There is insufficient evidence to indicate anything but the weakest, if any, tumor initiating activity for trihalomethanes. Chloroform in all the tests other than the low level of DNA binding did not indicate any possible tumor initiating activity. The only indication that the other trihalomethanes are tumor initiators was their marginal mutagenic activity in the Ames Salmonella Assay. The trihalomethanes would appear to possess tumor promoting activity as indicated by (1) induction of regenerative hyperplasia, (2) molecular markers such as induction of ODC and (3) the apparent weak promoting activity of chloroform in the rat liver foci bioassay. The dose-response relationship of the tumor promoting activity of chloroform could possess a threshold as indicated by the induction of ODC and regenerative hyperplasia. Therefore, the weight of the present evidence would indicate that the extrapolation model employed to estimate the low dose carcinogenic response for chloroform should be different than the model used for genotoxic carcinogens. However, it must be emphasized that confirming evidence is required to support this conclusion.