Summary Sex hormone binding globulin (SHBG) concentrations were measured by immunoradiometric assay in serum samples from 1,221 healthy female volunteers aged 34-79 who had never used oral contraceptives or hormone replacement therapy, had no history of thyroid disease or cancer, and had not used any drugs known to influence SHBG in the 14 days preceding blood sampling. There were 616 premenopausal and 605 naturally postmenopausal women. In both premenopausal and postmenopausal women, SHBG decreased with increasing weight (Quetelet's Index) and was lower in single nulliparous women than in married nulliparous women or parous women. In premenopausal women, SHBG was higher in women with late menarche, was higher in smokers than in non-smokers, and was higher in blood samples taken during the first 12 days of the luteal phase than during the rest of the menstrual cycle. In postmenopausal women, SHBG increased with years since the menopause. The possible biological importance of these findings is discussed with particular reference to risk factors for breast cancer.There is evidence that the risk of developing breast cancer is enhanced by increased exposure of breast epithelial cells to oestradiol (E2) (Henderson et al., 1982). The amount of E2 which reaches breast cells is determined by the concentration of E2 in the plasma and by the proportion of this E2 which is able to leave the plasma and enter the cells. About 98% of plasma E2 is bound either to albumin or to sex hormone binding globulin (SHBG). Anderson (1974) argued that only the small percentage of E2 which is not bound to albumin or to SHBG (i.e. the free E2 fraction) is able to cross cell membranes, but there is some evidence that E2 bound to albumin may also be biologically active (Pardridge, 1981). SHBG concentration is a major determinant of the proportions of E2 bound to SHBG, bound to albumin, and free (Siiteri et al., 1981). Severaf studies have included measurements of E2 binding and of the concentration (or binding capacity) of SHBG in breast cancer cases and controls. In most of these studies, and in one prospective study, women with breast cancer and women who were later diagnosed with the disease were found to have a higher percentage of free E2 than controls. This difference was often accompanied by a low SHBG concentration or SHBG binding capacity, and by an increase in the percentage of albumin bound E2 (see Moore et al., 1986).The mechanism of action of most of the known breast cancer risk factors (in particular, the reproductive and menstrual risk factors) is poorly understood. Bernstein et al. (1985) reported that SHBG binding capacity is higher in parous than in nulliparous women, and suggested that this difference in SHBG might be one mechanism by which parity reduces the risk of breast cancer. It is possible that other risk factors for breast cancer may also modify risk because they are related to SHBG. To investigate this we have measured the SHBG concentration in serum samples from almost 5,000 women living on the island of Guernsey...