Understory bat roosts, availability and occupation patterns in a Neotropical rainforest of Costa Rica

The strong link between bats and their roosts is widely recognized as being particularly significant. Despite this, roosting ecology of bats is poorly understood and much of the basic information is still unknown. In this study, we investigated the availability and occupation patterns of four roost types (trees, caves, termite nests and tents) used by bats at Tirimbina Biological Reserve (TBR), Costa Rica. To accomplish our aim, we systematically surveyed both sides of established trails and transects, looking for understory roosts. Potential roosts were examined for bat presence in order to establish occupation. Roost availability and density were estimated using traveled distances (km) and inspected area (10 m for trees/caves and 15 m for tents/termite nests) of each trail or transect sampled. For the tent roosts, data on taxonomic information of plant modified, type of architecture, condition and construction achievement were also recorded. The area surveyed represented 45.4 % of the total area of the TBR (345 ha). Tents were the most common roost (56.6 % of all roosts, N = 223), followed by trees (24.4 %, N = 96), termite nests (18.8 %, N = 74) and caves (0.2 %, N = 1). We detected only 27 roosts occupied by bats (6.8 % of all roosts, 0.17 occupied roosts/ha). Caves showed the highest occupation rate (100 %, N = 1), followed by trees (17.7 %, N = 17), tents (3.6 %, N = 8) and termite nests (1.3 %, N = 1). We found the roosts for 10 species, representing 33.9 % of the bat fauna documented at the reserve (62 species). Density of roosts per bat species varied between 0.017-0.138 roosts/ha. Phyllostomidae was the best-represented family with Micronycteris microtis representing the most common species encountered. Four distinct tent architectures were documented. Bifid architecture was the most common (133 tents), followed by Conical (47 tents), Apical (27 tents) and Inverted Boat (16 tents). Most of the tents found were healthy (76.7 %, N = 171) and totally constructed (88.8 %, N = 198). Our study demonstrated that occupied bat roosts are difficult to find in the forest. When compared to the roost availability, the low occupation rates suggested that, at least in our study area, roosts might not be a limiting resource. Nevertheless, to confirm this hypothesis, information about fidelity and selection process of the species is fundamental for understanding to what extent these roosts meet the requirements to be inhabited or modified. Worldwide conservation