ROOSTING BEHAVIOR AND ROOST-SITE PREFERENCES OF FOREST-DWELLING CALIFORNIA BATS (MYOTIS-CALIFORNICUS)

We followed nine radiotagged female California bats (Myotis californic us) to 19 roosts in trees at two study sites in southcentral British C olumbia. Bats regularly switched roosts and the number of bats emergin g from known roosts fluctuated widely. Logistic-regression analysis sh owed that reproductively active females preferred trees further away f rom other trees of the same height or greater, and closer to neighbori ng trees, relative to available trees in the immediate vicinity of the roost. Diameter at breast height and distance to the nearest tree of the same or greater height explained significant proportions of the va riation between roost and available trees found in other areas of the same forest stand. Roost trees had significantly larger diameters and were further away from trees of the same or greater height. Percentage of canopy closure also explained a significant proportion of the vari ation between roost and available trees, such that roost trees were si tuated in areas with lower canopy closure than available trees in othe r areas of the same stand. Roost and available trees were classified c orrectly >70% of the time based on the above tree characteristics. How ever, roost trees were only correctly classified 39% of the time in th e analysis of site characteristics. Our results, combined with those f rom other recent studies, lead to the general hypothesis that forest-r oosting bats require a number of large dead trees of specific species, in specific stages of decay, and that project above the canopy in rel atively open areas. For management and conservation reasons, there sho uld be a strong incentive to evaluate this hypothesis for a variety of species in a variety of locations to test its generality.