Background: It has been hypothesized that physical activity may reduce the risk of ovarian cancer by decreasing estrogen levels, reducing body fat, and reducing the frequency of ovulation. Epidemiologic studies of this relationship have obtained inconsistent results. The only prospective study to date reported a positive association between frequent vigorous exercise and ovarian cancer risk. We further evaluated this relationship in the Nurses' Health Study cohort. Methods: Participation in recreational physical activity was assessed by questionnaire in 1980, 1982, 1986, 1988, 1992, and 1994, with questions assessing exercise frequency, duration, and intensity. Results were adjusted for age, parity, oral contraceptive use, tubal ligation, and other risk factors for ovarian cancer. All statistical tests were two-sided. Results: During a 16-year follow-up (from 1980 to 1996), 1.2 million person-years were accrued by 92825 cohort members, and 377 cases of epithelial ovarian cancer were confirmed. The relative risk (RR) of ovarian cancer for women engaging in recreational physical activity for 7 hours or more per week compared with those reporting less than 1 hour per week was 0.80 (95% confidence interval [CI] = 0.49 to 1.32; Ptrend = .59). When both the frequency and intensity of activity were taken into account, activity level was also not associated with a reduced risk of ovarian cancer. Compared with inactive women, participants reporting high activity in terms of metabolic equivalent task hours (MET hours) were at greater risk of ovarian cancer (RR for 20 to <30 MET hours/week = 1.84 [95% CI = 1.12 to 3.02]; RR for >30 MET hours/week = 1.27 [95% CI = 0.75 to 2.14]). Conclusions: Overall, results did not suggest an inverse association between recreational physical activity and ovarian cancer. The possibility of a modest increase in risk with frequent vigorous activity requires further investigation.