Few prospective studies have reported dietary risk factors for ovarian cancer. A total of 71 histologically confirmed epithelial ovarian cancers occurred among 13,281 non-Hispanic white California Seventh-day Adventist women during follow-up. Participants were part of the Adventist Health Study (AHS) and had no history of cancer or hysterectomy at baseline in 1976 when they completed a detailed lifestyle questionnaire including a dietary assessment. The association of dietary variables with either all ovarian cancer cases or postmenopausal cases was tested using proportional hazards regression with adjustment for age and other covariates. The strongest hazardous risk factor associations among the food variables were found for meat intake with a risk ratio (RR) of 2.42 for intake > or = 1 time/week versus no meat (p for trend = 0.006), and cheese intake with a RR of 2.02 for intake of > 2 time/week versus < 1 per week (p for trend = 0.10), both of these being in postmenopausal cases. We found significantly reduced risk of all ovarian cancer with higher tomato consumption (RR = 0.32) comparing intakes > or = five times/week versus never to < 1 time/week (p for trend = 0.002), and also with higher fruit consumption (p < 0.01). A weak protective association was found with low fat, but not whole milk. Little confounding was observed between these foods.