We investigated the association of epithelial ovarian cancer with physical activity, waist-to-hip ratio, reproductive factors, and family history of cancer in a prospective cohort study of 31,396 postmenopausal women. Ninety-seven women developed incident epithelial ovarian cancer over 7 years. The number of livebirths was associated with lower risk (multivariate-adjusted relative risks for 1-2, 3-4, and > 4 livebirths compared with nulliparity were 0.64, 0.47, and 0.43, respectively). A family history of ovarian cancer in a first-degree relative was associated with a 2.5 times greater risk (95% confidence interval = 0.90-6.7). Multivariate-adjusted relative risks for the upper three quartiles of waist-to-hip ratio compared with the lowest quartile were 2.0, 1.6, and 2.3, respectively. Women with "moderate" and "high" levels of physical activity compared with those with "low" physical activity had relative risks of 1.4 and 2.1, respectively. Positive associations of physical activity and waist-to-hip ratio with ovarian cancer seem inconsistent with existing theories of ovarian cancer pathogenesis.