The results of two studies, one in Finland and one in the U.S.A., raise the possibility that fetal damage previously attributed to phenytoin and other anticonvulsant drugs, principally phenobarbitone, may be due to epilepsy itself. In the U.S.A., drug-exposure information was collected before delivery in a cohort of 50 282 mother/child pairs. The total malformation rate in 305 children born to epileptic mothers was 10.5%, as against 6.4% in the remainder (p less than 0.01); corresponding rates for major malformations were 6.6% and 2.7%. When the fathers had epilepsy, the malformation-rates in their children were intermediate. The rates did not vary significantly according to maternal anticonvulsants therapy. Mental and motor scores as 8 months of age, and intelligence quotient scores at 4 years were lower in children of epileptic mothers, but not in children of epileptic fathers. The scores showed only random variation according to maternal anticonvulsant therapy. In Finland, 2784 children with craniofacial anomalies were compared with an equal number of normal children; 8 and 2 mothers, respectively, received anticonvulsants, while pregnant, for epilepsy. In that study, the separate effects of the disease and its treatmet could not be evaluated. Both studies did not find evidence of fetal damage when phenobarbitone was taken for indications other than epilepsy.