Experiments examined how training dose of morphine modulates the agonist and antagonist effects of selected mu opioids. Saline and either 3.2 or 5.6 mg/kg morphine were established as discriminative stimuli for food-reinforced responses in two groups of rats. Etorphine and morphine evoked full morphine lever responding in both groups, but were more potent in subjects trained with 3.2 mg/kg morphine. Methadone, dezocine and dl-pentazocine were equieffective and equipotent in evoking full morphine-like stimulus effects in both groups. Nalbuphine evoked full morphine-like stimulus effects and decreased the dose of morphine required for stimulus effects in rats trained with 3.2 mg/kg morphine. In contrast, nalbuphine antagonized stimulus effects of morphine in rats trained with 5.6 mg/kg morphine, and antagonized rate-altering effects of morphine in both groups, with an apparent pA2 of 5.4 mol/kg. Nalorphine and naltrexone antagonized stimulus and rate-altering effects of morphine in both groups, with an apparent pA2 of 6.0 mol/kg for nalorphine and 7.8 mol/kg for naltrexone. For both compounds, apparent pA2 values did not discriminate between stimulus or rate-altering effects of morphine, or between training conditions. d-Amphetamine and ketamine enhanced the rate-suppressing effects of morphine. However, whereas d-amphetamine prevented stimulus effects of 3.2 mg/kg, but not 5.6 mg/kg morphine, ketamine increased sensitivity to stimulus effects of morphine in rats trained with 3.2 mg/kg morphine without altering sensitivity in rats trained with 5.6 mg/kg. These results demonstrate that training dose of morphine can modulate the likelihood that certain opioids will reproduce or antagonize discriminative stimulus effects of morphine. Furthermore, for nalbuphine and nalorphine, diminished morphine-like stimulus effects in subjects trained with the higher dose of morphine appeared to result from low efficacy mu agonist actions.