There have been few mortality studies conducted on working women despite their increasing presence in the work environment. This paper presents the findings of a cohort mortality study of 10,839 female employees with three or more days of service between 1940 and 1982 at the Midland or Bay City, Michigan, locations of the Dow Chemical Company. The cohort included substantial numbers of production and research personnel but was primarily comprised of clerical and office workers. Vital status was ascertained through 1982 for 89.4 percent of the cohort members, and death certificates were obtained for 94.0 percent of the 467 decedents. Comparisons of observed mortality with expected levels based on mortality rates for the U.S., Michigan, and a local 7-county area revealed consistently lower mortality in the cohort from the major causes of death, indicating a strong "healthy worker effect." Mortality from cancer of the cervix was significantly below expected levels, especially among women who were hired before 25 years of age. This observation, when combined with a nonsignificant excess of breast cancer, suggests a different distribution of maritally and sexually related risk factors between working and general populations of women. This and other findings are discussed relative to methodologic problems likely to accompany studies of working women.