In cancer tests with 1,3-butadiene (BD), the mouse is much more sensitive than the rat. This is considered to be related to the metabolism of BD to the epoxide metabolites, 1,2-epoxy-3-butene (EB), 1,2:3,4-diepoxybutane, and 1,2-epoxy-3,4-butanediol. This study evaluates whether the large difference in outcome in cancer tests with BD could be predicted quantitatively on the basis of the concentration over time in blood (AUC) of the epoxide metabolites, their mutagenic potency, and a multiplicative cancer risk model, which has earlier been used for ionizing radiation. Published data on hemoglobin adduct levels from inhalation experiments with BD were used for the estimation of the AUC of the epoxide metabolites in the cancer tests. The estimated AUC of the epoxides were then weighed together to a total genotoxic dose, by using the relative genotoxic potency of the respective epoxide inferred from in vitro hprt mutation assays using EB as standard. The tumor incidences predicted with the risk model on the basis of the total genotoxic dose correlated well with the earlier observed tumor incidences in the cancer tests. The total genotoxic dose that leads to a doubling of the tumor incidences was estimated to be the same in both species, 9 to 10 mmol/LÂh EB-equivalents. The study validates the applicability of the multiplicative cancer risk model to genotoxic chemicals. Furthermore, according to this evaluation, different epoxide metabolites are predominating cancer-initiating agents in the cancer tests with BD, the diepoxide in the mouse, and the monoepoxides in the rat.