; Dominance; Glucocorticoid; Stress
do not reproduce because unrelated breeding partners are
IntroductionIn cooperatively breeding vertebrates, individuals typical-
ly live in extended family groups in which all members
participate in the care of the young, though only the dominant
male and female tend to reproduce (Emlen, 1995; Solomon
and French, 1997; Stacey and Koenig, 1990). As a result,
researchers have attempted to understand why subordinate
males and females fail to breed within these cooperative
groups (e.g., Abbott et al., 1988; Creel et al., 1997; French,
1997; Moehlman and Hofer, 1997). Alternative hypotheses
include the possibility that (1) subordinate individuals are
reproductively suppressed and physiologically incapable of
breeding (reviewed by Creel and Creel, 1991; French, 1997;
Schoech et al., 1996), and (2) subordinate males and females
0018-506X/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.yhbeh.2004.01.009
* Corresponding author. CRES, Zoological Society of San Diego, PO
Box 120551, San Diego, CA 92112-0551. Fax: +1-619-557-3959.
E-mail address: [email protected] (A.A. Carlson).
lacking in their natal group (e.g., Clutton-Brock et al., 2001;
O’Riain et al., 2000; Reyer et al., 1986; Rickard and Bennett,
1997; Schoech et al., 1996).
Attempts to understand the proximate causes of the
reduced reproductive rates of subordinates are commonly
based on comparisons of the hormonal profiles of dominants
and subordinates. A number of studies have documented
significantly increased concentrations of luteinizing hor-
mone (LH) or sex steroids in dominants compared with
subordinates (Bennett et al., 1996; Creel et al., 1992, 1997;
Keane et al., 1994; Poiani and Fletcher, 1994; Schoech et
al., 1991; Ziegler et al., 1987), though others have found no
significant differences (Bennett et al., 1997; French et al.,
1989; Ginther et al., 2001; Khan et al., 2001; Mays et al.,
1991; Smith et al., 1997). In addition, studies have com-
monly compared levels of cortisol and other stress hor-
mones between dominants and subordinates due to
hypothesized links between glucocorticoid levels and repro-
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150142
ductive suppression (reviewed by Creel, 2001). Here, too,
results vary, with dominants showing similar levels to
subordinates in male Alpine marmots and male dwarf
mongooses, but higher levels of cortisol in female wild
dogs, male Alpine marmots, female dwarf mongooses, and
common marmosets (Creel, 2001).
Though differences in hormonal status between domi-
nants and subordinates are relatively common and may play
an important role in the control of reproduction, it is not yet
clear whether they are a consequence of contrasts in
dominance per se or of correlated variation in other traits
related to hormone levels. For example, dominant individ-
uals are often older and heavier than subordinates (Clutton-
Brock et al., 2001; Creel and Creel, 1991, 2002; O’Riain et
al., 2000). In addition, dominant individuals typically have
access to an unrelated opposite-sexed individual whereas
this may not be the case for subordinates (O’Riain et al.,
2000; Schoech et al., 1996). As a result, apparent differ-
ences in the hormonal levels of dominant and subordinate
individuals in cooperatively breeding species may be due to
the co-variance of other variables with dominance status
rather than an effect of dominance per se. Consequently,
controlling for age, weight, and access to unrelated breeding
partners is likely to be particularly important in any attempt
to disentangle the effects of social rank from those of
associated variables.
In this study, we investigate contrasts in sex hormones and
cortisol between dominant and subordinate meerkats, and
attempt to discriminate between the effects of social status
and those of associated variables, including age, weight, and
access to unrelated breeding partners. Meerkats are desert-
adapted mongooses (<1 kg) that live in groups of 3–45
members in Southern Africa (Russell et al., in press) in which
reproduction is usually limited to one dominant male and one
dominant female. However, subordinates of both sexes do
sometimes breed (Griffin et al., 2003; Russell et al., 2003a),
and their probability of doing so increases in the presence of
unrelated breeding partners within the group (Clutton-Brock
et al., 2001; O’Riain et al., 2000). The probability that
subordinate females breed also increases with their age and
body weight, though after age and weight effects are con-
trolled, dominant females still breed more frequently than
subordinates (Clutton-Brock et al., 2001). Subordinates of
both sexes contribute to pup care by babysitting pups at the
breeding burrow during their first month of life, and feeding
them with small prey items as they travel with the group
during their subsequent 2 months of dependence (Brotherton
et al., 2001; Clutton-Brock et al., 1998, 2002).
Previous research on hormonal variation in meerkats
shows that dominant females have higher levels of both
plasma LH (O’Riain et al., 2000) and fecal estrogen
metabolites (Clutton-Brock et al., 2001) than subordinate
females, though dominants and subordinates show similar
LH responses to a challenge with gonadotropin-releasing
hormone (GnRh) (O’Riain et al., 2000). Among subordinate
females, those with access to unrelated males in the same
group show higher levels of fecal estrogen metabolites than
those living in groups containing related males only (Clut-
ton-Brock et al., 2001). In contrast, dominant males do not
differ from subordinates in levels of plasma LH (O’Riain et
al., 2000). However, none of these analyses controlled for
variation in the age and weight of sampled individuals, and
no hormonal comparisons between dominant and subordi-
nate meerkats have ever controlled for whether these indi-
viduals had access to unrelated breeding partners in their
group. Here, we investigate whether there are consistent
differences in plasma levels of sex hormones and cortisol
between dominants and subordinates when the potential
confounding effects of age, weight, and access to unrelated
breeding partners are controlled for.
Methods
Animals and sample collection
We collected hormonal data between November 1997
and April 2002 on a total of 157 animals (69 females, 88
males) from 13 different meerkat groups in our study area
near Van Zyl’s Rus in the South African Kalahari Desert
(26j59VS, 21j50VE). Meerkats in our study population have
been monitored since 1993, are individually marked, and
have been habituated to close proximity and handling.
