The possibility of predicting the behavior of in vivo systems based on physical and chemical parameters determined by in vitro experiments is examined using benzoic acid. The physical and chemical parameters governing percutaneous absorption of benzoic acid--permeability, partition coefficient, and skin thickness--were determined by in vitro experiments as described in Ref. 1. These parameters were used, in combination with benzoic acid elimination kinetics, to predict the results of in vivo experiments using a comprehensive mathematical model. The in vivo system consists of a congenitally athymic (nude) rat with a surgically constructed human skin sandwich (HSSF) flap on which a donor cell is placed. To apply the in vitro parameters to an in vivo system requires a suitable pharmacokinetic model describing distribution and elimination for benzoic acid in the nude rat. Blood concentrations of benzoic acid following a bolus intravenous injection are closely described by a two-compartment open pharmacokinetic model with elimination occurring from only one compartment. The mathematical model of the rat-donor cell system combines this two-compartment model of the rat with a percutaneous absorption model to provide useful estimates of the measured in vivo blood levels. Comparisons of predicted and measured results suggest that the parameters determined by in vitro experimentation can be used to predict the behavior of complex in vivo systems, if a suitable mathematical model is available.