Causes and consequences of living in closed societies: lessons from a long-term socio-genetic study on Bechstein’s bats


Gerald Kerth, Fax: +49 (0)3834 86-4252; E-mail:


Understanding the ecological, behavioural and genetic factors influencing animal social systems is crucial to investigating the evolution of sociality. Despite the recent advances in population genetic methods and the analysis of social interactions, long-term studies exploring the causes and consequences of social systems in wild mammals are rare. Here, we provide a synthesis of 15 years of data on the Bechstein’s bat (Myotis bechsteinii), a species that raises its young in closed societies of 10–45 females living together for their entire lives and where immigration is virtually absent. We discuss the potential causes and consequences of living in closed societies, based on the available data on Bechstein’s bat and other species with similar social systems. Using a combination of observational and genetic data on the bats together with genetic data on an ecto-parasite, we suggest that closed societies in Bechstein’s bats are likely caused by a combination of benefits from cooperation with familiar colony members and parasite pressure. Consequences of this peculiar social system include increased sensitivity to demographic fluctuations and limits to dispersal during colony foundation, which have broad implications for conservation. We also hope to illustrate by synthesizing the results of this long-term study the diversity of tools that can be applied to hypothesize about the factors influencing a species’ social system. We are convinced that with the expansion of the number of social mammals for which comparably detailed socio-genetic long-term data are available, future comparative studies will provide deeper insights into the evolution of closed societies.


As animals differ strongly in the social composition of their groups (Kappeler & van Schaik 2002; Krause & Ruxton 2002), exploring the interaction between different behavioural strategies and patterns of genetic relatedness is crucial for understanding the evolution of animal social systems (Lott 1991; Emlen 1994). However, long-term studies that document the social and genetic structure of animal societies are rare (Kruuk & Hill 2008), as they require detailed information on both genetic and spatial structures of wild populations as well as their temporal dynamics. These problems are particular acute for studies that work with recognizable individuals, even though such individual-based studies are of particular importance for recording the structure of animal societies (Clutton-Brock & Sheldon 2010). Consequently, the few studies on social mammals that combine individual-based long-term field data with detailed knowledge of individual relatedness have had a significant impact on the fields of behavioural ecology, sociobiology, and evolutionary biology (Clutton-Brock 2002; Kruuk & Hill 2008; Clutton-Brock & Sheldon 2010).

Here, we synthesize published results and present new data analyses of a 15-year, individual-based study on the population genetic and social structure of the Bechstein’s bat (Myotis bechsteinii), a species where females live in closed societies with stable membership and (almost) no dispersers coming in. Subsequently, we compare these results to information available from other well-studied species living in similarly closed societies. Our aim is to explore: (i) the causes and consequences of living in closed societies and (ii) how a range of observational and genetic data and analyses can complement one another in studying the social system of vertebrates. Although we focus on a species with a closed society, we believe that the techniques and multifaceted approach used and discussed are applicable to most vertebrate societies.

Mammalian social systems

The social system of a species can be described as a function of three components: (i) social organization, describing the size, composition and cohesiveness of a social unit, (ii) social structure, encompassing social interactions and relationships among its members and (iii) the mating system, describing the social and genetic aspects of reproduction (Kappeler & van Schaik 2002). Most evidence suggests that social systems have developed as a by-product of individual strategies to optimize inclusive fitness (Clutton-Brock 2009; but see Wilson & Wilson 2007). In many mammals, this has led to the evolution of female philopatry and subsequently to the formation of groups (Clutton-Brock 1989). Sociality offers multiple benefits, e.g. improved foraging and protection against predators and inclement weather. If females form stable social groups, often consisting of matrilineal relatives, they frequently cooperate, e.g. in rearing young (Clutton-Brock 2009).

In many species of primates, carnivores and cetaceans, as well as some bats, females form societies with long-term social relationships (Dunbar & Shultz 2010; Kerth et al. 2011). In such societies, males are often able to defend sexual access to females or the resources required by females and as a result mixed-sex groups are formed with uni-male or multi-male polygyny (Clutton-Brock 1989). However, if males do not profit from being permanent group members, mating occurs in temporary harems or lekking aggregations (e.g. in temperate-zone bats; McCracken & Wilkinson 2000). Finally, in some mammals female natal philopatry is so strong that individual movements between societies become exceedingly rare, leading to the formation of closed societies, as in Bechstein’s bats (Kerth et al. 2000). We define a closed society as a social group where membership is stable (usually mediated via female philopatry) with no (or extremely rarely) dispersers coming in, and with social interactions between its members.

Bat sociality

Despite being largely ignored in discussions of mammalian societies (e.g. Clutton-Brock 2009), bats display one of the most diverse ranges of social systems among mammals (McCracken & Wilkinson 2000; Kerth 2008b). Most of this diversity is found in tropical species, whereas temperate-zone bats mainly have adopted a three-phase annual system known as the temperate cycle (Bradbury 1977). Each year, individuals cycle through mixed-sex aggregations in the winter (hibernation), a period of sexual segregation at maternity season in summer (usually only the females are social at this time), and a mating period with mixed-sex aggregations in autumn and/or spring (McCracken & Wilkinson 2000). In most temperate-zone bats, the males disperse and the females, despite their ability to fly long distances, are philopatric, as in most other mammals (Burland & Worthington Wilmer 2001). The social organization of bat maternity colonies ranges from aggregations with flexible membership to closed societies, where females are faithful to their natal colony, have individualized social bonds, and can actively reject intruding noncolony members (Kerth 2008b).

