Selective Feeding by Phyllostomid Fruit Bats in a Subtropical Montane Cloud Forest

Phyllostomid bats eat the fruits and disperse the seeds of a wide variety of neotropical plant species, particularly those typical of early successional stages (Fleming & Heithaus 1981, Heithaus 1982, Orozco-Segovia et al. 1985). In assessing the effects of frugivorous animals on plant populations and vice-versa, it is important to determine whether the animals are selective in the types of fruit they eat, or whether the proportional abundance of different fruit species in their diets simply reflects availability in the environment. The purpose of this study was to test experimentally the hypothesis that two common phyllostomid species, Sturnira ludovici and Dermanura tolteca (=Artibeus toltecus), feed selectively in the subtropical montane cloud forest of western México. Feeding selectivity of plant-visiting animals can be measured at several ecological scales. At the community level, fruit bats and other taxonomic guilds of vertebrate frugivores exploit only a limited subset of the available fruit resource base (Heithaus et al. 1975, Janson 1983, Gautier-Hion et al. 1985). At this level, the phyllostomid bats in our study area appear to feed selectively, insofar as seeds identified from the feces of mist-netted bats represent only a fraction of the potentially available cloud forest plant species. Fruits of second- growth shrubs accounted for the overwhelming majority of 117 fecal samples obtained from S. ludovici captured throughout the yearly cycle, 1991–1993. In particular, fruits of two common ground-story shrubs, Solanum nigricans and S. aphyodendron (Solanaceae), accounted for 40 and 43 percent of all samples, respectively. Fruits of Conostegia volcanalis (Melastomataceae), a subcanopy-level tree common in moist drainages, accounted for 7 percent. Of 29 fecal samples obtained from D. tolteca captured during this same time period, fruits of S. nigricans and S. aphyodendron accounted for 21 and 42 percent of all samples, respectively, whereas fruits of C. volcanalis accounted for 38 percent. All fecal samples contained seeds of only one species per sample. The relative importance of these three plants as fruit resources for bats is corroborated by rates of nocturnal fruit-removal from plants monitored in the field (Schöndube-F. 1994). Here we report the results of feeding experiments designed to test whether cloud forest bats feed selectively among a limited array of six fruit species actually known or suspected of being included in their diets. Compared to other fruits in its diet, S. ludovici handles and digests S. nigricans fruits more efficiently, and extracts more metabolizable energy per fruit. Solanum nigricans fruits are also highest in sugar and protein content (Schöndube-F. 1994). Because the overall ease of handling, digestibility, and nutritional quality of fruit are known to influence the diet choice of frugivores (Moermond & Denslow 1985, Bonaccorso & Gush 1987), we predicted that captive S. ludovici would feed selectively on S. nigricans fruits. In a premontane cloud forest in Costa Rica, both S. ludovici and D. tolteca fed on the fruits of second growth shrubs, although D. tolteca specialized on the fruit crops of canopy level trees during seasonal peaks in fruit abundance (Dinerstein 1986). Under the assumption that patterns of habitat use and feeding behavior are similar in our study area, we predicted that captive D. tolteca would feed selectively on the fruits of C. volcanalis