Long-term study of the groups allowed the origins of each
individual (immigrant versus natal) to be ascertained for
analyses dealing with dispersal status and access to unrelat-
ed breeding partners. All research protocols were approved
by the Research Ethics Committee at the University of
Cambridge and complied with regulations stipulated in the
Guidelines for the Use of Animals in Research.
Each blood sample was obtained by placing an animal
inside a capture bag and immobilizing it with an intramus-
cular injection of ketamine and dormitor at a 2:1 ratio and
administered as a weight-dependent dose (100 mg/ml ket-
amine hydrochloride concentration and 1 mg/ml meditomi-
dine hydrochloride) of 0.06 ml ketamine and 0.04 ml
dormitor (O’Riain et al., 2000). After the animal was
completely anesthetized, 0.3–1.5 ml of blood was taken
from the jugular vein using a 24 1/2-G needle and 1-ml
syringes. After blood collection, each animal received an
intramuscular injection of 0.05 ml antisedan (at 5 mg/ml
atipamezole hydrochloride/kg) to reverse the effects of the
dormitor. Blood samples were placed on ice until centri-
fuged for 30 min at 500 � g. Plasma samples were stored
frozen at �40jC until they were assayed. To minimize
potential diurnal variation in hormonal levels, blood sam-
ples were collected between 17 h and 20 h 30 min.
Hormonal assays
Hormonal assays for the determination of LH concen-
trations were completed at the University of Pretoria in
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150 143
South Africa. All hormonal assays for the determination of
estradiol, testosterone, and cortisol were carried out at the
Medical Research Council Human Reproductive Sciences
Unit laboratories at the University of Edinburgh in Scotland.
LH bioassay
LH concentrations were measured in an in vitro bioassay
based on the production of testosterone by dispersed mouse
Leydig cells (Van Damme et al., 1974). Harlow et al. (1984)
and Hodges et al. (1987) describe details of the assay, and
O’Riain et al. (2000) report on the validation of the bioassay
for meerkat plasma samples. Plasma samples were assayed
in duplicate at a dilution of 1:20 and 1:40 as a routine check
for parallelism and compared to the human LH pituitary
preparation (2nd International Standard 1988, code no. 80/
552, NIBSC, UK) over the range 400–1.4 miu/ml. The
testosterone produced was measured by radioimmunoassay
as described by Bennett (1994). Checks for parallelism to
the standard curve were carried out to validate the LH
bioassay for plasma taken before and after GnRh adminis-
tration. The curves were parallel and not significantly
different from the reference preparation. Sensitivity of the
assay was 12.7 miu/tube or 2.5 miu/ml. Intra- and inter-
assay coefficients of variation for repeated measurement of a
quality control were 7% and 12%, respectively.
Estradiol
Samples were assayed using Coat-a-Count Estradiol kits
(Diagnostic Products Corporation, Los Angeles, CA). Serial
dilutions of meerkat plasma were parallel to the standard
curve (n = 4, t test = �0.35, P > 0.05). Accuracy, calculated
as the percent recovery of known amounts of estradiol added
to each standard curve point, was 100.4% (n = 5). Assay
sensitivity was 10 pg/ml. Cross-reactivity of the Coat-a-
Count estradiol assay was 10.0% with estrone, and <5%
with estriol, estrone-h�D-glucuronide, estrone-3-sulfate, d-
equilenin, 17h-estradiol-3-monosulfate, testosterone, and
androsterone. Briefly, 25 Al meerkat plasma and 75 Al assaydiluent were added to the antibody-coated test tubes, the
assay was incubated for 3 h at room temperature after adding
1 ml 125I-estradiol to each tube, and then the assay was
decanted before counting each tube for 1 min in a gamma
counter. Intra- and inter-assay coefficients of variation for a
meerkat plasma pool were 5.2% and 6.7%, respectively (n = 9
assays). In repeated blood sampling (2–4 samples/animal) of
seven subordinate females 3–38 min after capture and
anesthesia, estradiol concentrations did not change signifi-
cantly with increasing capture-to-bleed time (Spearman rank
correlation = 0.106, n = 22, P = 0.64).
Testosterone
Samples were assayed using Coat-a-Count Testosterone
kits (Diagnostic Products Corporation). Serial dilutions of
meerkat plasma were parallel to the standard curve (n = 4, t
test = �0.10, P > 0.05). Accuracy was 101% (n = 7). Assay
sensitivity was 20 ng/dl. Cross-reactivity of the Coat-a-
Count testosterone antibody was 16% with 11-ketotestoster-
one, <5% with dihydrotestosterone and 19-hydroxyandros-
tendione, and <1% with aldosterone, androstendione,
cortisol, corticosterone, estrone, methyltestosterone, and
progesterone. Assay procedures were similar to those used
for estradiol, but with 10 Al meerkat plasma and 90 Al assaydiluent per tube and an incubation period of 3 h at 37jC.Intra- and inter-assay coefficients of variation were 4.1%
and 6.2%, respectively (n = 10 assays). In repeated blood
sampling (2–4 samples/animal) of 10 subordinate males 3–
38 min after capture and anesthesia, testosterone levels did
not change significantly with increasing capture-to-bleed
time (Spearman rank correlation = �0.007, n = 38, P =
0.97).
Cortisol
Samples were assayed using Coat-a-Count Cortisol kits
(Diagnostic Products Corporation). Serial dilutions of meer-
kat plasma were parallel to the standard curve (n = 3, t test =
1.36, P > 0.05). Accuracy was 103.4% (n = 5). Assay
sensitivity was 0.25 Ag/dl. Cross-reactivity with other hor-
mones was 76.0% with prednisolene, 11.4% with 11-deox-
ycortisol, 2.3% with prednisone, and <1% with aldosterone,
corticosterone, cortisone, estriol, estrone, and pregnenolone.