Potential causes of closed societies in bats

The existence of closed societies in some bats raises the question why females of these species do not switch colonies. The simple explanation that dispersal barriers restrict movement between colonies can be ruled out for many species with social groups that could be reached within minutes of flight, as is the case with Bechstein’s bats (Kerth et al. 2000) and spear-nosed bats (Phylostomus hastatus; McCracken & Bradbury 1981). More plausible ecological explanations for the existence of closed societies include benefits arising from familiarity with the local habitat (e.g. knowledge of suitable roosting and foraging sites), in combination with grouping benefits (e.g. through social thermoregulation), which together may explain the evolution of closed societies through selection for natal philopatry (Emlen 1994). Moreover, given their gregarious nature, bat colonies are host to a wide range of (often highly specialized) ecto- and endo-parasites, as well as bacteria and viruses (e.g. Allen 1962; Calisher et al. 2006). Thus, bats may also benefit from living in closed societies because it reduces exposure to (new) parasites and diseases by restricting their transmission between colonies (Kerth et al. 2002b; Altizer et al. 2003). In addition to ecological reasons, social benefits from living in groups with stable memberships, such as cooperation among related and/or familiar individuals, may explain the occurrence of closed societies in bats (Kerth 2008b), and other social animals (Emlen 1994; Clutton-Brock 2002).

Potential consequences of closed societies

Living in closed societies can have significant consequences for a species’ social and mating behaviour. A mating system providing sufficient opportunities for outbreeding with unrelated males becomes necessary once female philopatry is established, in order to avoid deleterious effects of inbreeding (Weatherhead & Forbes 1994; Perrin & Mazalov 1999). Due to the natal philopatry of females, mitochondrial (mt) DNA should be highly conserved among society members, with the only differences stemming from the diversity of the founding females, and mutations that arose afterwards. Nuclear (nuc) DNA, however, should be much less structured among societies if outbreeding leads to gene flow between them. If females mate promiscuously, this may even lead to an almost complete lack of nuclear genetic structure between societies, and high genetic diversity within them (Kerth et al. 2002a). It is also important to note that with promiscuity, apart from direct descendants, society members will only be half-sibs at best in species where females produce a single young at a time, as in many bats (Kerth et al. 2002b).

Finally, living in closed societies can have important consequences for dispersal. In obligate social animals, as in most female bats, this may only be possible through the establishment of completely new societies (group fission; Van Horn et al. 2007; Metheny et al. 2008a). This could limit the colonization capabilities of the species involved (Kerth & Petit 2005). Therefore, species with closed societies may be particularly vulnerable to extinction because of their potentially restricted dispersal options, the absence of rescue effects via immigration of individuals from other societies, and an increased vulnerability to Allee effects, resulting from the dependence on long-term social partners for cooperation, for example during information transfer.

Studying closed societies in Bechstein’s bats

For investigation of the social structure in Bechstein bats, four colonies have been extensively followed over 15 years, and all members had been implanted with microchips (PIT-tags), resulting in individual-based long-term field data of the bats (Kerth & Konig 1999; Kerth et al. 2011). The members of the four colonies were captured at least once a year to take biometric measurements, reproductive status and ecto-parasite infestation, as well as for the collection of a 3-mm2 piece of wing tissue of previously unmarked juveniles and subadults for genetic analysis. An extensive amount of genetic information is also available, ranging from large-range phylogeographic studies (Kerth et al. 2008), to local population dynamics and genealogy for individual colonies, including those for which long-term field data are available (Kerth et al. 2011). Finally, because additional insight into the social organization may be gained by studying the population structure of associated parasites (Criscione et al. 2005), the population genetic structure of the bats had been compared with that of their ecto-parasitic wing-mite Spintunix bechsteini from 24 bat colonies in Germany (Bruyndonckx et al. 2009; J. van Schaik et al. in Prep.).

