The effects of d-amphetamine on punished responding were studied in two experiments.In Experiment I, pigeons responded under a multiple fixed-ratio 30 response fixed-interval 5-min schedule of food presentation with 60-sec limited holds in both components. Each response was punished with electric shock, the intensity of which was varied systematically.In Experiment II, another group of pigeons responded under a multiple fixed-interval 5-min fixed-interval 5-min schedule of food presentation with 40-sec limited holds. Each response was punished with shock during one component, and every thirtieth response was punished in the other component. d-Amphetamine increased overall rates of punished responding only rarely under any of the punishment conditions; however, response rates within the fixed-interval when rates wvere low wvere increased by d-amphetamine when the shock intensity was low (Experinment I), or when responses produced shock intermittently (Experiment II). The data suggest that the effects of d-amphetamine on punished responding depend on the control rate of responding, the punishment intensity, the punishment frequency, and the schedule of food presentation.In a variety of situations, the effects of the amphetamines on behavior have been shown to depend on the pattern of responding (Dews, 1958a, b;Smith, 1964;Clark and Steele, 1966;McMillan, 1968McMillan, , 1969. In general, appropriate doses of the amphetamines increase low rates of responding and decrease high rates.Among several exceptions to the rate-dependent effects of the amphetamines is their effect on punished responding. Although punishing stimuli often reduce responding to very low rates, amphetamine has been reported not to increase rates of punished responding. On the contrary, it has been reported that the amphetamines reduce even further low rates of punished responding. Geller and Seifter (1960) observed the effects of amphetamine (0.5, 1.0, and 1.5 mg/kg) administered to rats working under a multiple schedule in which an unpunished variable-interval 2-min (VI 2-min) component alternated with a component in '