Opioid-maintained volunteers were trained to distinguish between a low dose of the opioid antagonist naloxone (0.15 mg/70 kg, i.m.; i.e., Drug A) and placebo (i.e., Drug B) under an instructed novel-response drug-discrimination procedure in which subjects identify the drug condition as "A," "B," or "N" (neither A nor B--'novel'). Once the discrimination was acquired, doses of naloxone (000.15 mg/70 kg, i.m.) and the mixed-action opioid agonist/antagonists butorphanol (0-1.5 mg/70 kg, i.m.) and nalbuphine (0-3.0 mg/70 kg, i.m.) were tested. Naloxone produced dose-related increases in naloxone-appropriate responding with little or no 'novel'-appropriate responding. Butorphanol produced a dose-related increase in naloxone- and 'novel'-appropriate responding, occasioning approximately 70% and 29%, respectively, at the highest dose tested. Nalbuphine produced 40-65% naloxone-appropriate responding at all doses tested and 33% 'novel'-appropriate responding at the highest doses. Self-reported effects produced by each agent differed only slightly. These results suggest that mixed-action opioid agonist/antagonists may be distinguished from the opioid antagonist naloxone based on their discriminative-stimulus effects under a novel-response naloxone discrimination procedure.