General life history of Bechstein’s bats

The Bechstein’s bat is a medium-sized bat species (±10 g) that occurs in large parts of Europe, Asia Minor and the Caucasus region (Dietz et al. 2007). From April to September, females form maternity colonies to raise young communally in tree cavities, preferentially in mature deciduous forests. If available, they also regularly occupy bird and bat boxes. Maternity colonies comprise between 10 and 45 females of different age, relatedness and reproductive status (Kerth et al. 2002b, 2011). Females switch day roosts frequently and use as many as 50 different communal roosts throughout the maternity period. During that time colonies also regularly split into two to six roosting groups (the observed maximum distance between roosts of the same colony was 2.8 km; Kerth & Konig 1999; Kerth & Melber 2009). Individual foraging areas are generally in close proximity to roosting sites (often <1 km), and stable across years (Kerth & Melber 2009). Despite their high fission–fusion dynamics, different colonies do not share roosts during summer, even if they live in close proximity (Kerth et al. 2000). The females profit from communal roosting through social thermoregulation (Pretzlaff et al. 2010) and coordinate themselves by means of information transfer and group decisions about suitable roosts (Kerth & Reckardt 2003; Kerth et al. 2006). Males are solitary during summer and disperse from their natal colony in the first year of life (Kerth & Morf 2004). In autumn, both sexes swarm in front of caves and other underground structures for mating (Kerth et al. 2003) and potentially to investigate possible hibernacula (Kerth 2008b). During winter, Bechstein’s bats are rarely observed, but if found they hibernate solitarily or in very small clusters in caves, mines and cellars (Dietz et al. 2007).


For this review, we took the opportunity to analyse the available genetic and field data sets using several different methods in order to investigate their efficacy in studies of social structure. Starting in 1996, more than 300 bat boxes and several accessible tree cavities located in the home ranges of four colonies have been monitored each summer for two to seven days per week, resulting in more than 42 000 individual roosting recordings. Five other colonies, also living in forests close to the city of Würzburg, Germany, were individually marked with aluminium split rings after 2004. Thus, overall, long-term capture–mark–recapture data were available from nine colonies.

The entire genetic data set comprised 56 colonies (Germany 45, Switzerland 1, Bulgaria 7, and Turkey 1; data from Kerth et al. 2008; plus 6 new colonies). All bats have been typed at eight nuclear and two mitochondrial di-nucleotide microsatellite loci (for details see Kerth et al. 2002a). Genetic variation within colonies (gene diversity for nucDNA and haplotype diversity for mtDNA was calculated for the whole data set using fstat 2.9.4 (Goudet 2001). From three different regions (Rhineland-Palatine, Germany (RP), Lower Frankonia, Germany (LF), and the Balkans, Bulgaria and European Turkey (BK)) we selected eight colonies each, with comparable numbers of bats sampled (mean per region: BK: 13.1; RP: 15.8; LF: 16.1). We compared nucDNA population structure among these 24 colonies using several approaches. First, an amova was performed using Arlequin 3.5 (Excoffier & Lischer 2010) to investigate the level of variation within and between colonies and regions. Second, differentiation between colonies was investigated using pairwise FST-values (Goudet et al. 1996), GST-values (Meirmans & Hedrick 2011) and DST-values (Jost 2008). Third, assignment tests were carried out in structure (Pritchard et al. 2000) using a burnin of 100 000 and a runtime of 1 000 000 iterations, using an admixture model with correlated allele frequencies. For the two mtDNA microsatellites we constructed a combined haplotype (Kerth & Petit 2005) and subsequently calculated FST-values and performed amova and structure analyses in the same fashion as for nucDNA. Overall, we analysed genetic structures on three different population levels: (i) within colonies, (ii) among colonies within regions, and (iii) among colonies between regions.


Genetic structure within colonies

The mtDNA microsatellites showed a low to moderate diversity within colonies (2.42 ± 1.29 haplotypes per colony for the combined microsatellites). The majority (36 of the 57 colonies) had one or two haplotypes, and only four colonies had more than four haplotypes. It is important to note that heteroplasmy is commonly observed in these microsatellites (Mayer & Kerth 2005). Therefore, multiple alleles per colony that differ by one repeat unit from each other, leading to a continuous, unimodal allele distribution within colonies (Fig. 1), are likely to result from frequency shifts in the alleles from one generation to another. In contrast, the presence of alleles that differ by more than one repeat unit, leading to a bimodal distribution within colonies (Fig. 1), probably is caused by the immigration of foreign bats into the colony, or alternatively, result from multiple mtDNA lineages during the foundation of the colony. Overall, based on a visual inspection, only 9 (16%) out of the 57 colonies had a bimodal allele distribution. Thus, the observed mtDNA variation within colonies is in agreement with strong female philopatry leading to the formation of few maternal lineages per colony (compare Kerth et al. 2000).

Figure 1.

 Distribution of AT-repeat numbers of a mtDNA microsatellite in nine selected Bechstein’s bats colonies. Of the altogether 57 colonies that had been analysed, 9 had a bimodal allele distribution (left graphs show three examples), 30 a unimodal distribution (middle graphs show three examples), and 17 were monomorphic for a given allele (right graphs show three examples).

In contrast to mtDNA, nuclear diversity was high within colonies (0.82 ± 0.06), indicating a high rate of gene flow between colonies and a low average relatedness among colony members (compare Kerth et al. 2002a, b). This is also supported by the amova, which shows 96% of nucDNA variance to be within colonies.

The combined genetic and mark–recapture data show that the four continuously monitored colonies, were genetically and demographically very heterogeneous. For example, in 2010 they comprised 112 adult females that were between 1 and 16 years old (average: 4.7 ± 3.8 years), and of which 69% reproduced in that year. In each colony, closely related females (mother–daughter, grandmother–granddaughter, aunt–niece pairs, and half-sisters) lived together with genetically largely unrelated females (Fig. 2; compare Kerth et al. 2002b, 2011).

