The pharmacodynamics of the neuromuscular blocking drug gallamine were investigated in 10 surgical patients using a constant-rate infusion regimen, and results are compared to those for d-tubocurarine (dTc). Gallamine effect (paralysis)-time data gathered during and following the infusion were fitted to a pharmacodynamic effect model, while paralysis-plasma concentration data gathered during (onset) and following (offset) the infusion were fitted separately to a nonlinear form of the Hill equation. The effect model was most appropriate in characterizing the combined (on and off infusion) effect data. While there was also an excellent characterization of onset and offset data with the Hill equation, the two effect-concentration curves were not superimposable. The mean (+/- S.D.) plasma concentration of gallamine at 50 per cent paralysis during onset of action (Cp50(onset) 8.0 +/- 1.8 micrograms/ml) or that predicted to exist at steady state using the effect model (Cp50(ss) 5.4 +/- 1.4 micrograms/ml). Cp50(offset) and Cp50(ss) did not differ significantly, and there was no significant difference in the power factor (lambda) estimates for the various model fits. Comparison of the pharmacodynamic parameters for gallamine and dTc using the effect model revealed no significant differences in keo, t1/2(keo), and lambda estimates. However, Cp50(ss) for gallamine (5.4 +/- 1.4 micrograms/ml) was nine times higher than that for dTc (0.61 +/- 0.15 micrograms/ml) in absolute terms and seven times higher when compared on a molar basis.