A case-control study of 229 ovarian cancer cases (including 172 epithelial tumors) and an equal number of population-based controls was conducted during 1984 to 1986 in Shanghai, China, a low-risk area for ovarian cancer. Similar to studies in high-risk areas, the risk of epithelial tumors was high for nulliparous women (odds ratio, 1.6; 95% confidence interval, 0.8 to 3.2) and decreased with increasing number of livebirths (P less than 0.01). Early menarche and late menopause were associated with increased risk, with the trend in risk for age at menarche being statistically significant. In contrast to other studies, oral contraceptive use was not associated with reduced risk, although there was some reduction in risk for those with a prior tubosterilization or intrauterine device use. Risk was also elevated among those reporting a prior ovarian cyst, medroxyprogesterone use, a first degree family history of cancer, and occupational exposure to paint. Risk factors for the nonepithelial tumors were similar to the other cancers, although the power to detect differences was limited.