Assay procedures were similar to those used for estradiol,
but with 25 Al meerkat plasma (and no diluent) added to
each tube and a 45-min incubation period at 37jC. Intra-and inter-assay coefficients of variation were 5.2% and
8.7%, respectively (n = 7 assays).
To assess the effects of the stress response on glucocor-
ticoid levels, we first determined that cortisol concentrations
did not vary according to sex (t test, n = 43 subordinate
females, n = 40 subordinate males, T = 0.19, P = 0.85).
Next, we combined different data from 10 subordinate
males and seven females during repeated blood sampling
(2–4 samples/animal) 3–38 min after capture and anesthe-
sia in an analysis that showed that cortisol levels increased
significantly with increasing capture-to-bleed time (Spear-
man rank correlation, n = 58, P < 0.001). Though we
factorize cortisol levels for final analysis, we still control
for capture-to-bleed time in the analyses.
Statistical analyses
The effects of individual, group, and environmental
characteristics on hormonal concentrations were investigat-
ed using linear models (GLM) in GENSTAT 4.2 (2000).
All terms were initially entered into these models and then
sequentially dropped until the model only included terms
whose elimination would have significantly reduced the
explanatory power of the model. The process was repeated
by the additive method (sequential adding of terms with
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150144
the retention of terms only if their addition significantly
increased explained variance) to confirm the structure of
the minimal model (Russell et al., 2002). All two-way
interactions between terms were tested but none was found
to be significant.
The primary term of interest in each analysis was a three-
level factor used to tease apart the effects of dominance
from those of access to unrelated breeding partners, classi-
fying individuals as either (i) dominant, (ii) subordinate
without access to unrelated breeding partners, or (iii) sub-
ordinate with access to unrelated breeding partners. All
sampled dominants had access to unrelated breeding part-
ners within their group. This was because dominant males
were typically immigrants, so dominants of both sexes
generally had access to unrelated breeding partners, and
rare samples collected from dominants whose mates had
died were excluded. In contrast, subordinate individuals of
both sexes often did not have access to unrelated breeding
partners in their group (i.e., 83 of 166 subordinate females in
11 study groups lived with related males only). Because
females do not transfer between established groups, but
males do, subordinate females were deemed to have access
to unrelated males if either (a) males had immigrated into
their group after the female was born or (b) the females had
founded a new group with foreign males. Subordinate males
were deemed to have access to unrelated breeding partners if
they had dispersed into or founded their current group
(‘immigrant subordinate’), and not to have access to unre-
lated breeding partners if they were born in their current
group (‘natal subordinate’).
Females designated as ‘dominant’ were the primary
breeders in the group (Clutton-Brock et al., 2001); their ages
ranged from 19 to 68 months. ‘Subordinate’ females were
behaviorally submissive to the dominant female in their
group (O’Riain et al., 2000) and bred at significantly lower
rates than the dominant female (Clutton-Brock et al., 2001);
their ages ranged from 12 to 47 months. Subordinate males
were behaviorally submissive to the dominant male in their
group (O’Riain et al., 2000). Dominant males had high rates
of anal marking and were the fathers of most litters born in
the group (Griffin et al., 2003). Dominant, subordinate natal,
and subordinate immigrant males ranged in age from 26 to
80, 12 to 44, and 15 to 53 months, respectively. Dominance
status is usually inherited by the oldest same-sex subordinate
after the dominant animal dies; as a result, dominant animals
can differ dramatically in age (14–68 months) as they need
only be the oldest of their sex class in the group at the time of
the breeding vacancy (Clutton-Brock et al., 2001).
We included a substantial number of additional variables
in our statistical models, for which definitions and details
follow. The age of each sampled individual was fitted as a
variate and calculated in days since their date of birth, all of
which were known. Rainfall and temperature were recorded
daily at the site; we fitted average maximum monthly
temperature and total monthly rainfall as variables. To
investigate the importance of seasonality, samples were
classified as either being taken during the conceptive season
(the period between June and January inclusive) or not
(Russell et al., 2003a). To control for a possible effect of
lactation on female hormone levels, samples were classified
as having been collected during lactation or not. Group size
was fitted as a variate and calculated as the number of
individuals in each group, excluding pups (i.e., those less
than 3 months of age) (Clutton-Brock et al., 2002). To
control for a possible effect of the breeding context of the
group, samples were classified as having been collected in
one of three contexts: (i) babysitting period, (ii) pup feeding
period, or (iii) currently without pups. Many of these
variables do not appear in the table detailing the minimal
models resulting from our analyses, as they did not have a
significant effect on the circulating levels of the hormone in
question (P > 0.05) and thus were dropped from the
statistical models (see above).
We excluded samples from all pregnant females from the
cortisol data set to control for reproductive state but used
samples from the first 2 weeks of pregnancy for the estradiol
analysis after confirming that estrogen levels in early
pregnancy samples did not differ from nonpregnant samples
(Moss et al., 2001; this study). Gestation in female meerkats
can be detected visually by the end of the first month, and
lasts approximately 70 days (Doolan and Macdonald, 1997),
so dates of conception could thus be backdated from the day
of parturition.
Finally, we compared the testes size for dominant and
subordinate (>1 year old) male meerkats. Measurements of
testes size (length and breadth) were obtained with cali-
pers during captures (see above), and we calculated a
testes size index by multiplying length by breadth. Testes
size index was fitted as a response term in a general linear
mixed model in which we fitted dominance status, age,
and body weight as fixed effects, and individual identity
as a random term, with the random term accounting for
repeated measures within males (1–11 measures per
individual). In this analysis, we were only able to compare
testes sizes of dominant and natal subordinates because
we were only able to collect testes size data from three
immigrant subordinates.
