BATS AND THE LOSS OF TREE CANOPY IN AFRICAN WOODLANDS

We studied the activity patterns, abundance, diversity, and diets of b ats, along with the abundance of nocturnal volant insects, at 30 sites in Miombo woodland in northern Zimbabwe. The woodland at 50% of the s ites had been disturbed by high elephant densities to the extent that the free canopy was greatly reduced The tree canopy was intact at the other sites. Intact and impacted sites differed significantly in tree (>3 m tall; >15 cm basal diameter) and shrub (1-3 m tall; (1 m tall) d iversity and cover. At each site we used ultraviolet lights to sample insects and mist nets and bat detectors to sample bats. To assess thei r diets we collected and analyzed feces from captured bats. We caught 343 bats representing the families Pteropodidae (1 species), Vespertil ionidae (11 species), and Molossidae (3 species). The molossids and ve spertilionids are all aerial feeders taking airborne insects. Bat spec ies richness, abundance, and activity were greater at intact than at i mpacted sites, but these differences were statistically significant on ly at adjacent sites (<5 km apart) not at more distant intact and impa cted sites (>20 km apart). At the adjacent sites we caught a significa ntly greater proportion of small (<10 g) bats at intact than at impact ed sites. These data and a significantly greater proportion of Scotoph ilus species (>10 g) caught during the early evening at intact than at impacted sites suggested that the removal of canopy trees affected ro ost availability for the bats. Although larger species may have commut ed between intact and impacted sites, smaller species did not. In cont rast the availability of prey did not appear to have been significantl y affected by the removal of the canopy trees, as indicated by the lig ht-trap catches of insects and the bats' diets Most bats ate mainly be etles and moths, the most abundant insects sampled at the ultraviolet lights. Our findings suggest that aerial-feeding bats such as vesperti lionids and molossids do not appear to be useful indicators of disturb ance in this habitat, even in the face of significant loss of tree can opy.