Abstract Limited testing resources, the need to limit animal use, and the demand for better knowledge about carcinogenic hazards require that the carcinogenicity testing paradigm based on lifetime cancer bioassays in rats and mice should be as efficient and reliable as possible. We have therefore reevaluated the rodent bioassay, particularly for nongenotoxic chemicals and conducted a rigorous examination of the 710 substances listed in the Carcinogenic Potency Database (CPDB) that were tested in both mice and rats. The CPDB is a web-based database that provides access to the literature and the results of 6540 bioassays on 1547 chemicals that have been published in the general literature through 2001 and by the National Cancer Institute/National Toxicology Program through 2004. Only three chemicals (o-benzyl-p-chlorophenol, Elmiron®, p-tolylurea) were identified as unequivocally non-genotoxic, mouse non-liver carcinogens. A careful analysis showed that their carcinogenicity in mice is irrelevant for assessment of human cancer hazards. This is consistent with data showing, with a few well-known exceptions, that non-genotoxic carcinogens in rodents are considered to be non-carcinogenic to humans. As a result, we propose that the inclusion of the mouse bioassay in the standard assessment scheme for non-genotoxic chemicals is no longer necessary.