Cooperative Fish and Wildlife Research Units Program: Virginia
Education, Research and Technical Assistance for Managing Our Natural Resources

Bat day-roost selection often is described by outcomes derived from comparisons of day-roosts with randomly selected, supposedly unused, trees. Relatively few studies, however, look at patterns of multi-year selection or compare day-roosts used across years. We explored the process determining day-roost selection using 2 years of roost selection data for northern long-eared bat (Myotis septentrionalis) maternity colonies on the Fort Knox Military Reservation, Kentucky, USA. We compared characteristics of random trees and day-roosts used in 2 different years using a multinomial logistic model and day-roost species selection using chi-squared tests. We found that the factors differentiating day-roosts from non-roosts and day-roosts between years varied substantially. Day-roosts differed from random trees in the first year of data in all measured factors, but only in diameter at breast height and decay stage in the second year. Between years, day-roosts differed in roost size and canopy position, but not decay stage. Day-roost species selection however, was non-random and did not differ significantly between years. Although bats use multiple trees within a colony, our results suggest that there were multiple unused trees that were suitable for roosts at any time. We believe day-roost selection pattern descriptions will be inadequate if based only on a single year of data, and that inferences of roost selection based on comparisons of roost to random trees should be limited.