The protective effect of early menopause shows that ovarian hormones increase the risk of breast cancer: it is likely that this is because they stimulate breast cell division. The mitotic rate of breast cells is higher during the luteal phase of the menstrual cycle than during the follicular phase, suggesting either that progesterone and oestrogen together induce more mitoses than oestrogen alone (the 'oestrogen plus progestagen hypothesis') or that oestrogen alone induces breast cell mitoses in a dose-dependent manner and that progesterone has no effect (the 'oestrogen alone hypothesis'). Both hypotheses are consistent with the known effects of reproductive history, obesity, combined oral contraceptives and oestrogen replacement therapy (ERT) on breast cancer risk, but while the oestrogen alone hypothesis predicts that hormone replacement therapy with oestrogen and a progestagen (HRT) will cause the same increase in risk as ERT, the oestrogen plus progestagen hypothesis predicts that HRT will cause a greater increase in risk than ERT. Both hypotheses suggest that the risk of breast cancer could be reduced by delaying the onset of regular ovulatory menstrual cycles and by minimizing the therapeutic use of oestrogens, and possibly of progestagens, in postmenopausal women. It may be possible to design hormonal contraceptives that will decrease breast cancer risk.