Ecological models predict a positive correlation between fruit production and primate abundance in the Neotropics. To test this relationship, I compiled information on primate abundance and calculated different indexes of fruit production for 30 Neotropical sites. These indexes can be grouped in three categories: (1) Fruit production estimates based on fruit traps, (2) basal area of endozoochorous trees and (3) density of these trees. The first estimate was the best predictor of both primate biomass (r2= 0.80) and species richness (r2= 0.64). The advantage of using fruit trap estimates is that they take into account production rates (which is not the case for basal area or density estimates), while the advantage of using basal area over density estimates is that it includes some of the expected variation due to tree size. However, using both basal area and density indexes I found a positive correlation between the basal area index and primate biomass for frugivorous monkeys and small platyrhines, but there was no correlation for folivorous and seed predator primates. I also found a positive correlation between pitheciine biomass and the abundance of Eschweilera trees. The analyses gave little support to the importance of suggested keystone resources such as figs and palms. Finally, when including climatic, geographic and plant diversity variables, fruit production continued to be a good predictor of primate biomass in the Neotropics, but primate species richness was best predicted by latitudinal gradients and plant species richness.