Figure 2.

 Social organizations of two wild Bechstein’s bat colonies in the year 2007. The last four digits of their PIT-tags identify individuals. The upper graph shows the genealogy and community membership of 42 adult females living in the colony GB2 (for details about communities see Kerth et al. 2011). PIT-tag numbers without female symbols symbolize nine mothers that lived in the colony in earlier years. Bats born after 1995 (the year of birth is given on the right side of the graph) are assigned to a specific mother or daughter using 11 polymorphic nuclear microsatellites (Kerth et al. 2002b). In this colony two mtDNA lineages occur (compare Fig. 1) indicated by orange and blue, respectively. The lower graph shows the genealogy of 18 adult females living in the colony BS, where no distinct communities occurred. In this colony also two distinct mtDNA lineages occur (compare Fig. 1) indicated by green and yellow, respectively.

Genetic differentiation among colonies

Mitochondrial DNA was strongly differentiated between colonies. High FST-values (0.74 ± 0.15) and differentiation was mainly found within regions rather than between them (50.15% and 9.95% variance; Table 1; compare Kerth & Petit 2005; Kerth et al. 2008). This, again, is consistent with (almost) complete demographic isolation of colonies and very strong female natal philopatry. These findings are confirmed by the long-term field data of marked females. No immigration of PIT-tagged females was observed in the four continuously monitored colonies, based on more than 42 000 individual re-sightings during 15 years. Moreover, we observed no immigration of marked females in any of the five colonies whose members had been marked with aluminium split rings.

Table 1. amova results of the three groups of eight colonies (Balkans, Rhineland-Palatine, Lower Frankonia) for (a) mtDNA, and (b) nucDNA. Values in parentheses indicate variance levels when the two German groups are considered one
Sourced.f.Sum of squaresVariance componentsPercentage of variation
 Among regions242.7340.0481 (0.0166)9.95 (3.46)
 Among colonies within regions21164.8570.2425 (0.2686)50.15 (56.19)
 Within colonies748144.3060.1929 (0.1929)39.9 (40.35)
 Among regions229.0870.0327 (0.0434)0.95 (1.26)
 Among colonies within regions21120.7520.0766 (0.0799)2.23 (2.31)
 Within colonies7482491.633.3311 (3.3311)96.82 (96.43)

For nucDNA populations were only weakly structured. An amova found very little differentiation both between colonies within regions and between regions (0.95% and 2.23%, Table 1; compare Kerth & Petit 2005; Kerth et al. 2008). Likewise, assignment tests carried out in structure, were not able to detect any underlying population subdivision, although a weak structuring was observed between Balkan and Central European samples (see supporting electronic information). Finally, the comparison between FST, GST and DST-values shows that all measures have similar trends but with some notable differences (see supporting electronic information).


Intracolony social structure

Based on genetic and mark–recapture data, we can conclude that Bechstein’s bat colonies are composed of several matrilines, which are strongly philopatric. Currently, however, it is unclear whether the existence of two very distinct mtDNA lineages in 16% of the observed colonies (with a bi-modal haplotype distribution) are the result of rare immigration events or mostly reflect the founding history of the respective colonies (i.e. multiple mtDNA lineages founded a colony). Nuclear genetic diversity is maintained through mating away from the maternity colonies (Kerth et al. 2003) leading to high levels of diversity and low relatedness.

Recently, a detailed analysis of the fission–fusion dynamics in two colonies revealed nonrandom individual roosting associations that could partially be explained by kinship, with closely related females roosting together preferentially (Kerth et al. 2011). Nevertheless, unrelated females also formed long-term social bonds, and the mtDNA haplotype of females did not explain the strength of their associations (Kerth et al. 2011). For comparison, in big brown bats (Eptesicus fuscus), the only other bat species with fission–fusion behaviour for which genetic data are available, related females do not roost preferentially together (Metheny et al. 2008b). Finally, in Bechstein’s bats neither age nor reproductive status affected the strength of roosting associations consistently, a finding that partially contrasts to previous results, where reproductive status had a stronger effect than genetic relatedness (Kerth & Konig 1999). The precise reason for this contrast is unknown. It may be related to differences in methodology and sample size of the two studies, or indicates some flexibility in the social organization of Bechstein’s bat colonies.

Intercolony social structure

Similar to within colony analyses, substantial mtDNA differentiation between colonies suggests strong natal philopatry in females, and low levels of nuclear genetic variation within regions suggest outbreeding at swarming sites. Additionally, only weak structuring was seen across Europe suggesting that male dispersal and mating at swarming sites is effective at maintaining genetic homogeneity over large ranges.

