Erschienen in: Journal of Mammalogy ; 89 (2008), 6. - S. 1379-1390 https://dx.doi.org/10.1644/08-MAMM-S-061.1

MY HOME IS YOUR CASTLE: ROOST MAKING IS SEXUALLY SELECTED IN THE LOPHOSTOMA SILVICOLUM

DINA K. N. DECHMANN* AND G. KERTH Leibniz Institute for Zoo and Wildlife Research, Alfred-Kowalke-Strasse 17, 10315 Berlin, Germany (DKND) Smithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Ancon, (DKND) Department of Ecology and Evolution, Biophore, University of Lausanne, 1015 Lausanne, Switzerland (GK)

Shelters are important for the survival and reproduction of many and this is particularly true for . Depending on the future use and effect of shelters on the fitness of individuals, not all members of a group of animals may contribute equally to shelter making. Thus, knowledge about the identity of shelter-making individuals may teach us much about the social system and mating strategy of . To exemplify this, we review what is known about the roost-making behavior and the social system of Lophostoma silvicolum, a neotropical bat that excavates roost cavities in active arboreal nests. Roosts in termite nests are highly beneficial for the bats because they offer improved microclimate and possibly are responsible for the lower parasite loads of L. silvicolum in comparison to bat species using other, more common, roost types. Examination of observational field data in combination with genetic analyses shows that roost cavities excavated by single males subsequently serve as maternity roosts for females and that this improves reproductive success of the male who excavated the roost. This suggests that roosts in termite nests serve as an extended male phenotype and roost making is a sexually selected behavior. Roost-making behavior is tightly linked to the species’ social organization (single-male multifemale associations that stay together year-round) and mating system (resource- defense polygyny). The case study of L. silvicolum shows that it is important to learn more about the implications of shelter making in bats and other animals from ongoing and future studies. However, differences in costs and benefits for each group member must be carefully evaluated before drawing conclusions about social systems and mating strategies in order to contribute to our current knowledge about the evolution of sociality in mammals.

Key words: extended phenotype, mating system, offspring dispersal, philopatry, reproductive success, resource defense polygyny, sexual dimorphism, social system

Shelter making is comparatively rare in mammals and the on bats (e.g., Chaverri et al. 2007; Kerth et al. 2001; Reckardt resulting structures are often temporary and relatively simple and Kerth 2007; Willis and Brigham 2007), have listed the (Hansell 1984). This is in contrast to other groups of animals advantages and disadvantages of refuges and linked them to the such as birds or , which frequently engage in the making social systems of the species in question. of elaborate refuges. Nonetheless, quality, availability, and Building or modifying a refuge requires time and energy, distribution of roosts or nesting sites may limit the geographic costs that must be compensated by later benefits. In social range of mammals including many bat populations (Kunz and animals, not all group members are expected to invest equally Lumsden 2003) and may influence their reproductive success in the cost associated with shelter construction, maintenance, (Racey 1973) as well as their population structure and social and defense (Collias 1964; Morrison 1979; Morrison and behavior (Kerth and Ko¨nig 1999; Kerth et al. 2000, 2001). Morrison 1981). Individual investment should depend on the Only a few studies on animals (Collias and Collias 1976; future use and indirect fitness benefits of a shelter by each Forsythe 1989; Hansell 1984; von Frisch 1974), and especially group member. Consequently, we should often see sex differ- ences in roost-making behavior. For example, females have to make shelters alone if males are living elsewhere or do not * Correspondent: dechmann@izw berlin.de contribute because paternal care is not required for the suc- cessful rearing of offspring (Dawkins 1976; Hansell 1984). A male may share parental care or provide other services such as shelter construction if he thus gains biased access to mating 1379

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-2-teweaueeza4v6 1380

(e.g., in case of postpartum estrus). In most terrestrial modify the nest itself (e.g., Rasa 1985). Some bats species vertebrates, including mammals, males compete for females, violate this general rule, including 1 vespertilionid (Murina whereas the latter choose their partners. Shelter making could florium Clague et al. 1999), 1 flying fox (Balionycteris serve as a form of extended male phenotype, enabling females maculate Hodgkison et al. 2003), and the entire neotropical to judge the quality of potential mates (Andersson 1994; genus Lophostoma (formerly Tonatia Lee et al. 2002) as far Dawkins 1999). Extended phenotype is the effect genes may as the roost choice is known (Goodwin and Greenhall 1961; have on the environment through an ’s behavior, a Handley 1976; Kalko et al. 2006; McCarthy et al. 1992). The famous example being beaver dams (Dawkins 1999). most common and well-known species of the genus Lophos- Bats form the 2nd largest mammalian order, with most toma, L. silvicolum, excavates active, arboreal nests of the species living in the tropics. The more than 1,200 currently termite corniger (Kalko et al. 2006). The termite recognized species display a tremendous diversity in ecology, nests, which are made from predigested wood, cemented body size, diet, and social system (Kunz and Fenton 2003; together with the ’ saliva, are very hard. Thus, the McCracken and Wilkinson 2000; Nowak 1994). Despite excavation of roosts is probably costly in terms of time and pronounced ecological differences, most bats live in groups energy. To offset these costs it seems likely that termite for at least some part of their annual life cycle. Size and mounds provide significant fitness benefits for the excavator. composition of groups differ between species (Bradbury and There were several possible, potentially nonexclusive Vehrencamp 1977; McCracken and Wilkinson 2000), but hypotheses to explain roost-excavation behavior in L. silvico- generally it is the females that are social and rear their young lum. First, we expected that roosts in termite nests would be communally. Male bats are often solitary, they may join female beneficial for the bats. One disadvantage of group living in bats maternity colonies, or, more rarely, form groups or colonies of is thought to be a higher transmission rate of ectoparasites. their own (Encarnac¸a˜o et al. 2005; Safi 2008; Safi and Kerth However, termites are known for their chemical defenses 2007). Limited roost availability could generate pressure on against other insects (Prestwich 1988) and we compared bats to aggregate. However, bats living in groups might also ectoparasite loads of L. silvicolum to those of a closely related gain individual benefits such as thermoregulatory advantages species, Tonatia saurophila, which occupies a similar ecolog- (e.g., Willis and Brigham 2007), decreased predation (e.g., ical niche, has similar group sizes, but lives in tree cavities. Fenton et al. 1994), or cooperation among group members Finally, based on what is known about Old World termites, (e.g., Wilkinson 1984). Given the clear importance of day whose mound temperatures are warm and stable (Korb and roosts in the lives of bats, it is surprising that relatively few bat Linsenmair 2000), we expected that the consistently round- species modify structures and construct their own roosts. shaped tree nests of N. corniger might also be advantageous Instead, most bats rely on naturally occurring cavities or concerning their microclimate. shelters built by other animals (Kunz 1982; Kunz and Lumsden Regarding the roost-making individuals, participation of 2003). This is surprising because in primates nest making is both sexes in roost construction was only expected if males and associated with nocturnality and offspring that are left behind females both provide parental care, which is extremely rare in by the parents during foraging (Kappeler 1999), both of which bats. Therefore, 2 modes of selection during the evolution of behaviors are typical for most bats. In this paper, we review roost making in L. silvicolum seemed possible: natural what is known about the roost-making behavior and the social selection or sexual selection. In the case of roost making by system of Lophostoma silvicolum, a neotropical bat that females, natural selection could have promoted the evolution of excavates active arboreal termite nests to roost in them. this behavior if females cooperate in the excavation of roosts in One strategy to improve roost availability and reduce com- termite nests. In contrast, sexual selection could have been the petition for limited shelters is to use one built by another species. main driving force for roost making if excavated termite nests The phenomenon of living together with species that provide are a resource provided by males to attract females. Of course, a refuge has been studied particularly well in . in the latter scenario the male himself also might profit directly Examples range from inhabiting the oothecas of spiders from the advantages a termite nest roost offers and natural (Dejean et al. 1999) to various arthropods inhabiting nests selection might be an additional force promoting the evolution (Wilson 1971). The latter are particularly interesting because of this behavior. In both scenarios, we would expect that roosts they are examples of species capable of inhabiting active ant in termite nests are beneficial for females for rearing their colonies, which normally defend themselves aggressively young. These 2 main different modes of selection on roost- against intruders. However, not only arthropods live in the nests making behavior result in different and testable sex-specific of social insects. For example, the use of termite nests as shelters predictions, concerning group composition, behavior, mor- occurs in a large variety of birds (reviewed in Brightsmith 2000), phology, reproductive success, and dispersal (Table 1). some reptiles (e.g., Varanus niloticus Cowles 1928), and If males were solely responsible for excavating termite mammals (e.g., Herpestes Rasa 1985). Termites occur nearly nests, this would suggest that they benefit by attracting females worldwide and many species build elaborate nests or mounds, to the resource they are providing. In this type of mating but secondary users of termite nests can only occur wherever the system, a resource-defense polygyny, females choose breeding required termite host species is distributed. roosts provided and defended by males and aggregate there. Most mammals, such as mongooses that use termite nests, Possession of such roosts would improve male reproduc- usually live in the ventilating channels of mounds and do not tive success, thereby balancing the costs of roost making. 1381

