Behavioral Ecology and Sociobiology

, Volume 57, Issue 1, pp 91–100

Selected polyandry: female choice and inter-sexual conflict in a small nocturnal solitary primate (Microcebus murinus)

Authors

    • Abteilung SoziobiologieDeutsches Primatenzentrum
    • Institut für Zoologie & Zoologisches MuseumUniversität Hamburg
  • Peter M. Kappeler
    • Abteilung SoziobiologieDeutsches Primatenzentrum
Original Article

DOI: 10.1007/s00265-004-0823-4

Cite this article as:
Eberle, M. & Kappeler, P.M. Behav Ecol Sociobiol (2004) 57: 91. doi:10.1007/s00265-004-0823-4

Abstract

Sex-specific interests over the maximization of reproductive success lead to an inter-sexual conflict over the optimal mating system in a species. Traditionally, the outcome of this inter-sexual conflict has been studied from the male perspective but it also depends on female mating strategies, such as manipulating the temporal distribution of sexual activity, advertisement, and mate choice. We used a small nocturnal primate, the gray mouse lemur (Microcebus murinus) to determine the relative importance of female mating strategies on the outcome of this conflict in a species where females are solitary during their activity period. We studied their mating behavior over three consecutive annual mating seasons and determined the genetic relationships among more than 300 study animals to quantify individual reproductive success. We found that most females were receptive asynchronously. Females did not exhibit any obvious direct mate choice, probably due to a highly male-biased operational sex ratio and the corresponding costs of choosiness. However, females exercised indirect choice for multiple matings. They mated with 1–7 males up to 11 times during their single night of receptivity. As a result, mixed paternity was common but heavier males sired more offspring, meaning that indirect female choice for superior males cannot be excluded. Females exhibited a mixed mating strategy, avoiding costly direct mate choice but still counteracting male efforts to monopolize mating, successfully increasing genetic variability among offspring. Thus, females had a major influence on the outcome of the inter-sexual conflict despite male monopolization attempts.

Keywords

Female choiceSexual conflictPromiscuityMixed paternityMicrocebus

Introduction

When potential reproductive rates are male-biased, the reproductive success of males is primarily limited by the number of females they can fertilize, favoring the evolution of a male “copulatory imperative” (Ghiselin 1974) that leads males to tune their mating strategies towards monopolization of as many receptive females as possible (Clutton-Brock and Parker 1992). Female fitness, in turn, is mainly determined by their offspring’s survival, selecting for female choosiness: females should prefer high-quality males in order to obtain good genes and/or various male services (Eberhard 1998; Jennions and Petrie 2000; Gavrilets et al. 2001; Fedorkaf and Mousseau 2002; Kokko et al. 2003). These fundamentally different interests lead to a conflict between the sexes over the optimal mating system (e.g., Davies 1992; Moore et al. 2001; Chapman et al. 2003; East et al. 2003).

Traditionally, the outcome of this inter-sexual conflict has been studied from the male perspective (reviewed by Gowaty 2004). However, the male potential to monopolize females largely depends on female strategies, which include a range of behavioral and physiological mechanisms that alter the costs and benefits of male reproductive strategies (Ahnesjö et al. 1992; Zeh and Zeh 2003; reviewed for primates by Setchell and Kappeler 2003). We use the term “strategy” for a genetically fixed norm of reaction (“behavioral genotype”) containing a more or less variable set of “tactics” specifically expressed by environmental conditions (“behavioral phenotypes”) (Dominey 1984; Hazel et al. 1990; Gavrilets and Scheiner 1993; Brockman 2001). These female strategies include altering the temporal distribution of sexual activity, advertisement, and mate choice. We use the term “female choice” in a broad sense, referring to any female mating bias influencing male mating success (Wiley and Poston 1996; Kokko et al. 2003). Accordingly, on a proximate level, females can exercise choice either directly, through preference or resistance, or indirectly, e.g., through estrous advertisement that increases male-male competition. On an ultimate level, females can choose for direct benefits such as paternal care, or indirect benefits such as good genes. Finally, females can exercise choice before, during, and even after matings (Birkhead and Møller 1993; Eberhard 1996; Birkhead 2000b; Paul 2002).

