Risk assessment methodology, particularly pertaining to potential human carcinogenic risks from exposures to environmental chemicals, is undergoing intense scrutiny from scientists, regulators, and legislators. The current practice of estimating human cancer risk is based almost exclusively on extrapolating the results of chronic, high-dose studies in rodents to estimate potential risk in humans. However, many scientists are questioning whether the logic used in this current risk assessment methodology is the best way to safeguard public health. A major tool of human cancer risk assessment is the linearized multistage (LMS) model. The LMS model has been identified as an aspect of risk assessment that could be improved. One way to facilitate this improvement is by developing a way to incorporate a carefully derived, more biologically relevant mechanism of action data on carcinogenesis. Recent data on chloroform indicate that the dose-response relationship for chloroform-induced tumors in rats and mice is nonlinear, based upon events secondary to cell necrosis and subsequent regeneration as the likely mode of action for the carcinogenic effects of chloroform. In light of these data, there is a sound scientific basis for removing some of the uncertainty that accompanies current cancer risk assessments of chloroform. The following points summarize the critical data: (1) a substantial body of data demonstrates a lack of direct in vivo or in vitro genotoxicity of chloroform; (2) chloroform induces liver and kidney tumors in long-term rodent cancer bioassays only at doses that induce frank toxicity at these target sites; (3) the chloroform doses required to produce tumors in susceptible species exceed the MTD, often by a considerable margin; (4) cytotoxicity and compensatory cell proliferation are associated with the chloroform doses required to induce liver or kidney tumors in susceptible rodent species; (5) there are no instances of chloroform-induced tumors that are not preceded by this pattern of dose-dependent toxic responses; (6) it is biologically plausible that cytolethality leads to chronically stimulated cell proliferation and related events such as inflammation and growth stimulation which act to initiate and promote the carcinogenic process; and (7) the consistently linked cellular events of cytolethality and subsequent cell proliferation, for which doses of no adverse effect have been clearly shown to exist, are one of the biological prerequisites for chloroform-induced tumors in animals. Based on these data, it is inappropriate to extrapolate cancer risk from high doses that produce necrosis and regenerative cell proliferation to low doses that do not with a model that presumes genotoxicity and a linear dose-response relationship. The weight of the scientific evidence concerning chloroform-induced tumors in rodents is consistent with and supports a cancer risk assessment methodology based on mode of action as the basis for establishing regulatory standards for this compound.