TABLE 1. Predictions concerning male and female behavior, genetic relatedness, and morphology depending on whether roost making is under sexual or natural selection.

Sexual selection for access to mating partners Natural selection for communal breeding Both sexes Group composition Single male multifemale Only females, or brief male tenure Excavation of roosts in termite nests Males Females Males Sexual dimorphism Males larger No prediction Condition of males Successful males in better condition No prediction Reproductive success of nest owning males High; mating takes place in the roost No prediction; mating may or may not take place in the roost Reproductive skew among males High Low Females Social interactions among females No prediction; female groups may be anonymous Social interactions; cooperation during roost aggregations with no or few social interactions making; individual or group recognition among females Female philopatry Low; female offspring leave the group and avoid High; female philopatry stabilizes cooperation mating with the father if male tenure is long during roost making via or familiarity. If low, at least very stable groups are expected and cooperation during roost making should be stabilized via familiarity Genetic diversity within female groups No prediction Low Genetic structuring among female groups Absent or low High

According to Clutton-Brock (1989b) variance in male mating picked up from the vegetation and the ground. The bats hunt in success is predominantly influenced by 3 variables: the effect small individual foraging areas in the proximity of their roosts of the male’s contribution to parental care and thus the survival (Bockholdt 1998; Kalko et al. 1999; Servatius 1997), are of his offspring; the defensibility of groups of females and strictly forest dwelling, and are distributed throughout tropical degree to which a male is able to monopolize copulations; and Central America and part of South America (Reid 1997). the size, stability, and defensibility of female groups. Theoreti- Exclusive day roosts are arboreal, active termite nests of the cally, the conditions favoring a polygynous mating system species N. corniger, which are excavated by the bats (Kalko would be the following: little or no effect of paternal care, high et al. 2006). The resulting cavities are inhabited by small defensibility of females, and high stability of female groups. groups of an average of 4 8 bats (maximum 19 Dechmann However, many species do not fulfill this pattern so clearly and, et al. 2005; Ueberschaer 1999). In spite of their roosting in depending on the availability of mating partners or resources termite nests, DKND found no evidence in the feces of L. such as food, roosts, or territories, there also may be variation silvicolum that termites are part of the bats’ diet. In fact, the within species (Schaal and Bradbury 1987). cavity inhabited by the bats is sealed off by the termites, who Alternatively, females might cooperatively excavate termite never enter it while the bats are present, as could be seen on nests to create roosts as suitable shelters to rear their offspring. video recordings. Instead, the termites close all tunnels damaged In this case, sociality would be the result of females sharing by the excavation of the nest, creating a surface similar to the a roost they made together, facilitating female philopatry and outer skin of the nest. maternal structuring of groups (Kerth et al. 2000; Ko¨nig 1997). All data that we review here were collected at 2 sites In this paper, we review recent research on L. silvicolum, between 1998 and 2003. The 1st site was Barro Colorado which made it possible to test the relevant parameters (Table 1) Natural Monument, Panama. Most of the fieldwork there was against each other and to put our conclusions in the context of done on Barro Colorado Island at the center of Barro Colorado mammalian mating systems in general. Natural Monument. This 1,560-ha island is located in Gatun Lake (98109N, 7981509W) and borders the Panama Canal in central Panama. Barro Colorado Natural Monument is covered BIOLOGY OF LOPHOSTOMA SILVICOLUM AND with semideciduous tropical lowland rain forest (Foster and STUDY SITES Brokaw 1982). Rainfall averages 2,600 mm per year, and about The white-throated round-eared bat, L. silvicolum (Phyllo- 90% of this falls during the rainy season from mid-April to stomidae), is medium sized (about 30 g at our study site in December (Windsor 1990). The 2nd study site was located in Panama Dechmann et al. 2005), and characterized by the 22,000-ha Soberania National Park (98079N, 798429W) and extremely round-tipped, broad wings and large ears. This was covered with similar vegetation as the Barro Colorado morphology is ideally suited for their gleaning foraging mode Natural Monument. Soberania National Park stretches along (Norberg and Rayner 1987). L. silvicolum feeds on large the mainland border of the Panama Canal, east of Barro arthropods, particularly katydids (Tettigoniidae), which are Colorado Island. 1382

TABLE 2. Comparison of ectoparasites on Lophostoma silvicolum and a closely related and morphologically as well as ecologically similar species living in tree cavities (Tonatia saurophila).

Mann Whitney U test Lophostoma silvicolum Tonatia saurophila (X 6 SD) n (X 6 SD) n UU9 P Wing mites 3.62 6 5.84 210 9.93 6 13.4 91 6,497.0 6,497.0 ,0.0001 Streblid flies 0.79 6 1.67 244 1.92 6 2.99 60 5,531.5 9,108.5 0.0029