Sample sizes
We used a total of 63 plasma samples from 11 groups (9
dominant females and 36 subordinate females) to assess the
effects of dominance status on female LH levels after a
natural logarithm transformation (Kolmogorov–Smirnov
normality tests: raw data, P = 0.021, transformed data, P >
0.2). We used a total of 73 plasma samples from 11 groups
(7 dominant females and 50 subordinate females) to assess
the effects of dominance status on estradiol levels. As 48 of
the 73 samples had undetectable levels of plasma estradiol,
statistical models based on continuous error structures were
inappropriate (normality test, P = 0.001). Estradiol levels
were factorized into values of 0 or 1 signifying undetectable
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150 145
and detectable levels of plasma estradiol, respectively,
allowing the use of a generalized linear model with a
binomial error structure. We used a total of 114 plasma
samples from 13 groups (10 dominant males and 72
subordinate males) to assess the effects of dominance status
on testosterone levels (normality test, P = 0.20), and
obtained 55 measures of testes size from 21 dominants
and 170 measures from 98 subordinate natal males.
We conducted separate cortisol analyses for male and
female meerkats. For females, we used a total of 48 plasma
samples from 10 groups (from 8 dominant females and 36
subordinate females) to assess the effects of dominance
status on cortisol levels. For males, we used a total of 100
plasma samples from 12 groups (from 7 dominant males and
60 subordinate males). As the distributions of plasma
cortisol concentrations were skewed, it was not possible to
use a continuous error structure for statistical modeling
(normality tests for males and females: P = 0.001). Cortisol
levels were factorized into values of 0 or 1 signifying
undetectable and detectable levels of plasma cortisol, re-
spectively, allowing the use of a generalized linear model
with a binomial error structure.
Results
Females
In female meerkats, plasma levels of LH were significant-
ly influenced by age and the presence or absence of unrelated
males in the group (Table 1). After controlling for the effects
of age, we found that dominant females had comparable
levels of plasma LH to subordinate females with access to
unrelated males (P = 0.30, Fig. 1a), but significantly higher
levels than subordinate females without access to unrelated
males (P = 0.002, Fig. 1a). In contrast, plasma levels of
Table 1
Factors affecting hormone levels in female and male meerkats
Hormone analysis Model terms Estimate
Female LH Female type See Fig. 1a
Age (days) �0.00057
Constant 2.13
Female estradiol Female type See Fig. 1b
Body weight (g) 0.0014
Constant �0.7
Female cortisol Female type See Fig. 1c
Male testosterone Age (days) 0.39
Conceptive season (yes > no) 291.2
Constant 397.4
Male cortisol Male type See Fig. 2c
Age (days) 0.0020
Capture-to-bleed time (min) 0.93
Constant �4.07
Minimal models for each of the five multivariate hormonal analyses are presente
significant terms have been presented; for all other terms (see Methods) P > 0.05. ‘
with access to unrelated males, and subordinates without access to unrelated male
natal, and subordinate immigrant. All status effects are presented in Fig. 1 (fema
estradiol were significantly and positively influenced by
body weight and by the presence or absence of unrelated
males. After controlling for the effects of body weight, we
found that dominant females were significantly more likely
to have detectable levels of plasma estradiol than subordinate
females irrespective of whether unrelated males were present
(P < 0.05 in both cases; Fig. 1b). Means and ranges of raw
estradiol were 1202 pg/ml (4.5–4139), 160.7 pg/ml (0–
1300), and 28.7 pg/ml (0–377.4), respectively, for dominant
females, subordinate females with access to unrelated males,
and subordinate females without access to unrelated males.
The probability of having detectable levels of plasma
cortisol was unaffected by female age or body weight (P
values > 0.2). However, the probability of having detectable
cortisol levels differed significantly according to female
status (Table 1). Dominant females were significantly more
likely to have detectable cortisol levels than subordinate
females regardless of whether subordinates had access to
unrelated males in the same group or not (P < 0.05; Fig. 1c).
Means and ranges of raw cortisol were 1.6 Ag/dl (0.5–4.2),0.85 Ag/dl (0–5.8), and 1.2 Ag/dl (0–8.8), respectively, fordominant females, subordinate females with access to un-
related males, and subordinate females without access to
unrelated males.
Males
Plasma testosterone levels were positively correlated
with male age and were higher during the conceptive
season (Table 1). After controlling for these effects,
dominant males did not have higher levels of circulating
testosterone than either subordinate natal males (P = 0.37)
or subordinate immigrant males (P = 0.36) (Fig. 2a).
Additionally, subordinate immigrant males did not have
higher levels of testosterone than subordinate natal males
(P = 0.85).
S.E. F statistic df P value
7.76 2,61 0.001
0.00020 7.89 1,60 0.007
0.18
7.91 2,59 <0.001
0.0066 5.27 1,58 0.022
21.3
6.80 2,47 0.001
0.095 16.62 1,112 <0.001
69.5 17.55 1,112 <0.001
83.5
5.65 2,97 0.004
0.0010 4.14 1,96 0.042
0.29 14.31 1,96 <0.001
2.01
d above. Estimates indicate the direction and strength of each effect. Only
Female type’ refers to three types of adult females: dominants, subordinates
s. ‘Male type’ refers to three types of adult males: dominants, subordinate
les) and 2 (males).
Fig. 2. Endocrine and anatomical correlates of dominance status in male
meerkats. In each case means (F SE) are presented. (a) Dominant (white
bars), subordinate natal (hatched bars), and subordinate immigrant males
(black bars) had equivalent concentrations of plasma testosterone. (b)
Dominant (white bars) and subordinate natal (hatched bars) males had
similarly sized testes. (c) Dominant males (white bars) were significantly
more likely to have detectable levels of plasma cortisol than immigrant
males (black bars), while the cortisol levels of dominant and subordinate
natal males (hatched bars) did not differ. Subordinate natal males were
significantly more likely to have detectable levels of plasma cortisol than
subordinate immigrant males.
