To examine relations between the risk of epithelial ovarian cancer and time-related effects of pregnancies, we analyzed data from a prospective study of 1,145,076 women in Norway aged 20 to 56 years. The mean follow-up time per woman was 16.4 years and a total of 1,694 women were diagnosed with epithelial ovarian cancer. Incidence rate ratios (IRR) were estimated by Poisson regression analysis of person-years at risk. The risk of epithelial ovarian cancer decreased with an increasing number of full-term pregnancies (IRR = 0.56, 95 percent confidence interval = 0.48-0.67 for three pregnancies cf one). However, no further reduction in risk was seen after the third pregnancy. The association with parity became weaker with increasing age at last birth. Further, the reduction in risk among parous women compared with nulliparous women was more pronounced shortly after birth. Both higher age at last birth and shorter time since last birth were associated with a reduction in risk, although these relations were seen mainly for the first and second births. Increasing age at first birth was associated with a decrease in risk among uniparous women but not among multiparous women. Our results indicate that the relations between the incidence of epithelial ovarian cancer and reproductive factors are more complex than previously believed.