Physiologically based pharmacokinetic (PB-PK) models describe the dynamic behavior of chemicals and their metabolites in individual tissues of living animals. Because PB-PK models contain specific parameters related to the physiological and biochemical properties of different species as well as the physical chemical characteristics of individual chemicals, they are useful tools for performing high dose/low dose, dose route, and interspecies extrapolations in hazard evaluations. An example of such extrapolation has been presented by M. E. Andersen, H. J. Clewell III, M. L. Gargas, F. A. Smith, and R. H. Reitz (Toxicol. Appl. Pharmacol. 87, 185-205, 1987), who employed a PB-PK model for methylene chloride (CH2Cl2) to estimate the chronic toxicity of this material. However, one limitation of this PB-PK model was that the metabolic rate constants for the glutathione-S-transferase (GST) pathway in humans were estimated by allometric scaling rather than from experimental data. In this paper we report studies designed to estimate the in vivo rates of metabolism of CH2Cl2 from in vitro incubations of lung and liver tissues from B6C3F1 mice, F344 rats, Syrian Golden hamsters, and humans. A procedure for calculating in vivo metabolic rate constants from the in vitro studies is presented. This procedure was validated by making extrapolations with mixed function oxidase enzymes (MFO) acting on CH2Cl2, where both in vitro and in vivo rates of metabolism are known. The in vitro rate constants for the two enzyme systems are consistent with the hypothesis presented by Andersen et al. that metabolism of CH2Cl2 occurs in vivo by two competing pathways: a high-affinity saturable pathway (identified as MFO) and a low-affinity first-order pathway (identified as GST). The metabolic rate constants for GST obtained from these studies are also consistent with the hypothesis of Andersen et al. that production of large quantities of glutathione/CH2Cl2 conjugates in vivo may increase the frequency with which lung and liver tumors develop in some species of animals (e.g., B6C3F1 mouse). When in vivo studies in humans are unavailable, in vitro enzyme assays provide a reasonable method for estimating metabolic rate constants.