These findings are confirmed by the long-term field data of marked females. No immigration of individually marked females into foreign colonies was observed. It is important to note that five of the altogether nine marked colonies lived together in the same forest with at least one other marked colony within potential daily roost-switching distance (2.8 km). In the permanently monitored BS colony, however, two unmarked females appeared, whose mtDNA haplotypes did not fit to that of the BS females. Instead they matched the haplotypes present in the colony IB3 that lived within 1 km of BS. Both immigrants arrived as subadults and established themselves successfully, changing status of BS from a colony with a unimodal haplotype distribution to one with a bimodal distribution. There was no evidence that these immigrating females and their offspring were less associated within the colony’s social network than the original colony members (Kerth et al. 2011). Interestingly, BS females behaved aggressively in previous experiments when confronted with members of the colony UH, carrying haplotypes similar to those of the later immigrants (Kerth et al. 2002b).

Comparison of host and parasite genetic structure

An analysis of a 513-bp cytochrome b sequence of mtDNA of the parasitic wing-mite S. bechsteini revealed strong genetic differentiation between bat colonies (FST: 0.68) but found no evidence for isolation by distance. Moreover, temporal differentiation analysed in five colonies between samples from 2002 and 2007, showed strong genetic differentiation of mites within colonies between years (Bruyndonckx et al. 2009). Together with the lack of isolation by distance, this high temporal turnover suggests strong genetic drift in combination with substantial dispersal over large distance, as would be expected through transfer at swarming sites or hibernacula.

Bechstein’s bat social system

Using the observational, genetic, and parasitic data available for Bechstein’s bat, as well as supporting evidence from other species, we can begin to hypothesize about the origins and evolution of their peculiar closed societies.

We start with the observation that temperate-zone bats are much less variable in their social organization and mating system than tropical bats (McCracken & Wilkinson 2000; Kerth 2008b). The fact that in almost all temperate-zone bats, the females form maternity colonies suggests that in order to raise young in colder climates, females are forced to aggregate for social thermoregulation. However, it does not explain why males are rarely part of these maternity colonies, and why females are usually philopatric in temperate-zone bats, leading, in extreme cases, to closed societies observed in Bechstein’s bats.

The absence of males from the colonies probably has a twofold explanation. First, as in other mammals that cannot directly defend females or the resources they require (Clutton-Brock 1989), males resort to a promiscuous mating system, such as temporary harems or lekking. As female bats are highly mobile in three dimensions, are able to store sperm for several months, and their colonies often show high fission–fusion dynamics, in many species it becomes impossible for males to monopolize females (McCracken & Wilkinson 2000). Consequently, cohabiting with female bats during and directly after parturition is unlikely to strongly increase reproductive output in temperate-zone bats. Nonetheless, this does not explain why males rarely join the females in the maternity colonies.

We propose that in many species incompatibility in thermoregulatory requirements exists between males and females, although alternative explanations are possible (Senior et al. 2005; Safi et al. 2007). While females only infrequently go into daily torpor during gestation and lactation in spring and early summer in order to facilitate the growth of their offspring, males only require such investment during spermatogenesis in late summer and autumn, at times when females frequently use torpor to save energy (Racey & Entwistle 2000; for Bechstein’s bat see Pretzlaff et al. 2010). In species that roost in large cavities such as underground sites or buildings, sexual segregation may be present on fine scale only, with males and females roosting separately but in proximity to each other (and together at times when energetic requirements are similar). This incompatibility in seasonal timing of daily torpor of the two sexes in combination with the inability of males to monopolize females within their colonies, could explain female-only maternity colonies in most temperate-zone bat species. Upon segregation, further partitioning of roosting and hunting grounds is sometimes seen in species where females occupy optimal habitat (Senior et al. 2005; Safi et al. 2007). However, it is unclear whether males are forced from the optimal habitat by females, altruistically move away for the benefit of their kin in the colony, or simply avoid competition with the females by dispersing to available suboptimal habitats that are sufficient to sustain the lower requirements of the males throughout the summer.

The mating system of bats

The absence of males from the maternity colonies produces a new challenge, as the sexes now have to find a way to meet for mating. In temperate-zone bats, effectively three types of mating systems are observed: local recruitment, seasonal harems, and lek-like aggregations. Local recruitment at colony sites is obviously by far the easiest method of finding a mate, as auditory, olfactory, and perhaps even visual cues would be effective methods at such short ranges. On the other hand, local recruitment of mates carries the risk of inbreeding, especially in spatially isolated colonies where the immigration of dispersing males from other colonies would be rare. Nonetheless, it has been shown that local males sire substantial percentages of offspring in some species (e.g. Myotis daubentoni, Senior et al. 2005), but not in Bechstein’s bats (Kerth & Morf 2004), and interestingly in both cases inbreeding is not found. In case of the seasonal formation of harems, a single male recruits multiple females who then mate and roost together during several days. Here again, the possibility for inbreeding is present if females are recruited locally, but this has apparently been avoided in many species, such as Nyctalus noctula, where recruitment takes place during or after a seasonal migration of the females (Petit et al. 2001). Finally, males and females may assemble in leks, where promiscuous mating can take place. Indeed, many of the bat species found in the temperate-zone, including Bechstein’s bats, have a lek-like system known as autumn swarming (e.g. Parsons et al. 2003; Kerth 2008b).

