Key pressing by rats was maintained under spaced-responding and random-ratio schedules of food delivery, and rates of responding were reliably different for each schedule. When responding was maintained under a multiple schedule, appropriate doses of d-amphetamine (1.0 and 1.7 mg/kg) markedly increased low rates of spaced responding while markedly decreasing high rates of ratio responding. These drug-produced changes in response rate resulted in decreased food presentation during both schedule components. When 1.0 mg/kg d-amphetamine was given daily, tolerance developed to initially decreased ratio responding in six to nine sessions, but did not develop to initially increased spaced responding. However, when the ratio schedule was removed, tolerance developed very quickly to increases under the spaced-responding schedule, and associated food frequency returned to control levels. When the ratio schedule was reinstated, spaced responding was once again increased, and its associated frequency of food delivery was again decreased. The development of tolerance to the behavioral effects of d-amphetamine was influenced more by global changes in response consequences during entire experimental sessions than by local changes in response consequences in single components of those sessions. Whenever the concept of "response cost" is used, it should be understood in terms of total "cost".