A case-control study is presented that estimates ovarian cancer risk for various factors, including diet. Data collected by interview between 1957 and 1965 for 274 white women aged 30-79 years with epithelial carcinoma of the ovary are compared to data similarly collected for 1,034 hospital controls. Relative risk estimates are presented for the total group as well as for premonopausal (ages 30-49) and postmenopausal (ages 50-79) are groups. In the total group, cancer risk increased with increasing age at first marriage (P less than .01) and previous history of benign breast disease (P less than 0.1), and risk decreased with increasing number of previous pregnancies (P less than .01). In the 50- to 79-year age group, a marginally significant trend for decreasing risk with increasing obesity was observed (P less than .10). There was no significant risk (i.e., P less than .10) associated with the consumption of alcohol, cigarettes, coffee, tea, total dietary protein, vitamin C, or fat at any age. In the 30- to 49-year age group only, increased risk (P less than .01) was seen in women reporting diets low in fiber and vitamin A from fruit and vegetable sources. Multiple regression analysis demonstrated that the apparent protective effect of vitamin A in the 30- to 49-year age group (but not dietary fiber) was independent of the nondietary factors analyzed in this study (P less than .05).