Steady-state plasma antiepileptic drug (AED) concentrations were measured at intervals throughout pregnancy and during the postnatal period in 105 women who underwent 134 pregnancies. Phenytoin (PHT) dosage had to be increased in 85% of pregnancies in which the drug was received, carbamazepine (CBZ) dosage in 70%, and phenobarbital (PB) or methylphenobarbital (MPB) dosage in 85%, in an attempt to prevent or correct a fall in plasma concentrations of the respective drugs as pregnancy progressed. The altered disposition of the AEDs usually began in the first 10 weeks of pregnancy (often before epileptic pregnant women are referred for neurological supervision), and had returned to baseline value within 4 weeks of childbirth in two thirds of the women receiving PHT. The return to the nonpregnant situation appeared to be slower for CBZ, PB, and MPB. In women studied during more than one pregnancy, the changes in AED dosage to plasma concentration ratios tended to be greater in the first than in the subsequent pregnancies. Full seizure control prior to pregnancy was associated with a more favorable outcome for freedom from seizures during pregnancy. However, the plasma level monitoring-dosage adjustment policy produced no marked improvement in overall seizure control in pregnancy. This may have occurred because some patients were seen too late in their pregnancies for the policy to have been applied optimally.