Multicompartment models have been fitted to experimental data on plasma lead and blood lead concentrations of subjects studied by de Silva (1981, Brit. J. Industr. Med. 38, 209-217) and one subject studied by Manton and Malloy (1983, Brit. J. Industr. Med., 40, 51-57). Nonlinear models for plasma lead vs. blood lead in populations provide significantly better descriptions than does a linear model. Short-term kinetic data do not clearly resolve the precise nonlinear mechanism, however. Parameters of plasma and erythrocyte distribution in a single subject were significantly different on two occasions.