Epithelial ovarian cancer is fairly common with high rates in Scandinavia, intermediate rates in western Europe and North America and low rates in the developing countries and in Japan. The 5-year survival rate is less than 40%. Increasing parity consistently gives a strong protection against epithelial ovarian cancer. A lesser degree of protection is probably derived from incomplete pregnancies and lactation. Ages at menarche and menopause are most probably weak predictors of epithelial ovarian cancer risk. Ever users of oral contraceptives (OC) have 30% lower risk compared to never users. The protection increases with duration of OC use, being about 50% after 5 years. The reduced risk among past OC users persists for at least 10 years after cessation of use. Results concerning hormone replacement therapy (HRT) and epithelial ovarian cancer risk are conflicting, but most data point to a weak or no association, but as an increasing number of women use HRT it still seems important to resolve any potential effect. Infertility adds to epithelial ovarian cancer risk in nulliparous women, while temporary fertility problems in parous women do not appear to increase risk. A possible independent risk effect of fertility drug use has not been easy to assess and remains unresolved. It has been particularly difficult to separate the effects of fertility drugs from those of infertility. Tubal ligation and hysterectomy convey protection against epithelial ovarian cancer, possibly through a suppressed ovarian hormone production. The causes of epithelial ovarian cancer are poorly understood, but reproductive hormones are thought to be involved in the aetiology. For a long time the 'incessant' and 'gonadotrophin' hypotheses have been promoted in relation to carcinogenesis. Both hypotheses find support in ovarian cancer epidemiology, and recent progress in molecular biology adds to the understanding of possible aetiological mechanisms. Another hypothesis focuses on the retrograde transport of contaminants or carcinogens through the Fallopian tubes. It is important to establish if the same risk factors apply to the various histological types of ovarian cancer, as particularly the mucinous ovarian tumours seem to present with different risk factors. Another question to resolve is if sporadic vs. inherited cases carry distinct risk profiles. As the hypotheses above do not explain all of the results derived from ovarian cancer epidemiology, there is a need to test additional hypotheses to possibly define preventive programmes and to come closer to the cause of ovarian cancer.