Content uploaded by Oliver Patrick Höner
Author content
All content in this area was uploaded by Oliver Patrick Höner
Content may be subject to copyright.
LETTERS
Female mate-choice drives the evolution of
male-biased dispersal in a social mammal
O. P. Ho
¨
ner
1
, B. Wachter
1
, M. L. East
1
, W. J. Streich
1
, K. Wilhelm
1
, T. Burke
2
& H. Hofer
1
Dispersal has a significant impact on lifetime reproductive suc-
cess
1
, and is often more prevalent in one sex than the other
2
.In
group-living mammals, dispersal is normally male-biased and in
theory this sexual bias could be a response by males to female mate
preferences, competition for access to females or resources, or the
result of males avoiding inbreeding
2–7
. There is a lack of studies on
social mammals that simultaneously assess these factors and mea-
sure the fitness consequences of male dispersal decisions. Here we
show that male-biased dispersal in the spotted hyaena (Crocuta
crocuta) most probably results from an adaptive response by males
to simple female mate-choice rules that have evolved to avoid
inbreeding. Microsatellite profiling revealed that females pre-
ferred sires that were born into or immigrated into the female’s
group after the female was born. Furthermore, young females
preferred short-tenured sires and older females preferred
longer-tenured sires. Males responded to these female mate pre-
ferences by initiating their reproductive careers in groups contain-
ing the highest number of young females. As a consequence, 11%
of males started their reproductive career in their natal group and
89% of males dispersed. Males that started reproduction in groups
containing the highest number of young females had a higher
long-term reproductive success than males that did not. The
female mate-choice rules ensured that females effectively avoided
inbreeding without the need to discriminate directly against close
kin or males born in their own group, or to favour immigrant
males. The extent of male dispersal as a response to such female
mate preferences depends on the demographic structure of breed-
ing groups, rather than the genetic relatedness between females
and males.
Why is dispersal in most group-living mammals heavily biased
towards males and which social groups offer males the best repro-
ductive prospects? Answers to these questions are important because
dispersal influences crucial components of lifetime reproductive suc-
cess and is a major source of variance in fitness
1
. In mammals with
polygynous mating systems, females are assumed to incur higher
costs from breeding with close relatives than males
8,9
. These sexual
asymmetries in costs are thought to cause sex-biased dispersal
8,10
.
High costs of inbreeding for females may favour female mate-choice
towards immigrant males and discrimination against male kin
3,11,12
and, in theory, female mate-choice can cause male-biased dispersal
3
.
We are unaware of any study that has assessed the impact on fitness
of the decision by males about where to initiate their reproductive
career and simultaneously evaluated the ultimate causes proposed
for male-biased dispersal in social mammals. Here we tested whether
male-biased dispersal in spotted hyaenas is driven by female mate-
choice or by one of the other three main factors proposed to
explain male-biased dispersal: male–male competition for access to
females
2,4,5
, inbreeding avoidance by males
2,6
, or competition for
resources
2,7
. We used ten years of detailed demographic data from
the entire hyaena population (eight social groups) in the Ngorongoro
Crater (hereafter referred to as ‘Crater’) in Tanzania, a habitat where
processes of natural selection are still intact. To assess fitness benefits
in terms of reproductive success of males after they initiated their
reproductive career in a group we used microsatellite profiling of 426
offspring.
The spotted hyaena is a large carnivore that lives in social groups or
‘clans’ in which females socially dominate males
13
. Most but not all
natal males disperse (that is, males leave the clan in which they were
born and immigrate into a new clan)
14
, whereas female dispersal is
very rare
15
. Immigrant male social status increases with length of
tenure (time spent living in one group) because males observe strict
social queueing conventions
14
. Because of the unusual anatomy of the
female genitalia
16
, female cooperation is a prerequisite for intromis-
sion
13,17
, and as a result females exercise considerable mate-choice
18
and mate promiscuously in clans with numerous reproductively
active males
18,19
. Females are likely to incur far higher costs of
inbreeding than males because only females care for offspring, lact-
ating for an exceptionally long period and producing milk with a high
protein, fat and energy content
20
. Therefore, females would be
expected to avoid breeding with close kin and to be choosier than
males when selecting a mate
9,21
.
