The hypothesis that mesolimbic dopamine is selectively involved in cocaine reinforcement was investigated in the rat. Animals were trained under a multiple schedule in which responding was reinforced by intravenous cocaine (0.75 mg/kg/injection) or food (45-mg pellets) under fixed-ratio 15 schedule requirements in alternate 30-min components of a 2-hr daily session. Infusion of the catecholaminergic neurotoxin 6-hydroxydopamine, but not the vehicle solution, into the region of the nucleus accumbens and olfactory tubercle produced selective reductions in cocaine self-administration without significantly altering responding maintained by food within the same sessions. This effect was reproduced in intact animals by substituting saline for cocaine in the self-administration component. These results support the hypothesis that the reinforcing effects of cocaine are dependent upon mesolimbic dopamine and demonstrate that cocaine self-administration can be disrupted in animals without altering behavior maintained by a nondrug reinforcer.