The effects of the dopamine receptor D1 partial agonist, SKF 38393, on behavior maintained by cocaine was assessed in squirrel monkeys (Saimiri sciureus). One group of subjects was trained to press a key under a fixed-ratio 30-response schedule of cocaine injection; when green stimulus lamps were illuminated each 30th response produced an injection (17 micrograms/kg) followed by a 1-min period during which the lights were out and responses had no scheduled consequences. Another group of squirrel monkeys was trained under an identical schedule with food reinforcement. SKF 38393 produced dose-related decreases in rates of responding maintained by either cocaine injection or food presentation. Rates of responding maintained by cocaine were decreased to a greater extent than those maintained by food. The ED50 value for SKF 38393 for responding maintained by cocaine was 2.53 mg/kg (95% CL: 1.22-5.23), whereas that value was 15.63 mg/kg (95% CL: 2.83-86.33) for responding maintained by food. Rates of responding maintained by cocaine were an inverted-U-shaped function of dose. Pretreatment with 3.0 mg/kg SKF 38393 shifted the ascending limb of the cocaine dose-effect curve to the right. These findings suggest that indirect D1-receptor activation plays a role in the reinforcing effects of cocaine, and that drugs acting at D1 receptors may show promise as therapeutic agents in the treatment of cocaine abuse.