Reversibility of forestomach lesions induced by genotoxic and non-genotoxic carcinogens was compared histopathologically. Groups of 30 to 33 male F344 rats were given dietary 0.1% 8-nitroquinoline, dietary 0.4-0.2% 2-(2-furyl)-3-(5-nitro-2-furyl)acrylamide, an intragastric dose of 20 mg/kg body weight N-methyl-N'-nitro-N-nitrosoguanidine once a week, or 20 ppm N-methylnitrosourethane in the drinking water as a genotoxic carcinogen, or 2% butylated hydroxyanisole, 2% caffeic acid, 2% sesamol or 2% 4-methoxyphenol in the diet as a non-genotoxic carcinogen for 24 weeks. Ten or 11 rats in each group were killed at week 24. Half of the remainder were maintained on basal diet alone for an additional 24 weeks and the other half were given the same chemical for 48 weeks, and then killed. Forestomach lesions induced by genotoxic carcinogens did not regress after removal of carcinogens. In contrast, simple or papillary hyperplasia (SPH), but not basal cell hyperplasia (BCH), induced by non-genotoxic carcinogens clearly regressed after cessation of insult. SPH labeling indices in the non-genotoxic carcinogen-treated cases decreased after removal of the carcinogenic stimulus whereas BCH values were low irrespective of treatment. Atypical hyperplasia (AH), observed at high incidences in rats treated with genotoxic carcinogens, was also evident in animals receiving non-genotoxic agents, even after their withdrawal, albeit at low incidences. AH labeling indices remained high even without continued insult. These results indicate that even with non-genotoxic carcinogens, heritable alterations at the DNA level could occur during strong cell proliferation and result in AH development. This putative preneoplastic lesion might then progress to produce carcinomas.