The cannabinoid receptor agonist Δ(9)-tetrahydrocannabinol (THC) enhances the antinociceptive effects of µ-opioid receptor agonists, raising the possibility of using a combination of THC and opioids for treating pain. This study examined the effects of noncontingent and contingent administration of THC on intravenous heroin self-administration in rhesus monkeys. Self-administration of different unit doses of heroin (0.0001-0.1 mg/kg/infusion) generated a typical inverted U-shaped dose-response curve. In one experiment (n=4), noncontingent THC (0.1-1.0 mg/kg) dose dependently shifted the heroin dose-response curve downward in three monkeys and slightly leftward in one monkey. In a second experiment (n=4), monkeys could self-administer THC alone (0.0032-0.032 mg/kg/infusion), heroin alone, or a mixture of THC and heroin. THC alone did not maintain responding above that obtained with saline; however, increasing the THC dose with heroin dose dependently decreased the number of infusions received and the rate of responding, as compared with data that were obtained with heroin alone. These results indicate that THC does not significantly enhance the positive reinforcing effects of heroin, further supporting the view that combining cannabinoid and opioid receptor agonists (e.g. for treating pain) does not increase, and might decrease, the abuse liability of individual drugs.