Using mist nets, we compared phyllostomid bat ensembles of continuous mature forest in Tikal National Park, Guatemala, and of forest fragments in the nearby farming landscape. Of 20 species captured, 13 were shared between treatments, 4 were unique to continuous forest, and 3 were unique to forest fragments. Dominance–diversity curves were similar for the two treatments except that Sturnira lilium comprised 43 percent of captures in the forest fragments, resulting in greater dominance there. Capture rates (and presumably relative abundance) differed significantly between continuous forest and forest fragments, both in terms of species and feeding guilds. Sturnia lilium and Dermanura sp. were captured significantly more often in forest fragments than in continuous forest, whereas Artibeus jamaicensis, A. lituratus, and Centurio senex were taken significantly more often in continuous forest. Large frugivores accounted for a higher proportion of total captures in continuous forest than in forest fragments, whereas small frugivores showed the opposite pattern. By their abundances, Carollia perspicillata and S. lilium are indicators of forest disturbance. The relative abundances of large frugivores, which feed on large fruits of mature forest trees, and small frugivores, which feed on small-fruited plants occurring in early succession, are an indicator of forest disturbance. Other groups, such as large insect- and vertebrate-eating bats, because of their low capture rates, are impractical as indicators for rapid assessment of forest disturbance based on mist netting, but may prove especially vulnerable to forest fragmentation.