The lack of procedures which can unequivocally demonstrate cannabinoid self-administration in animals has been an obstacle to the study of the neural basis for the reinforcing effects of this drug class. Because delta(9)-tetrahydrocannabinol (delta(9)-THC) produce a relatively slow-onset, long-lasting behavioral effect, a self-administration procedure with widely spaced drug deliveries was evaluated as an alternative to fixed-ratio schedules which typically require frequent, closely spaced injections to demonstrate reinforcing effects. Three adult male rhesus monkeys were surgically implanted with intravenous catheters and trained to self-administer phencyclidine (PCP) under a 10min fixed-interval schedule of reinforcement. Three injections were available each day, separated by 2h periods during which responding had no programmed consequences. In an attempt to link the effect of the drug with the response which produced it, each 20s injection was paired with a red light which remained illuminated for 10min. PCP (100µg/kg/injection) maintained steady rates of responding during each availability period, ranging from approximately 0.2 to 0.7 responses/s. During 7 day substitution periods, Delta(9)-THC (17-100µg/kg/injection) maintained low rates of responding which occasionally surpassed those during vehicle substitutions, but fell far below rates maintained by PCP. Substitution tests with the potent Delta(9)-THC analog CP 55,940 also resulted in low rates of responding. These results demonstrate that Delta(9)-THC is a poor reinforcer in animals, even under conditions where some of its unfavourable biodispositional properties are taken into consideration.