In mammals with solitarily active females (as defined by Kappeler and van Schaik 2002), a situation characterizing the majority of mammals, male mating strategies are more complex than in gregarious species, and only little is known about corresponding female mating strategies. Males must either defend large territories to monopolize several receptive females or opportunistically search for females. Females are therefore expected to desynchronize their sexual activity and advertise estrus in order to increase their potential for mate choice. Desynchronizing sexual activity increases the costs of male territory defense as it extends the period during which territory defense is necessary (Emlen and Oring 1977). In addition, both estrous desynchronization and estrous advertisement increase male bias in the operational sex ratio (OSR) and, in turn, male-male competition. Desynchronization of estruses and estrous advertisement can therefore also be interpreted as mechanisms of indirect female choice for superior males. Especially in species with more than one pup per litter, females should choose multiple high-quality mates to increase genetic diversity among offspring (Berteaux et al. 1999; Jennions and Petrie 2000; Fedorkaf and Mousseau 2002).

Female choice and multiple mating also incur costs and are constrained by male tactics, in particular male monopolization success and its consequences for the risk of infanticide (Hrdy 1979; Agrell et al. 1998; van Noordwijk and van Schaik 2000). Especially when the OSR is highly male-biased, i.e., when single males cannot control access to females, females may become targets of harassment and coercion by numerous males (Smuts and Smuts 1993), which makes female choice costly. Costs of multiple mating include risk of predation during mating and transmission of sexual diseases (Reynolds and Gross 1990; Smuts and Smuts 1993; Clutton-Brock and Parker 1995; Nunn and Altizer 2004). High costs of choosiness through male harassment, or the risk of male infanticide together with the benefits of good genes, necessitate female strategies that confuse and bias paternity, namely multiple mating and cryptic female choice (van Schaik 2000; van Schaik et al. 2000; Zeh and Zeh 2003; Wolff and Macdonald 2004). Multiple mating can also be driven by the necessity to increase paternal care (Kleiman and Malcolm 1981; Clutton-Brock 1991; Huber et al. 2002). Thus, female mating strategies are constrained by various, partly opposing determinants.

The precise conditions favoring specific forms of female choice and multiple mating have largely remained obscure. Only a few studies have quantified actual male-female encounters, mating behavior and their consequences for reproductive success in solitary mammals, perhaps because the majority of them are nocturnal and difficult to study in the field (Sterling et al. 2000). The gray mouse lemur (Microcebus murinus) is a small (60 g), solitarily active nocturnal primate endemic to Madagascar. The mating system of gray mouse lemurs in Kirindy Forest in western Madagascar is characterized by highly seasonal reproduction, seasonally fluctuating dimorphism in body mass, relatively large testes, a high degree of home-range overlap within and between the sexes, and female grouping during diurnal resting, while males usually sleep alone (reviewed in Kappeler 2000). This combination of traits suggests a low male monopolization potential, pronounced sperm competition, and scramble competition polygyny as the main male mating tactic (Eberle and Kappeler 2002).

Predictions about female choice are less straightforward, partly because information about mating behavior in mouse lemurs is not available from previous field studies. Based on short-term local variation of the monopolization potential, males could successfully defend access to females, thereby preventing females from choosing other males, although in gray mouse lemurs, as in most other lemurs, females dominate males (Richard 1987; Kappeler 1993; Radespiel and Zimmermann 2001). Furthermore, female receptivity is limited to only a few hours during a single night (in captivity: Glatston 1979; Perret 1982; Wrogemann et al. 2000; in the field: Eberle and Kappeler 2004), constraining the potential for female choice on the time axis. However, most females in a population were receptive on different nights (Eberle and Kappeler 2002, 2004), which increases their potential to choose, and captive gray mouse lemurs advertise estrus by specific vocalizations and urine marking, also increasing the potential for female choice (Buesching et al. 1998).

The main aim of this study was to collect detailed behavioral and genetic data to determine more precisely the relative importance of female strategies for the outcome of this sexual conflict in a small solitary mammal, using gray mouse lemurs as an example. We predict that receptive females should be located by several males. Relatively free female choice should occur but females should mate with multiple males (Berteaux et al. 1999). Female choice is expected to be an important factor in reproductive outcomes, despite possible male monopolization.

Methods

Study area and trapping

We studied a free-living population of gray mouse lemurs (M. murinus) in Kirindy Forest, located about 60 km northeast of Morondava in western Madagascar (Sorg et al. 2003). Members of this population have been regularly captured and individually marked with sub-dermal transponders within a 9-ha study area since 1994, referred to below as the central study area, and within additional 21 ha surrounding the central study area since 1999. It should be noted that the “central study area” was not derived from a certain population structure but merely from the history of the coverage of our study area.

