Previous research has demonstrated that the effects ofd-amphetamine on punished lever pressing of squirrel monkeys are modified by an avoidance history in which lever pressing postpones shock. In the present experiment generalization of behavioral history across responses was assessed by determining the effects ofd-amphetamine and cocaine on punished lever pressing of squirrel monkeys before and after exposure to an avoidance procedure in which a chain-pulling response postponed shock. The punishment schedule consisted of a fixed-interval 5-min schedule of food delivery in which every 30 lever presses produced a 5-mA electric shock. During avoidance sessions each chain pull postponed shock delivery for 25 s; in the absence of chain pulling, shocks occurred every 5 s. Only a single response manipulandum was present in each phase. Punished lever pressing was initially unaltered or decreased byd-amphetamine and cocaine. Following the chain-pull avoidance history, however,d-amphetamine produced dose-dependent increases in the punished lever pressing of all three monkeys at several doses that formerly did not alter or reduce responding; a similar pattern of results was obtained when cocaine was administered to two of the subjects. The effects ofd-amphetamine and cocaine on punished lever pressing were subsequently determined within the context of a multiple schedule of lever-press punishment and chain-pull avoidance, with both manipulanda present simultaneously. The effects of the drugs on punished lever pressing within the multiple schedule were similar to their initial, pre-avoidance effects for the two monkeys whose responding was increased by both drugs in the preceding, post-avoidance phase. Chain pulling during the punishment component was dose-dependently increased, suggesting that chain pulling during punishment reduced the opportunity to exhibit increases in punishing lever pressing. The remaining monkey punished lever pressing was increased by both drugs within the context of the multiple schedule. This experiment demonstrates that avoidance-dependent upward shifts in the dose-response curves ofd-amphetamine and cocaine can occur when the punishment and avoidance responses differ, and that original effects can be partially restored when both responses are available simultaneously. The results suggest that generalization across responses of the effects of a critical behavioral history may be a general property of behavioral history phenomena within behavioral pharmacology. These findings underscore the generality and importance of behavioral history as a modulatory influence on the effects of abused drugs.