During pregnancy in goats the concentration of endogenous oestrone sulphate in milk increased more than twofold, and that in arterial and mammary venous plasma 10- and 20-fold respectively. The concentration in milk was higher than that in arterial plasma, particularly in lactating goats during mid-gestation. This was partly related to mammary production of oestrone sulphate (or of a closely related steroid which cross-reacted in the radioimmunoassay) since in tracer infusion studies the specific activity of oestrone sulphate in milk was significantly lower than that in arterial or mammary venous plasma. It was also related to the existence of a mechanism within the gland which concentrates oestrone sulphate in milk since when infused close-arterially into the mammary gland of a non-pregnant goat with undetectable levels of the endogenous compound in the circulation, a concentration ratio of 7.4:1.0 was reached for oestrone sulphate in milk:arterial plasma. Tracer kinetic studies showed that mammary extraction of [3H]oestrone sulphate was variable (up to 41.3 +/- 30.6%, mean +/- S.E.M.). During intravenous or close-arterial infusion, radioactivity in arterial and mammary venous plasma at steady state was mainly in the form of [3H]oestrone sulphate (range, 64 +/- 10.6 to 80.2 +/- 5.9% of total radioactivity in plasma). The remainder was in the form of compounds chromatographically similar to oestradiol-17 beta-3-monosulphate, oestradiol-17 alpha-3-monosulphate and unconjugated oestrogens. The distribution of radioactivity between these different steriods was similar in arterial mammary venous plasma indicating a low level of selective mammary metabolism or extraction. The amount of labelled oestrone sulphate transferred into milk was low, and it was significantly less in pregnant (range, 0.11 +/- 0.07 to 0.27 +/- 0.16% of total infusate) than in non-pregnant animals (3.23 +/- 0.50%). Studies of the rate of transfer of [3H]oestrone sulphate from blood to milk indicated the presence of a transcellular route with peak activity in milk occurring about 110 min after the start of the infusion.