| Abstract: | It has been suggested that the anthropogenically driven loss of herbivorous mammals might lead to changes in the recruitment patterns of tropical rain forest plants, but few studies have examined the explicit mechanisms that might account for this effect. Here we propose a conceptual model linking differential mammalian defaunation and differential plant recruitment. We posit that in the absence of medium/large herbivores but with small rodent granivores still present (i.e. differential defaunation), predation pressure will be greater on small-seeded species than on large-seeded species. We tested such differential predation hypothesis (DPH) by means of a series of experiments directed to assess: 1) seed consumption by small rodents in laboratory cages; 2) seed-removal rates in small rodent enclosures in the field and 3) removal of seeds placed on the forest floor and exposed to either the full complement of mammals or only small rodents. Seeds used in the experiments were arranged in pairs consisting of species from the same taxonomic family but with a contrasting size (large, small). We found: 1) a significantly greater consumption of smaller seeds (a 2.3- to 20.5-fold difference) in cages and 2) a significantly greater removal of small-seeded species (a 3.7- to 65-fold difference) in field enclosures. Results of seed removal experiments in free-access plots and selective exclosures were more complex, with a general absence of significant differences among treatments but we found that predation was in general concentrated on small-seeded species and small rodents were the predominant visitors to the plots. This, together with the overall short distance of dispersal among large seeds suggest that in differentially defaunated forests large-seeded species are more likely to escape predation. We posit that such size related differential predation may lead to the floristic impoverishment observed in some defaunated forests. |