In species such as the spotted hyaena, where females mate with
several males in one oestrus cycle and males do not care for their
young
18,19
, females may not be able to distinguish their own father
from other potential mates. A simple female mate-choice rule—
‘avoid males that were members of your group when you were born
and favour males that were born into or immigrated into your group
after your birth’—would reduce the chance of costly inbreeding of
females with their father or with older brothers. Female Crater hyae-
nas generally adhered to this rule, choosing sires that were born into
or immigrated into their clan after their birth more often than
expected from the mean proportion of candidate males that fulfilled
this requirement (Wilcoxon signed-rank test, N 5 64 females of
known age, P 5 0.0001). Most females (81.3%, N 5 64) always
applied the rule; only a single female did not do so more than once.
As a result, most litters (89.6%, N 5 134) were sired by males that
were born into or immigrated into the female’s group after her birth
(Fig. 1).
Young females (less than five years of age) produced litters sired by
males with significantly shorter tenures than older females (Mann–
Whitney U test, U 5 1,416.5, N
1
5 82 litters by young females,
N
2
5 52 litters by older females, P 5 0.001). These results are con-
sistent with the previously reported greater tolerance by young
females of short-tenured males (with less than three years tenure)
than longer-tenured males
14
, and the greater probability of
offspring of young females being sired by short-tenured rather than
1
Leibniz Institute for Zoo and Wildlife Research, Alfred-Kowalke-Strasse 17, D-10315 Berlin, Germany.
2
Department of Animal and Plant Sciences, University of Sheffield, Western
Bank, Sheffield S10 2TN, UK.
Vol 448
|
16 August 2007
|
doi:10.1038/nature06040
798
Nature
©2007
Publishing
Group
long-tenured males (assortative mating)
18
. Given these female pre-
ferences, we would expect males to initiate their reproductive career
in the clan with the highest number of young females, irrespective of
whether this was their natal or another clan.
Spotted hyaena males are likely to assess potential dispersal desti-
nations by undertaking short-term excursions into territories of
other clans
15,22
. Before natal adult males in the Crater initiated their
reproductive career in their natal or a non-natal clan, they were
frequently observed on excursions in territories of non-natal clans
(mean proportion of 0.2 6 0.02 of all sightings, N 5 114 males, of
which 62.3% were observed on such excursions). Furthermore, they
were more often observed on such excursions than their twin sisters
during the same period (Wilcoxon signed-rank test, N 5 20 brother–
sister twins, exact P 5 0.013). Thus, males in the Crater are unlikely
to be constrained in assessing potential dispersal destinations.
Of 142 males that were reared in Crater clans and reached adult-
hood, 114 males (80.3%) initiated their reproductive career in a clan
on the Crater floor and 28 males (19.7%) died or dispersed elsewhere.
Of the 114 males, 101 dispersed to a non-natal clan and 13 males
(11.4%) initiated their reproductive career in their natal clan. Eleven
males immigrated into Crater clans from elsewhere. For the 114
Crater-born males that initiated their reproductive career in a
Crater clan, we assessed the key factors hypothesized to influence
male dispersal (Table 1) for all eight clans on the Crater floor. As
predicted, clan selection was influenced by the number of young
females per group (Table 1), and males indeed preferred clans with
the highest number of young females (x
2
5 22.15, degrees of free-
dom, d.f. 5 1, P , 0.001; Fig. 2).
Males that initiated their reproductive tenure in clans with the
highest number of young females obtained fitness benefits because
male long-term reproductive success increased with the number of
young females present at clan selection (stepwise backward regres-
sion, final model: ln(y) 521.102 1 0.120x, F
1,23
5 20.563, r
2
5
0.472, P , 0.001; Fig. 3) after considering and removing from the
model the annual rate of mortality of these females (full model:
t 521.07, P 5 0.298). Furthermore, the long-term reproductive
success of such males was higher than that of other males (U 5 8,
N
1
5 9 males that initiated their reproductive career in clans with the
highest number of young females, N
2
5 16 males that initiated their
reproductive tenure in clans that did not contain the highest number
of young females, exact P , 0.0001; Fig. 4).
Males that initiated reproductive activity in the clan with the high-
est number of young females were likely to secure long-term access to
numerous mating partners because survivorship of these females was
above 75% during the first six years of male tenure. Thus, a judicious
clan selection provides males with a high number of females with
which they can develop long-term ‘friendly’ associations as both male
and female tenure increases—a male tactic actively preferred by
females that promotes male reproductive success
18
.
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
91011
Age of mother (years)
Tenure of father (years)
Figure 1
|
The relationship between the age of the mother on the date of
conception and the tenure of the father. (N 5 134 litters.) Filled circles,
litters sired by males that were born into or immigrated into the mother’s
clan after the mother’s birth. Open circles, litters sired by males that were
present when the mother was born.
