Conservation of cavity-roosting bats must take into account their frequent movements among multiple roosts within a forest. One such species, Rafinesque's big-eared bat (Corynorhinus rafinesquii) is apparently rare or declining or both over its range. From 2001 through 2004 we captured C. rafinesquii at roosts located under bridges and used radiotelemetry to monitor movements of bats among roosts in southeastern Mississippi. We calculated an index of roost fidelity for individuals and compared indices based on sex, age-class, and location of capture (2 different sites). Of 25 different roost structures that we located, 14 were hollow trees and 11 were human-made (bridges, abandoned houses, and an oil tank). Overall, bats switched roosts every 2.1 days, changed roosts 2.6 ± 2.0 times (mean ± SD) per tracking period, and used 2.5 ± 1.2 roosts per tracking period. Bats that were captured together sometimes reunited at subsequent roosts. Dissimilarity in roosting opportunities between the 2 localities could have explained differences in roost fidelity among bats in the 2 areas. Roosting behavior of C. rafinesquii was comparable to that of other bats that primarily roost within tree cavities. The roosting strategy of C. rafinesquii appeared flexible; bats showed low day-to-day fidelity to roosts that were relatively common but not exceptionally stable (trees) and, in the apparent absence of tree roosts, higher fidelity to human-made roosts that were of higher structural integrity. Because C. rafinesquii moves frequently among roosts and alters fidelity by roost type, researchers should design sampling protocols and/or interpret data from surveys accordingly.