We studied interrelationships between maternal and neonatal thyroid function, TSH receptor binding inhibiting immunoglobulins (TBII), and dose of thionamide antithyroid drugs in 44 women with active Graves' disease presenting during 46 pregnancies, and their 48 infants. The women were treated with propylthiouracil (PTU) or carbimazole (CBZ). In 30 pregnancies (30 infants) treatment was withdrawn from 3 to 18 weeks before delivery (Group A). Drug treatment (PTU, n = 10, dose 50-400 mg/day or CBZ, n = 6, dose 5-45 mg/day) was continued throughout pregnancy and delivery in 16 pregnancies producing 18 infants (Group B). The maternal TBII at delivery was well correlated with maternal free thyroxine index (FTI) averaged over the third trimester (r = 0.603, P less than 0.001) and umbilical venous serum TBII (r = 0.940, P less than 0.001). Neonatal FTI was independently related to umbilical vein TBII (t = 2.29, P = 0.03) and maternal dose of antithyroid drug (t = -2.21, P = 0.03). Neonatal thyrotoxicosis was seen in all four infants (8% of births) of women whose TBII levels at delivery exceeded 70%. No child was born with a subnormal FTI but 7/18 infants in group B had raised TSH at birth. This was more likely to occur (P = 0.05) if maternal TBII was less than 30% (6/10) than if maternal TBII was greater than 30% (1/8). Four Group B women with FTI in the lower half of the reference range delivered infants with raised TSH compared with 3/14 (21%) women whose FTI was in the upper half of the reference range or above (P = 0.05). In pregnant women with active Graves' disease TBII levels reflect stimulatory TSH receptor antibody activity. TBII measurements are of use in the prediction of neonatal thyrotoxicosis and impaired neonatal thyroid function in infants of women treated with antithyroid drugs.