Fig. 1. Endocrine correlates of female status in meerkats. In each case
means (F SE) are presented. (a) After controlling for significant effects of
age, we found that dominant female meerkats (white bars) had similar
levels of plasma LH to subordinate females with access to unrelated males
(hatched bars), but significantly higher levels than subordinate females
without access to unrelated males (black bars). (b) Dominant female
meerkats (white bars) were significantly more likely to have detectable
levels of estradiol than both types of subordinate female. Subordinate
females with access to unrelated males (hatched bars) were significantly
more likely to have detectable levels of plasma estradiol than subordinate
females without access to unrelated males (black bars). (c) Dominant
female meerkats (white bars) were significantly more likely to have
detectable levels of plasma cortisol than subordinate female meerkats with
(hatched bars) and without (black bars) access to unrelated males.
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150146
Testosterone levels alone may not yield a complete
picture of a male’s reproductive potential as dominant
males may have larger testes, and hence larger ejaculates,
than subordinates. A crude comparison by dominance
class alone suggested that dominant males had larger
testes than subordinate males (v2 = 4.64, df = 1, P =
0.031). However, testes size increased with male age (v2 =6.16, df = 1, P = 0.013) and weight (v2 = 16.14, df = 1,
P < 0.001). After these effects were controlled, the
significant difference between the testes size of dominant
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150 147
and subordinate males disappeared (v2 = 0.79, df = 1, P =
0.37; Fig. 2b).
Like dominant females, cortisol levels were not influ-
enced by age (P > 0.2) and dominant males usually had
higher cortisol levels than subordinate males. However, in
this case, the difference was largely due to differences
between dominant males and subordinate immigrant
males: dominant males were significantly more likely to
have detectable levels of cortisol than subordinate immi-
grant males (P = 0.014) but not natal subordinates (P =
0.17; Fig. 2c). Subordinate natal males had a significantly
higher probability of having detectable levels of cortisol
than immigrant subordinate males (P = 0.014). Means
and ranges of raw cortisol were 1.1 Ag/dl (0–2.2), 1.0 Ag/dl (0–7.1), and 0.3 Ag/dl (0–1.8), respectively, for dom-
inant males, subordinate natal males, and subordinate
immigrant males.
Discussion
Previous studies of meerkats have shown that dominant
females have higher levels of both plasma LH (O’Riain et
al., 2000) and fecal estrogen metabolites (Clutton-Brock et
al., 2001) than subordinates. This study significantly
extends these findings, because neither of the previous
studies controlled for female age, weight, or access to
unrelated males. Here, we show that after controlling for
these confounding effects, rank-related differences in LH
levels disappear, whereas those for estradiol remain. In
female meerkats, circulating levels of LH and estradiol are
positively affected by the presence of unrelated males and
are also influenced by female age and weight, respectively.
These results suggest that although the hypothalamic and
pituitary activity of dominant and subordinate female meer-
kats may be equivalent, subordinate females appear to
experience downregulation of their ovarian activity, depress-
ing circulating levels of estradiol below those of dominants.
This rank-related disparity in ovarian activity may result
either from suppression of subordinate female fertility by
dominants (e.g., through aggression, Young, 2003) or phys-
iological restraint by subordinates, due to the high proba-
bility of their litters being killed by the dominant (Clutton-
Brock et al., 2001; Russell et al., 2003b). Either way, the
reduced ovarian activity of subordinate females offers a
physiological explanation for their lower breeding rates
(even when they have access to unrelated males), relative
to dominants (Clutton-Brock et al., 2001; O’Riain et al.,
2000). In addition, the fact that heavier females are more
likely to have detectable levels of plasma estradiol offers a
physiological explanation for the higher breeding rates of
heavier females (Clutton-Brock et al., 2001).
Though there are rank-related differences in ovarian
activity, the primary determinant of sex hormone levels in
female meerkats is the presence or absence of unrelated
males in their group. The results presented here confirm
previous findings that access to unrelated males increases
both female LH and estrogen levels (Clutton-Brock et al.,
2001; O’Riain et al., 2000), and extends them by demon-
strating that this difference remains after controlling for
the potentially confounding effects of age and body
weight. The effect of the presence or absence of unrelated
males is presumably related to mechanisms that inhibit
reproduction with close relatives, and provides a physio-
logical explanation for the dramatic effect of unrelated
males on the breeding rates of subordinate females (Clut-
ton-Brock et al., 2001). Schoech et al. (1996) have rightly
emphasized the importance of distinguishing the absence
of a necessary stimulus from the presence of an inhibitor
when investigating causes of breeding suppression. Our
results indicate that, for female meerkats, both factors may
play a role: a stimulatory effect of unrelated males is
apparently necessary for the attainment of physiological
fertility, but an inhibitory effect of subordination appears
to suppress estradiol levels which in turn may compromise
breeding rates.
Dominant female meerkats were significantly more
likely to have detectable levels of cortisol than both types
of subordinate females, and similar differences between
dominants and subordinates have been found in several
other studies of social mammals (reviewed by Creel,
2001). Because dominant female meerkats breed signifi-
cantly more often than subordinate females (Clutton-Brock
et al., 2001; Griffin et al., 2003) and heightened levels of
reproductive hormones have a stimulatory effect on adre-
nal steroidogenesis (reviewed by Saltzman et al., 2000), it
is possible that frequent pregnancies permanently alter the
basal adrenal steroidogenesis patterns of dominant females.
Alternatively, increased levels of glucocorticoids in dom-
inant animals may be related to the stress of dominating
other group members (Creel, 2001). Either way, we found
no evidence that the reduced levels of plasma estradiol in
subordinate female meerkats were associated with high
levels of circulating glucocorticoids.