The entire study area is equipped with a rectangular system of foot trails at 25-m intervals. To trap mouse lemurs, we baited Sherman live traps with small pieces of banana and set them in the late afternoon near trail intersections on three consecutive nights per month. Since 1999, captures have been conducted in the central study area (160 trap locations) between March and December, and within the entire 30 ha (492 trap locations) in April and November. Additionally, during three mating seasons between 1999 and 2001, we trapped once per week in the central study area between mid-October and mid-November, to determine the females’ reproductive state (for details see below), and to equip proestrous females with radio-collars and to remove those of postestrous females. To minimize interference with the animals’ nocturnal mating activities, these traps were set during the second half of the night.

Captured animals were collected in the early morning, individually marked with subdermal transponders (or re-identified in case of recaptures), subjected to standard morphometric measurements and released at the site of capture in the following late afternoon. Body size was measured as the inter-orbital distance. For individual recognition of non-collared animals, we marked all captured animals using a system of rings shaved on the tail. On average, 91% of the males and 81% of females present during a given mating season could be marked this way. Tissue samples for genetic analyses were taken from all captured animals in the form of small (2–3 mm2) ear biopsies during brief anesthesia induced by applying 0.01 ml Ketanest 100 subdermally (see Rensing 1999). To mark as many juveniles per generation as possible, we equipped lactating females with radio-collars to determine their sleeping sites during the day, and intensified trapping locally around these sleeping sites.

Study population

We captured a total of 505 individuals 5,882 times between March 1994 and November 2001 within the entire study area (Table 1). All these animals were individually marked with a passive transponder. During the presented study, between 1999 and 2001, we captured 313 animals a total of 3,812 times. One hundred and sixty-three out of these 313 individuals had their center of activity within the central study area, and were captured a total of 3,079 times. During the 3 mating seasons between 1999 and 2001, 34, 27, and 32 females, and 39, 32, and 38 males, respectively, had their center of activity within the central study area. A center of activity was calculated for each animal as the mean xy-coordinates of an animal’s independent locations using trapping and observational data. Virtually all animals could be marked before maturity and later identified visually. Thus, our records of animals present in the central study area are as complete as possible for a small, nocturnal mammal (for further details, see Eberle and Kappeler 2004).
Table 1

Number of individuals captured in the study area. Note that the sum of individuals captured in 1999, 2000, and 2001 in the central study area is greater than the total of the 3-year study period (202 vs 163) because the total for the 3-year study period includes each individual only once, whereas those individuals that survived to the next year were included in 2–3 annual counts

Period

Area

Total

Females

Males

1994–2001

Entire area (30 ha)

505 (all transpondered)

223

282

This study (1999–2001)

Entire area

313 (all genotyped)

159

154

Central study area (9 ha)

163

74

89

1999

Central study area

73

34

39

2000

59

27

32

2001

70

32

38

Genetic analyses

Tissue samples from all 313 animals captured since 1999 were used for genetic analyses. DNA was isolated from ear biopsies following standard protocols (Qiagen QIAmp DNA Mini Kit no. 51306). For detailed microsatellite analysis, we used 17 loci with an average of 17 alleles. For 12 of them we developed new primers (Hapke et al. 2003); the remaining primers were developed by Wimmer (2000). We typed all 313 animals at these 17 loci. Parentage analysis based on combined mismatch and likelihood analysis was performed with Cervus 2.0 (Marshall et al. 1998). Candidate parents were excluded through at least two homozygous mismatches or one heterozygous mismatch. The likelihood analysis for non-excluded candidates was based on detailed parentage simulation (100,000 runs, 70 candidate parents, assumptions: 0.95 sampling rate, 0.92 average loci typing rate, 0.01 error rate, 1 close relative of the true parent among the other candidate parents) to estimate the resolving power of all loci and to estimate critical values to evaluate the parentage analysis statistically. Parentage analysis was then performed for each generation separately. Additionally, pair-wise relatedness R of all animals was calculated, using Relatedness 5.08 (Queller and Goodnight 1989).

We genotyped all 313 animals captured since 1999, 163 from the central study area and 150 from the surrounding area. In total 72 out of 74 females captured in the central study area since 1999 could unambiguously be assigned to 1 of 12 matrilineal pedigrees that continuously led back to 12 females born between 1994 and 1997. We determined the mothers of all juvenile females (n=42), and of most juvenile males (n=39 out of 67). The difference in the proportion of determined maternities in female and male juveniles is because females are philopatric whereas most males disperse already as juveniles, leading to a high proportion of immigrated males whose parents are unknown. We could determine the fathers of 62 out of 67 young with a mother from the central area (for further details, see Eberle and Kappeler 2004).

