The oxygenate tert-amyl methyl ether (TAME) is a gasoline fuel additive used to reduce carbon monoxide in automobile emissions. To evaluate the relative health risk of TAME as a gasoline additive, information is needed on its pharmacokinetics and toxicity. The objective of this study was to use a physiologically-based pharmacokinetic (PBPK) model to describe the disposition of TAME and its major metabolite, tert-amyl alcohol (TAA), in male Fischer-344 rats. The model compartments for TAME and TAA were flow-limited. The TAME physiological model had 6 compartments: lung, liver, rapidly perfused tissues, slowly perfused tissues, fat, and kidney. The TAA model had 3 compartments: lung, liver, and total-body water. The 2 models were linked through metabolism of TAME to TAA in the liver. Model simulations were compared with data on blood concentrations of TAME and TAA taken from male Fischer-344 rats during and after a 6-hour inhalation exposure to 2500, 500, or 100 ppm TAME. The PBPK model predicted TAME pharmacokinetics when 2 saturable pathways for TAME oxidation were included. The TAA model, which included pathways for oxidation and glucuronide conjugation of TAA, underpredicted the experimental data collected at later times postexposure. To account for biological processes occurring during this time, three hypotheses were developed: nonspecific binding of TAA, diffusion-limited transport of TAA, and enterohepatic circulation of TAA glucuronide. These hypotheses were tested using three different model structures. Visual inspection and statistical evaluation involving maximum likelihood techniques indicated that the model incorporating nonspecific binding of TAA provided the best fit to the data. A correct model structure, based upon experimental data, statistical analyses, and biological interpretation, will allow a more accurate extrapolation to humans and, consequently, a greater understanding of human risk from exposure to TAME.