The effects of daily treatment with buprenorphine (0.237-0.70 mg/kg/day), naltrexone (0.32-3.20 mg/kg/day) and saline on cocaine self-administration were compared in rhesus monkeys. Cocaine (0.05 or 0.10 mg/kg/injection) and food (1-g banana pellets) self-administration were maintained on a fixed-ratio 4 (variable ratio 16:S) schedule of reinforcement. Buprenorphine, naltrexone or an equal volume saline control solution were infused slowly over 1 hr through one lumen of a double lumen i.v. catheter at the same time each day. Saline and each dose of buprenorphine (0.237, 0.40 and 0.70 mg/kg/day) or naltrexone (0.32 and 3.20 mg/kg/day) were studied for 60 sessions over 15 consecutive days. Buprenorphine significantly suppressed cocaine self-administration (P less than .001-.0001) in comparison to saline in all monkeys. Cocaine self-administration decreased by 49 to 95% in five of six monkeys on the 1st day of buprenorphine administration (0.237 and 0.40 mg/kg/day) and remained suppressed by an average of 72 to 93% during buprenorphine treatment. After abrupt termination of buprenorphine treatment (0.237 and 0.70 mg/kg/day), cocaine self-administration remained suppressed for an average of 16 +/- 4.4 and 28 +/- 6.6 days, respectively. Buprenorphine (0.237 and 0.40 mg/kg/day) initially suppressed food self-administration in some monkeys (P less than .01), but tolerance developed to buprenorphine's effects on food-maintained responding whereas cocaine self-administration remained significantly suppressed. During treatment with 0.70 mg/kg/day of buprenorphine, food self-administration returned to or significantly exceeded (P less than .01) base-line levels in three animals. Daily patterns of food self-administration were not disrupted by buprenorphine treatment. Naltrexone (0.32 mg/kg/day) initially suppressed cocaine self-administration by an average of 28% over 15 days (P less than .0009). During high-dose naltrexone treatment (3.20 mg/kg/day), cocaine-maintained responding was suppressed by 25% over 15 days (P less than .01). Cocaine-maintained responding was not significantly changed by naltrexone in one of the five subjects. Food self-administration decreased by 24% (P less than .05) after 5 days of 0.32 mg/kg of naltrexone administration, then exceeded baseline levels during 3.20 mg/kg of naltrexone administration. These data suggest that buprenorphine decreases cocaine's reinforcing properties more effectively than naltrexone across the dose-range studied. Buprenorphine may be an effective pharmacotherapy for treatment of cocaine abuse as well as dual abuse of cocaine plus heroin.