Intravenous cocaine self-injection behaviour was studied using a new discrete-trials procedure in monkeys. Daily sessions consisted of two sets of 40 discrete trials, separated by a 5-min time out (TO) period; the intertrial interval during each segment of the session was equal to 50 s. In the first experiment, cocaine self-injection behaviour (15 μg/kg per injection) was suppressed in a dose-dependent manner by morphine pretreatment (1–5.6 mg/kg, im). When cocaine self-injection responding was suppressed by morphine pretreatment (3 or 5.6 mg/kg, im), the morphine antagonist, naloxone (0.01–1.0 mg/kg, im), reinstated drug-taking behaviour; the effect depended upon the dose of the antagonist and upon the amount of morphine given to suppress cocaine self-administration behaviour. When monkeys were pretreated with chlorpromazine (0.03–5.6 mg/kg, im) before a cocaine self-injection session, small doses marginally increased responding and larger doses exerted a suppressive effect. These effects were observed when cocaine was available at a unit dose of 15 μg/kg per injection. When the unit dose of cocaine was increased to 300 μg/kg per injection, small doses of chlorpromazine (0.03–1 mg/kg, im) exerted no clearly detectable effects, though responding for drug injections was profoundly suppressed when sessions were preceded by chlorpromazine pretreatment at doses of 3 and 5.6 mg/kg, im. In this experiment, drug intake per session was greater when a large injection dose (300 μg/kg per injection) was available for self-administration than when a small dose (15 μg/kg per injection) could be self-injected; the frequency of drug self-injection was reduced, however, when the larger unit dose of cocaine replaced the smaller one. When considered in relation to the results of other experiments, it is evident that control of cocaine self-injection responding by a discrete-trials procedure results in schedule-dependent behaviours that differ in relative sensitivity to drug pretreatments from cocaine-reinforced responding controlled by fixed-ratio schedules of drug reinforcement.