The effects of morphine, naltrexone, and nalorphine were studied in rats trained to lever-press for intravenous heroin and then tested under conditions of non-reinforcement. Animals were reinforced for lever-pressing on a continuous reinforcement schedule (100 micrograms/kg per infusion) for 2-3 h each day following which reinforcement was terminated and animals were studied under extinction conditions for the remainder of the session. Each day following the termination of responding under extinction conditions, animals were given a single injection of saline, morphine, nalorphine, or naltrexone; lever-pressing under the extinction conditions was then observed for several hours. When animals adapted to this regimen, very low levels of responding were seen following saline injections; morphine (2 or 10 mg/kg) reinstated vigorous responding that lasted 1-4 h. Naltrexone (2 mg/kg) suppressed responding below the levels seen after saline, and nalorphine (10 mg/kg) had the same effect as saline. These observations support the view that opioid-seeking behavior is primed by the proponent or opioid-like actions of opioids and not by the opponent or drug-opposite effects associated with opioid withdrawal.