To establish whether the plasma concentration of pancuronium reflects magnitude of neuromuscular blockade, the authors determined times of recovery from neuromuscular blockade and associated plasma concentrations following equipotent doses of pancuronium using three methods of pancuronium administration: the isolated-arm technique in conscious volunteers (n = 4), and the bolus intravenous injection (n = 7) and continuous-infusion methods (n = 3) in anesthetized patients. Although maximum depressions of twitch tension were similar (85 ± 11,91 ± 6, and 92 ± 4 per cent, respectively) with the three techniques, times to recovery from neuromuscular blockade differed significantly, being 10 ± 2 min with the isolated-arm technique, 23 ± 7 min with the bolus-injection technique, and 46 ± 5 min with the continuous-infusion method. The plasma concentration of pancuronium necessary for neuromuscular blockade was least with the isolated-arm technique and greatest with continuous infusion. At 25 and 75 per cent recovery, mean plasma concentrations were 0.07 ± 0.01 and 0.04 ± 0.01 μg/ml in the isolated arm; 0.13 ± 0.04 and 0.09 ± 0.02 μg/ml after bolus injection, and 0.20 ± 0.04 and 0.11 ± 0.01 μg/ml during continuous infusion, respectively. It is concluded that the relationship between plasma concentration of pancuronium and magnitude of neuromuscular blockade depends on the method of pancuronium administration.