Abstract: When foraging together, animals are often observed to feed from food discoveries of others. The producer-scrounger (PS) game predicts how frequently this phenomenon of food parasitism should occur. The game assumes: (1) at any moment all individuals can unambiguously be categorized as either playing producer (searching for undiscovered food resources) or scrounger (searching for exploitation opportunities), and (2) the payoffs received from the scrounger tactic are negatively frequency dependent; a scrounger does better than a producer when the scrounger tactic is rare, but worse when it is common. No study to date has shown that the payoffs of producer and scrounger conform to the game's assumptions or that groups of foragers reach the predicted stable equilibrium frequency (SEF) of scrounger, whereby both tactics obtain the same payoff. The current study of three captive flocks of spice finches, Lonchura punctulata, provides the first test of the PS game using an apparatus in which both assumptions of the PS game are met. The payoffs to the scrounger, measured as feeding rate (seeds/s), were highly negatively frequency dependent on the frequency of scrounger. The feeding rate for scrounger declined linearly while the rate for producer either declined only slightly or not at all with increasing scrounger frequency. When given the opportunity to alternate between tactics, the birds changed their use of each, such that the group converged on the predicted SEF of scrounger after 5-8 days of testing. Individuals in this study, therefore, demonstrated sufficient plasticity in tactic use such that the flock foraged at the SEF of scrounger. Copyright 2000 The Association for the Study of Animal Behaviour.