At least 20 million women in developed countries are estimated to be currently using hormone replacement therapy (HRT). Almost 100 epidemiological studies have reported on the relationship between the use of HRT and the risk of cancer of female reproductive organs, namely the breast, uterus or ovary. Cancer at these sites is common and there are a priori reasons why the use of hormonal therapy to 'replace' the endogenous production of ovarian hormones after the menopause might increase the risk of these cancers. The available evidence indicates that the risk of breast cancer or endometrial cancer is increased while women are using HRT, the risk increasing with increasing duration of use. Most of the evidence about these cancers relates to use of HRT preparations containing oestrogens alone. The limited evidence about combination therapy, with oestrogens and progestogens, suggests that, compared to oestrogens alone, the effect on the breast is similar, but the effect on the endometrium is diminished, the diminution in risk being greater the more days each month that progestogens are used. The effect of HRT on breast cancer wears off after use ceases and has disappeared largely, if not wholly, within 5 years, whereas the effects on endometrial cancer take longer to wear off, if at all. The breast and endometrial cancers that are diagnosed in HRT users are less aggressive clinically than cancers in never-users but, as yet, there is little reliable information about the relationship between use of HRT and mortality from these cancers. For other cancer sites, the existing data about the effects of HRT are inconclusive. The longer the period of use of HRT, the greater the excess incidence of cancer of the breast and endometrium is likely to be. Use of HRT for short periods of time should have little effect on the incidence of these cancers. The cumulative excess incidence in 1000 women who used HRT for 10 years, beginning at age 50, is estimated to be six for breast cancer, 42 for endometrial cancer in women with an intact uterus using oestrogen therapy alone and about 20 for endometrial cancer in women with an intact uterus using oestrogen-progestogen combinations. The estimate for combined therapy is based on small numbers and may well vary with the type of preparation used. The overall balance between the excess incidence of these cancers and other effects of HRT needs to be evaluated carefully and will require more reliable data than exist at present.