The present study investigated the ability of drug-associated cues to reinstate extinguished responding following an extended period of withdrawal from cocaine self-administration. Rats self-administered cocaine (0.33mg/infusion) for 2 weeks of daily 3-h limited-access sessions under a fixed-interval (FI) schedule of reinforcement, in which responding resulted in simultaneous illumination of a stimulus light and drug infusions. Rats were then exposed to 20 daily extinction sessions. Noncontingent presentation of the stimulus light and infusion pump sound on day 21 of extinction resulted in a significant increase in responding. Twenty days later (43 days after cocaine withdrawal), rats returned to the operant chamber and exposed to drug-associated cues responded significantly more than animals exposed to extinction conditions. In a second experiment, using a form of variable-interval schedule in which the stimulus light was presented prior to drug infusions, stimulus-cued recovery of responding was similar to that obtained under the FI schedule. A third experiment showed that noncontingent presentation of the stimulus light alone on day 21 failed to reinstate extinguished responding, suggesting that stimulus-cued reinstatement of responding was due to a compound stimulus or preferential conditioning of the infusion pump sound. The present paradigm may serve as a useful model for the investigation of drug abuse and relapse, since it allows for the independent examination of the reinforcing and conditioned effects of a drug.