BENEFITS OF ROOSTING IN TERMITE NESTS silvicolum has significantly lower ectoparasite loads of both types of parasites than a similarly sized, closely related, and For our study, regardless whether males or females make ecologically similar species, T. saurophila, roosting in tree roosts, it was important to determine 1st whether termite nests cavities (Table 2). The lower ectoparasite loads may be an are advantageous in comparison to roost types more commonly indirect benefit of the chemical defenses against parasites for used by other bat species, such as tree holes, to justify the time which termites are known (Prestwich 1988). and energy investment required for the excavation. There are Microclimate. It can be crucial for reproductive female many potential advantages to roosting in termite nests, in- bats to minimize daily energy expenditure through the selection cluding protection from predators, proximity to foraging areas, of warm roosts (Racey and Speakman 1987; Tuttle 1975; reduced parasite loads, reduction of intra- and interspecific com- Wilde et al. 1999). Despite this important role of roost micro- petition, and, most of all, a beneficial microclimate, which is climate for bats, little is known about the thermoregulatory particularly important for reproductive female bats (Kunz 1982; advantages that may be gained by roost making. Old World Kunz and Lumsden 2003; Lewis 1995). In L. silvicolum, the 3 termites are well known for intricate manipulation of mound parameters that we investigated availability of suitable termite microclimate through both their nest architecture and metab- nests (Kalko et al. 2006), ectoparasite load (data included below), olism within mounds (Korb and Linsenmair 2000). If the same and roost microclimate (Dechmann et al. 2004) suggest that were true for nests of the neotropical N. corniger, Dechmann excavated termite nests are highly desirable shelters. et al. (2004) predicted that temperature in the cavities made by Nest availability. The obligate tie of L. silvicolum to its L. silvicolum would be higher and more stable than in tree host may limit the bats’ geographical distribution to that of the cavities used by other bats. Energy saved through this roost termites. However, the criteria bats use to determine nest suit- choice might play a critical role in the evolution of roost ability and whether availability of such nests was limited making by L. silvicolum and other animals living in termite remained unclear. In a census of termite nests on 2 plots on nests. In addition, we hoped that the temperature regime would Barro Colorado Island, Kalko et al. (2006) recorded a set of offer an explanation for the observation that the bats leave their 15 descriptive nest parameters. They then compared nests from excavated termite nests whenever the colony dies. the census with 44 excavated nests according to these param- Although temperatures were very stable in both tree cavities eters and found that suitable nests were active, medium sized and termite nests, temperatures inside active termite nests were (.30 30 cm), shaded, have few or no branches leading 2.1 2.88C warmer than in tree cavities (Fig. 1). In addition, through them, and are free of vegetation immediately below the temperatures were both more stable and higher in active than in opening of the cavity at the bottom of the nest. Thirty-nine inactive termite nests (Dechmann et al. 2004). This significant nests in the census met all criteria, but only 5 of them had a bat- difference was independent of location of the roost in the study made cavity. Thus, availability was much greater than use by area or season. After we established that microclimates of the bats and density of L. silvicolum on Barro Colorado Island roosts in termite nests are indeed beneficial and superior to is probably limited by other factors. those of tree cavities, but only while the termite colony was Ectoparasite load. There are 2 major groups of bat ecto- alive, it became necessary that we assess who was actually parasites at our study site: streblid flies (Streblidae) and wing excavating the roosts males, females, or both in order for us mites (Acarina). Streblids are transmitted via pupae on the roost to understand the link between this roost choice and the social wall and may be the reason that some bat species frequently system, and perhaps also to understand the mating strategy of switch roosts (Reckardt and Kerth 2007; ter Hofstede and these bats. In Table 1 we have summarized and outlined pre- Fenton 2005). Wing mites are transmitted through direct dictions concerning behavior of both sexes, behavior of males, physical contact between bats. Generally, loads of contact- morphology, and physical condition, as well as female group transmitted ectoparasites increase with group size (Cote´ and composition, genetic structure, and dispersal of offspring, Poulin 1995). This may be particularly true for reproductive depending on the main selective pressure promoting roost female bats, which have depressed immune systems, often making: sexual selection or natural selection. accompanied by an increase of ectoparasites (Christe et al. 2000; Lourenc¸o and Palmeirim 2007). The energetic or repro- WHO ARE THE ROOST MAKERS AND WHAT IS ductive costs of higher ectoparasite loads in bats have rarely THEIR FITNESS BENEFIT? been quantified but they can result in higher grooming costs in nonreproductive bats (Giorgi et al. 2001). Examination Behavior of both sexes. Group composition reflects the of previously unpublished data presented here shows that L. social organization and often mating system of a species. In 1383

return to the roost for nocturnal breaks from foraging (Lang et al. 2006). Thus, the male not only invests physical energy in roost making, but he also spends less time foraging than the females. This is an additional indicator of a high male investment, which is typical for a resource-defense polygynous mating system. Morphology, behavior, and reproductive success of males. In most bat species investigated to date, there is no sexual dimorphism or, if there is sexual dimorphism, females are slightly larger than males. This is probably due to the fact that most studies were conducted in the temperate zone, where many species mate more or less promiscuously and females give birth to very large young that must reach adulthood before FIG.1. Temperatures inside (open symbols) and outside (filled the next hibernation period (Myers 1978; Ralls 1977; Rossiter symbols) active termite nests (squares, 27.98C 6 1.08C, mean 6 SD, et al. 2006; Williams and Findley 1979). A very interesting n ¼ 10) and tree holes (triangles, 25.18C 6 0.58C, n ¼ 5) measured tropical example is the flying fox Cynopterus sphinx, where over 1 week. Temperatures in termite nests are significantly higher sexual dimorphism switches from larger females to larger than in tree cavities inhabited by species closely related to males along a latitudinal gradient. This reversal is inversely Lophostoma silvicolum. Mean inside temperatures of both roost types correlated with the distribution of polygyny in this species are more stable than ambient temperatures. Figure adapted from Dechmann et al. (2004). (Storz et al. 2001). In polygynous mammals with a strong reproductive skew among males, females are usually the smaller sex (Ralls 1977). This is especially the case when the males have to invest a large effort into courtship or into the a polygynous mating system typically only 1 reproductively creation and defense of a resource or a harem or both active male would be present in a social group during the (Lindenfors et al. 2002; Weckerly 1998). In L. silvicolum, mating season, although subordinate or satellite males also can males were significantly larger than the females in 3 standard be present (Ortega and Arita 2002; Voigt and Streich 2003). In size parameters (length of forearm, length of tibia, and body a social system more reminiscent of maternity colonies of mass Dechmann et al. 2005). Consistent with this result temperate bat species and with roost making by females, one was the finding that 21 males who successfully excavated a would expect any reproductively active males to be absent in termite nest were in significantly better body condition (relative the roost or to be present only during the mating season. body mass corrected for size measured by forearm length) than Dechmann et al. (2005) captured groups of L. silvicolum 61 33 bachelor males, the former being .2 g heavier on average times from 34 excavated nests during all seasons. Capture data (6 7% Dechmann et al. 2005). This may have several non- were always consistent with a single-male multifemale social exclusive reasons. Only males in good physical