Table 1
|
Test of predictions from the main hypotheses for the evolution of male-biased dispersal
Hypothesis Variable predicted to influence the likelihood of males
to select a clan
Model coefficient Standard error t-ratio P
Avoidance of competition with
other males for access to females
Intensity of male
–
male competition* 0.001 0.029 0.024 (0. 122 *) 0.981 (0.903*)
Female mate-choice Number of females most likely to breed with
males (‘young females’ as defined in Methods)
0.072 0.034 2.102 (2.064*) 0.036 (0.039*)
Avoidance of breeding with close
female relatives
Number of unrelated adult females with
relatedness of , 0.5
0.005 0.024 0.224 (0.298*) 0.823 (0.766*)
Avoidance of competition for
resources
Number of main prey animals per adult or
yearling spotted hyaena
0.001 0.001 0.464 (0.456*) 0.643 (0.649*)
Discrete choice regression model with the identity of the clan selected by 114 males as dependent variable; log-likelihood of whole model 52229.988.
* ‘Intensity of male
–
male competition’ refers to the number of reproductively active natal males plus immigrant males. In an alternative model (t-ratios and P values given in parentheses), the number
of reproductively active males per adult female was chosen instead.
–15
–10
–5
0
5
10
15
20
12345678
Clan rank based on number of
y
oun
g
females
Difference between observed and
expected number of clan selections
*
*
***
Figure 2
|
Preference of male spotted hyaenas for clans with the highest
number of young females.
Clans were ranked in relation to the number of
young female clan members on each date of clan selection by 114 males; in
each case the clan with the highest number of young females had rank 1.
*P , 0.05; ***P , 0.001.
0
1412108642016
0.5
1.0
1.5
2.0
Number of
y
oun
g
females, x
Long-term reproductive success, y
Figure 3
|
The influence of the number of young female clan members at
clan selection on the long-term reproductive success of male spotted
hyaenas.
Long-term reproductive success was the mean number of cubs
produced per year of tenure for 25 males with a minimum tenure of four
years in a clan. The line shows regression of long-term reproductive success
by number of young females, y 5 0.332 3 e
0.120x
.
NATURE
|
Vol 448
|
16 August 2007 LETTERS
799
Nature
©2007
Publishing
Group
As male tenure in a clan increased, the number of females in the
clan that were young at the time the males initiated their reproductive
career in the clan declined because of female mortality; after seven
years of tenure only 59.3% and after eight years only 45.8% of these
females remained alive. This may explain the decrease in reproduct-
ive performance of long-tenured males towards the end of their
tenures
18,19
, and (secondary) dispersal of 16.7% of Crater males with
tenures exceeding six years, despite the fact that these males had
obtained a high social status in the male hierarchy.
There was little evidence that females produced offspring sired by
close male relatives, thereby risking fitness costs of inbreeding. Only
four of 426 cubs in two of 309 litters (0.6% ) produced by one of 110
females (0.9%) resulted from daughter–father matings. Females that
conceived when their father was a member of their clan produced
only two of 88 litters with their father. None of the five litters that
mothers conceived when their sons were reproductively active in
their clan resulted from mother–son matings. None of the males that
were reproductively active in their natal clan had a sister that con-
ceived during their tenure, so breeding between females and their
brothers could not occur.
Males did not appear to avoid the chance of breeding with close
female relatives, because clan selection was independent of the
number of unrelated females in a clan (Table 1). Furthermore, of
76 males that consorted with (shadowed
14
) females, the 13 males that
had the opportunity to shadow daughters did not shadow daughters
less often than expected from the mean proportion of daughters
in the pool of adult females (Wilcoxon signed-rank test, exact
P 5 0.542).
Males did not select clans with respect to the likely level of male–
male competition because clan selection was independent of two
measures of competition: the total number of male competitors
and the number of male competitors per adult female (Table 1).
Males thus did not prefer clans with short male social queues, a result
consistent with the idea that when the number of potential mating
partners available to males in a long queue is greater than that in a
shorter queue, males benefit more by joining longer queues
23
.
There was also no evidence that clan selection was influenced by
competition for food because selection was independent of the per
capita number of main prey animals in a clan territory (Table 1).
Our findings suggest that female mate-choice is the main factor
determining the clan in which males initiate their reproductive
career. Males that responded best to the observed female mate
preferences had the highest long-term reproducti ve success. We
conclude that female mate-choice represents a sufficient cause for
the evolution of sex-biased dispersal in social mammals.
