In a series of field experiments using Costa Rican rain forest plants, we examined the effect of accessibility on fruit removal rates. We compared the effects of fruit placement in terminal and axillary infructescences on diurnal and nocturnal removal rates, visitation rates, and incidence of fruit damage. We used three different species of berries (Phytolacca rivinoides, Psychotria brachiata, and Psychotria pitteri) and worked in three different habitats (fallow fields, treefall gaps, and forest understory) and in two different seasons (July-September, a season of fruit abundance and December-January, a season of fruit scarcity.)We found that in oldfields especially, diurnal removal rates by birds were significantly greater from axillary than from terminal infructescences. Nocturnal removal from axillary infructescences-presumably by rodents-is also occassionally significant. From these data, from observations on climbing ability and fruit use in captive rodents, and from reports in the literature, we suggest that rodents are significant sources of fruit and seed loss in tropical shrubs. We hypothesize that placement of the infructescence on the plant affects fruit removal by both seed-dispersing birds and by the less agile, often seed-destroying rodents. The balance between the two rates is an important component of a plant's dispersal success.Diurnal fruit removal rates were higher during the season of fruit scarcity than during the season of fruit abundance and higher in old fields than in forest gaps or understory. Fruit damage rates-probably due to orthopterans-were slightly greater in gaps and understory than in old fields.