Respiratory depression is a limiting factor in the therapeutic use of opioid analgesics. It has been suggested that respiratory depression is mediated by mu rather than kappa receptors and may involve a decrease in central nervous system sensitivity to hypercapnia. This study investigated opioid receptor mechanisms underlying respiratory depression in unanesthetized rhesus monkeys (n = 3) breathing air or 5% CO2 in air into a pressure displacement head plethysmograph. Apparent pA2 analyses of s.c. quadazocine (a mu-selective antagonist) were carried out on the effects of cumulative doses of s.c. bremazocine, ethylketocyclazocine (EKC) and (+/-)-(1-R/S,5-R/S,2 = R/S)-5,9-dimethyl-2'-hydroxy-2- tetrahydrofurfuryl-6,7-benzomorphan (Mr2033) (compounds with kappa agonist effects in other in vivo assays), alfentanil and etonitazene (compounds with mu agonist effects in other in vivo assays). Alfentanil, bremazocine, EKC and Mr2033 were approximately equipotent in causing dose-dependent depression of respiratory minute volume of CO2-stimulated and air respiration, whereas etonitazene was approximately 10-fold more potent than the above compounds. Dose-effect curves for respiratory frequency, tidal volume and respiratory minute volume for all of the agonists except bremazocine were shifted to the right by increasing quadazocine doses. Together with data previously obtained in drug discrimination and analgesia assays, results of the present study demonstrating homogeneous pA2 values for quadazocine with alfentanil, etonitazene, EKC and Mr2033 strongly suggest that the latter two compounds decrease respiratory function in rhesus monkeys by acting on mu receptors.