Psychostimulants lead to withdrawal from social interactions and to a decline of affective behavior in squirrel monkeys [Saimiri sciureus]. These changes, in addition to motor stereotypies, may be related to stimulant-induced psychosis in humans. In the 1st of 2 series of experiments, 1 mg/kg d-amphetamine or 10 mg/kg cocaine, administered orally 3 times over 24 h to 1 adult male member of an established group (n = 6-9), engendered stereotyped movements of the head and hands, reduced rest postures and greatly reduced all forms of social initiatives. Chlorpromazine (0.25-1.0 mg/kg), haloperidol (0.25, 0.5 mg/kg) and physostigmine (0.04, 0.08 mg/kg), administered before the 3rd amphetamine or cocaine injection, blocked the motor stereotypies and hyperactivity. Chlorpromazine, haloperidol and physostigmine did not reliably antagonize the pronounced reduction in social behavior. The 2nd series of experiments focused on agonistic behavior in the context of resident-intruder confrontations and on affiliative behavior toward group members. d-Amphetamine (3 .times. 0.5 mg/kg) and, to a lesser extent, cocaine (3 .times. 10 mg/kg) decreased affiliative and agonistic behavior. Chlorpromazine (0.5, 1.0 mg/kg) and haloperidol (0.1, 0.25 mg/kg) did not block the severe disruption of the affiliative and agonistic behavior in amphetamine-treated monkeys; physostigmine (0.06 mg/kg) reversed the decline in time spent close to the familiar monkey in amphetamine-treated monkeys. Stimulant-induced stereotypies were effectively antagonized by chlorpromazine, haloperidol and physostigmine. Psychostimulant-induced changes in primate social behavior may be mediated by mechanisms other than those underlying motor stereotypies.