Determination of female reproductive states

The mating season of gray mouse lemurs in Kirindy Forest is limited to a 4-week period between mid-October and mid-November. During this period, an individual female mates only during a few hours in a single night (Eberle and Kappeler 2004). The reproductive state of females was determined by external examination of vulval morphology. The vulva of mouse lemurs is sealed with a membrane during most of the year (Petter-Rousseaux 1964). In addition, the vulval area is flat and inconspicuous. It opens only briefly around estrus and at birth. During the early estrous cycle, the vulva reddens and begins to swell up for 5–15 days. Eventually, the closing membrane ruptures and the vulva remains open for an average of 6 days during which a female is receptive for only a few hours, typically during 1 of the first 3 days of vulval opening. Soon afterwards, the vulval area collapses and the vulva starts to reseal, showing a characteristic scar several days following estrus (Glatston 1979; Eberle and Kappeler 2002).

We use the terms receptivity for the period between a female’s first and last mating (Beach 1976), estrus for the single night during which receptivity occurred, proestrus for the period before estrus starting with vulval swelling, postestrus for the period after estrus ending with completed vulval resealing, and anestrus for the remaining time. If a female could not be observed mating, the likely day of estrus was reconstructed using the assessment of female reproductive status at the days she was captured. This method, combined with our observations, allowed us to define the day of estrus of most of the females known to be present, without unduly interrupting their mating activities.

Observations

Focal animal sampling (Altmann 1974) was performed by M.E. between 1999 and 2001 during the three annual mating seasons between mid-October and mid-November and during the lactation season between January and April 2000. We equipped a total of 39 proestrous females during the mating seasons and 17 anestrous females during lactation with radio-collars (Table 2). Spatial data on the position of females were taken every minute. All occurrences of the following behavioral elements were recorded: Encounter: being within 10 m of the focal animal; Approach: locomote from beyond to within 1 m of the focal animal and begin interacting; Chase: sprint (>1 m) in pursuit of fleeing partner; Fight: holding and biting partner while emitting characteristic threat vocalization (Zimmermann 1995); Mate guarding: a male stays close (<0.5 m, >20 min) to a receptive female; Escape: a guarding male cannot follow the female and cannot be observed any more in her vicinity for at least 15 min; Sheltered: staying immobile in a hollow tree; Roaming: moving around in the vegetation (as opposed to being sheltered).
Table 2

Number of individuals observed between 1999 and 2001

Females

Males

Focal observation

56

14a

During proestrus

39

During estrus

27 (out of 39)

During postestrus

7 (out of 27)

During anestrus

17a

aIn 2000

We observed the 39 proestrous females for a total of 161 h. The number of days and the daily time span during which a proestrous female could be observed were determined by the timing of estrus of all females and, consequently, varied widely among individuals. Twenty-five females were observed during their complete night of receptivity and 2 during the first half of their night of receptivity. Seven females were observed for a total of 22 h during the first night of postestrus and 17 anestrous females for a total of 233 h during lactation. Additionally, 14 males were observed by focal sampling for 177 h parallel to females during the mating season in 2000 by an assistant (Table 2).

Proestrous females were followed alternately during the first hours of the night. Depending on the number of proestrus females that were equipped with radio-collars at a time, an individual female was observed for 30–60 min before switching to the next female. Most females started mating within the first hours of the night. The first female found mating was followed and observed exclusively for the rest of the night. During the day, we determined the sleeping sites of all radio-collared animals and the composition of their sleeping groups with the help of radio-tracking and a transponder-reading device.

Data sources and statistical analyses

The study was carried out over the course of three mating seasons, and one season of lactation. We combined observational data from focal observations with trapping data and genetic data. Observational data were available only from a small fraction of the population. Genetic data and trapping-based morphological and spatial data were available from most animals but most litters could not be determined (completely) because early offspring mortality was very high. Reproductive success could therefore not be related to all aspects of the observed mating behavior. The analyses of reproductive success and mixed paternities were based on genetic, morphological, and spatial data from all animals. Criteria used for inclusion of data in analyses are given with the results.

In order to identify determinants of female choice, data on male traits and genetic relationships between the actors were analyzed from the females’ point of view, leading to comparisons of data samples that are matched female-wise, although these samples stem from two groups of different males. For instance, for each female, the mean body mass of all males that mated only once with her were compared with the average body mass of all males that mated several times with her, leading to a test of matched-pairs with the number of females defining the sample size. Advanced behavioral analyses of the mating histories of the involved males were not possible because all matings (n=98) were observed during female focal observations when data on individual males are opportunistic.

