Mice exposed to various chemicals have been shown to respond by decreasing their core body temperature. To examine what effect such a response might have on the determination of in vivo metabolism, core body temperatures of B6C3F1 mice were recorded with temperature telemetry devices during exposure to chloroform (CHCl3) in a closed, recirculating chamber (100 to 5500 ppm). Significant decreases in body temperature occurred in all mice exposed to greater than 100 ppm CHCl3, with the greatest decrease of 14 degrees C occurring at 5500 ppm. A starting CHCl3 concentration of 4000 ppm had no effect on the 7-ethoxycoumarin O-deethylase (ECOD) activity or P450 levels determined at the end of a 5-hr gas uptake exposure. A physiologically based pharmacokinetic (PB-PK) model was developed to describe the effects of decreased body temperature on the analysis of metabolic data. In vitro ECOD activity as a measure of in vivo P450 metabolism was determined for temperatures ranging from 24 to 40 degrees C. In vitro enzyme activity decreased linearly from a maximum at 37 degrees C to one-third of this activity at 24 degrees C. A linear equation describing this enzymatic activity-temperature correlation was incorporated into the PB-PK model structure to describe decreases in metabolic activity resulting from decreases in core body temperature. In vitro blood/air and tissue/air partition coefficients were determined for CHCl3 at temperatures ranging from 24 to 40 degrees C. All blood/air and tissue/air partitions increased with decreasing temperature, while the tissue/blood partition coefficients calculated from the tissue/air and blood/air partitions decreased with decreasing temperature. Adding these temperature corrections to the model greatly improved the overall fit of the gas uptake curves at all concentrations. Incorporation of a first-order metabolic rate constant was also required to provide an adequate representation of the data at high concentrations. The analysis of gas uptake data by the use of a PB-PK computer model is a very powerful technique for determining in vivo metabolism of many volatile compounds, but the incorporation of significant deviations from a generally used model structure (i.e., Ramsey-Andersen model) to account for shortcomings of the model's ability to adequately analyze a gas uptake data set should be based on data collection when possible.