Morphine-dependent rats can be trained to discriminate between s.c. injections of saline and 0.1 mg/kg of naltrexone. The discriminative effects of naltrexone, measured by the number of trials completed on the naltrexone-appropriate choice lever in a 20-trial avoidance paradigm, derive from stimuli associated with morphine withdrawal. Opiate and nonopiate drugs were injected s.c. and examined for their ability to block naltrexone-induced discriminative effects and loss of body weight in morphine-dependent rats. Seven opiates blocked dose dependently the discriminative effects of naltrexone and loss of body weight. Potency ranged from fentanyl (330 X morphine) to meperidine (less than 1 X morphine); effects were stereoselective for levorotatory isomers. Loperamide, an opiate that does not readily enter the brain, blocked loss of body weight but not discriminative effects, suggesting that discriminative effects are mediated centrally. Nonopiate behavioral depressants, diazepam, haloperidol and pentobarbital, did not substantially affect either dependent variable, but clonidine (0.01-1.0 mg/kg) blocked discriminative effects of naltrexone partially and weight loss completely. The blockade by morphine (30 mg/kg) of naltrexone-induced discriminative effects and weight loss was surmounted by increasing the dose of naltrexone whereas the blockade by clonidine (0.1 mg/kg) was not. Thus, blockade by opiates of effects of naltrexone appears to be due to a competitive interaction at the mu opioid receptor; clonidine has a different mechanism of action. This discrimination paradigm may afford a specific animal model for studying fundamental processes underlying physical dependence on opiates and for evaluating novel pharmacologic approaches for treating opiate withdrawal in humans.