Results

Spatio-temporal distribution of receptive females

In Kirindy Forest, receptivity of M. murinus females is limited to a few hours of a single night during a 3- to 4-week-long mating season between mid-October and mid-November. The following analyses include only females for whom the time span that included the single night of receptivity could be narrowed down to 1–7 days. During this short mating season, two-thirds of the females in the home range of a given male were receptive on different nights (median: 66%, range: 53–80, n=14 males and 30 females that had their center of activity in the home ranges of these males, including females from outside of the central study area) (for details see, Eberle and Kappeler 2004). The temporal distance of estruses among females sharing a daytime sleeping site was smaller than that between members of different sleeping groups (median: within 3.8 days (range: 0–11) vs between 4.5 days (range: 0–24), Mann-Whitney U-test (MWU), P=0.04, n=44 females with known sleeping associations) and the greater the distance between female centers of activity was, the greater was the time elapsed between their estruses (distance range: 9–405 m, time range: 0–32 days, MatMan 1.0 matrix correlation test (de Vries et al. 1993), 10,000 Permutations, Mantel’s Z=2,004,715, Pearson’s r=0.10, t=2.02, P=0.03, n=49 females).

Multiple matings

Between 2 and 15 males approached a female during her estrus (median: 7), maximally 2–6 at a time (median: 3, n=25 females observed during the entire night of estrus). We observed a total of 98 matings. A female mated with 1–7 males in total 2–11 times (both medians: 3, n=25 females as before). Only three females were monogamous. Eight females mated twice with one of the males, two females three times. Males that mated twice or three times with one female differed neither in body mass nor age from males that mated only once with the same female (Wilcoxon matched-pairs test (WMP), P=0.66 and 0.40, n=10 females for whom the considered traits could be determined for all of their mates). Females mated over a time span of 0.6–8.5 h during their night of receptivity (median: 4.4 h, n=25 females observed during the entire night of estrus). A copulation lasted 1.0–7.1 min (mean: 2.6 min, n=98 matings) and was always terminated by either female aggression towards the mating male or by an attack of the mating male by a rival.

Female choice

Direct choice—rejecting males

Between the first and the last mating, a female accepted almost all approaching males (Fig. 1). Out of 27 females, 23 accepted all males (median 3, range: 1–6), 1 female rejected 2 out of 9, 2 females 1 out of 7 and 1 female 1 out of 4 courting males. In this context “rejection” refers to physical attacks by the female of a male attempting to mount. The small number of rejected males did not permit reliable tests of the relationships between particular male traits and rejection. Two of these five rejected males sired offspring with other females during the same mating season. The average degree of genetic relatedness between a female and successful males did not differ from that between that female and males that were observed in her vicinity but did not mate (WMP, P=0.94, n=20 females for whom all mates could be determined as well as all males in her vicinity during estrus). Successful and unsuccessful males did also not differ in age or any morphological trait (body size, absolute or relative body mass or testes volume, WMP, P=0.24–0.58, n=20 females as before).
Fig. 1

Number of approaching males that a receptive female accepted (mated with, gray bars) or rejected (did not mate with, black bars). Receptivity was defined as the time span between a female’s first and last mating

Indirect choice—attracting males

Captive female gray mouse lemurs are known to advertise estrus by specific (ultrasonic) vocalizations and ritualized urine marking (Buesching et al. 1998). During this study of wild mouse lemurs, only data on ritualized urine marking could be collected. We found no evidence for olfactory estrous advertisement. Females marked more often during anestrus, compared to proestrus/estrus (anestrus: median 1.0 times/h (0.0–2.9), proestrus/estrus: median: 0.2 times/h (0.2–1.9), MWU, P<0.04, n=9 vs 8 females visible more than 4 h during minimum 3 different days of observation per female).

Indirect choice—escaping from males

Males tried to guard receptive females, i.e., to stay close to a female and to chase away rivals (Eberle and Kappeler 2004). However, immediately following most of the observed matings (68 of 98), the females escaped the mating male, as well as all other males present. In most cases after escapes, the next copulation of a female occurred with a different male than the former mate (56 of 62; in 6 of the 68 cases the escaping female did not mate anymore). Following the other 30 out of 98 matings, the females stayed in a shelter, such as a hollow tree. Following the first mating, females spent 60% of the rest of the night in 2 different shelters (range: 25–100% in 1–5 trees, n=25 females observed during the entire night of estrus). Most of the time the females spent in shelters was recorded during the first half of the night (median: first half of the night 3.9 h (1.9–5.6 h) vs second half 1.1 h (0.0–5.0 h), WMP, P<0.01, n=25 females as before).

