The epidemiological studies performed thus far have presented only limited evidence for the carcinogenicity of trichloroethylene (TRI) to humans. However, these studies had drawbacks such as insufficient size of cohort, short observation period, and inadequate TRI exposure assessment; therefore, no concrete conclusion has been reached concerning TRI carcinogenicity to humans. Despite the limited epidemiological evidence as to the carcinogenicity of TRI, the International Agency for Research on Cancer (IARC) has changed the carcinogenicity classification of TRI from Group 3 (not classifiable as to carcinogenicity to humans) to Group 2A (probably carcinogenic to humans). In regard to the new classification by the IARC, the committee for occupational exposure limits of the Japan Society for Occupational Health has made a proposal that it is too early to classify the carcinogenicity of TRI as Group 2A and that it is proper to promote exposure control with the carcinogenicity being classified as 2B for the moment. There are species differences in TRI carcinogenicity, particularly between rats and mice. Although experimental studies have found no evidence that TRI induces liver cancer in rats, there is ample evidence that TRI promotes the development of liver cancer in mice, particularly in B6C3F1 mice. The carcinogenicity of TRI in this strain of mice may be based on an epigenetic mechanism rather on a genotoxic mechanism and the liver cancer may be induced only after TRI has been inhaled for a long period of time at concentrations high enough to cause cytotoxicity. Conversely, with no reports showing TRI-induced renal tumors in mice, the possibility has been suggested that this chemical induces such tumors in male rats. The species differences are mainly accounted for by differences in the metabolism of TRI between rats and mice. From a general survey of the literature, it can be concluded that TRI itself is not mutagenic. However, the conjugation of TRI with glutathione (GSH), a minor pathway of TRI metabolism, results in mutagenic metabolites in the kidney of rats. The acute toxicity of TRI is neurotoxicity based on its anesthetic action. An exposure to extremely high levels of TRI may cause the liver and kidney disorders. Repeated exposures to high levels of TRI may result in neuro-, hepato-, and/or nephrotoxicity. The main symptoms appearing after chronic exposure at low levels are neurological changes represented by subjective symptoms relating to central and autonomic nervous systems, or by a lowered conduction velocity of the nerves or a prolonged latency of the nerve responses. For the present, it is reasonable to use the neurological findings for establishing the reference values of TRI for both work and general environments. A value of 25 ppm (135 mg/m3) is proposed as a reference value for work environments, and 25-50 ppb (135-270 micrograms/m3) for the general environment (1/1,000 of the value for work environment).