Experimental test of the mechanism underlying sexual segregation at communal roosts of harvestmen (Prionostemma spp.)

Closely related, ecologically similar species often roost in distinctly different habitats, and roosting patterns also vary within species in relation to sex, age and season. The causes of such variation are not well understood at either a proximate or ultimate level. We studied communal roosting in two congeneric species of Prionostemma harvestmen at a rainforest site in Nicaragua. Previous research showed that Prionostemma sp. 1 forms male‐biased communal roosts in tree‐root cavities, while Prionostemma sp. 2 forms communal roosts of variable but temporally stable sex ratios on spiny palms. Here, we investigate potential mechanisms underlying variation in roosting site choice between and within these syntopic species. First, we present the results of a field experiment designed to probe the mechanism underlying skewed roost sex ratios in Prionostemma sp. 2. Previous studies have suggested that these harvestmen use conspecific scent to locate communal roosts and that new roosts can be established via group translocation. Therefore, to test the hypothesis that skewed roost sex ratios in this species arise from sex differences in scent marks, we translocated single‐sex groups of ca. 30 individuals to each of 20 previously unoccupied spiny palms. Female release sites attracted new recruits of both sexes, while male release sites attracted almost exclusively males. We infer that Prionostemma sp. 2 females preferentially roost in sites scent‐marked by females and that this mechanism is sufficient to explain the skewed roost ratios. Further adding to knowledge of Prionostemma roosting behavior, we show that Prionostemma sp. 1 forms female‐biased communal roosts on spiny palms, that some roosts contain both species, and that the species composition is stable on a time scale of at least 2 weeks. To the best of our knowledge, this study is the first experimental test of mechanisms underlying sexual segregation at communal roosts in any taxon.     

Mechanisms for the formation and maintenance of traditional night roost aggregations in a territorial damselfly

Communal roosting has been studied extensively in birds, but the mechanisms and functions of this taxonomically widespread behaviour pattern remain poorly understood. We studied the roosting behaviour of rubyspot damselflies, Hetaerina americana, in relation to sex and territorial status, and conducted field experiments to test for specific mechanisms of roost formation and maintenance. Both sexes tended to return close to their previous night's roost, but only males were significantly more roost site faithful than chance expectations based on individual day ranges. Males were more roost site faithful when they held mating territories. After acquiring a territory, males usually began roosting closer to the territory after a delay of a few days. Roosts were not located at sites that reduced the daily commuting distance between hunting areas and territories; males generally hunted closer to their territories than to their roosts. In field experiments, sites 'seeded' with synthetic models of male rubyspots attracted more recruits than vacant control sites and control sites seeded with nonrubyspot (clear-winged) damselfly models. Sites seeded repeatedly with rubyspot models often remained popular for roosting after the models were removed, suggesting that the models established new traditional roosts. These results indicate that conspecific attraction and individual spatial memory together may be sufficient to explain, at a proximate level, the traditional night roost aggregations of this species. We discuss these results in relation to functional hypotheses for roost site choice and fidelity. Copyright 2000 The Association for the Study of Animal Behaviour.     

Roosting behavior of a Neotropical migrant songbird, the northern waterthrush Seiurus noveboracensis, during the non-breeding season

Several species of Nearctic-Neotropical migratory songbirds appear to form roosting aggregations while on their wintering grounds but little is understood about the ecology of this behavior. We studied roosting behavior and patterns of roost habitat selection in the northern waterthrush Seiurus noveboracensis , during three winter years (2002–2004) in Puerto Rico using radio telemetry. Overall, red mangrove was selected for roosting disproportionately to its availability. Regardless of diurnal habitat used, 87% (n=86) of northern waterthrush selected dense stands of coastal red mangrove for roost sites. Individuals traveled up to 2 km to access roost sites in this habitat on a daily basis. The majority (8 of 14) of individuals roosted alone, while others roosted in loose aggregations near communal roosts of gray kingbirds Tyrannus dominicensis . Patterns of roost site selection did not vary by sex. Individuals showing aggressive response to playback during the day, however, selected roost sites significantly closer to the coast. Several additional migratory and resident bird species also used red mangrove for night-time roosting habitat. Red mangrove may be a critical nocturnal roosting habitat for bird populations that live in proximity to coastal areas in the Neotropics. The benefits of nocturnal roosting behavior as well as why individuals appear to select red mangrove remain poorly understood.     

