Are Large Predators Keystone Species in Neotropical Forests? The Evidence from Barro Colorado Island

Large cats have been hypothesized to act as keystone predators in Neotropical forests by limiting mid-sized terrestrial mammals, which in turn control forest regeneration. The evidence for this is based predominantly on cross-site comparisons involving Cosha Cashu (CC), Peru, with its complete predator complement, and Barro Colorado Island (BCI), Panama, where puma and jaguar are no longer residents. We reevaluate the evidence in light of 725 km of transect censuses performed on BCI, comparisons with additional sites throughout the Neotropics, and historical changes in the BCI mammal fauna. Several mid-sized species showed habituation to humans near the BCI laboratory compound (where previous censuses had been carried out), indicating that cross-site comparisons, despite using similar methodologies, may be biased by differentially meeting technique assumptions, and that densities of these species at BCI and CC may not differ as greatly as previously assumed. Analyses of additional sites also indicate that individual species population density estimates for BCI, while high for several species, are not extreme. Cross-site analyses indicate that most, but not all, species show no significant difference in mean densities at sites with and without large cats. On the other hand, an evaluation of the qualitative historical evidence indicates that as cats were eliminated from BCI by poaching, many potential prey species population densities increased. These increases were not sustained, however, and may reflect natural population variability rather than the status of large cats. Finally, we examine the evidence for prey species controlling forest regeneration. While extreme removal of herbivorous and frugivorous mammals will profoundly affect forest regeneration, the effects of slight decreases in densities, or the effects of increased densities (as implicated by the keystone predator hypothesis) are equivocal.