Group Foraging Revisited: Information Sharing or Producer-Scrounger Game?

Group foraging allows some individuals to join resources discovered by others. Analyses of joining decisions in groups have taken two approaches: information sharing and the producer-scrounger (PS) game. Traditionally, information-sharing models assume that in a group of G foragers all individuals earch for food independently while, at the same time, monitoring the behavior of other group members, which thereby allows unsuccessful foragers to join and exploit food clumps discovered by others (Clark and Mangel 1984; Ranta et al. 1993; Ruxton et al. 1995). In an information-sharing group, a successful forager is joined by the remaining G 1 group members. Because each resource clump discovered yields one producing event and G 1 joining events, the predicted proportion of joining events is (G 1)/G. Information-sharing models thus predict similar frequencies of joining for all group members and suggest hat the proportion of joining events increases with the number of foragers available to join. Ruxton et al. (1995) also show that increased frequency of joining can decrease the individual foraging rate in large groups that rely on information sharing. Although less widely recognized, a "producer-scrounger" (PS) game model (Barnard and Sibly 1981) can also be used to analyze the occurrence of joining. This model is an n-person alternative-option scramble in which individuals use producer to find resources and scrounger to join resources discovered by the producers. In this type of model, the use of tactics can be genetically determined or modified through experience. A stable mixture of producer and scrounger alternatives, characterized by equality of payoffs to each, is possible when the following conditions are met: there are strong negative frequencydependent actic payoffs, so that when scrounger is rare, it does better than producer, but when scrounger is common, it does worse than producer; and the payoffs are not phenotypically imited and depend on the tactic adopted by each individual. In a PS game model, the stable proportion of joining events is expected to vary with group size along with a host of ecological factors, such as patch richness and density (Giraldeau et al. 1994). In contrast to the fixed policy of joining envisaged in information-sharing models, the main feature of PS game models is the assumption that joining behavior can vary adaptively. Despite its appeal, the assumption that all unsuccessful group members will join resources discovered by others, instead of searching further individually, has not received critical attention. For instance, it is conceivable that such a high tendency to join others provides a maximum reduction in foraging variance but at a potential, and as yet unknown, cost to mean intake. Moreover, empirical