Roost use by long-tailed bats in South Canterbury: examining predictions of roost-site selection in a highly fragmented landscape

We studied the roosting ecology of the long-tailed bat (Chalinolobus tuberculatus) during the springautumn months from 1998–2002 at Hanging Rock in the highly fragmented landscape of South Canterbury, South Island, New Zealand. We compared the structural characteristics and microclimates of roost sites used by communally and solitary roosting bats with those of randomly available sites, and roosts of C. tuberculatus occupying unmodified Nothofagus forest in the Eglinton Valley, Fiordland. Roosting group sizes and roost residency times were also compared. We followed forty radio-tagged bats to 94 roosts (20% in limestone crevices, 80% in trees) at Hanging Rock. Roosts were occupied for an average of 1 day and 86% were only used once during the study period. Colony size averaged 9.8 ± 1.1 bats (range 2–38) and colonies were dominated by breeding females and young. Indigenous forest, shrubland remnants and riparian zones were preferred roosting habitats. Communally roosting bats selected roosts in split trunks of some of the largest trees available. Selection of the largest available trees as roost sites is similar to behaviour of bat species occupying unmodified forested habitats. Temperatures inside 12 maternity roosts measured during the lactation period were variable. Five roosts were well insulated from ambient conditions and internal temperatures were stable, whereas the temperatures inside seven roosts fluctuated in parallel with ambient temperature. Tree cavities used by bats at Hanging Rock were significantly nearer ground level, had larger entrance dimensions, were less well insulated, and were occupied by fewer bats than roosts in the Eglinton Valley. These characteristics appear to expose their occupants to unstable microclimates and to a higher risk of threats such as predation. We suggest that roosts at Hanging Rock are of a lower quality than those in the Eglinton Valley, and that roost quality may be one of the contributory factors in the differential reproductive fitness observed in the two bat populations. The value of introduced willows (especially Salix fragilis) as bat roosts should be acknowledged. We recommend six conservation measures to mitigate negative effects of deterioration of roosting habitat: protection and enhancement of the quality of existing roosts, replanting within roosting habitat, provision of high quality artificial roosts, predator control, and education of landowners and statutory bodies.     

Predation risk and resource availability explain roost locations of Whimbrel Numenius phaeopus in a tropical mangrove delta

Roosts are important sites for shorebirds in non‐breeding areas at night and during high tides. How the spatial configuration of food and risk of predation and disturbance influence roost site use in tropical locations remains poorly known. We analysed the locations of nocturnal roosts of Whimbrel Numenius phaeopus in mangroves of Sanquianga National Park, Colombia, with respect to variation in spatial variables related to food resources and risk of predation and disturbance. We contrasted characteristics of all 13 known nocturnal roost locations with those of all other mangrove islands (n = 209) within the limits of the park. We estimated the distance from roosts and other mangrove islands to foraging sites, and sources of predators and human disturbance. Larger areas of feeding habitat surrounded nocturnal roosts than other mangrove islands, and the average distance to individual feeding patches was shorter. Roosts were also more isolated than other islands, but proximity to sources of human disturbance did not differ. We conclude that Whimbrel roost site use in Sanquianga was best explained by a combination of access to feeding territories and isolation from potential sources of mainland predators, but not by avoidance of human disturbance. Beyond identifying factors influencing roost site selection, the large aggregations of individuals in single locations may suggest that presence of conspecifics itself also plays a role in the formation of Whimbrel roosts. We highlight the interaction of food and risk landscapes with intraspecific attraction on the roost site selection by Whimbrels and the importance of mangroves as roosting sites in tropical regions.     

The benefit of being a social butterfly: communal roosting deters predation

Aposematic passion-vine butterflies from the genus Heliconius form communal roosts on a nightly basis. This behaviour has been hypothesized to be beneficial in terms of information sharing and/or anti-predator defence. To better understand the adaptive value of communal roosting, we tested these two hypotheses in field studies. The information-sharing hypothesis was addressed by examining following behaviour of butterflies departing from natural roosts. We found no evidence of roost mates following one another to resources, thus providing no support for this hypothesis. The anti-predator defence hypothesis was tested using avian-indiscriminable Heliconius erato models placed singly and in aggregations at field sites. A significantly higher number of predation attempts were observed on solitary models versus aggregations of models. This relationship between aggregation size and attack rate suggests that communally roosting butterflies enjoy the benefits of both overall decreased attack frequency as well as a prey dilution effect. Communal roosts probably deter predators through collective aposematism in which aggregations of conspicuous, unpalatable prey communicate a more effective repel signal to predators. On the basis of our results, we propose that predation by birds is a key selective pressure maintaining Heliconius communal roosting behaviour.     

