Epidemiologic evidence on the relation between reactive chemicals and cancer is reviewed. These highly reactive chemicals (acrylonitrile; bis[chloromethyl]ether and chloromethyl methyl ether; 1,3-butadiene, ethylene oxide; formaldehyde; mustard gas; sulfuric acid; and vinyl chloride) vary in use and exposure. All are animal carcinogens that also have received considerable epidemiologic attention. Acrylonitrile is a chemical of current economic importance. The epidemiologic evidence is quite weak, but the available studies were very small. Epidemiologic studies clearly demonstrate that bis (chloromethyl) ether and chloromethyl methyl ether cause lung cancer. Continued follow-up of exposed workers is encouraged to provide information on risks for other cancers. Results from epidemiologic studies of butadiene-exposed workers are somewhat inconsistent, but the largest study with the best exposure assessment found the largest relative risk for leukemia. The failure of several larger studies to replicate the early Swedish findings of a very strong association between leukemia and ethylene oxide has not been adequately explained. Epidemiologic studies of formaldehyde provide limited evidence for an association with cancer of the nasopharynx and possibly with nasal cancer. These very rare tumors, however, are difficult to study epidemiologically. Mustard gas is a well-established lung carcinogen, but a recent follow-up of the English cohort suggests that other sites also may be affected. Sulfuric acid appears to cause laryngeal cancer. A suggested relationship with lung cancer in a few studies is of concern because of the widespread opportunity for exposure from ambient air pollution. Vinyl chloride causes angiosarcoma of the liver, but a large, multi-country study provided no clear evidence that other sites are affected.