Immunoactive inhibin (ir-inhibin) concentrations in maternal serum during normal human pregnancy have been established in two separate studies employing cross-sectional and longitudinal sampling regimes. Ir-inhibin concentrations rose from the mid-luteal phase (geometric mean + 95% confidence intervals 1.490 (1.086.2.028) U mL-1) to peak at week 11 of gestation (3.77 (3.26-4.35) U mL-1), declined to a plateau from 14 to 25 weeks with means ranging from 1.8 to 2.3 U mL-1, and subsequently rose slowly to a peak of 6-53 U mL-1 at 41 weeks. In the longitudinal study, similar results were obtained and no differences were found in maternal inhibin levels in women carrying male or female fetuses. Paired cord blood and maternal samples showed no significant difference in ir-inhibin concentrations irrespective of the sex of the fetus. However, in all such pregnancies amniotic fluid ir-inhibin levels were 2-3 fold greater than maternal or fetal levels raising the possibility that the amnion may secrete inhibin. In 12 women without functional ovaries in whom a singleton pregnancy was achieved by donation of oocytes and in vitro fertilization, the ir-inhibin levels showed a similar pattern in the first trimester of pregnancy but the concentrations achieved were markedly lower (peak 1.1 U mL-1 at 9 weeks). In five women from the group in whom samples were available late in gestation, three showed greater than normal levels and two had subnormal levels. These results indicate that although a first trimester rise of ir-inhibin can occur in pregnancies in women without ovaries, there appears to be a significant component due to ovarian secretion. Furthermore, the third trimester rise in ir-inhibin levels can occur in the absence of the ovary.