In Ateles, individuals forage for fruit in subgroups of unstable membership. The size and composition of these foraging parties are highly variable. Data from my long-term field study of a population of black spider monkeys in Peru show a strong positive correlation between the average size of these parties and the number and size of trees currently producing fruit in the environment. This correlation appears to be driven by intragroup feeding competition, since individuals in larger parties showed substantially lowered feeding efficiencies, as measured by increased daily ranges and decreased time spent feeding. Patch size and patch density, which jointly determine fruit abundance, interact to determine foraging party size, apparently by the following mechanism. The number of spider monkeys that can feed within a particular patch is limited by the size of that patch. When patches are close together (i. e. high patch density) individuals in excess of this limit are able to feed in nearby patches and thus maintain contact with the other members of their foraging party. When patch density is low, the increased travel costs associated with such a strategy become prohibitive and foraging parties are smaller than or equal to the number of individuals able to feed within a single patch. Females that are routinely excluded from the best feeding sites may have reduced reproductive success because of their low priority of access to resources. These data provide a quantitative link between two ecological variables (resource patch size and density) and fission-fusion social organization in the spider monkey.