The observed female preferences were simple mate-choice rules
that radically reduced the chance of costly inbreeding for females.
These rules do not require direct kin discrimination, nor indirect
location-based kin discrimination (such as preference for immigrant
males or discrimination against natal males)
3,11,12
, both important
parameters in theoretical models that seek to explain male dispersal
in social mammals
3
. Instead, they are consistent with indirect time-
based kin discrimination cues
24
. This means that an intrinsically
demographic property—fluctuations in the number of young
females in different clans—can lead to male dispersal in the majority
of cases. Thus, changes in the demographic structure of groups will
alter the likelihood of males dispersing, and the demographic struc-
ture of a group in relation to other groups will set the level of emig-
ration from and immigration into that group.
METHODS SUMMARY
Study area and groups. All approximately 370 hyaenas of the eight Crater clans
were individually known
15
and studied between April 1996 and April 2006. Natal
adult males that attempted to mate with or shadowed females from their natal
clan or that excluded competing males from access to a female
25
were termed
‘reproductively active natal males’. Immigrant males were considered members
of the new clan if they initiated non-aggressive interactions with its members
over a period of at least three months. Date of clan selection was the date of first
sighting in the new clan’s territory (immigrant males) or of first observation of
mating, shadowing, or defending (reproductively active natal males). Male
tenure was calculated as the number of days from the date of clan selection until
the date of the event in question.
Social status, clan selection and paternity analysis. To test whether males
preferred to initiate their reproductive career in the clan with the highest number
of young females, the eight clans (including each male’s natal clan) were placed in
a linear rank order on each date a male selected a clan. Rank 1 was the clan with
the highest, and rank 8 the clan with the lowest number of young females. This
expected pattern of clan selection was compared with the observed pattern by
calculating Manly’s standardized selection ratio
26
B 5 Chesson’s a. Number of
young females included females that were between one and five years of age
18
.
Relatedness between individuals was calculated from known pedigrees based on
genetic paternity analyses; first-degree relatives were referred to as ‘closely
related’. Paternity analyses were based on amplification of six highly poly-
morphic microsatellite loci using genetic material from 575 Crater individuals
collected as previously described
18
. Results are quoted as means 6 standard
error, and probabilities are for two-tailed tests.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 21 March; accepted 21 June 2007.
1. Alberts, S. C. & Altmann, J. Balancing costs and opportunities: dispersal in male
baboons. Am. Nat. 145, 279
–
306 (1995).
2. Greenwood, P. J. Mating systems, philopatry and dispersal in birds and mammals.
Anim. Behav. 28, 1140
–
1162 (1980).
3. Lehmann, L. & Perrin, N. Inbreeding avoidance through kin recognition: choosy
females boost male dispersal. Am. Nat. 162, 638
–
652 (2003).
4. Dobson, F. S. Competition for mates and predominant juvenile male dispersal in
mammals. Anim. Behav. 30, 1183
–
1192 (1982).
5. Moore, J. & Ali, R. Are dispersal and inbreeding avoidance related? Anim. Behav.
32, 94
–
112 (1984).
6. Cockburn, A., Scott, M. P. & Scotts, D. J. Inbreeding avoidance and male-biased
natal dispersal in Antechinus spp. (Marsupialia: Dasyuridae). Anim. Behav. 33,
908
–
915 (1985).
7. Johnson, M. L. & Gaines, M. S. Evolution of dispersal: theoretical models and
empirical tests using birds and mammals. Annu. Rev. Ecol. Syst. 21, 449
–
480
(1990).
8. Smith, R. H. On selection for inbreeding in polygynous animals. Heredity 43,
205
–
211 (1979).
9. Parker, G. A. in Mate Choice (ed. Bateson, P. P. G.) 141
–
166 (Cambridge Univ.
Press, Cambridge, 1983).
10. Parker, G. A. in Sexual Selection and Reproductive Competition in Insects (eds Blum,
M. S. & Blum, N. A.) 123
–
166 (Academic Press, New York, 1979).
11. Pusey, A. E. & Packer, C. in Primate Societies (eds Smuts, B. B., Cheney, D. L.,
Seyfarth, R. M., Wrangham, R. W. & Struhsaker, T. T.) 250
–
266 (Univ. Chicago
Press, Chicago, 1986).
12. Keane, B. Dispersal and inbreeding avoidance in the white-footed mouse,
Peromyscus leucopus. Anim. Behav. 40, 143
–
152 (1990).
