theories of resistance to extinction such as the generalized decrement theory (gdt) have been largely based on research involving isolated individuals. in the natural environment however, extinction, like acquisition takes place in a social context. very few studies have addressed this issue and those that have either failed to create a stable social environment or have not addressed the effect of population density, an important aspect of sociality, on resistance to extinction. to study this effect a group of eight rats was exposed to two daily 30-minute vi-120 schedules of reinforcement in an eight station operant conditioning arena. during each of four phases of the experiment, the population was consistently exposed to the vi-120 reinforcement schedule at a given number of stations (n=l, 2, 4, 8). following each condition the rats were exposed to a session of extinction in which bar pressing at a station no longer produced food. to avoid the confounding effect of amount of reinforcement, since only the number of work stations available and not the reinforcement schedule changed, a second experiment was conducted in which the vi schedule was varied to compensate for the amount of reinforcement. according to the gdt the results should have produced an increased resistance to extinction during trials when the conditions of reinforcement were similar to conditions of extinction. this similarity occurred when fewer bars were available for reinforcement, which according to the gdt, should have created a negative relationship between responses in extinction and the number of stations presented. exactly the opposite effect occurred; that is, a decrease in resistance to extinction was observed when fewer bars were presented to the rats. the data support an ecological theory of social foraging called the skill pool effect, which predicts a positive relationship between the number of responses in extinction and the number of stations available. this suggests that social learning and transmission mechanisms exist and should be considered when evaluating the effectiveness of various learning and foraging theories.