This case-control study evaluated reproductive and other factors in relation to epithelial ovarian cancer (EOC) risk. Between 1993 and 1995, the authors recruited 655 EOC cases and 3,899 population controls aged 50-74 years who were born in and residents of Sweden. Data were collected through mailed questionnaires. Odds ratios were estimated by unconditional logistic regression. Parity reduced EOC risk (odds ratio = 0.61, 95% confidence interval (CI): 0.46, 0.81) for uniparous compared with nulliparous women. The risk of EOC decreased with incomplete pregnancies, early menopausal age, late age at first birth, and unilateral oophorectomy; increased with family history of ovarian cancer; and was not associated with menarcheal age, lactation, irregular menses, and menopausal symptoms. Histology-specific odds ratios of EOC for ever compared with never users of oral contraceptives were: serous, 0.56 (95% CI: 0.42, 0.74); mucinous, 1.96 (95% CI: 1.04, 3.68); endometrioid, 0.71 (95% CI: 0.49, 1.03); clear cell, 0.66 (95% CI: 0.31, 1.43); and all EOCs, 0.73 (95% CI: 0.59, 0.90). Prolonged oral contraceptive use reduced EOC risk, with persistent protection up to 25 years after the last use. Ever use of hormone replacement therapy increased EOC risk (odds ratio = 1.41, 95% CI: 1.15, 1.72). Among etiologic hypotheses, the retrograde transportation hypothesis accommodates most epidemiologic findings concerning EOC risk.