Following copulations, females roamed or stayed in a shelter, depending on the age of the male with whom they mated last. When females roamed, the mating males were younger, compared to when they stayed in a shelter (roamed: median 1.9 years (1.0–4.0), stayed: 3.5 years (1.0–6.5), WMP, P=0.04, n=10 females that roamed and stayed following matings and for whom all mates could be determined). Male absolute body mass did not predict this female behavior (roamed: median 61 g (53–72), stayed: 65 g (53–83), WMP, P=0.18, n=10 females as before), although these two variables are highly correlated (Spearman r=0.58, P<0.001, n=127 males; only one data point per male taken from his last mating season in the area).

Males tried to guard receptive females but whether a female could be guarded or not depended on her roaming behavior. When females roamed, they were almost never guarded. When females stayed in a shelter, they were guarded most of that time (median: roaming 0% (0–20) vs sheltered 75% (0–100), WMP, P<0.001, n=27 estrous females). Despite the high proportion of time receptive females stayed in a shelter, most males encountered the female they subsequently mated with unguarded (80 of 98 matings, Chi2=38.0, P<0.001), and most of these unguarded females were roaming at the time (68 of 80 matings, Chi2=37.8, P<0.001). The less a female was guarded, the more mating partners she had (Fig. 2; Spearman’s r=−0.44, P=0.03, n=25 females observed during the entire night of estrus).
Fig. 2

Total time a female was guarded during her night of receptivity versus number of mating partners she had

Guarding was terminated for various reasons: 31 of 55 times the male just left the female, whereas in 24 cases the female escaped, in 8 of these cases while the guarding male fought with a rival. During four guarding episodes (with different males and females), the males aggressively tried to deter the female from escaping. When the females appeared in the entrance of the tree-shelter, the males attacked her, but three out of these four guarded females finally escaped. Females that escaped were next encountered by a different male than the former guard in most cases (23 of 24, Chi2=18.38, P<0.001, n=24 cases).

Indirect choice—plug removal

Most males transferred mating plugs, which may serve as an indirect post-copulatory mate guarding (Eberle and Kappeler 2004). Captive gray mouse lemur males can successfully displace plugs of previous males (unpublished data, see also Parga 2003) and we directly observed such a displacement on six occasions in the field. Females seemed unable to remove mating plugs. Plug removals always occurred after mounting and we frequently observed females that needed up to 2 days to remove a plug with their teeth.

Genetic outcome

Mixed paternities

For analyses of mixed paternity, parenthood data from all animals captured since 1994 were used. In Kirindy, gray mouse lemurs give birth to 1–3 young once per year (Eberle and Kappeler 2004). In total, 26 litters with more than 1 young could be determined from the 12 pedigrees, and 17 of these had mixed paternities. Fourteen litters with twins and two litters with triplets had two different fathers each, and in one litter a set of triplets had three different fathers. Given the data on number of mates per female, the assumption of twin births and that fertilization occurs at random with respect to mating order, male traits, or female choice, an expected probability M of mixed paternity can be calculated: M=(nyn)/ny, where n is the number of mates and y the number of young per litter (here: 2). The average stochastic probability across all females is 0.60, which does not differ from the empirical proportion of mixed paternity of 0.65 (17/26; Proportion difference test, P=0.72). Relatedness between mothers and exclusive fathers did not differ from that between mothers and fathers who shared paternity with at least one rival (MWU, P=0.50, n=9 vs 17 litters, relatedness between a mother and paternity-sharing fathers averaged per mother).

Female reproductive success

Offspring survival depended on infant body mass, as well as their mothers’ body mass, but not on the date of birth. For analyses of reproductive success, offspring body mass was averaged over an age of 8–12 weeks, and a mother’s body mass was averaged over 12 weeks when her offspring was between 8 and 20 weeks old. Offspring that survived to more than 12 months were heavier at an age of 8–12 weeks than offspring that died earlier (survived: mean=43.5, SD=6.5; not survived: mean=37.9, SD=7.4, t-test, P=0.03, n=15 vs 17 female non-sibling juveniles). Offspring body mass did not depend on birth date, i.e., on the mother’s time of estrus (Spearman’s r=−0.01, P=0.97, n=20 mothers and their female offspring). However, heavier females had heavier pups (Spearman r=0.54, P=0.03, n=16 twin rearing females) whereas offspring body mass did not depend on the father’s body mass during the corresponding mating season, although heavier males sired more offspring (Eberle and Kappeler 2004).

