The effects of desipramine treatment (0.56-10.0 mg/kg per day) on cocaine self-administration were compared to saline treatment in five rhesus monkeys. Cocaine (0.050 or 0.100 mg/kg per inj) and food (1 g banana pellets) self-administration were maintained on an FR 4 (VR 16:S) reinforcement schedule. Desipramine (or an equal volume saline control solution) was infused over 1 h each day through the second lumen of an intravenous catheter. After 5 days of baseline saline treatment, seven doses of desipramine each were administered for 5 days in an ascending order. Cocaine self-administration increased (P less than 0.01) or remained equivalent to base-line levels in 4 of 5 subjects during the first 15 days of desipramine treatment (0.56 to 1.78 mg/kg per day). Three monkeys continued to self-administer cocaine equivalent to or significantly above base-line levels (P less than 0.01) during days 16-30 of desipramine treatment (3.2-7.86 mg/kg per day). The highest desipramine dose (10 mg/kg per day) significantly suppressed cocaine self-administration in only one of these three monkeys (P less than 0.01). Desipramine treatment (3.2-10.0 mg/kg per day) suppressed cocaine self-administration (P less than 0.01) without a concomitant suppression of food-maintained responding in one of five subjects. A generalized suppression of both cocaine and food-maintained responding (P less than 0.01) during desipramine treatment occurred in one monkey that took the highest base-line levels of cocaine (6.3 +/- 1.03 mg/kg per day). Food-maintained responding remained equivalent to or significantly above (P less than 0.01) base-line levels in four of five monkeys during desipramine treatment. A transient decrease in food self-administration at desipramine doses of 1.78-5.62 mg/kg per day occurred in one monkey (P less than 0.01). Thirty days of desipramine treatment at the highest doses (5.62, 7.86 and 10.0 mg/kg per day for 10 days each) also did not suppress cocaine self-administration in a monkey that took an average of 4 mg/kg per day of cocaine during saline base-line treatment. These primate data are concordant with the extant clinical literature on the inconsistent effects of desipramine treatment; i.e., both stimulation of cocaine use and inconsistent or incomplete attenuation of cocaine abuse during desipramine maintenance have been reported.