Discriminative stimulus effects of opioid agonists were studied in morphine-dependent (10.0 mg/kg/day) pigeons discriminating among i.m. injections of morphine (10.0 mg/kg), saline and naltrexone (0.032 mg/kg) while responding under a fixed ratio schedule of food presentation. Morphine and naltrexone occasioned responding on the respective correct keys in a dose-related manner with complete generalization (greater than or equal to 90% correct) occurring with doses of morphine larger than 10.0 mg/kg and doses of naltrexone larger than 0.001 mg/kg. Several opioid mu and kappa agonists were studied in dependent pigeons and in morphine-abstinent pigeons (i.e., 30-hr morphine-deprived). The opioid mu agonist etonitazene substituted completely for morphine in dependent pigeons; in abstinent pigeons a small dose of etonitazine reversed withdrawal and larger doses substituted for morphine. Some opioid agonists did not substitute completely for morphine in dependent pigeons and partially (nalbuphine, meperidine) or completely (butorphanol, buprenorphine) reversed withdrawal in abstinent pigeons. In contrast, the opioid kappa agonists trans-3,4-dichloro-N-methyl-N(2-[1-pyrrolidinyl]cyclohexyl) benzeneacetamide methanesulfonate and bremazocine did not substitute for morphine or for naltrexone in dependent pigeons and did not reverse withdrawal in abstinent pigeons. Moreover, in dependent pigeons buprenorphine, butorphanol and nalbuphine, but not trans-3,4-dichloro-N-methyl-N(2-[1-pyrrolidinyl]cyclohexyl)benzeneace tam ide methanesulfonate, antagonized the discriminative stimulus effects of morphine and of naltrexone. These results confirm in morphine-dependent pigeons qualitative differences in discriminative stimulus effects among mu and kappa opioid agonists and support the notion that differences among agonists under some conditions can result from variations in efficacy.