Discussion

The most important findings of our study revealed that females did not perform direct mate choice; rather, they exhibited indirect choice for multiple mates. As a result, male monopolization potential was limited and mixed paternity was common. Below, we discuss different forms and mechanisms of female choice and their consequences for the inter-sexual reproductive conflict in this small solitary primate.

Spatio-temporal distribution as a mating strategy

The spatio-temporal distribution of receptive females is a major determinant of male mating strategies and provides a potential mechanism for female manipulation of male behavior (e.g., Ims 1988; Eberle and Kappeler 2002). The results of our study revealed that receptive females were dispersed not only in space but also in time, despite pronounced seasonality of reproduction, but that synchronization among spatially clustered related females occurred.

The females’ dispersed spatial distribution during activity is probably necessitated by the nature of important food sources. Gray mouse lemurs feed on arthropods, homopteran secretions, fruits, and gum (Martin 1973; Hladik et al. 1980; Barre et al. 1988; Corbin and Schmid 1995), resources that cannot be exploited simultaneously by several animals (Kappeler 1997). Consequently, there is only little potential for females to tune their spatial distribution in terms of a mating strategy.

Seasonal environments with a short rainy season during which offspring can be raised, generally necessitate annual clustering of female reproductive activity (e.g., Ford and Pitelka 1984; Keller 1984; Ims 1990a). Thus, seasonal reproduction is probably also an adaptation to the ecological context rather than a strategy to limit male monopolization potential. However, seasonal reproduction still allows for female control of male monopolization potential because seasonality is not necessarily an operational synchrony (see also Pereira 1991). The term “operational” is used here to point out that the biological significance of a certain degree of temporal clustering largely depends on how many females a male can reach, and which states of an estrous cycle males can discriminate (Eberle and Kappeler 2002). Numerous males actually encountered a female around estrus but they spent only little time with proestrous females, indicating their ability to discriminate reproductive states (Eberle and Kappeler 2004). Despite pronounced seasonality, the observed temporal distribution of estrous females within the mating season represents an operational asynchrony for males, because for males it is still feasible to achieve multiple matings by roaming.

The impact of the degree of synchrony on the number of mating opportunities, however, remains puzzling. In birds, where pair-living and paternal care is more common, synchrony promotes extra-pair paternity (Stutchbury 1998) although many studies provide mixed results (e.g., Kempenaers 1997; Weatherhead 1997; Yezerinac and Weatherhead 1997; reviewed in Griffith et al. 2002). In solitary mammals, asynchrony facilitates both male monopolization and multiple matings (e.g., Ims 1988). When population density is high and range overlap pronounced, however, de-synchronization of estruses increases the potential for multiple mates for females but the male monopolization potential is limited by a highly male-biased OSR. The observed operational asynchrony could therefore be a female strategy in order to promote increased genetic variability among offspring. In gray mouse lemurs, however, the extent of asynchrony is not only limited by the seasonal environment but also by other constraints: closely related females form sleeping groups and rear their offspring cooperatively (M. Eberle and P.M. Kappeler, unpublished work), which necessitates a certain amount of temporal coordination of reproduction.

Female choice

Females can exercise choice either directly or indirectly, as well as for direct or indirect benefits. They can choose before, during, and even after matings. In species in which females are dominant, relatively free direct female choice should be possible. Female gray mouse lemurs dominate males at feeding and sleeping sites (Pagès-Feuillade 1988; Kappeler 1993; Radespiel et al. 1998; Schmid 1998; Radespiel and Zimmermann 2001) and aggressively reject male attempts to mate outside the period of receptivity (Andrès et al. 2003). However, they are not able to reject males during receptivity. Compared to the defense of feeding and sleeping sites, males might be more motivated to mate, as mating opportunities are particularly crucial to fitness, but rare. Accordingly, coercion by males plays an important role in various taxa (e.g., Smuts and Smuts 1993; Clutton-Brock and Parker 1995; Soltis et al. 1997; Pizzari and Birkhead 2000). For similar reasons, females may also not be able to prevent a male from repeated matings. Repeated matings could therefore be interpreted as a result of male mating strategies in order to increase the probability of paternity.