condition may system. It was not always possible to catch all bats present in be able to spare the energy required to excavate, maintain, and the roosts, but there was never more than 1 adult male (except defend a cavity (e.g., in fallow bucks [Dama dama] only strong in bachelor groups), and in all cases where there was no male, males can defend a harem McElligott et al. 2003). Females at least 1 individual had escaped. also may choose a mate according to his physical condition. Dechmann et al. (2005) also recorded nocturnal behavior in Finally, there could be an additional age effect, with older 2 bat-made cavities in termite nests with infrared video to males being heavier and thus better able to excavate and defend determine sex and identity of nest excavators. All bats in the a roost. study were marked using passive integrated transponders (PIT The data presented above support the hypothesis of the tags; Euro ID, Weilerswist, Germany) implanted subcutane- evolution of roost making in L. silvicolum under sexual ously. Automatic antennae (handmade by DKND) placed selection for at least 2 reasons. First, single males excavate the around the roost entrance, and attached to a logger (Euro ID) termite nests, investing time and effort to provide a beneficial recorded the identity of each bat entering and exiting. In a total resource for females. Second, females seem to choose a male of 19 nights, nest excavation was recorded on 5 nights (2 35 according to his physical condition as well as his possession of min each). In addition, an unmarked male was filmed working a roost. The roosts appear to act as an extended male on the termite nest and on the stump of a branch that had led phenotype, making it easier for the females to judge male through it for .2.5 h. In all nest excavation events, the male quality, but also to locate a high-quality male via his nest. This excavated only when females and juveniles were not present in scenario only makes sense if the females then also mate with the roost (Dechmann et al. 2005). On each of those occasions the male owning a roost. If roost making improves the mating the single adult male in the nest repeatedly bit into the cavity success of males, the evolution of such a trait would be wall with his canines and then pushed himself off with his enhanced (Andersson 1994). Even though sexual selection wrists, breaking off small pieces of nest material in the process. seems to be the dominant driving force for the evolution of this Many bat species remain in their foraging areas throughout the behavior, natural selection cannot be completely excluded. night. However, both male and female L. silvicolum always Females and young, and also the males, probably profit from 1384 the warm and stable microclimate and reduced ectoparasite loads, which may directly improve fitness of both sexes. Female L. silvicolum exhibit postpartum estrus (i.e., they become fertile shortly after they give birth to a young). If the females are very mobile and exhibit low or no group stability, an unlucky male might be joined only by pregnant females, which then give birth to another male’s young in the roost he so elaborately provided (Fig. 2). Termite nests, in spite of all their advantages, can be destroyed by anteaters (Tamandua mexicana), tree falls, decay as a consequence of death of the FIG.2. Three potential outcomes regarding reproductive success termite colony, or other natural causes, and nest longevity can of a nest making male Lophostoma silvicolum. Same shading indicates vary from a few months to several years (Dechmann et al. genetic relatedness. In roost 1, all young born in his roost are sired by 2007). Females are thus fairly often forced to switch roosts the nest male (mothers did not leave since the last reproductive and males must excavate new ones. Reproduction by females season); in roost 2, 1 young was sired by the nest male, and 2 by other is highly synchronized and seasonal (Dechmann et al. 2005); males (their mothers mated outside the roost or recently joined the thus infanticide by males would not speed up the female’s group already pregnant); and in roost 3, all young were sired by other reproductive cycle and the male’s access to mating. Conse- males. quently, even though sheltering another male’s offspring in the roost does not incur additional cost, only a relatively high reproductive success at the next postpartum estrus can explain increase the reproductive success of harem males, was not the evolution of roost making by the males. assessed for L. silvicolum and cannot be distinguished from Dechmann et al. (2005) determined reproductive success matings with previous, uncaptured harem males. of males with 2 measures: number of young they sired and There was always only 1 adult male in each of the groups average relatedness of each male to young in his own nest caught from termite nests, implying that male offspring of L. compared to the same male’s relatedness to young in other silvicolum invariably disperse. In some species of polygynous nests. They were able to assign 21 of 46 young to the bats subordinate or satellite males, or both, can be found, predicted father, corresponding to a reproductive success of especially in larger harems (Ortega and Arita 2002; Voigt and 46%. Males sired 0 4 of 5 young in their own roost per Streich 2003). They are usually related to the dominant harem reproductive season. This is probably a low estimate, as for male and take over the group of females when he is removed paternity assignments it is necessary to predict putative fathers. (Voigt and Streich 2003). It could have been possible that Most mothers were captured only once and thus it was young males of L. silvicolum disperse but then establish a roost impossible to determine whether they had switched roosts and territory near those of their father to indirectly profit from since the last reproductive season. However, 17 of the 21 his defense of the territory. However, Dechmann et al. (2007) mothers were still roosting with the father of their young. This conducted a Mantel test that investigated the relationship shows that, although females switch roosts quite easily, they between distance between roosts and pairwise relatedness of commonly stay with the same male for several reproductive males. The results showed that there is no correlation between seasons if undisturbed, thus providing a stronger motivation the 2 factors. This is in spite of the fact that roosts of for the male to invest in roost making. According to anecdotal (nonrelated) males can be as close as 25 m to each other. Males evidence, females may also move to a new roost together probably randomly settle wherever they find a suitable nest with a male when the old roost is destroyed (Dechmann et al. for excavation. The male’s resource ensuring his reproduc- 2007). tive success is the cavity he makes and not a territory (forest In the 2nd analysis, Dechmann et al. (2005) compared patch). In fact, foraging areas of males from neighboring roosts average relatedness of males (n ¼ 12) with all young born in have been found to overlap in a telemetry study (Bockholdt their roost with all young born in other males’ roosts, thus 1998). testing for scenarios such as roosts 2 and 3 in Fig. 2. Males Another parameter typical for polygyny is strong reproduc- were significantly more closely related to young caught in the tive skew among males. This was not assessed directly for L. same roost than to young from other roosts. silvicolum. However, roost captures revealed significantly A consequence of polygyny is that a few males monopolize higher numbers of females. In contrast, numbers of males the majority of matings, whereas most males gain little or no and females caught in mist nets at random sites in the forest did access to females. A reproductive success of 46% and a 25% not differ (Dechmann et al. 2007). Consequently, the female- average relatedness to young in their roost thus seems to justify biased sex ratio in the roosts is not due to different survival a high investment of time and energy by the males. These rates of the sexes. No young were assigned to males from values also fall well into the range of relatedness found in other bachelor roosts and roost males accounted for almost 50% of polygynous bat species (e.g., 29% in S. bilineata [Heckel et al. the young. All of the above are consistent with a strong 1999] and 69% in A. jamaicensis [Ortega et al. 2003]). reproductive skew, further supporting the hypothesis of roost Extraharem reproductive success, which plays an important making evolving under sexual selection, resulting in a unique role in the bat species mentioned above and which may further form of resource-defense polygyny. 1385