Effects of Hierarchical Roost Removal on Northern Long-Eared Bat (Myotis septentrionalis) Maternity Colonies

Forest roosting bats use a variety of ephemeral roosts such as snags and declining live trees. Although conservation of summer maternity habitat is considered critical for forest-roosting bats, bat response to roost loss still is poorly understood. To address this, we monitored 3 northern long-eared bat (Myotis septentrionalis) maternity colonies on Fort Knox Military Reservation, Kentucky, USA, before and after targeted roost removal during the dormant season when bats were hibernating in caves. We used 2 treatments: removal of a single highly used (primary) roost and removal of 24% of less used (secondary) roosts, and an un-manipulated control. Neither treatment altered the number of roosts used by individual bats, but secondary roost removal doubled the distances moved between sequentially used roosts. However, overall space use by and location of colonies was similar pre- and post-treatment. Patterns of roost use before and after removal treatments also were similar but bats maintained closer social connections after our treatments. Roost height, diameter at breast height, percent canopy openness, and roost species composition were similar pre- and post-treatment. We detected differences in the distribution of roosts among decay stages and crown classes pre- and post-roost removal, but this may have been a result of temperature differences between treatment years. Our results suggest that loss of a primary roost or ≤ 20% of secondary roosts in the dormant season may not cause northern long-eared bats to abandon roosting areas or substantially alter some roosting behaviors in the following active season when tree-roosts are used. Critically, tolerance limits to roost loss may be dependent upon local forest conditions, and continued research on this topic will be necessary for conservation of the northern long-eared bat across its range.     

Roosting behavior and group stability of the big fruit-eating bat Artibeus lituratus (Chiroptera: Phyllostomidae)

The roosting behavior of the big fruit-eating bat, Artibeus lituratus (Phyllostomidae, Stenodermatinae) in an Andean region of Venezuela is described. Sixty-four video recordings made at three separate foliage roosts during 1 year showed that group size varied between two and 14 individuals. One male was regularly observed roosting with more females than others, and this male was associated with the highest quality roost, defined here as the highest, most structurally stable, and least disturbed. Males invariably occupied exclusive roosts during the study, whereas females frequently moved among adjacent roosts. The high roost fidelity of males appears to be related to the defense of the highest quality roosts. Because females were observed roosting with each of the three males present at the study site, female groups were considered unstable. We suggest that shuttling movements of individuals may reflect a commonly observed adaptation of foliage roosting bats related to the avoidance of predators and/or parasites. The type of male–female association observed in A. lituratus is consistent with a resource defense polygyny hypothesis.     

Reconsidering the importance of harvested forests for the conservation of tree-dwelling bats

Intensively managed forests are often seen as of low priority to preserve forest bats. The main conservation strategy recommended, i.e. saving unmanaged “habitat islands” from logging to preserve some suitable habitat, detracts conservationists’ attention from ameliorating conditions for bats in harvested sites. We studied the threatened bat Barbastella barbastellus, mostly roosting in snags, in two beech forests: an unmanaged forest—the main maternity site—and a nearby, periodically logged area. We compared roost availability, roost use, capture rates, food availability and movement between these areas. The managed forest had a greater canopy closure, fewer dead trees, a smaller tree diameter and trees bearing fewer cavities than the unmanaged one. These differences helped explain the larger number of bats recorded in the unmanaged forest, where the sex ratio was skewed towards females. Prey availability was similar in both areas. We radiotracked bats to 49 day roosts. Five individuals caught in the managed area roosted in the unmanaged one at 6.7–8.2 km from the capture site. Few bats roosted in the managed forest, but those doing so proved flexible, using live trees and even rock crevices. Therefore, bats utilise areas in the matrix surrounding optimal roosting sites and sometimes roost there, highlighting the conservation potential of harvested forests. Besides leaving unmanaged patches, at least small numbers of dead trees should be retained in logged areas to favour population expansion and landscape connectivity. Our findings also question the validity of adopting presence records as indicators of forest quality on a site scale.     

