In contrast to cancers of non-hormone-dependent sites, cancers of the breast, ovary and endometrium show a slowing down of the rate of increase in incidence at around age 50, as if ceasing menstrual activity reduced risk. Also nulliparous women appear more prone to these three cancers as compared to parous women, thus suggesting that pregnancies also represent a 'protected' time. Epidemiological studies on breast cancer, the only ones sufficiently large to try to disentangle meaningfully the effects of collinear reproductive variables such as parity and ages at first and last birth, show, however, that the effect of pregnancy is not simple and depends on how many births take place and at what age. Larger population-based investigations able to obtain with greater precision information not only on reproductive factors but also on possible confounding variables (e.g. socio-economic status, dietary habits, etc.) are mandatory, particularly as regards ovarian cancer and endometrial cancer. The lesson from the recent studies on pregnancy-related events and breast cancer is, however, that initially a decrease of old certainties must be expected to derive from the accumulation of new, better epidemiological data.