Abstract A six compartment physiologically based pharmacokinetic (PB-PK) model was developed to describe the disposition of diethylene-1,4-dioxide (dioxane) and its principal metabolite β-hydroxyethoxyacetic acid in rats, mice, and humans. The model was developed from experimentally measured partition coefficients (reported here for the first time) as well as pharmacokinetic data previously reported. The completed PB-PK model adequately described data from gavage and intravenous studies in rats, as well as inhalation studies in rats and humans. Substantial nonlinearities were observed in the kinetic behavior of dioxane under high exposure conditions (water concentrations greater than 0.1% dioxane and atmospheric concentrations greater than 300 ppm dioxane). The PB-PK model was subsequently used to prepare quantitative estimates of the “plausible upper bounds” on carcinogenic risk for human populations exposed to dioxane in air or water. Based on these quantitative estimates, it appears that human populations continuously exposed to 740–3700 ppb dioxane in air or 20,000–120,000 ppb dioxane in water would be unlikely to experience increased frequencies of tumors.