Behavior and genetic structure of female groups, and Thus, females may profit from familiarity with their natal area dispersal of offspring. The evidence outlined above is when dispersing. consistent with the hypothesis that males excavate roosts in Both sexes of offspring disperse in L. silvicolum, probably as termite nests to improve reproductive success, as expected in a consequence of the species’ mating system. This dispersal resource-defense polygyny. This in turn may have implications pattern results in high genetic diversity within social groups, for the social organization of the females (Table 1). There which should be reflected by the behavior among group may or may not be philopatry of female offspring, depending members (Kerth 2008). In groups of flexible composition few on the degree of cooperation between female group members or no social interactions among group members are expected or other benefits of natal philopatry such as familiarity with (e.g., Fleming et al. 1998). In contrast, cooperation is expected the habitat. In fact, many associations of bats are charac- in stable female groups (e.g., Kerth et al. 2003; Ortega and terized by philopatry of female offspring or both male and Maldonado 2006). In the latter species, females engage in female offspring, although 1 case where only male offspring allogrooming. are philopatric also is known (male and female philopatry in We also made video recordings from the same roosts of Plecotus auritus [Burland et al. 2001], female philopatry in L. silvicolum as were observed for nest excavation. These Myotis myotis [Castella et al. 2001], Myotis bechsteinii [Kerth recordings were made during all seasons of the year to exclude et al. 2000], Desmodus rotundus [Wilkinson 1985b], and male potential seasonal effects, and they were analyzed (n ¼ 12 full philopatry in S. bilineata [Nagy et al. 2007]). However, nights from 1st emergence to sunrise) to determine social matings take place outside the roost in most of those species. interactions between females in the roosts. The 1st roost Philopatry of female offspring in outbreeding bats should contained 1 adult male, 3 adult females, and 2 juveniles. The be reflected by a high diversity in nuclear DNA but a very low 2nd roost consisted of 1 adult male and 3 5 females. Both diversity in mitochondrial DNA (Kerth et al. 2000, 2002). groups were thus typical harems. The only social interaction Genetic diversity in L. silvicolum does not follow this between group members occurred when individuals returned to pattern. Instead, both nuclear and mitochondrial DNA diversity the roost, especially during the first 60 s, and consisted of the was close to 1 and thus very high (n ¼ 75 adult females returning bat presenting its belly to be sniffed by bats already from 14 groups; tested with 10 polymorphic microsatellite present in the nest. This interaction was particularly intensive [¼ nuclear] loci and the mitochondrial D-loop Dechmann between the male and each female. The resident male once et al. 2007). In addition, a genetic assignment of all adult expelled a strange male from the roost, although a female was L. silvicolum caught in roosts (n ¼ 159) to each other allowed to join the group after having been sniffed particularly revealed only 8 possible parent offspring pairs that had been intensively by the resident bats in the roost. Social interactions caught in the same roost, only 5 of which were caught from indicating cooperation or any kind of close bond such as the same roost simultaneously. Examination of these data allogrooming between females were never observed. indicates that both sexes of offspring disperse from the natal roost before they reach adulthood. This is in accordance with GENERAL DISCUSSION the resource-defense mating system in conjunction with relatively long male tenure (up to 36 months) compared to This review provides evidence that roost making in L. the time it takes females to become sexually mature (at an age silvicolum is strongly influenced by sexual selection by females of 6 months or more Dechmann et al. 2007). Young females for male-excavated roost cavities in termite mounds. In may have to disperse to avoid inbreeding with their own father addition, natural selection acting on males may have further (Clutton-Brock 1989a), whereas young males may disperse enhanced the evolution of the use of this roost type through to gain access to breeding partners and possibly also because a beneficial microclimate and reduced parasite loads. Males they are expelled by the resident male (Dobson 1982; excavate the termite nests and thus pay the entire costs of this Greenwood 1980). behavior. Harems are composed of females, with each harem In addition, Dechmann et al. (2007) tested genetic structure aggregating in the roost of 1 male. Indeed, one of us (DKND) between the same social groups. Here the prediction would observed that almost all females usually leave a roost after have been the opposite from within-group diversity: if there a capture attempt, probably a similar experience to an attack on was female philopatry, female groups should be strongly the nest by a predator, whereas the male usually returns. Thus, differentiated genetically. Pairwise differentiation between the female roost fidelity is lower than that of males, showing that FST-values of mitochondrial haplotypes of groups of females roost availability through male excavating behavior is very always was nonsignificant. Similarly, 98 of 105 pairs in important for the social system of this species. a comparison of microsatellite genotypes showed no significant Many tropical bats, including all known roost-making difference before Bonferroni correction, indicating high levels species, live in single-male multifemale associations (Kunz of gene flow between groups. However, there was a negative et al. 1994; McCracken and Wilkinson 2000) and probably influence of distance between roosts on the average FST-values have polygynous mating systems. Is roost-making behavior for nuclear DNA of female groups (Dechmann et al. 2007). a sexually selected male trait in all of these species? Tent Female offspring left the parental roost but settled nearby, making is the most common form of roost making in this order, because the danger of accidentally choosing a roost occupied but the process of the actual making of a tent has very rarely by a related male was independent of distance (see above). been observed in the field (Kunz et al. 1994). However, the few 1386 available studies indicate that male roost making is usually, but cooperation seems to be so important in this species that not always, the case (Balasingh et al. 1995; Rodrı´guez-Herrera females have found alternative, yet unexplained, ways to avoid et al. 2006). For example, in Ectophylla alba from the same inbreeding. Female offspring are philopatric even though male family as L. silvicolum (Phyllostomidae), which also lives in tenure is longer than it takes them to reach sexual maturity. In single-male multifemale societies, females have been recorded contrast, male offspring are forced by the dominant male to manipulating the roost leaf to construct a tent with no observed leave the roost (Wilkinson 1985a, 1985b). The opposite is the contribution of males. This indicates that at least in this species case in S. bilineata, a well-studied species with resource- roost making might be a cooperative behavior of females and defense polygyny (see also Voigt et al. 2008). Here, females thus naturally selected (Rodrı´guez-Herrera et al. 2006). Knowl- also form long-term stable groups, but no cooperative behavior edge about the advantages of roosting in leaf tents, as well as the is known (Heckel and von Helversen 2003). However, female mating system and genetic group composition of such species, offspring disperse, whereas males can be philopatric and will help us to better understand roost making in general. queue for harem access (Nagy et al. 2007; Voigt and Streich Lophostoma silvicolum exclusively roosts in excavated 2003). Males of this species do not provide roosts, but exhibit active termite nests and thus is an example of a species with very elaborate courtship behavior. Nonetheless, extraharem a single social system associated with the roost choice. paternity is high. Extraharem mating may be a strategy of the However, this is not true for all roost-making bats. For females, which have been in the harem for several years, to example, A. jamaicensis regionally switches between the use of avoid mating with their sons. leaf tents, probably made by the bats (Kunz and McCracken The more we learn about roost-making behavior and the 1996), to naturally occurring tree cavities or solution cavities in social systems of bats, the more it becomes clear that an caves (Ortega and Arita 1999, 2000). A. jamaicensis always enormous amount of work remains to be done. Roost making lives in harems, but the stability and composition of female can be a sexually selected male trait as in L. silvicolum, or, it groups is known only from cave-dwelling populations. appears, a naturally selected female behavior as in E. alba.It However, in all roost types, defense of roosts has been will be intriguing to test other species according to the charac- observed and it will be interesting to see if the mating system teristics listed in Table 1 to determine their social systems and of A. jamaicensis is always a female-defense polygyny. This mating strategies. Detailed studies on additional species and in turn depends on high stability of female groups, which may variation within species will teach us more about the evolution be decreased in more temporary roosts such as tents (Lewis of the making of shelters and contribute to our general under- 1995). Another group of bats with facultative tent making are standing of the evolution of sociality in mammals. the Old World flying foxes of the genus Cynopterus.In Cynopterus the use of naturally occurring or bat-made tents CONCLUSIONS depends on local availability of resources. When resources are clumped and spaced, tent making is more common and female Here, we review the literature on the social and mating groups are more stable, but wherever resources allow, naturally system of L. silvicolum, a roost-making bat, which modifies an occurring, unmodified roosts are used (Campbell et al. 2006a, unusual type of structure, active termite nests. We try to 2006b). In C. sphinx there is additional variation in the mating determine, in particular, whether natural or sexual selection system, and the direction of sexual dimorphism depends on was the main driving force during the evolution of roost- latitude (Storz et al. 2001). Nonetheless, when tent making excavating behavior. We show that this roost choice is closely occurs in C. sphinx and C. brachyotis it is done by single males linked to the bats’ social organization, single-male multifemale and the process can be very elaborate and time consuming. Thus, associations, and to their mating system, resource-defense roost making does seem to be a sexually selected trait in this polygyny. Single males invest large amounts of time and genus (Balasingh et al. 1995; Bhat and Kunz 1995; Tan et al. energy to provide a roost, gaining access to matings with the 1997). Perhaps roost making evolved several times in polygy- females who join them. Females in turn choose the male nous systems and due to different reasons in each species of according to his physical condition among roost-owning males, roost-making bat. but do not form closed groups or show any kind of cooperative In addition, a number of non roost-making phyllostomids behavior. Both sexes of offspring disperse, resulting in a seem to live in resource- or female-defense polygynous genetically homogenous population, where the choice about systems (Heckel and von Helversen 2003; McCracken and where to settle for both males and females depends on suitable Bradbury 1977, 1981; Ortega and Arita 2000; Ortega et al. termite nests (and possibly other ecological factors such as 2003). But even here the degree of stability of female groups foraging areas), but not on social context. Roost making in varies and the females of some of these species use long-lasting L. silvicolum is clearly under sexual selection and the roosts roosts, where they form long-term stable associations (see also can be regarded as an extended male phenotype. Our compari- above), which can be combined with cooperation among group son with other roost-making species shows variation in social members (Ortega and Maldonado 2006; Wilkinson and systems and dispersal behavior from species to species. This Boughman 1998). The most extreme case is the common indicates multiple potential origins and courses of evolution of vampire bat, D. rotundus, where females have been observed roost-making bats. Future studies on bats and other mammals regurgitating blood and otherwise cooperating with nonrelated will help us to better understand the implications and thus the group members (Wilkinson 1984, 1986, 1990). Female evolution of making of shelters for social and mating systems. 1387

