This industry-wide epidemiologic study to evaluate the human carcinogenicity of the chloromethyl ethers (CME) included 1827 CME-exposed workers and 8870 controls. Duration and relative intensity of exposure were classified by job description in the personnel records. This information permitted assigning relative magnitude of exposure scores for each job category at several plants, with allowance for temporal changes in the plant processes. Social security records were used for the ascertainment of deaths among separated employees. Death certificates were obtained for virtually all known deaths, and hospital pathology reports were obtained where possible for the cancer-related deaths. No differences in noncancer death rates were found. An increased risk of respiratory cancer death in CME-exposed workers was found at only one firm where high exposures are known to have occurred. A clear dose-response relationship with risk ratios exceeding ten for the longest duration and greatest exposure subgroups was demonstrated for this firm.