The discriminative stimulus effects of acute morphine followed by naltrexone have been described previously in nonhuman primates. The purposes of this study were to 1) extend the pharmacological characterization of the discrimination by testing mu-opioid agonists other than morphine and opioid-like compounds other than naltrexone and 2) to examine further the relationship between agonist pretreatment time and manifestation of the cue produced by morphine followed by naltrexone. Subjects were trained to discriminate 1.7 mg/kg morphine --> 0.1 mg/kg naltrexone (MOR --> NTX) versus saline followed by 0.1 mg/kg naltrexone. When combined with 0.1 mg/kg naltrexone, all agonists tested, save buprenorphine, meperidine, and nalbuphine, produced dose-dependent increases in MOR --> NTX-appropriate responding, culminating in criterion levels of responding. Comparing agonist ED50 values revealed a rank order of potency of etorphine >> fentanyl >> levorphanol > heroin > or = methadone > or = nalbuphine > or = morphine. ED50 values for buprenorphine and meperidine could not be calculated. MOR --> NTX-appropriate responding after doses of agonist that produced criterion or near criterion levels of responding was also a function of naltrexone dose. After morphine pretreatment, diprenorphine and nalorphine, but not buprenorphine, dose-dependently substituted for naltrexone. The MOR --> NTX discrimination also depended upon the interval between morphine and NTX administration. Finally, 1-h pretreatment with morphine and etorphine, but not buprenorphine, followed by naltrexone generalized to 4 h MOR --> NTX. These results suggest a minimum efficacy requirement of acutely administered agonists together with the naltrexone training dose for stimulus control of behavior. However, in some cases this requirement can be overcome with higher doses of naltrexone.