Patas monkeys acquired a different four-response chain each session by responding sequentially on three keys in the presence of four geometric forms. The response chain was maintained by food presentation under a fixed-ratio schedule. Errors produced a brief timeout but did not reset the chain. Each day there were four 15-min sessions, with a 10-min intersession interval. Cumulative dose-effect curves for phencyclidine were obtained by giving an IM injection before each of the four sessions; successive injections increased the cumulative dose by 1/4 log-unit steps. When phencyclidine was administered alone, overall response rate decreased and percent errors increased with increasing doses. When cocaine was injected IM before the first session at a dose that was ineffective when given alone, the phencyclidine dose-effect curves for both rate and accuracy tended to shift to the left. After pretreatment with the lowest effective dose of cocaine, which decreased rate without affecting accuracy when given alone, the rate-decreasing and error-increasing effects of phencyclidine were generally even more pronounced in two of three subjects. The results indicate that cocaine potentiates the disruptive effects of phencyclidine on complex operant behavior in monkeys.