Diphenylhydantoin (DPH) elimination was studied in 4 overdosed children who presented with serum concentrations ranging from 44 to 76 mg/l. Serum was assayed for DPH and urine was assayed for both DPH and 5-(p-hydroxyphenyl)-5-phenylhydantoin (HPPH). The serum and urine data were subjected to simultaneous computer nonlinear regression analysis using a one-compartment pharmacokinetic model, which accounts for much of the known disposition kinetics of DPH. Computed values for the apparent in vivo Michaelis-Menten constants, K-M and V max, were compared with values derived from data in the literature for normal adult subjects. A trend toward relatively lower K-M and higher V max/K-M values was seen in children. Patients with higher V max values had greater urinary excretion rates of HPPH which, at high serum levels of DPH, were relatively constant except for an apparent diurnal rhythm. The time of onset of DPH toxicity in the children was related to the magnitude by which the rate of DPH administration exceeded the V max values.