P < 0.0001
Most
2.0
1.5
1.0
0.5
0.0
2.5
Fewer
Youn
g
females in selected clan
Long-term reproductive success
Figure 4
|
The fitness benefits of male spotted hyaenas that selected clans
with the highest number of young females.
Long-term reproductive success
calculated as in Fig. 3. The box indicates the interquartile range around the
median (line inside the box), and the vertical error bars represent values plus
or minus 1.5 times the interquartile range.
LETTERS NATURE
|
Vol 448
|
16 August 2007
800
Nature
©2007
Publishing
Group
13. Kruuk, H. The Spotted Hyena. A Study of Predation and Social Behavior (Univ.
Chicago Press, Chicago, 1972).
14. East, M. L. & Hofer, H. Male spotted hyenas (Crocuta crocuta) queue for status in
social groups dominated by females. Behav. Ecol. 12, 558
–
568 (2001).
15. Ho
¨
ner, O. P., Wachter, B., East, M. L., Runyoro, V. A. & Hofer, H. The effect of prey
abundance and foraging tactics on the population dynamics of a social, territorial
carnivore, the spotted hyena. Oikos 108, 544
–
554 (2005).
16. Matthews, L. H. Reproduction in the spotted hyaena, Crocuta crocuta, (Erxleben).
Phil. Trans. R. Soc. Lond. B 230, 1
–
78 (1939).
17. East, M. L., Hofer, H. & Wickler, W. The erect ’penis’ is a flag of submission in a
female-dominated society: greetings in Serengeti spotted hyaenas. Behav. Ecol.
Sociobiol. 33, 355
–
370 (1993).
18. East, M. L., Burke, T., Wilhelm, K., Greig, C. & Hofer, H. Sexual conflicts in spotted
hyenas: male and female mating tactics and their reproductive outcome with
respect to age, social status and tenure. Proc. R. Soc. Lond. B 270, 1247
–
1254
(2003).
19. Engh, A. L. et al. Reproductive skew among males in a female-dominated
mammalian society. Behav. Ecol. 13, 193
–
200 (2002).
20. Hofer, H. & East, M. L. in Serengeti II. Dynamics, Management, and Conservation of
an Ecosystem (eds Sinclair, A. R. E. & Arcese, P.) 332
–
363 (Elsevier, Chicago,
1995).
21. Trivers, R. L. in Sexual Selection and the Descent of Man (ed. Campbell, B.) 136
–
179
(Aldine, Chicago, 1972).
22. Smale, L., Nunes, S. & Holekamp, K. E. in Advances in the Study of Behavior Vol. 26
(eds Slater, P. J. B., Rosenblatt, J. S., Milinski, M. & Snowdon, C. T.) 181
–
250
(Academic Press, London, 1997).
23. Kokko, H. & Sutherland, W. J. Optimal floating and queuing strategies:
consequences for density dependence and habitat loss. Am. Nat. 152, 345
–
366
(1998).
24. Sherman, P. W., Reeve, H. K. & Pfennig, D. W. in Behavioural Ecology—An
Evolutionary Approach (eds Krebs, J. R. & Davies, N. B.) 69
–
96 (Blackwell Science,
Oxford, 1997).
25. East, M. L. & Hofer, H. Loud calling in a female-dominated mammalian society: II.
Behavioural contexts and functions of whooping of spotted hyaenas, Crocuta
crocuta. Anim. Behav. 42, 651
–
669 (1991).
26. Manly, B. F. J., McDonald, L. L. & Thomas, D. L. Resource Selection by
Animals. Statistical Design and Analysis for Field Studies (Chapman & Hall,
London, 1993).
Acknowledgements We thank the Tanzania Commission for Science and
Technology for permission to conduct the study, the Tanzania Wildlife Research
Institute, the Ngorongoro Conservation Area Authority, A. Francis, L. Kimaay,
T. Ndooto, G. Orio, H. Richner, D. Thierer, C. Trout, L. Trout, C. Voigt and
W. Wickler for their assistance and suggestions. This study was financed by the
Leibniz Institute for Zoo and Wildlife Research, the Fritz-Thyssen-Stiftung, the
Stifterverband der deutschen Wissenschaft, the Max Planck Society, the German
Academic Exchange Service (DAAD) and the Messerli Foundation.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Correspondence and requests for materials should be addressed to O.P.H.
(hoener@izw-berlin.de).