RESUMEN BALASINGH, J., J. KOILRAJ, AND T. H. KUNZ. 1995. Tent construction by the short nosed fruit bat Cynopterus sphinx (Chiroptera, Pteropo Los refugios son importantes para la sobrevivencia y didae) in southern India. Ethology 100:210 229. reproduccio´n de muchos animales, y esto es particularmente BHAT, H. R., AND T. H. KUNZ. 1995. Altered flower fruit clusters of the cierto para los murcie´lagos. Dependiendo del uso futuro y el kitul palm used as roosts by the short nosed fruit bat, Cynopterus efecto de los refugios en la adecuacio´n de los individuos, no sphinx (Chiroptera, Pteropodidae). Journal of Zoology (London) todos los miembros de los animales podrı´an contribuir 235:597 604. equitativamente a hacer el refugio. Por lo tanto, el conoci- BOCKHOLDT, C. 1998. Hangplatzwahl, Aktivita¨tsrhythmik und miento de la identidad de los individuos que hacen el refugio Aktionsraum der neotropischen Fledermaus Tonatia silvicola podrı´a ensen˜ar mucho sobre del sistema social y del sistema de (d’Orbigny, 1836). Diploma thesis, Universita¨t Freiburg, Freiburg, apareamiento. Para evaluar esto, se hizo´ una revisio´n de lo que Germany. se conoce acerca la conducta para hacer refugios y el sistema BRADBURY, J. W., AND S. L. VEHRENCAMP. 1977. Social organization and foraging in emballonurid bats. 3. Mating systems. Behavioral social de Lophostoma silvicolum, un murcie´lago neotropical Ecology and Sociobiology 2:1 17. que excava sus refugios en nidos de termitas activos. Los BRIGHTSMITH, D. J. 2000. Use of arboreal termitaria by nesting birds in refugios en nidos de termitas son de gran beneficio para los the Peruvian Amazon. Condor 102:529 538. murcie´lagos porque ofrecen un mejor microclima y esta´n BURLAND, T. M., E. M. BARRATT,R.A.NICHOLS, AND P. A. RACEY. probablemente relacionados con una carga de para´sitos ma´s 2001. Mating patterns, relatedness and the basis of natal philopatry baja que en especies que usan otros tipos de refugios mas in the brown long eared bat, Plecotus auritus. Molecular Ecology comunes. Observaciones de campo y ana´lisis gene´ticos 10:1309 1321. mostraron que los refugios, que subsecuentemente sirven de CAMPBELL, P., Z. AKBAR,A.M.ADNAN, AND T. H. KUNZ. 2006a. refugios de maternidad, son excavados por un so´lo macho y Resource distribution and social structure in harem forming Old que esto aumenta el e´xito reproductivo del macho que hizo´el World fruit bats: variations on a polygynous theme. Animal refugio. Esto sugiere que los refugios en nidos de termitas Behaviour 72:687 698. CAMPBELL, P., N. M. REID,A.ZUBAID,A.M.ADNAN, AND T. H. KUNZ. sirven como una extensio´n del fenotipo del macho y que la 2006b. Comparative roosting ecology of Cynopterus (Chiroptera: construccio´n de refugios es una conducta seleccionada Pteropodidae) fruit bats in peninsular Malaysia. Biotropica 38:725 sexualmente. La conducta de construccio´n del refugio esta 734. fuertemente ligada a la organizacio´n social de la especie CASTELLA, V., M. RUEDI, AND L. EXCOFFIER. 2001. Contrasted patterns (asociaciones de un macho con varias hembras que se of mitochondrial and nuclear structure among nursery colonies mantienen juntos todo el an˜o) y el sistema de apareamiento of the bat Myotis myotis. Journal of Evolutionary Biology 14:708 (poliginia por defensa del recurso). El caso de L. silvicolum 720. muestra que es importante entender mejor las implicaciones de CHAVERRI, G., O. E. QUIROS,M.GAMBA RIOS, AND T. H. KUNZ. 2007. la construccio´n de refugios en murcie´lagos y otros animales en Ecological correlates of roost fidelity in the tent making bat estudios futuros. Sin embargo, diferencias en costos beneficios Artibeus watsoni. Ethology 113:598 605. para cada miembro del grupo debe ser evaluadas cuidadosa- CHRISTE, P., R. ARLETTAZ, AND P. VOGEL. 2000. Variation in intensity mente antes de inferir conclusiones acerca del sistema social y of a parasitic mite (Spinturnix myoti) in relation to the reproductive cycle and immunocompetence of its bat host (Myotis myotis). las estrategias de apareamiento, para contribuir al conocimiento Ecology Letters 3:207 212. de la evolucio´n de la sociabilidad en mamı´feros. CLAGUE, C. I., R. B. COLES,O.J.WHYBIRD,H.J.SPENCER, AND P. FLEMONS. 1999. The occurrence and distribution of the tube ACKNOWLEDGMENTS nosed insectivorous bat (Murina florum) in Australia. Proceedings Many people who have contributed to fieldwork, data analysis, and of the Linnean Society of New South Wales 121:175 191. manuscript quality are acknowledged in previous publications or are CLUTTON BROCK, T. H. 1989a. Female transfer and inbreeding coauthors on them. We nonetheless specially thank E. K. V. Kalko avoidance in social mammals. Nature 337:70 72. and B. Ko¨nig, who were essential to the entire project. We also thank CLUTTON BROCK, T. H. 1989b. Mammalian mating systems. Proceed S. Spehn, who provided ectoparasite data for T. saurophila, as well as ings of the Royal Society of London, B. Biological Sciences R. Furrer and P. Saner, who quantified social behavior on videotapes. 236:339 372. J. Altringham, R. Bixenmann, K. Murray, K. Safi, and an anonymous COLLIAS, N. E. 1964. The evolution of nests and nest building in birds. reviewer provided feedback on drafts of the manuscript and G. American Zoologist 4:175 190. Ceballos and J. Ortega wrote the Spanish resumen. Our thanks also COLLIAS, N. E., AND E. C. COLLIAS. 1976. External construction by go to J. Ortega and J. Wilkinson for organizing the symposium, which animals. Vol. 4. Dowden, Hutchington & Ross, Inc., Stroudsburg, resulted in this special issue, to J. Ortega for coordinating the writing Pennsylvania. process, and to the Journal of Mammalogy for inviting us to review COTE´ , I. M., AND R. POULIN. 1995. Parasitism and group size in social our work. Funding was provided by the German Science Foundation animals: a meta analysis. Behavioral Ecology 6:159 165. (Deutsche Forschungsgemeinschaft; DKND:VO 890/11 1; GK: KE COWLES, R. B. 1928. The life history of Varanus niloticus. Science 746, 2 1). 67:317 318. DAWKINS, R. 1976. The selfish gene. Oxford University Press, Oxford, United Kingdom. LITERATURE CITED DAWKINS, R. 1999. The extended phenotype: the long reach of ANDERSSON, M. 1994. Sexual selection. Princeton University Press, the gene. 2nd ed. Oxford University Press, Oxford, United Princeton, New Jersey. Kingdom. 1388

