We studied the removal of seeds of three species of large-seeded tree (Astrocaryum standleyanum, Attalea butyracea and Dipteryx oleifera) from three different heights within six study plots in a lowland forest in central Panama. Fresh fruits with intact seeds fitted with industrial sewing bobbins were placed within semi-permeable exclosures. Removed seeds were tracked to deposition sites, and seed fate was determined. Removals were likely perpetrated by two small rodents, the strictly terrestrial Proechimys semispinosus and the scansorial Sciurus granatensis, because they were the most abundant small rodents in the study site during the study period and were of sufficient size to remove large seeds. Rodent abundance and fruit availability were estimated by conducting censuses. Nine microhabitat variables were measured at each deposition site to determine if these two rodents were preferentially depositing seeds in sites with certain characteristics or were randomly depositing seeds. During the study, rodents handled 98 seeds, 85 of which were not predated upon and could potentially germinate. Removal rates were not influenced by rodent abundance or fruit availability. Seeds were most frequently moved <3 m and deposited with the fruit eaten and the seed intact. However, some seeds did experience relatively long-distance dispersal (>10 m). Rodents preferentially deposited seeds in locations with large logs (>10 cm diameter), dense herbaceous cover, and an intact canopy. The number of large logs was different from random locations. Despite not being able to determine long-term fate (greater than c. 1 y), we show that these small rodents are not primarily seed predators and may in fact be important mutualists by dispersing seeds relatively long distances to favourable germination sites.
