Forest management activities influence habitat suitability for bats, and knowledge of the roosting ecology of bats is fundamental to developing strategies for conserving bats in managed forests. Information on use of roosts by multiple species of bats in a given area may provide insight into interspecific ecological patterns and could improve management prescriptions to provide habitat for bats through time across diverse ownerships and over multiple spatial scales. We investigated use of conifer snags as roosts by females of 3 species of forest-dwelling bats in Douglas-fir (Pseudotsuga menzesii) forests in the western Oregon Cascade Range. We radiotagged 29 female big brown bats (Eptesicus fuscus), 55 long-legged myotis (Myotis volans), and 27 long-eared myotis (Myotis evotis) and located 42, 105, and 24 snag roosts for the 3 species, respectively. All 3 species most frequently used Douglas-fir snags and in similar proportions to their availability. Big brown bats and long-legged myotis rarely roosted in stands <40 years old but age of stands used by female long-eared myotis did not differ from those randomly available. Odds of a snag being used as a roost by big brown bats increased with diameter at breast height and decreased with distance from the capture site. Diameter of snags used for roosting and the number of small (10–50 cm dbh) snags within the 20-m radius plot were variables in the best model for roost use by long-legged myotis. The best model for long-eared myotis included distance to the capture site. Odds of a snag being used by female long-eared myotis decreased with increasing distance from the capture site. There was considerable overlap in structural characteristics and the physical context of roost snags among the 3 species, but the types of roosts used among landscapes with differing densities of snags differed among the 3 species. Although big brown bats and long-legged myotis used only snags and live trees as roosts, long-eared myotis used a diversity of structures and the frequency of use of these structures differed with density of snags in the landscape. Relative to other roost types, frequency of use of snags by long-eared myotis was nearly twice as high in landscapes with high densities of snags as in those with low densities. We found that some species of bats alter selection of roosts depending on landscape context and availability of different types of roosts. Our findings demonstrate that forest managers must consider the needs of multiple bat species and the distribution of roosts in the landscape, especially where densities of snags are low and at low elevations in intensively managed landscapes.