Comparative Roosting Ecology of Cynopterus (Chiroptera: Pteropodidae) Fruit Bats in Peninsular Malaysia1

Although the use of modified roosts has been reported in more than 20 species of bats in the tropics, comparative studies of the roosting ecology of congeneric tent‐roosting species are notably lacking. In the Paleotropics, this unique behavior has been described in two species belonging to the genus, Cynopterus: C. sphinx and C. brachyotis. However, it is not known whether tent roosting is an essential component of their roosting ecology, or whether the behavior is found in other members of the genus. In this study we characterize the roosting ecology of four sympatric species of Cynopterus in peninsular Malaysia and use these data to address two main questions. (1) Do all four species use modified roosts and, in those that do, is tent‐roosting obligate or opportunistic? (2) Do species pairs overlap in roost preferences and roosting habitat and, if so, is there evidence for interspecific interactions in relation to these resources? We radio‐tracked bats at two floristically distinct sites and located a total of 249 roosts. Interspecific roost niche overlap was minimal at both sites and we found no evidence for interspecific competition for roost resources at the local level. Species differences in roosting ecology were defined primarily by spatial separation of roosting habitats and secondarily by within‐habitat differences in roost selection. Importantly, we found that although periodic use of modified roosts was a characteristic shared by all four species, most roosts were unmodified, indicating that tent roosting is a facultative behavior in Malaysian Cynopterus.     

Roosting Requirements of Two Frugivorous Bats (Sturnira lilium and Arbiteus intermedius) in Fragmented Neotropical Forest1

As tropical forest fragmentation accelerates, scientists are concerned with the loss of species, particularly those that play important ecological roles. Because bats play a vital role as the primary seed dispersers in cleared areas, maintaining healthy bat populations is critical to natural forest regeneration. Observations of foraging bats suggest that many Neotropical fruit‐eating species have fairly general habitat requirements and can forage in many different kinds of disturbed vegetation; however, their roosting requirements may be quite different. To test whether or not general foraging requirements are matched by equally broad roosting requirements, we used radiotelemetry to locate roost sites of two common frugivorous bat species (Sturnira lilium and Artibeus intermedius) in a fragmented forest in southeastern Mexico. Sturnira lilium roosted inside tree cavities and selected large‐diameter roost trees in remnant patches of mature forest. Fewer than 2 percent of trees surveyed had a mean diameter equal to or greater than roost trees used by . S. lilium, Artibeus intermedius roosted externally on branches and vines and under palm leaves and selected roost trees of much smaller diameter. Compared to random trees, roost trees chosen by A. intermedius were closer to neighboring taller trees and also closer in height to these trees. Such trees likely provide cryptic roosts beneath multiple overlapping crowns, with sufficient shelter from predators and the elements. While males of A. intermedius generally roosted alone in small trees within secondary forest, females roosted in small groups in larger trees within mature forest and commuted more than three times farther than males to reach their roost sites. Loss of mature forest could impair the ability of frugivorous bats to locate suitable roost sites. This could have a negative impact on bat populations, which in turn could decrease forest regeneration in impacted areas.     

Hunting Activity Effects on Roost Selection by Male Wild Turkeys

Roosting is an important component of wild turkey (Meleagris gallopavo; turkey) ecology as roosts provide security from predators and inclement weather. Males call (gobble) from roosts during the reproductive season, and roost locations are important for maximizing access to females and transmission of calls across the landscape, while also minimizing predation risk. Spring hunting of male turkeys occurs during the reproductive season, and hunting activity influences male behaviors and calling. Because roost sites are important for wild turkey ecology, we evaluated roost site selection and fidelity of male turkeys relative to land cover types, vegetative characteristics, and the presence of hunting activity during 2017–2018 in Georgia, USA. Prior to onset of hunting, males selected roosts nearest to hardwood and pine (Pinus spp.) forests. Roost site fidelity was low and distances between roosts were large. After onset of hunting, males selected pine forests less and exhibited greater plasticity in roost selection while fidelity remained minimal, suggesting that males may have altered selection to mitigate risk from hunting while maintaining the strategy of moving about their ranges and roosting at different sites on consecutive nights. Future research should examine potential effects of hunting-induced shifts in resource selection on other aspects of male turkey behavior and ecology. © 2019 The Wildlife Society.     