NATURE
|
Vol 448
|
16 August 2007 LETTERS
801
Nature
©2007
Publishing
Group
METHODS
Study area and population. The floor of the Ngorongoro Crater in northern
Tanzania covers 250 km
2
and is inhabited by eight spotted hyaena clans with
between 24 and 65 members each that defended territories of 24 6 4km
2
(ref.
15). The Crater population is linked to the neighbouring Serengeti population by
individual movements and gene flow
27
. Both populations contain reproductively
active natal males and both have a similar incidence of inbreeding (Crater: 0.9%;
Serengeti
18
: 0.8%) and expected heterozygosity (Crater: 0.824, this work;
Serengeti
18
: 0.856). Sex, age and social status of individuals were determined
as previously described
15
. Individuals less than 12 months of age were classified
as cubs, those aged between 12 and 24 months as yearlings, and those 24 months
of age or older as adults. The date of conception was calculated from litter birth
dates on the basis of a gestation period of 110 days
16
.
Selection of clans. We assessed hypotheses for the evolution of male-biased
dispersal using a discrete choice (multinomial logistic) regression model
28
by
asking which of four variables predicts the clan in which males started their
reproductive career (this may have been their natal clan). The variables were
(1) intensity of male–male competition, (2) number of young females, (3)
number of unrelated females, and (4) mean number of main prey animals per
hyaena (adults and yearlings) on the dates of clan selection (Table 1). Intensity of
male–male competition was the length of the male social queue (that is, the
number of reproductively active natal males plus immigrant males), or the
number of reproductively active males per adult female. In spotted hyaenas,
the length of the male queue may be the more appropriate measure of male–
male competition because males need to build friendly relationships with
females to reproduce and queue for social status, and as a result, levels of aggres-
sion between males are low
14
.
From the perspective of each male, young females in non-natal clans were
those between one and less than five years of age on the date of clan selection,
since recent immigrant males rarely have contact with female cubs less than 12
months of age
25
. Young females in his natal clan were those that were born before
the male’s birth and less than five years of age. The number of unrelated females
was all adult females with a coefficient of relatedness r , 0.5. The mean number
of main prey animals per hyaena was determined from data on mean main prey
density and territory size
15
divided by the mean clan size (adults and yearlings).
Genetic analysis and survivorship of females. Methods for the collection and
processing of genetic material for paternity analysis have been previously
described
18,29
. Microsatellite loci were typed for 575 Crater individuals including
434 offspring (65.2% of all offspring born during the study period). Paternity
was assessed using maximum-likelihood methods as implemented in Cervus
30
.
All immigrant and reproductively active natal males that were clan members
when a litter was conceived were considered to be putative fathers. The mean
proportion of candidate males that were typed was 0.979; for 386 (88.9%) off-
spring all candidate males were typed. Hence, for 426 (98.2%) of the 434 off-
spring from which DNA was isolated, paternity was determined with 95%
confidence. The mean expected heterozygosity was 0.824, total exclusionary
power was 0.999, the mean proportion of individuals typed was 0.992, and the
error rate was 0.0052 and was set at 1.0%. The survivorship of young females was
calculated as the mean proportion of young females present at clan selection that
survived to the end of each year of male tenure.
Statistical analysis. Nonparametric tests, the discrete choice regression model
and the stepwise regression model were performed using Systat 11.0 (Systat
Software Inc.). For the stepwise regression model, natural-logarithm trans-
formation was applied to the dependent variable to satisfy the requirement of
normal distribution of residuals as judged by the Lilliefors test. The significance
of Wilcoxon signed-rank and Mann–Whitney U tests with sample sizes below 30
were based on exact P-values calculated with StatXact 7.0 (Cytel Inc.).
27. Albert, R. Genstruktur und Genfluß in ausgewa
¨
hlten Populationen der Tu¨pfelhya
¨
ne
(Crocuta crocuta). PhD thesis, Freie Universita
¨
t Berlin (2002).
28. McFadden, D. in Frontiers in Econometrics (ed. Zarembka, P.) 105
–
142 (Academic
Press, New York, 1974).
29. Wilhelm, K. et al. Characterization of spotted hyena, Crocuta crocuta,
microsatellite loci. Mol. Ecol. Notes 3, 360
–
362 (2003).
30. Marshall, T. C., Slate, J., Kruuk, L. E. B. & Pemberton, J. M. Statistical confidence for
likelihood-based paternity inference in natural populations. Mol. Ecol. 7, 639
–
655
(1998).
doi:10.1038/nature06040
Nature
©2007
Publishing
Group
