Habitat differences and roost selection between two disparate landscapes by a neotropical bat (Artibeus lituratus) in the Atlantic Forest of Paraguay

1. Species that range across heterogeneous environments should optimally modify their behavior to local conditions. We explored this adaptive habitat-use concept by examining a primary resource: bat roosts. If landscape types differ in roosting characteristics, changes in roost use should manifest. 2. We tested this relationship between bats and landscape by contrasting bat roost use between contiguous forest and fragmented matrix. We posited that if forest structure is different and roosts are limited, fragmentation may affect roost use and selection. 3. The neotropical great fruit-eating bat (Artibeus lituratus) is a common bat species in Alto Parana & PRIME; Atlantic Forests of eastern Paraguay. We collected habitat use data by following 39 telemetered bats, and compared tree and stand characteristics from roost trees and random trees in both contiguous and matrix landscapes. 4. Diameter at breast height (DBH) of trees used as roosts was greater than the DBH of random trees, but only in the contiguous landscape. Tree height and crown cover density decreased in tandem and sequentially: contiguous landscape (roost trees taller with denser crowns than random trees), then matrix landscape (random trees taller with denser crowns than roost trees). 5. Across both landscapes, roost trees had greater vine species richness, and more vine knots and surrounding vines than random trees. Also, the nearest taller tree occurred closer to random trees than roost trees. Furthermore, canopy height around random trees was taller than around roost trees, and random trees tended to have more surrounding basal area than roost trees. 6. When A. lituratus roost selection in the Atlantic Forest is considered across disparate landscapes, a tall, large-boled tree is not the most important predictor. The strongest indicators of an A. lituratus roost were a vine-covered tree, in an uncluttered stand of smaller trees. While roost selection by A. lituratus also differed between contiguous and matrix landscapes, the aforementioned indicators remained consistent. Ergo, behavioral differences between these two A. lituratus groups cannot be explained through roost-use alone. Further exploration of tropical bat roosting is needed if a true comprehensive and holistic synthesis of bat roosting ecology is to be achieved.