In morphine-treated (3.2 mg/kg/day) rhesus monkeys discriminating between s.c. injections of 0.01 mg/kg of naltrexone and saline while responding under a fixed-ratio schedule of stimulus-shock termination, the apparent affinity of antagonists for opioid receptors was estimated using two different methods: 1) substitution studies in which cumulative doses of drug were administered to monkeys that received morphine 3 hr before the session and 2) studies in which single doses of antagonists were administered before cumulative doses of agonists in monkeys that had received saline 3 hr before the session. The ED50 values from the substitution study were compared to apparent pA2 values determined when antagonists were administered before agonists. Apparent affinities (pA2) for antagonists, estimated by their capacity to prevent the effects of the opioid agonist alfentanil, were: naltrexone = 8.69, naloxone = 7.78, quadazocine = 7.55, nalorphine = 7.31 and naltrindole = 5.29. The pA2 values were highly correlated (r = 0.94) with ED50 values determined in substitution studies. These results demonstrate that substitution by opioid antagonists for the naltrexone discriminative stimulus in morphine-treated monkeys provides a good estimate of affinity for opioid mu receptors as supported by independent pA2 analyses. These data also suggest affinity of opioid antagonists for opioid mu receptors is not changed in monkeys treated daily with 3.2 mg/kg of morphine and, furthermore, support the notion that binding to opioid receptors is the mechanism, by which antagonists precipitate withdrawal in opioid-dependent monkeys.