Our understanding of the hormonal control of the proliferation of normal human breast epithelium is still surprisingly meager. However, the results of a number of recent studies have confirmed that estrogen is the major steroid mitogen for the luminal epithelial cell population (the usual targets for neoplastic transformation). Estrogen seemingly exerts its effects on cell division indirectly as there is complete dissociation between the population of luminal epithelial cells expressing the estrogen receptor (ER)4 and those that proliferate. We suggest that the ER-negative proliferating cells represent a precursor or stem cell population that differentiates to ER-containing, nonproliferative cells. In turn, these ER-positive cells act as 'estrogen sensors' and transmit positive or negative paracrine growth signals to the precursor cells depending on the prevailing hormonal environment. As yet there is no direct evidence supporting this hypothesis but we suggest ways in which it may be obtained. The implication of these studies is that inhibition of luminal epithelial proliferation with tamoxifen or pure antiestrogens or by preventing ovarian steroid secretion should be an effective strategy for the prevention of breast cancer. In addition, we may be able to predict the risk of breast cancer in an individual by measuring the intrinsic estrogen sensitivity of her breast epithelium. Finally, study of the paracrine mechanisms of growth control in the normal human breast may provide new, more specific, therapeutic targets for breast cancer prevention.