DECHMANN, D. K. N., E. K. V. KALKO, AND G. KERTH. 2004. Ecology bats, Tonatia silvicola and Trachops cirrhosus (Phyllostomidae). of an exceptional roost: energetic benefits could explain why the bat Biotropica 31:344 353. Lophostoma silvicolum roosts in active termite nests. Evolutionary KALKO, E. K. V., K. UEBERSCHAER, AND D. K. N. DECHMANN. 2006. Ecology Research 6:1037 1050. Roost structure, modification, and availability in the white throated, DECHMANN, D. K. N., E. K. V. KALKO, AND G. KERTH. 2007. All round eared bat, Lophostoma silvicolum (Phyllostomidae) living in offspring dispersal in a tropical mammal with resource defense active termite nests. Biotropica 38:398 404. polygyny. Behavioral Ecology and Sociobiology 61:1219 1228. KAPPELER, P. M. 1999. Nests, tree holes, and the evolution of primate DECHMANN, D. K. N., E. K. V. KALKO,B.KO¨ NIG, AND G. KERTH. 2005. life histories. American Journal of Primatology 46:7 33. Mating system of a neotropical roost making bat: the white throated, KERTH, G. 2008. Causes and consequences of sociality in bats. round eared bat, Lophostoma silvicolum (Chiroptera: Phyllo BioScience 58:737 746. stomidae). Behavioral Ecology and Sociobiology 58:316 325. KERTH, G., B. ALMASI,N.RIBI,D.THIEL, AND S. LU¨ POLD. 2003. Social DEJEAN, A., A. GRIMAL,M.C.MALHERBE, AND J. P. SUZZONI. 1999. interactions among wild female Bechstein’s bats (Myotis bechstei From specialization in spider egg predation to an original nesting nii) living in a maternity colony. Acta Ethologica 5:107 114. mode in a ‘‘primitive’’ ant: a new kind of lestobiosis. Naturwissen KERTH, G., AND B. KO¨ NIG. 1999. Fission, fusion and nonrandom schaften 86:133 137. associations in female Bechstein’s bats (Myotis bechsteinii). DOBSON, F. S. 1982. Competition for mates and predominant juvenile Behaviour 136:1187 1202. male dispersal in mammals. Animal Behaviour 30:1183 1192. KERTH, G., F. MAYER, AND B. KO¨ NIG. 2000. Mitochondrial DNA ENCARNAC¸A˜ O, J. A., U. KIERDORF,D.HOLWEG,U.JASNOCH, AND V. (mtDNA) reveals that female Bechstein’s bats live in closed WOLTERS. 2005. Sex related differences in roost site selection by societies. Molecular Ecology 9:793 800. Daubenton’s bats Myotis daubentonii during the nursery period. KERTH, G., F. MAYER, AND E. PETIT. 2002. Extreme sex biased Mammal Review 35:285 294. dispersal in the communally breeding, nonmigratory Bechstein’s FENTON, M. B., I. L. RAUTENBACH,S.E.SMITH,C.M.SWANEPOEL, bat (Myotis bechsteinii). Molecular Ecology 11:1491 1498. J. GROSELL, AND J. VAN JAARSVELD. 1994. Raptors and bats: threats KERTH, G., K. WEISSMANN, AND B. KO¨ NIG. 2001. Day roost selection and opportunities. Animal Behaviour 48:9 18. in female Bechstein’s bats (Myotis bechsteinii): a field experi FLEMING, T. H., A. A. NELSON, AND V. M. DALTON. 1998. Roosting ment to determine the influence of roost temperature. Oecologia behavior in the lesser long nosed bat Leptonycteris curasoae. 126:1 9. Journal of Mammalogy 79:147 155. KO¨ NIG, B. 1997. Cooperative care of young in mammals. Natur FORSYTHE, A. 1989. The architecture of animals: the equinox guide wissenschaften 84:95 117. to wildlife structures. Camden House Publications, Rochester, KORB, J., AND K. E. LINSENMAIR. 2000. Thermoregulation of termite New York. mounds: what role does ambient temperature and metabolism of the FOSTER, R. B., AND N. V. L. BROKAW. 1982. Structure and history colony play? Insectes Sociaux 47:357 363. of the vegetation of Barro Colorado Island. Pp. 151 172 in KUNZ, T. H. 1982. Roosting ecology of bats. Pp. 1 55 in Ecology of The ecology of a tropical forest (E. G. Leigh, Jr., A. S. Rand, and bats (T. Kunz, ed.). Plenum Press, New York. D. M. Windsor, eds.). Smithsonian Institution Press, Washington, KUNZ, T. H., AND M. B. FENTON. 2003. Bat ecology. University of D.C. Chicago Press, Chicago, Illinois. GIORGI, M. S., R. ARLETTAZ,P.CHRISTE, AND P. VOGEL. 2001. The KUNZ, T. H., M. S. FUJITA,A.P.BROOKE, AND G. F. MCCRACKEN. energetic grooming costs imposed by a parasitic mite (Spinturnix 1994. Convergence in tent architecture and tent making behavior myoti) upon its bat host (Myotis myotis). Proceedings of the Royal among neotropical and paleotropical bats. Journal of Mammalian Society of London, B. Biological Sciences 268:2071 2075. Evolution 2:57 78. GOODWIN, G. G., AND A. M. GREENHALL. 1961. A review of bats of KUNZ, T. H., AND L. F. LUMSDEN. 2003. Ecology of cavity and foliage Trinidad and Tobago. Bulletin of the American Museum of Natural roosting bats. Pp. 3 90 in Bat ecology (T. H. Kunz and M. B. History 122:187 302. Fenton, eds.). University of Chicago Press, Chicago, Illinois. GREENWOOD, P. J. 1980. Mating systems, philopatry and dispersal in KUNZ, T. H., AND G. F. MCCRACKEN. 1996. Tents and harems: birds and mammals. Animal Behaviour 28:1140 1162. apparent defence of foliage roosts by tent making bats. Journal of HANDLEY, C. O., JR. 1976. Mammals of the Smithsonian Venezuelan Tropical Ecology 12:121 137. project. Brigham Young University Science Bulletin, Biological LANG, A. B., E. K. V. KALKO,H.RO¨ MER,C.BOCKHOLDT, AND D. K. N. Series 20:1 91. DECHMANN. 2006. Activity levels of bats and katydids in relation to HANSELL, M. H. 1984. Animal architecture and building behaviour. the lunar cycle. Oecologia 146:659 666. Longman, London, United Kingdom. LEE, T. E., S. R. HOOFER, AND R. A. VAN DEN BUSSCHE. 2002. HECKEL, G., C. C. VOIGT,F.MAYER, AND O. VON HELVERSEN. 1999. Molecular phylogenetics and taxonomic revision of the genus Extra harem paternity in the white lined bat Saccopteryx bilineata Tonatia (Chiroptera: Phyllostomidae). Journal of Mammalogy (Emballonuridae). Behaviour 136:1173 1185. 83:49 57. HECKEL, G., AND O. VON HELVERSEN. 2003. Genetic mating system and LEWIS, S. E. 1995. Roost fidelity of bats a review. Journal of the significance of harem associations in the bat Saccopteryx Mammalogy 76:481 496. bilineata. Molecular Ecology 12:219 227. LINDENFORS, P., B. S. TULLBERG, AND M. BIUW. 2002. Phylogenetic HODGKISON, R., S. T. BALDING,Z.AKBAR, AND T. H. KUNZ. 2003. analyses of sexual selection and sexual size dimorphism in Roosting ecology and social organization of the spotted winged pinnipeds. Behavioral Ecology and Sociobiology 52:188 193. fruit bat, Balionycteris maculata (Chiroptera: Pteropodidae), in LOURENC¸O, S. I., AND J. M. PALMEIRIM. 2007. Can mite parasitism a Malaysian lowland dipterocarp forest. Journal of Tropical affect the condition of bat hosts? Implications for the social structure Ecology 19:667 676. of colonial bats. Journal of Zoology (London) 273:161 168. KALKO, E. K. V., D. FRIEMEL,C.O.HANDLEY, AND H. U. SCHNITZLER. MCCARTHY, T. J., A. I. GARDNER, AND C. O. HANDLEY,JR. 1992. 1999. Roosting and foraging behavior of two neotropical gleaning Tonatia carrikeri. Mammalian Species 407:1 4. 1389