When direct female choice is too costly, females can still increase their fitness and bias matings through indirect choice (Wiley and Poston 1996). Female gray mouse lemurs can escape particular males or choose post-copulatory via cryptic female choice (Eberhard 1996; Dixson 1998; Birkhead 2000a; Birkhead and Kappeler 2004).

Whether gray mouse lemurs performed such indirect female choice remains unknown. It is not clear whether offspring-siring males were only more successful in male-male competition or whether they were also selected by the females. The relationship between female escapes and male age could be a result of the older males’ ability to follow a female more successfully because spatial experience is an important male trait when searching is the predominant mating tactic (Schwagmeyer 1988, 1994; Ims 1990b; Schwagmeyer et al. 1998). Accordingly, older males mated earlier and had increased reproductive success (Eberle and Kappeler 2004).

Whether there is potential for cryptic female choice, e.g., through genetic incompatibility, cannot be answered with our present data. In general, cryptic female choice is still poorly studied in mammals (Tregenza and Wedell 2000) and in primates in particular (Setchell and Kappeler 2003; Birkhead and Kappeler 2004). Female mouse lemurs could not efficiently remove mating plugs, which is in contrast to observations in other species (e.g.: muriquis (Brachyteles arachnoides), Strier 1999, ring-tailed lemurs (Lemur catta), mandrills (Mandrillus sphinx): Setchell and Kappeler 2003; tree squirrels (Sciurus sp.): Koprowski 1992). It remains obscure why female gray mouse lemurs have not developed this ability, as it would increase their potential for post-copulatory choice.

Another form of indirect female choice is to mate rather indiscriminately and multiply in order to increase the genetic diversity of offspring or to ensure fertilization. Attracting males through estrous advertisement would increase the number of potential males. Why we could not confirm olfactory estrous advertisement as observed in captivity remains puzzling. Genetic benefits of multiple mating, however, remain controversial (Simmons 2001), and other potential direct female benefits of multiple matings, such as soliciting paternal care or reducing the risk of infanticide, appear unlikely in this case. As in most mammals, paternal care does not occur in gray mouse lemurs and infanticide is only beneficial to males when it increases the probability of future fertilizations (van Schaik and Janson 2000). In Kirindy, the reproductive season is too short to raise two litters per year and the probability of survival to the next year is only 0.5 for adult animals (unpublished data). Multiple matings and repeated matings with individual males may be the result of “making the best of a bad job” in the face of male harassment. However, females have various possibilities to reduce the costs of multiple mating: a short period of receptivity and temporary hiding in shelters during nights of receptivity. The fact that females did not hide permanently after the first mating implies that they chose to mate multiply, actively counteracting male monopolization.

Conclusions

Detailed predictions about the impact of female mating strategies in this solitary small mammal on the outcome of the inter-sexual conflict are difficult because of the various determinants of the costs of female choice and multiple mating, as well as the mutual interactions between these determinants. Despite expected selection for female choosiness, female mouse lemurs had only a few options to increase their fitness because of a highly male-biased OSR but, at the same time, females successfully limited male monopolization potential. Females exhibited a mixed mating strategy, avoiding costly direct mate choice but still counteracting male mating efforts by successfully increasing genetic variability among offspring. Thus, females also have a major influence on the outcome of the inter-sexual conflict despite male monopolization attempts.

Acknowledgements

We thank Berthe Rakotosamimanana (Département de Paléontologie et d’Anthropologie Biologique de l’Université d’Antananarivo), Olga Ramilijaona and Daniel Rakotondravony (Département de Biologie Animale, Université d’Antananarivo), Lucien Rakotozafy (Parc Botanique et Zoologique Tsimbazaza Antananarivo), the Commission Tripartite and the CAFF of the Direction des Eaux et Forêts, the CFPF Morondava, and Jörg U. Ganzhorn, Hans Zischler and Andreas Hapke for their authorization or support of this study. Thanks go to Tiana Andrianjanahary, Karoline Franz and Wiebke Plästerer for assistance in the field, and to Dietmar Zinner, Patty Gowaty, Lee C. Drickamer and an anonymous reviewer for their helpful comments on an earlier version of the manuscript. This article is part of a doctoral study by M. Eberle in the Faculty of Biology, University of Hamburg, with financial support from the Deutsches Primatenzentrum (DPZ) and the Deutsche Forschungsgemeinschaft (DFG, Ka 1082/5-1, 2).

Copyright information

© Springer-Verlag 2004