During swarming, many males and females assemble at potential hibernacula in the months prior to hibernation and mate promiscuously. How swarming sites are established and how juveniles localize them in their first year of life remains a mystery (Kerth 2008b). However, once established, bats should have no trouble returning to such traditional sites in subsequent years because of their good navigational abilities (Holland et al. 2006). During swarming in Bechstein’s bats, siring occurs primarily by unrelated males, as suggested by the almost complete lack of structure between colonies in nucDNA (Kerth et al. 2002a, 2003). Thus, despite living in closed societies, Bechstein’s bats avoid inbreeding through promiscuous and probably random mating at swarming sites away from their summer habitat where they meet with large numbers of unrelated males.

Causes for closed maternity colonies in Bechsteins’s bats

As mentioned in the introduction, dispersal barriers cannot be the reason why colonies are closed, due to their close proximity and the bats’ ability to fly. Indeed, solitary males regularly disperse between the home ranges of different colonies (Kerth et al. 2000; Kerth & Morf 2004). Other explanations such as the defence of local resources, although a common reason for intercolony female competition in many other species (Stockley & Bro-Jørgensen 2011), also do not appear to apply in the case of Bechstein’s bats. Foraging grounds of females sometimes overlap with those of other colonies (Kerth & Melber 2009), and do so regularly with those of males (Kerth & Morf 2004). Indeed, it seems highly unlikely that the females could defend individual foraging ranges that are known to cover several hectares of forest. Exclusion to protect the microclimate of the roosts is also an implausible benefit, as more bats in a roost would provide energetic benefits for all females during pregnancy and lactation (Pretzlaff et al. 2010). Indeed, maternity colonies appear to provide an egalitarian environment. None of the typical types of intra-group female competition, as reviewed by Stockley & Bro-Jørgensen (2011), ranging from open aggression to more subtle forms such as reproductive inhibition have not been observed within Bechstein’s bat colonies. Nevertheless, the observation that colony members do not allow immigrants to join during confrontation tests (Kerth et al. 2002b), suggests that the pressure against the immigration of foreign bats is strong.

A possible explanation for xenophobic behaviour and the evolution of closed societies are fitness benefits from cooperation with close kin. In Bechstein’s bats, a strong influence of kin selection on the evolution of closed societies however seems unlikely for several reasons. For one, common cooperative behaviours such as allogrooming and information transfer among roosts are not restricted to close kin (Kerth & Reckardt 2003; Kerth 2008b). Furthermore, as social network analysis has shown, individualized roosting associations occur between unrelated females as well as related females, suggesting it is not essential for individuals to be related for social relationships to be established (Kerth et al. 2011). Moreover, the promiscuous mating system combined with the birth of maximally one offspring per female per year mean that genetic relatedness dilutes quickly within colonies (Kerth et al. 2002b). This can be seen by the high within-colony diversity and low differentiation between colonies in nucDNA. Nevertheless, colonies may be closed because of benefits arising from cooperation with familiar individuals. It remains to be tested, whether female Bechstein’s bats indeed benefit from living together with familiar colony members.

The argument that females do not switch colonies because of familiarity with the habitat cannot be fully excluded but seems unlikely to be the crucial selective pressure towards the evolution of closed societies in Bechstein’s bats. Females are faithful to individual foraging areas over many years (Kerth & Melber 2009), suggesting that long-term information about profitable foraging sites could be important. However, dispersing males are able to establish new foraging areas (Kerth & Morf 2004) and there seems no reason why females should not also be able to do so if dispersal becomes beneficial. Within colonies, information about roosts is shared among members through leading and following each other to roosts in combination with calling and swarming at roost entrances (Kerth & Reckardt 2003; Schöner et al. 2010). Theoretically, potential immigrants could exploit this way of information transfer and gain knowledge of roosts relatively easily, if only the local females would accept them in their roosts.

Finally, a plausible explanation for closed societies in Bechstein’s bats is parasite pressure. Bechstein’s bats have seemingly minimized contact with foreign conspecifics throughout their annual cycle. For example, mating at swarming sites produces high levels of outbreeding while minimizing body contact with conspecifics when compared to the mating harems seen in several temperate-zone bats, where bats from different colonies roost together for some days (Dietz et al. 2007). Also, Bechstein’s bats are generally observed to hibernate solitarily in rock crevices, or rarely, in small clusters of a few individuals, again more isolated from each other than observed in many other bat species (Dietz et al. 2007). While female Bechstein’s bats cannot completely avoid body contact with noncolony members during mating, in keeping their maternity colonies closed during summer, they effectively minimize any possibility for parasites to enter.