Roosting Behavior of Colonial and Solitary Flying Foxes in American Samoa (Chiroptera: Pteropodidae)1

We examined characteristics of roosting sites utilized by two flying fox species (Pteropus tonganus and P. samoensis) in American Samoa. The colonial roosting sites of P. tonganus were observed over a ten‐year period, including two years when severe hurricanes devastated bat populations and destroyed roost trees. Prior to the hurricanes, roosts were located on cliff faces above the ocean or steep mountainsides, locations that were either inaccessible to people or in protected areas where hunting was not allowed. In the years immediately following the hurricanes, P. tonganus colonies split into smaller groups that moved frequently to different locations. Four years after the second hurricane, colonies had coalesced and returned to many of the traditional roosting sites used before the hurricanes. Common tree species in upland and coastal forest were selected as roosts. The isolated locations selected for P. tonganus roosts were apparently the result of hunting pressure on the colonies. The solitary roosts of P. samoensis were observed during 29 months. Roosting bats were well concealed and hard to detect within the forest; even bats on exposed branches were cryptic. Mature primary forest was favored as roosting habitat. Individual bats used specific branches or trees as roosts and returned to them for up to 29 months. Unlike P. tonganus, people did not alarm roosting P. samoensis easily and some roosts were located near houses and along roads.     

Common Myna Roosts Are Not Recruitment Centres

We studied communal roosting in the Common Myna (Acridotheres tristis) in the light of the recruitment centre hypothesis and predation at the roost. The number and sizes of flocks departing from and arriving at focal roosts were recorded over a two year period. We also recorded the sizes and behaviour of foraging flocks. We found that flock sizes of birds departing from roosts at sunrise were larger than those at the feeding site, suggesting that there was no recruitment from the roosts. Flocks entering the roosts during sunset were larger on average than those leaving the following sunrise, suggesting no consolidation of flocks in the morning. Flocks entering the roosts at sunset were also larger on average than those that had left that sunrise, although there was no recruitment at the feeding site. There was no effect of group size on the proportion of time spent feeding. Contrary to expectation, single birds showed lower apparent vigilance than birds that foraged in pairs or groups, possibly due to scrounging tactics being used in the presence of feeding companions. Thus, the recruitment centre hypothesis did not hold in our study population of mynas. Predation at dawn and dusk were also not important to communal roosting: predators near the roosts did not result in larger flocks, and resulted in larger durations of arrival/departure contrary to expectation. Since flock sizes were smallest at the feeding site and larger in the evening than in the morning, but did not coincide with predator activity, information transfer unrelated to food (such as breeding opportunities) may possibly give rise to the evening aggregations.     

Bats relocate maternity colony after the natural loss of roost trees

Understanding the ephemerality of trees used as roosts by wildlife, and the number of roost trees needed to sustain their populations, is important for forest management and wildlife conservation. Several studies indicate that roosts are limiting to bats, but few studies have monitored longevity of roost trees used by bats over several years. From 2004–2007 in Cypress Hills Interprovincial Park, Saskatchewan, Canada, several big brown bats (Eptesicus fuscus) from a maternity group roosted in cavities in trembling aspen (Populus tremuloides) trees approximately 7 km southeast away from their original known roosting area (RA1). Using a long-term data set of the roost trees used by bats in this area from 2000–2007, we evaluated whether the movement of bats to the new roosting area (RA4) corresponded with annual and cumulative losses of roost trees. We also determined whether longevity of the roosts from the time we discovered bats first using them differed between the 2 roosting areas based on Kaplan-Meier estimates. Bats began using RA4 in addition to RA1 in 2004, when the cumulative loss of roost trees in RA1 over 3 consecutive years reached 18%. Most bats exclusively roosted in RA4 in 2007, when the cumulative loss of roost trees over 6 consecutive years had reached 46% in RA1. Annual survival for roost trees, from when we first discovered bats using them, was generally lower in RA1 than in RA4. Our results suggest that the movement of bats to the new roosting area corresponded with high losses of roost trees in RA1. This provides additional evidence that to maintain high densities of suitable roost trees for bats in northern temperature forests over several decades, management plans need to recruit live and dead trees in multiple age classes and stages of decay that will be suitable for the formation of new cavities. © 2019 The Wildlife Society.     

Roost and hunting site fidelity of female and juvenile Daubenton's bat Myotis daubentonii (Kuhl, 1817) (Chiroptera: Vespertilionidae)

We investigated roosting and hunting site fidelity of Daubenton's bats Myotis daubentonii (Kuhl, 1817) in the Forêt de Soignes, an old-stand forest dominated by 150–200-year-old beeches, during the summers of 2003 and 2004. Roosting behaviour and hunting activity over ponds of adult females and juveniles were monitored using radio-telemetry. Eighteen roosts were located, all in natural cavities. The bats occupied a limited number of trees located in a specific and small roosting area. This roost aggregation was not linked to the distribution of hollow trees. Furthermore, whereas all 11 ponds in the study area were exploited by Daubenton's bats, monitored individuals were limited to two ponds to which they showed high fidelity. These two ponds were not the closest to the roosting area. Overall, these results show that at least for the time we conducted our experiment, female and juvenile Daubenton's bats are highly faithful to specific roosting and hunting grounds.     

