Roost characteristics have been described for most North American bats, but debate continues over the ultimate mechanisms behind roost selection. Hypotheses include the need for a stable microclimate, protection from predators, proximity to foraging habitat, and availability of alternative roosts, among others. Our objective was to weigh evidence for hypotheses regarding selection of diurnal summer roosts using southeastern myotis (Myotis austroriparius) as a model. We used transect searches and radiotelemetry to locate 25 roosts at eight study sites across the Coastal Plain of Georgia, USA. We measured 22 characteristics of trees, at all occupied roosts and at randomly selected unoccupied trees. We evaluated 10 hypotheses using single-season occupancy models. The best supported model predicted bat presence based on the variables tree species, solid wood volume, and canopy cover. Because these characters affect heat retention and insolation, our results are consistent with the hypothesis that bats select roosts that provide a favorable microclimate. However, data on roost temperature and humidity are needed for a conclusive determination. Occupancy was greatest at the study area closest to caves occupied by southeastern myotis. Water tupelo trees appear to be an important resource for this species, although proximity to suitable caves also seems to affect presence and should be considered in conservation planning.