There are several reasons why parasite avoidance may be especially relevant in the case of bats. Most notably, bats harbour a large diversity of highly specialized parasites, bacteria, and viruses (Allen 1962; Calisher et al. 2006). This is likely due to the favourable conditions in maternity colonies, where parasites are provided with warm, immuno-compromised, and clustered hosts for them to exploit (Lourenço & Palmeirim 2008). The frequent roost switching of Bechstein’s bats and other bats has been interpreted as avoidance of parasites that deposit their larval stages in the roosts, such as bat flies (Lewis 1995; Reckardt & Kerth 2006). But roost switching does not help against contact transmitted parasites, such as wing mites, that only leave their individual hosts to move to another bat (Reckardt & Kerth 2009). For such parasites, transmission between colonies can be limited by living in closed societies, along with minimizing contact with noncolony members during mating and hibernation. Although the fitness consequences of wing-mites and other common ecto-parasites are generally considered to be minor (Giorgi et al. 2004; Lourenço & Palmeirim 2007), they may act as vectors and models for more costly endo-parasites, bacteria, viruses and fungi. A poignant example, which illustrates the effect that contact and/or close proximity with foreign bats can have, is the current spread of a pathogenic fungus (Geomyces destructans) causing the ‘White Nose Syndrome’ in North-American bats (Blehert et al. 2009). This fungus has wiped out up to 90% of the members of certain bat populations, affecting many colonies at the same time. Although living in closed societies during summer cannot completely protect against rapidly advancing infectious diseases such as G. destructans, it does have the potential to minimize parasites and pathogens transmission.

In addition to avoiding initial infection, prevention of local parasite adaptation may also be necessary (Gandon & Michalakis 2002). From the investigation of the genetic structure of the ecto-parasitic wing-mite of Bechstein’s bats (Bruyndonckx et al. 2009; J. van Schaik et al. in Prep.), we suppose that the bats successfully prevent this adaptation by limiting their colony size and keeping colonies isolated from each other during summer as well as by minimizing transmission though the mating and hibernation phase. Thereby the bats are causing strong genetic drift in the mites, which eliminates opportunities for local adaptation in the parasites (J. van Schaik et al. in Prep.).

To summarize, no single explanation can currently explain the existence of closed societies in Bechstein’s bats. The conclusion most consistent with the current data is that Bechstein’s bats benefit from living in closed societies because stable groups may facilitate cooperation and to reduce parasite pressure. We must however underscore that despite the large amount of knowledge gained about this system during the last 15 years, there remain several critical and untested assumptions in the explanation presented above. This includes possible benefits of stable long-term relationships among colony members as well as fitness consequences of infestation with different parasites. Also, it remains to be identified whether the rare acceptance of immigrants is an active choice (behavioural plasticity in response to critically low population size), or simply due to a lack of recognition among some colonies. Perhaps the largest gaps in our knowledge are details concerning the social interactions outside of the maternity colony. Swarming dynamics as well as potential hibernation associations have yet to be studied in detail, and could significantly contribute to the understanding of this interesting social system.

Consequences of living in a closed society

Living in a closed society has several positive as well as negative consequences. We already discussed the potentially positive influence on the evolution of cooperation, either through reciprocity among long-term associates, or through shared genes (kin-selection), and even by-product mutualism may be favoured in closed societies (Clutton-Brock 2002; West et al. 2002; Sachs et al. 2004). Interestingly, it has been shown that cooperation (in bacteria) is favoured via the budding dispersal of colonies as this maintains high relatedness, while reducing local competition (Kummerli et al. 2009). Thus possibly the budding of colonies as observed in big brown bats (Metheny et al. 2008a), and supposed for Bechstein’s bats, is not only a product of living in closed societies, but may also strengthen their advantage by increasing within colony relatedness and subsequently cooperation (Kerth 2008a). At the same time, living in closed societies may substantially limit the dispersal capabilities and genetic resilience of the species. For example, expansion and dispersal is only possible in the event that societies reach the critical size for allowing them to fission permanently. Also, rescue effects, where critically low population sizes in individual societies are prevented through immigration from other societies, are highly unlikely. This means that not only may the species have difficulties expanding its range, but it will also have difficulties maintaining declining populations. Finally, limiting dispersal to one sex may be risky, especially at the fringe of the distribution, or in the case of low population densities, due to the difficulty of locating individuals of the opposite sex outside of the natal society.

Comparison with other species living in closed societies

Unlike in bats and few other mammals, closed societies seem often to be connected with high reproductive skews among group members (a single breeding pair; Kerth et al. 2000). For example, the structure of eusocial insect colonies shows many parallels to those of the closed societies seen in cooperatively breeding mammals (Boomsma 2009). Whether due to direct kin selection (Helanterä & Bargum 2007) or multi-level selection (Wilson & Wilson 2007), it is clear that natal philopatry and demographic isolation is critical to the success of the colonies in eusocial insects. In mammals, a similar example can be found in two mole rat species, which live eusocially in large colonies with highly related males and females. Here, not only cooperation among kin but also ecological constraints such as restrictions in food availability and the high costs of dispersal are hypothesized to be the reasons for the (almost) strict natal philopatry of both sexes and the evolution of closed societies (Burda et al. 2000). Nonetheless, even in naked mole rats substantial outbreeding and occasional dispersal of males is observed in the wild (Braude 2000) and negative effects of inbreeding have been documented following the outbreak of a novel disease in the lab (Ross-Gillespie et al. 2007).