Patterns of roost site selection and use by Southern Ground-Hornbills in north-eastern South Africa

Different habitats may be used for the needs of various aspects of an animal’s life. Southern Ground-Hornbill Bucorvus leadbeateri groups announce their presence within year-round territories by calling at dawn from their overnight roost sites. Knowledge on ground-hornbill roosting habits is limited. Groups roost in large trees, apparently close to where they end up after daily foraging. We investigated patterns of roost site selection and use for four Southern Ground-Hornbill groups in the Associated Private Nature Reserves, north-eastern South Africa, based on data from GPS-satellite transmitters. The number of roost sites used per month averaged 15.4 ± 4.7 across all groups, indicating little evidence of strong preferences for specific sites. This number was least when groups were breeding, decreasing throughout the early wet season (October–December) and was lowest during the late wet season (January–March) when actively breeding groups frequently roosted close to the nest (54–83% of roosts <1 000 m of the nest). As might be expected, the mean monthly number of nights per roost peaked during the breeding season (December–January). Riparian habitats were preferred for roosting during the breeding season, whereas disturbed areas, as well as Combretumand mopane-dominated habitats were preferred during the dry non-breeding season. Adequate large trees not only for nesting, but also for roosting, particularly in riparian habitats, may therefore be an important and potentially limiting factor for the successful reproduction of Southern Ground-Hornbills.     

Summer tree roost selection by Rafinesque's big-eared bat

Roost requirements of most North American forest bats are well-documented, but questions remain regarding the ultimate mechanisms underlying roost selection. Hypotheses regarding roost selection include provision of a stable microclimate, space for large colonies, protection from predators, and proximity to foraging habitat, among others. Although several hypotheses have been proposed, specific mechanisms likely vary by species and geographic region. Rafinesque's big-eared bat (Corynorhinus rafinesquii) commonly roosts in trees with large basal hollows in the Coastal Plain of the southeastern United States. Our objective was to weigh evidence for hypotheses regarding selection of diurnal summer roosts by Rafinesque's big-eared bat at 8 study sites across the Coastal Plain of Georgia, USA. We used transect searches and radiotelemetry to locate roosts and measured 22 characteristics of trees, tree cavities, and surrounding vegetation at all occupied roosts and for randomly selected unoccupied trees. We evaluated 10 hypotheses using single-season occupancy models and used Akaike's information criterion to select the most parsimonious models. We located 170 tree roosts containing approximately 870 bats for our analysis. The best supported model predicted bat presence from cavity size, interior wall texture, and number of entrances. Because large cavities allow bats to fly and smooth walls impede attacks by terrestrial predators, our results are consistent with the hypothesis that bats select roosts that allow them to evade predators. However, data on predation rates are needed for a conclusive determination. Because trees suitable as roosts for Rafinesque's big-eared bat are rare in the landscape, protection of suitable forested wetland habitat is essential to provide current and long-term roost tree availability. © 2012 The Wildlife Society.     

Roost site selection by Little Owls Athene noctua in relation to environmental conditions and life‐history stages

Roost site selection is a state‐dependent process, affected by the individual's costs and benefits of roosting at a specific site in the available environment. Costs and benefits of different roost sites vary in relation to intrinsic factors and environmental conditions. Thus, the cost–benefit functions of roost sites are expected to differ between seasons and life‐history stages, resulting in adjustments in roost site selection. Studying roost site selection throughout the year therefore provides information about year‐round habitat requirements at different life‐history stages. However, little is known about the roosting behaviour of birds. Here, the roost site selection of Little Owls Athene noctua was studied by repeated daytime location of 24 adult and 75 juvenile radiotagged individuals from July to November. Little Owls preferred sheltered roost sites such as tree cavities with multiple entrances. They increasingly used sheltered sites from summer to winter and preferentially used sheltered roost sites with low ambient temperatures. Juveniles used significantly less sheltered sites during dispersal than before and afterwards, and used less sheltered sites than adults within their home‐range. The survival probability of birds roosting frequently at exposed sites was reduced. Roost site selection is probably driven by the two mechanisms of predator avoidance and thermoregulation, and the costs of natal dispersal may include increased predation threat and higher energy expenditure for thermoregulation. We suggest that adequate roost sites, such as multi‐entrance tree cavities, are an important habitat requirement for Little Owls and that habitat quality can be affected by manipulating their availability.