in two separate experiments, an 8 rat 8 station operant arena was used to study resistance to extinction to test learning and optimal foraging theories of behaviour. the learning theory, called the generalized decrement theory (gdt) assumes that increased resistance to extinction will occur with increased partial reinforcement schedules. it becomes more difficult to distinguish the acquisition phase from the extinction phase as the time interval increases and as a result less generalized decrement is observed during extinction. an extention of optimal foraging theory called the skill pool effect (spe) describes a foraging group as composed of either active producers (active participants in locating food) or scroungers (passive participants who specialize in following others to food sites). the producers determine when to leave a food patch and move on. the gdt and spe were tested by allowing the population access to all stations with some bars capable of producing food reward and others as not producing food reward. for three consecutive sessions the population was exposed to either 1,2,4 or 8 bars functioning and then extinguished under the same conditions. in experiment 1, an attempt was made to provide an equal distribution of food as the number of functioning bars decreased. this was done by enriching schedules of reinforcement as the functioning bars decreased. in experiment 2, the schedule of reinforcement was held constant at a vi of 120 in order to control for any schedule differences. according to the gdt, the population should manifest the greatest number of responses with the least amount of bars working, during extinction. however, the spe would anticipate the greatest number of responses occurring with the most amount of bars functioning. the results showed the greatest number of responses occurring with 4 bars functioning. this occurred in both experiments. the findings did not support either the gdt or the spe. it appears that additional work is needed to account for the above phenomenon.