Previous studies suggest that the greater sensitivity of mice, compared to rats, to the carcinogenicity of 1,3-butadiene (BD) is linked to higher rates of BD metabolism to butadiene diepoxide (BDO2) by mice than rats. The purpose of this study was to determine the tumorigenicity of BDO2 in mice and rats exposed by inhalation to the same concentrations of the agent. Female B6C3F1 mice and Sprague-Dawley rats, 10-11 weeks old, 56/group, were exposed to 0, 2.5, or 5.0 ppm BDO2, 6 h/day, 5 days/week for 6 weeks. At the end of the BDO2 exposure, 8 animals/group were evaluated for toxicity. The remainder of the exposed rats and mice were held for up to 18 months for observation of tumor development. At the end of the exposure, rats had no biologically significant alteration in standard hematological parameters, but mice had a dose-dependent increase in neutrophils and decrease in lymphocytes. Most of the significant lesions in both species were in the nose, concentrated around the main airflow pathway. Necrosis, inflammation, and squamous metaplasia of the nasal mucosa, as well as atrophy of the turbinates, were all present in animals exposed to 5.0 ppm. In mice, necrosis and inflammation subsided within 6 months, but squamous metaplasia remained. In rats that died after exposure, squamous metaplasia was seen in areas of earlier inflammation and extended beyond those areas with time. The metaplasia was severe enough to restrict and occlude the nasopharyngeal duct. Later, keratinizing squamous-cell carcinomas developed from metaplastic foci in rats, but these were not seen in mice. At the end of 18 months, the only significant increase in neoplasia in the exposed rats was a dose-dependent increase in neoplasms of the nasal mucosa (0/47, 12/48, and 21/48 for the control, 2.5 ppm, and 5.0 ppm exposures, respectively). Neoplasia of the nasal mucosa did not increase significantly in the mice. Neoplastic lesions in the mice were observed in reproductive organs, lymph nodes, bone, liver, Harderian gland, pancreas, and lung, but the only significant increase in neoplasms in a single organ in the mice was in the Harderian gland (0/40, 2/42, and 5/36 for the control, 2.5 ppm, and 5.0 ppm exposures, respectively). This tumor accounts for the apparent trend toward an increase in total neoplastic lesions in mice as a function of dose (10/40, 7/42, and 16/36 for control, 2.5 ppm, and 5.0 ppm, respectively). These findings indicate that the metabolite of BD, BDO2, is carcinogenic in the upper respiratory tract of rats. An increase in upper respiratory tract tumors was not observed in similarly exposed mice, despite the fact that preliminary studies indicated mice should have received twice the dose to tissue than did the rats. Higher cytosolic activity of detoxication enzymes has been reported in the liver and lung cells of the mouse compared to the rat, and this may account, in part, for the differences in response. The transport of externally administered BDO2, into the cell and through the cytoplasm, might allow detoxication of the molecule before it reaches critical sites on the DNA. The results indicate that the site of formation of the BDO2 is important for tumor induction.