The basis for the apparent teratogenicity of maternal epilepsy is controversial. Is the critical factor the anticonvulsant drugs taken by the pregnant woman, or the genes which cause the mother's epilepsy? We describe a study design developed to assess these competing theories in a cohort study of newborn infants. We show the feasibility of ascertaining exposed and unexposed infants at several birthing hospitals in one urban area. Between 1986 and 1988, we identified 180 drug-exposed and 218 epilepsy-history infants among 49,403 infants. The rate of exposure to seizure medication was 0.36% and of maternal history of epilepsy was 0.44%. A significant number of infants could not be evaluated because they were missed, ineligible, or either the doctor, nurse or parent refused to participate. Overall, there was a significant increase in major malformations, microcephaly or growth retardation among the drug-exposed infants in comparison to both the epilepsy-history and the unexposed infants. The types of epilepsy and the apparent etiology were the same among women who took anticonvulsants and women with a history of epilepsy but no anti-convulsants during pregnancy. This study must be extended to include a sufficient number of infants exposed to each commonly used drug as monotherapy to allow for a comparison of the effect of each drug on pregnancy outcomes; to provide a comparison of infants whose mothers had a strong family history of epilepsy with infants whose mothers had trauma-induced epilepsy; and to assess the possible impact of the severity of the mothers' disease on the infants.