Male testosterone levels varied seasonally and increased
with age, whereas testes sizes increased with both male
weight and age. After controlling for these effects, dominant
and subordinate males were found to have comparable
levels of testosterone and similarly sized testes, irrespective
of their dispersal status (whether they were in the presence
of unrelated females). Furthermore, O’Riain et al. (2000)
showed that dominant and subordinate males have similar
levels of LH. Though there is no physiological disparity
been dominant and subordinate male meerkats, subordinate
males never breed in their natal group, and even after
dispersal they breed at lower rates than dominants (Griffin
et al., 2003). Consequently, the low breeding rates of
subordinate male meerkats likely result from inbreeding
avoidance in the presence of related females only (while
males remain in their natal group) and behavioral interfer-
ence by dominants when in the presence of unrelated
females (after males have dispersed).
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150148
A lack of physiological suppression among males fits
the pattern emerging for males of other cooperative
species, where dominants and subordinates commonly
have comparable hormonal profiles (e.g., Bennett, 1994;
Bennett et al., 1993; Creel et al., 1992), but see naked
mole rats (Heterocephalus glaber) for an exception
(Faulkes et al., 1991, 1994). Selection may have favored
behavioral rather than physiological suppression in male
cooperative mammals, for several reasons. If dominants
can monopolize reproduction through mate-guarding alone,
selection may not favor more complex means of suppres-
sion. Alternatively, subordinate males may be selected to
maintain their fertility if they are able to mate outside their
group by conducting extraterritorial prospecting forays
(e.g., meerkat males: Doolan and Macdonald, 1996; Grif-
fin et al., 2003).
Cortisol levels of dominant males and natal subordinate
males did not differ from one another, but both dominant
and natal subordinate males were significantly more likely
to have detectable levels of cortisol than immigrant subor-
dinate males. In a review of the effects of dominance status
on stress hormones in cooperatively breeding species, Creel
(2001) found that the glucocorticoid levels of dominant
males were either significantly higher than or equal to the
glucocorticoid levels of subordinate males. While our
results are generally consistent with the findings from the
review of Creel (2001), this is the first time to our knowl-
edge that immigrant males have been found to have signif-
icantly lower levels of cortisol than dominant and
subordinate natal males in a cooperative breeder. While
one might expect increased levels of glucocorticoids in
dominant males relative to subordinate males if intragroup
dominance hierarchies were unstable (Sapolsky, 1992) or
dominant males fought more frequently than subordinate
males (Creel, 1996), we have no reason to believe that either
of these factors accounts for the observed patterns of cortisol
secretion in our study groups. Instead, plasma cortisol levels
of subordinate immigrant males appear reduced in compar-
ison with cortisol levels of dominant and subordinate natal
males, offering a curious parallel to the discovery of Saltz-
man et al. (2000) that the plasma cortisol concentrations of
female common marmosets dropped when females became
subordinate in a new group and then remained chronically
suppressed (compared with the basal cortisol levels of
dominant females) (see also Ziegler et al., 1995). However,
while Saltzman et al. (2000) believe that these patterns are
caused by a combination of the effects of social subordina-
tion and the absence of reproductive hormones in female
common marmosets, there is no evidence that subordinate
male meerkats experience chronic reproductive suppression.
As yet, we are unable to explain the observed variation in
cortisol secretion among male meerkats of different status.
Our comparisons of the hormonal levels of dominant
and subordinate meerkats have implications for all
attempts to understand the proximate causes of rank-
related differences in breeding rates in cooperative socie-
ties. At present, it is difficult to tell whether these
reproductive disparities are determined physiologically or
behaviorally, despite the considerable number of studies on
the topic. This difficulty arises in part because few studies
have controlled for factors that may confound physiolog-
ical comparisons of dominant and subordinate group
members. Here we show that circulating levels of repro-
ductive hormones may vary significantly with individual
age (LH and testosterone), weight (estradiol), and access to
unrelated breeding partners (LH and estradiol), all of
which are likely to confound rank-related comparisons in
most cooperative species. It is consequently important that
future correlational studies in both birds and mammals
employ multivariate analyses to control for such effects,
allowing the physiological characteristics associated with
dominance per se to be identified.
Acknowledgments
We are grateful to members of the Mammal Research
Institute at the University of Pretoria, without whom this
study would not have been possible, including Professors
John Skinner and Johan du Toit, as well as Martin Haupt.
We thank Mr. and Mrs. H. Kotze for generously allowing us
to work on their land at Vanzyl’s Rus in the southern
Kalahari, and the Northern Cape Conservation Trust for
granting us permission to conduct research in South Africa.
We are grateful to R. Carlson and G. McIlrath for collecting
blood samples. D. Abbott, T. Coulson, J. McNeilly, and W.
Saltzman provided much-appreciated discussion, and two
anonymous reviewers made valuable suggestions for
improvements to the manuscript. The National Institute of
Biological Standard and Controls (2nd International Stan-
dard 1988 Code no. 80/552, Hertfordshire, England) is
acknowledged for the donation of the LH reference
preparation. This research was funded by grants from the
National Research Foundation, South Africa, and the
University of Pretoria awarded to N.C. Bennett and the
Natural Environmental Research Council and the Biotech-
nology and Biological Sciences Research Council awarded
to T. H. Clutton-Brock.
References
Abbott, D.H., Hodges, J.K., George, L.M., 1988. Social status controls LH
secretion and ovulation in marmoset monkeys (Callithrix jacchus).
J. Endocrinol. 117, 329–339.
Bennett, N.C., 1994. Reproductive suppression in eusocial Cryptomys
damarensis colonies: a lifetime of socially-induced sterility in males
and females. J. Zool. (London) 234, 25–39.
Bennett, N.C., Jarvis, J.U.M., Faulkes, C.G., Millar, R.P., 1993. LH
responses to single doses of exogenous GnRH by freshly captured
Damaraland mole-rats, Cryptomys damarensis. J. Reprod. Fertil. 99,
81–86.
