Tubocurarine (dTC) was administered to twenty patients by intravenous infusion at a predetermined constant rate, and the twitch responses recorded at stimulus intervals of 10 seconds (0.1 Hz). In half the group dTC plasma concentrations were measured both during and following the infusion. This provided data for derivation of the pharmacokinetic parameters, two plasma concentration-response curves and the dose-response curve. The mean effective dTC plasma concentration for 50% paralysis during the infusion (onset) was 0.93 micrograms/ml, but was lower post-infusion (offset) (0.54 micrograms/ml, p less than 0.001). The steady state dTC plasma concentration which should produce 95% paralysis was predicted to lie between 0.95 and 1.67 micrograms/ml. In the range 20 to 80% paralysis the slopes of the two (log) concentration-response lines were similar, although steeper than the associated (log) dose-response regression line, the test for parallelism showed no statistical difference. Combination of the plasma concentration-response curve with pharmacolkinetic models was used to predict the time course of paralysis in patients with normal and impaired renal function. Dose-response curves for the other 10 patients were also constructed to stimulus frequencies from 0.02 to 5 Hz. In the range 20 to 80% paralysis the regression lines appeared to be parallel. The results were pooled for all patients at a stimulation rate of 0.1 Hz, giving a mean effective dose of dTC at 95% paralysis of 0.53 mg/kg, and an ED50 of 0.22 mg/kg.