MCCRACKEN, G. F., AND J. W. BRADBURY. 1977. Paternity and genetic REID, F. A. 1997. A field guide to the mammals of Central America heterogeneity in the polygynous bat, Phyllostomus hastatus. and southeast . Oxford University Press, New York. Science 198:303 306. RODRI´GUEZ HERRERA, B., R. A. MEDELLI´N, AND M. GAMBA RI´OS. 2006. MCCRACKEN, G. F., AND J. W. BRADBURY. 1981. Social organization Tent building by female Ectophylla alba (Chiroptera: Phyllostomi and kinship in the polygynous bat Phyllostomus hastatus. dae) in Costa Rica. Acta Chiropterologica 8:557 560. Behavioral Ecology and Sociobiology 8:11 34. ROSSITER, S. J., R. D. RANSOME,C.G.FAULKES,D.A.DAWSON, AND G. MCCRACKEN, G. F., AND G. S. WILKINSON. 2000. Bat mating systems. JONES. 2006. Long term paternity skew and the opportunity for Pp. 321 357 in Reproductive biology of bats (E. G. Crichton and selection in a mammal with reversed sexual size dimorphism. P. H. Krutzsch, eds.). Academic Press, London, United Kingdom. Molecular Ecology 15:3035 3043. MCELLIGOTT, A. G., F. NAULTY,W.V.CLARKE, AND T. J. HAYDEN. SAFI, K. 2008. Social bats: the males’ perspective. Journal of 2003. The somatic cost of reproduction: what determines re Mammalogy 89:1342 1350. productive effort in prime aged fallow bucks? Evolutionary SAFI, K., AND G. KERTH. 2007. Comparative analyses suggest that Ecology Research 5:1239 1250. information transfer promoted sociality in male bats. American MORRISON, D. W. 1979. Apparent male defense of tree hollows in the Naturalist 170:465 472. fruit bat, Artibeus jamaicensis. Journal of Mammalogy 60:11 15. SCHAAL, A., AND J. W. BRADBURY. 1987. Lek breeding in a deer MORRISON, D. W., AND S. H. MORRISON. 1981. Economics of harem species. Biology of Behaviour 12:28 32. maintenance by a neotropical bat. Ecology 62:864 866. SERVATIUS, A. 1997. Das Jagd und Echoortungsverhalten von MYERS, P. 1978. Sexual dimorphism in size of vespertilionid bats. d’Orbigny’s Rundohrenfledermaus Tonatia silvicola im Flugraum. American Naturalist 112:701 711. Diploma thesis, University of Tu¨bingen, Tu¨bingen, Germany. NAGY, M., G. HECKEL,C.C.VOIGT, AND F. MAYER. 2007. Female STORZ, J. F., ET AL. 2001. Clinal variation in body size and sexual biased dispersal and patrilocal kin groups in a mammal with dimorphism in an Indian fruit bat, Cynopterus sphinx (Chiroptera: resource defence polygyny. Proceedings of the Royal Society of Pteropodidae). Biological Journal of the Linnean Society 72:17 31. London, B. Biological Sciences 274:3019 3025. TAN, K. H., A. ZUBAID, AND T. H. KUNZ. 1997. Tent construction and NORBERG, U. M., AND J. M. V. RAYNER. 1987. Ecological morphology social organization in Cynopterus brachyotis (Muller) (Chiroptera: and flight in bats (Mammalia, Chiroptera) wing adaptations, flight Pteropodidae) in Peninsular Malaysia. Journal of Natural History performance, foraging strategy and echolocation. Philosophical 31:1605 1621. Transactions of the Royal Society of London, B. Biological TER HOFSTEDE, H., AND M. B. FENTON. 2005. Relationships between Sciences 316:337 419. roost preferences, ectoparasite density, and grooming behaviour of NOWAK, R. M. 1994. Walker’s bats of the world. 1st ed. Johns neotropical bats. Journal of Zoology (London) 266:333 340. Hopkins University Press, London, United Kingdom. TUTTLE, M. D. 1975. Population ecology of the gray bat (Myotis ORTEGA, J., AND H. T. ARITA. 1999. Structure and social dynamics of grisescens): factors influencing early growth and development. harem groups in Artibeus jamaicensis (Chiroptera: Phyllostomidae). Occasional Papers of the Museum of Natural History of the Journal of Mammalogy 80:1173 1185. University of Kansas 36:1 24. ORTEGA, J., AND H. T. ARITA. 2000. Defence of females by dominant UEBERSCHAER, K. 1999. Untersuchung zur Hangplatzo¨kologie der males of Artibeus jamaicensis (Chiroptera: Phyllostomidae). neotropischen Fledermaus Tonatia silvicola unter Beru¨cksichtigung Ethology 106:395 407. der Sozialstruktur und des Reproduktionsstatus. Diploma thesis, ORTEGA, J., AND H. T. ARITA. 2002. Subordinate males in harem University of Wu¨rzburg, Wu¨rzburg, Germany. groups of Jamaican fruit eating bats (Artibeus jamaicensis): VOIGT, C. C., ET AL. 2008. Songs, scents, and senses: sexual selection satellites or sneaks? Ethology 108:1077 1091. in the greater sac winged bat, Saccopteryx bilineata. Journal of ORTEGA, J., AND J. E. MALDONADO. 2006. Female interactions in harem Mammalogy 89:1401 1410. groups of the Jamaican fruit eating bat, Artibeus jamaicensis VOIGT,C.C.,AND W. J. STREICH. 2003. Queuing for harem access in (Chiroptera: Phyllostomidae). Acta Chiropterologica 8:485 495. colonies of the greater sac winged bat. Animal Behaviour 65:149 156. ORTEGA, J., J. E. MALDONADO,G.S.WILKINSON,H.T.ARITA, AND VON FRISCH, K. 1974. Animal architecture. Harcourt Brace Jovano R. C. FLEISCHER. 2003. Male dominance, paternity, and relatedness vich, Inc., New York. in the Jamaican fruit eating bat (Artibeus jamaicensis). Molecular WECKERLY, F. D. 1998. Sexual size dimorphism: influence of mass Ecology 12:2409 2415. and mating systems in the most dimorphic mammals. Journal of PRESTWICH, G. D. 1988. Die chemischen Waffen der Termiten. Pp. 78 Mammalogy 79:33 52. 86 in Spektrum der Wissenschaft Biologie des Sozialverhaltens WILDE, C. J., C. R. KNIGHT, AND P. A. RACEY. 1999. Influence of (D. Franck, ed.). D. Franck, Heidelberg, Germany. torpor on milk protein composition and secretion in lactating bats. RACEY, P. A. 1973. Environmental factors affecting the length of Journal of Experimental Zoology 284:35 41. gestation in the pipistrelle bat, Pipistrellus pipistrellus. Journal of WILKINSON, G. S. 1984. Reciprocal food sharing in the vampire bat. Reproduction and Fertility Supplement 19:175 189. Nature 308:181 184. RACEY, P. A., AND J. R. SPEAKMAN. 1987. The energy costs of WILKINSON, G. S. 1985a. The social organization of the common pregnancy and lactation in heterothermic bats. Symposia of the vampire bat. 1. Pattern and cause of association. Behavioral Zoological Society of London 57:107 125. Ecology and Sociobiology 17:111 121. RALLS, K. 1977. Sexual dimorphism in mammals: avian models and WILKINSON, G. S. 1985b. The social organization of the common unanswered questions. American Naturalist 111:917 938. vampire bat. 2. Mating system, genetic structure, and relatedness. RASA, A. 1985. Mongoose watch: a family observed. John Murray Behavioral Ecology and Sociobiology 17:123 134. Ltd., London, United Kingdom. WILKINSON, G. S. 1986. Social grooming in the common vampire bat, RECKARDT, K., AND G. KERTH. 2007. Roost selection and roost Desmodus rotundus. Animal Behaviour 34:1880 1889. switching of females Bechstein’s bats (Myotis bechsteinii)as WILKINSON, G. S. 1990. Food sharing in vampire bats. Scientific a strategy of parasite avoidance. Oecologia 154:581 588. American 262:64 70. 1390

WILKINSON, G. S., AND W. J. BOUGHMAN. 1998. Social calls coordinate WILSON, E. O. 1971. The insect societies. Belknap Press of Harvard foraging in greater spear nosed bats. Animal Behaviour 55:337 350. University, Cambridge, Massachusetts. WILLIAMS, D. F., AND J. S. FINDLEY. 1979. Sexual size dimorphism in WINDSOR, D. M. 1990. Climate and moisture variability in a vespertilionid bats. American Midland Naturalist 102:113 127. tropical forest: long term records from Barro Colorado Island, WILLIS, C. K. R., AND R. M. BRIGHAM. 2007. Social thermoregulation Panama. Smithsonian Contributions to the Earth Sciences 29: exerts more influence than microclimate on forest roost preferences 1 145. by a cavity dwelling bat. Behavioral Ecology and Sociobiology.