Bennett, N.C., Faulkes, C.G., Molteno, A.J., 1996. Reproductive suppres-
sion in subordinate, non-breeding female Damaraland mole-rats: two
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150 149
components to a lifetime of socially-induced fertility. Proc. R. Soc.
London, Ser. B 263, 1599–1603.
Bennett, N.C., Faulkes, C.G., Spinks, A.C., 1997. LH responses to single
doses of exogenous GnRh by social Mashona mole-rats: a continuum of
socially-induced infertility in the family Bathyergidae. Proc. R. Soc.
London, Ser. B 264, 1001–1006.
Brotherton, P.N.M., Clutton-Brock, T.H., O’Riain, M.J., Gaynor, D.,
Sharpe, L.L., Kansky, R., McIlrath, G.M., 2001. Offspring food alloca-
tion by parents and helpers in a cooperative mammal. Behav. Ecol. 12,
590–599.
Clutton-Brock, T.H, Gaynor, D., Kansky, R., MacColl, A.D.C., McIlrath,
G., Chadwick, P., Brotherton, P.N.M., O’Riain, J.M., Manser, M., Skin-
ner, J.D., 1998. Costs of cooperative behaviour in suricates (Suricata
suricatta). Proc. R. Soc. London, Ser. B 265, 185–190.
Clutton-Brock, T.H., Brotherton, P.N.M., Russell, A.F., O’Riain, M.J.,
Gaynor, D., Kansky, R., Griffin, A., Manser, M., Sharpe, L., McIlrath,
G.M., Small, T., Monfort, S., 2001. Cooperation, control, and conces-
sion in meerkat groups. Science 291, 478–481.
Clutton-Brock, T.H., Russell, A.F., Sharpe, L.L., Young, A.J., Balmforth,
Z., McIlrath, G.M., 2002. Evolution and development of sex differences
in cooperative behavior in meerkats. Science 297, 253–256.
Creel, S., 1996. Social stress and dominance. Nature 379, 212.
Creel, S., 2001. Social dominance and stress hormones. Trends Ecol. Evol.
16, 491–497.
Creel, S.R., Creel, N.M., 1991. Energetics, reproductive success and obli-
gate communal breeding in carnivores. Behav. Ecol. Sociobiol. 28,
263–270.
Creel, S.R., Creel, N.M., 2002. The African Wild Dog: Behavior, Ecology
and Conservation. Princeton Univ. Press, Princeton.
Creel, S., Creel, N., Wildt, D.E., Monfort, S.L., 1992. Behavioural and
endocrine mechanisms of reproductive suppression in Serengeti dwarf
mongooses. Anim. Behav. 43, 231–245.
Creel, S., Creel, N.M., Mills, M.G.L., Monfort, S.L., 1997. Rank and
reproduction in cooperatively breeding African wild dogs: behavioral
and endocrine correlates. Behav. Ecol. 8, 298–306.
Doolan, S.P., Macdonald, D.W., 1996. Dispersal and extra-territorial
prospecting by slender-tailed mongooses (Suricata suricatta) in the
south-western Kalahari. J. Zool. (London) 240, 59–73.
Doolan, S.P., Macdonald, D.W., 1997. Breeding and juvenile survival
among slender-tailed meerkats (Suricata suricatta) in the south-west-
ern Kalahari: ecological and social influences. J. Zool. (London) 242,
309–327.
Emlen, S.T., 1995. An evolutionary theory of the family. Proc. Natl. Acad.
Sci. U. S. A. 92, 8092–8099.
Faulkes, C.G., Abbott, D.H., Jarvis, J.U.M., 1991. Social suppression
of reproduction in male naked mole-rats (Heterocephalus glaber).
J. Reprod. Fertil. 91, 593–604.
Faulkes, C.G., Trowell, S.N., Jarvis, J.U.M., Bennett, N.C., 1994. Investi-
gation of sperm numbers and motility in reproductively active and
socially suppressed males of two eusocial African mole-rats, the naked
mole-rat (Heteropcephalus glaber), and the Damaraland mole-rat
(Cryptomys damarensis). J. Reprod. Fertil. 100, 411–416.
French, J.A., 1997. Proximate regulation of singular breeding in Callitri-
chid primates. In: Solomon, N.G., French, J.A. (Eds.), Cooperative
Breeding in Mammals. Cambridge Univ. Press, Cambridge, England,
pp. 34–75.
French, J.A., Inglett, B.J., Dethlefs, T.M., 1989. The reproductive status of
nonbreeding group members in captive golden lion tamarin social
groups. Am. J. Primatol. 18, 73–86.
Ginther, A.J., Ziegler, T.E., Snowdon, C.T., 2001. Reproductive biology of
captive male cotton-top tamarin monkeys as a function of social envi-
ronment. Anim. Behav. 61, 65–78.
Griffin, A.S., Pemberton, J.M., Brotherton, P.N.M., McIlrath, G., Gaynor,
D., Kansky, R., Clutton-Brock, T.H., 2003. A genetic analysis of breed-
ing success in the cooperative meerkat (Suricata suricatta). Behav.
Ecol. 14, 472–480.
Harlow, C.R., Gems, S., Hodges, J.K., Hearn, J.P., 1984. The relationship
between plasma testosterone and the timing of ovulation and early
embryonic development in the marmoset monkey (Callithrix jacchus).
J. Zool. (London) 201, 272–282.
Hodges, J.K., Cottingham, P., Summers, P.M., Yingnan, L., 1987. Con-
trolled ovulation in the marmoset monkey (Callithrix jacchus), with
human chorionic gonadotropin following prostaglandin induced luteal
regression. Fertil. Steril. 48, 299–305.
Keane, B., Waser, P.M., Creel, S., Creel, N.M., Elliott, L.F., Minchella,
D.J., 1994. Subordinate reproduction in dwarf mongooses. Anim.
Behav. 47, 65–75.