Some other mammals show closed societies without strong reproductive skew (having multiple breeders). One example is the killer whale (Orcinus orca), which forms philopatric pods of males and females, and mating only takes place during temporary interactions with other pods (Pilot et al. 2010). In this species, a strong limitation to dispersal is provided by clear foraging specializations (ecotypes) (Dahlheim et al. 2008). As learning of such specializations requires a large individual investment, it is supposed that individuals do not disperse to pods specialized in other foraging strategies. Surprisingly, dispersal between pods within ecotypes is also rare, and cannot be explained as a function of either cooperation or dispersal barriers so far (Pilot et al. 2010).

Outside of bats, perhaps the most similar society to those of Bechstein’ bats is that of the white-nosed coati (Carnivora: Nasua narica). In this species females are philopatric (Gompper et al. 1997), and males only join female bands during brief mating periods (Booth-Binczik et al. 2004). Males disperse from their natal bands, probably to reduce resource competition (Gompper et al. 1998), and female bands are only formed via fissioning of existing bands. Nonetheless, dispersal of females away from their natal bands does occur, and seems to be most successful when done together with a partner (Russell 1983; Gompper et al. 1997). Social behaviours such as allogrooming are observed (Gompper 1997) and female bands also maintain and defend well-defined home ranges. Therefore, it seems likely that cooperation among kin (and potentially long-term band-mates) acts as the driving force leading to extensive natal philopatry and almost closed bands in coatis.

Finally, although a large variety of female-bonded societies can be observed in primates (Kappeler & van Schaik 2002), virtually all primate societies comprise males and females, and usually at least one sex is able to regularly disperse between societies. Hence, to our knowledge no examples can be shown where societies of primates are as closed as in Bechstein’s bats.

In conclusion, the closed female societies observed in Bechstein’s bat (and potentially several other bat species) appears to be unique within the mammals. This suggests that the pressures selecting for this unusual social system may be relatively specific to bats, yet at this point no definitive cause can be identified. Knowledge of the causes for closed female societies however may be highly relevant towards understanding the factors shaping the evolution of mammalian societies. Unfortunately, the long term data sets required for detailed characterization of social systems, especially in combination with analysis of additional pressures such as parasitism are still rare, making comparative studies unfeasible at this point.

Studying closed societies: an evaluation of the available methods

As we have shown, a large amount of detailed knowledge is required to adequately characterize and hypothesize about the causes and consequences of the social system of a species. With respect to field data, it is especially important to observe social interactions that cannot be investigated genetically because they do not result in gene flow. This includes occasional visits between social groups and individual roosting associations as well as behavioural mechanisms, such as xenophobic behaviour, that keep societies closed. Moreover, field data may provide an independent measure for dispersal rates and distances as well as estimates of individual fitness parameters. Hence field data shed light on the benefits and costs of living in particular social organizations that genetic data alone cannot fully address.

Regarding genetic approaches, it is evident that selecting the suitable genetic markers and techniques is critical. It has long been known that the analysis of mtDNA and nucDNA will shed light on different parts of the social system of a species, but we would like to emphasize that this is in particularly the case for species with extreme sex-specific dispersal rates such as Bechstein’s bats. As a number of reviews and comparisons have illustrated (de Meeus et al. 2007; Meirmans & Hedrick 2011; Wilkinson et al. 2011), the last decade has seen remarkable advances in the analysis of population genetic data. Consequently, it is becoming increasingly important to carefully select the proper analytical tools (Excoffier & Heckel 2006). In our study, use of any of the measures for analysing nucDNA (FST, GST, DST, Assignment tests, amovas) have resulted in similar conclusions, making it difficult to decide which measure agreed best with the field data. However, it must be emphasized that none of the measures used is optimal for all types of studies (e.g. dispersal, drift, demography, within/between population; Meirmans & Hedrick 2011).

As a final note, it is important to not only consider the species at hand, but also those interacting with it. Parasites, while often ignored, can offer extra insights into the interactions of a species with conspecifics and their environment (Nieberding & Olivieri 2007). Likewise, apart from observational and genetic data, which must be gathered on the species in question, comparison with related species can be a powerful tool. Often species within a genus/order show variations in social system, and contrasting such variations can give insight into the selection pressures that caused them.


We thank Nelly Ménard, Nicolas Perrin and Eric Petit for inviting us to write this paper and Alain Frantz, Maria van Noordwijk, Carel van Schaik and two anonymous referees for helpful comments on the manuscript. We are very grateful to a large number of collaborators providing data and ideas during the course of this long-term study. Samples for five of the six new colonies have been kindly provided by Markus Melber, Beytullah Özkan, Boyan Petrov, and Serbülent Paksuz. Capture and handling of bats was carried out under license from the responsible nature conservancy departments. Among many other funding organizations, the national science foundations of Germany (DFG) and Switzerland (SNF) supported this study over the past 15 years.

GK’s research focuses on the behavioural ecology, conservation and socio-genetics of bats. JvS is interested in the population structure and ecology of bats and their parasites.

Data accessibility

Allele frequencies and locations of the six newly analysed Bechstein′s bat colonies can be found at Dryad entry doi:10.5061/dryad.dh8tt.