The separate effects of graded doses of morphine, naloxone, d,l-cyclazocine, and d-amphetamine on responding maintained by a differential reinforcement of low rate schedule of food presentation were examined in rats. Morphine did not alter response rates at doses of 1--5.6 mg/kg; at 10 mg/kg a 57% decrease in responding was observed and behaviour was even more severely depressed by 30 mg of morphine per kilogram. Naloxone did not affect responding at doses ranging from 0.1 to 10 mg/kg. d,l-Cyclazocine at doses of 3 and 5.6 mg/kg induced substantial increases in responding not observed when the dose was increased to 10 mg/kg. Cyclazocine, as well as morphine, produced dose-dependent decreases in the number of reinforcements per session. d-Amphetamine exerted a biphasic effect on responding; small doses increased response rates (0.3--3 mg/kg) and responding was suppressed by the drug at a dose of 10 mg/kg. Behaviourally active doses of d-amphetamine caused a dose-dependent reduction in the number of reinforcements per session. Naloxone at otherwise inactive doses (1--10 mg/kg) was found, in separate experiments, to antagonize the rate-decreasing effects of morphine, and to reduce the rate-increasing effects of d-amphetamine. The latter effect is not easily interpreted but confirms and extends other research employing rats in which naloxone was found to reduce the rate-increasing effects of small doses of d-amphetamine upon locomotor activity and responding maintained by a continuous electric-shock postponement procedure. In additional experiments morphine was given daily for 25 consecutive sessions at a dose of 30 mg/kg, 5 min preceding each test session. Responding was suppressed throughout this period and the dose of morphine given before each session was reduced to 10 mg/kg for 35 further sessions. Tolerance to the rate-decreasing effects of morphine was demonstrated; naloxone given in conjunction with morphine (10 mg/kg) in morphine-tolerant rats restored to control values the number of reinforcements per session without causing significant change in overall rates of responding. Few experiments have dealt previously with the development of tolerance to the behavioural effects of morphine under comparable dose regimens, time-course relationships, or behavioural testing procedures. Systematic analyses of these interrelated variables are needed since it is now evident that the schedule employed to maintain responding itself exerts significant effects.