Khan, M.Z., McNabb, F.M.A., Walters, J.R., Sharp, P.J., 2001. Patterns of
testosterone and prolactin concentrations and reproductive behavior of
helpers and breeders in the cooperatively breeding red-cockaded wood-
pecker (Picoides borealis). Horm. Behav. 40, 1–13.
Mays, N.A., Masters Vleck, C., Dawson, J., 1991. Plasma luteinizing
hormone, steroid hormones, behavioral role, and nest stage in coop-
eratively breeding Harris’ hawks (Parabuteo unicinctus). Auk 108,
619–637.
Moehlman, P.D., Hofer, H., 1997. Cooperative breeding, reproductive sup-
pression and body mass in canids. In: Solomon, N.G., French, J.A.
(Eds.), Cooperative Breeding in Mammals. Cambridge Univ. Press,
Cambridge, England, pp. 76–128.
Moss, A.M., Clutton-Brock, T.H., Monfort, S.L., 2001. Longitudinal go-
nadal steroid excretion in free-living male and female meerkats (Sur-
ricata suricatta). Gen. Comp. Endocrinol. 122, 158–171.
O’Riain, M.J., Bennett, N.C., Brotherton, P.N.M., McIlrath, G., Clutton-
Brock, T.H., 2000. Reproductive suppression and inbreeding avoidance
in wild populations of co-operatively breeding meerkats (Suricata sur-
icatta). Behav. Ecol. Sociobiol. 48, 471–477.
Poiani, A., Fletcher, T., 1994. Plasma levels of androgens and gonadal
development of breeders and helpers in the bell miner (Manorina mel-
anophrys). Behav. Ecol. Sociobiol. 34, 31–41.
Reyer, H.-U., Dittami, J.P., Hall, M.R., 1986. Avian helpers at the nest: are
they psychologically castrated? Ethology 71, 216–228.
Rickard, C.A., Bennett, N.C., 1997. Recrudescence of sexual activity in a
reproductively quiescent colony of the Damaraland mole-rat, by the
introduction of a genetically unrelated male—A case of incest avoid-
ance in ‘queenless’ colonies. J. Zool. (London) 241, 185–202.
Russell, A.F., Clutton-Brock, T.H., Brotherton, P.N.M., Sharpe, L.L., McIl-
rath, G.M., Dalerum, F.D., Cameron, E.Z., Barnard, J.A., 2002. Factors
affecting pup growth and survival in co-operatively breeding meerkats
Suricata suricatta. J. Anim. Ecol. 71, 700–709.
Russell, A.F., Brotherton, P.N.M., McIlrath, G.M., Sharpe, L.L., Clutton-
Brock, T.H., 2003a. Breeding success in cooperative meerkats: effects
of helper number and maternal state. Behav. Ecol. 14, 486–492.
Russell, A.F., Brotherton, P.N.M., McIlrath, G.M., Sharpe, L.L., Clutton-
Brock, T.H., 2003b. Cost minimization by helpers in cooperative verte-
brates. Proc. Natl. Acad. Sci. U. S. A. 100, 3333–3338.
Russell, A.F., Carlson, A.A., McIlrath, G.M., Jordan, N.R. and Clutton-
Brock, T.H., in press. Adaptive modification by dominant female meer-
kats. Evol.
Saltzman, W., Prudom, S.L., Schultz-Darken, N.J., Abbott, D.H., 2000.
Reduced adrenocortical responsiveness to adrenocorticotropic hormone
(ACTH) in socially subordinate female marmoset monkeys. Psycho-
neuroendocrinol 25, 463–477.
Sapolsky, R.M., 1992. Neuroendocrinology of the stress response. In:
Becker, J.B., et al. (Eds.), Behavioral Endocrinology. Massachusetts
Institute of Technology Press, Massachusetts, pp. 287–324.
Schoech, S.J., Mumme, R.L., Moore, M.C., 1991. Reproductive endocri-
nology and mechanisms of breeding inhibition in cooperatively breed-
ing Florida scrub jays (Aphelocoma c. coerulescens). Condor 93,
234–244.
Schoech, S.J., Mumme, R.L., Wingfield, J.C., 1996. Delayed breeding in
the cooperatively breeding Florida scrub-jay (Aphelocoma coerules-
cens): inhibition or the absence of stimulation? Behav. Ecol. Sociobiol.
39, 77–90.
Smith, T.E., Schaffner, C.M., French, J.A., 1997. Social and developmental
A.A. Carlson et al. / Hormones and Behavior 46 (2004) 141–150150
influences on reproductive function in female Wied’s black tufted-ear
marmosets (Callithrix kuhli). Horm. Behav. 31, 159–168.
Solomon, N.G., French, J.A., 1997. Cooperative Breeding in Mammals
Cambridge Univ. Press, Cambridge, England.
Stacey, P.B., Koenig, W.D., 1990. Cooperative Breeding in Birds:
Long-term Studies of Ecology and Behavior. Cambridge Univ.
Press, Cambridge.
Van Damme, M.P., Robertson, D.M., Diczfalusy, E., 1974. An improved in
vitro bioassay method for measuring luteinising hormone (LH) activity
using mouse Leydig cell preparations. Acta Endocrinol. 77, 655–671.
Young, A.J., 2003. Subordinate tactics in cooperative meerkats: helping,
breeding and dispersal. PhD thesis, Cambridge University, England.
Ziegler, T.E., Savage, A., Scheffler, G., Snowdon, C.T., 1987. The endo-
crinology of puberty and reproductive functioning in female cotton-top
tamarins (Saguinus oedipus) under varying social conditions. Biol.
Reprod. 37, 618–627.
Ziegler, T.E., Scheffler, G., Snowdon, C.T., 1995. The relationship of
cortisol levels to social environment and reproductive functioning in
female cotton-top tamarins, Saguinus oedipus. Horm. Behav. 29,
407–424.