Under two titration schedules, responses of squirrel monkeys adjusted the intensity of a continuous electric shock. In one schedule, responses produced food pellets but also increased a shock intensity which otherwise was decreased at a fixed rate (punishment titration). In another schedule, responses decreased a shock intensity which otherwise was increased at a fixed rate (escape titration). Responding maintained under the punishment titration schedule was only decreased by morphine, methadone, nalorphine and naloxone. This decrease in response rate was associated with a decrease in intensity of continous electric shock. Comparable rates of responding maintained under the escape titration schedule were either unaffected or increased over the same dose range of the same drugs. These effects on rate of responding were associated with either no effect or a slight decrease in intensity of continuous electric shock. When control rate of responding under the escape titration schedule was then increased by requiring five responses to decrease shock intensity, morphine, but not d-amphetamine, decreased responding at doses which had no effect or even increased responding when only one response was required to decrease shock intensity. These findings indicate that titration schedules are not always useful for studying the analgesic effects of drugs, but that specific effects of drugs are determined by specific details under which responding is maintained.