Many studies have shown that cultured rat liver epithelial cells can be neoplastically transformed by repeated or long-continued exposure to chemical carcinogens. These cells also may transform spontaneously in the absence of carcinogen treatment after long-term, continuous passage in culture or after chronic maintenance in a confluent state in vitro. In this study, we have compared the times of emergence and rates of accumulation of transformed cells in populations of rat hepatic epithelial cells exposed either to a single dose of N-methyl-N'-nitro-N-nitrosoguanidine (MNNG, 3 micrograms/ml culture medium for 30 minutes) or to acetone vehicle alone (3 microliters/ml culture medium for 30 minutes). Transformation was compared in cell populations that were passaged continuously once a week as they attained a confluent density (nonselective growth conditions), or that were maintained at a confluent density for 3 weeks between passages once a month (selective growth conditions). Emergence of both spontaneous transformants and transformants induced by MNNG was facilitated by selective growth conditions, as compared with non-selective growth conditions. Transformants were detected both in cultures exposed to MNNG and in cultures exposed only to acetone (solvent controls), but they always emerged earlier in cultures exposed to MNNG (nine population doublings earlier when grown under selective growth conditions and 22 population doublings earlier when grown under nonselective growth conditions). Once transformants were detected, they replaced the nontransformed population more quickly under selective than under nonselective conditions of culture. Cells possessing the ability to grow in soft agar and to produce tumors in syngeneic rats were detected (at about 12 population doublings after treatment) under selective conditions much earlier than under nonselective growth conditions (at about 90 population doublings after treatment). Among MNNG-treated cultures, the fraction of aneuploid cells in the population was correlated significantly with tumorigenicity. In contrast, among acetone-treated control populations, aneuploidy and tumorigenicity were not correlated; populations of aneuploid acetone-treated cells often were not tumorigenic. These observations suggest that MNNG treatment produced a specific type of aneuploidy that was associated with tumorigenicity.