In a prospective study, 121 children of mothers with epilepsy (study group) and 105 control children were examined in a blinded fashion at age 5 1/2 years for 80 minor physical anomalies, including nine typical features previously reported characteristic of fetal hydantoin syndrome. Of the study group, 106 children had been exposed to antiepileptic drugs (82 to phenytoin) during pregnancy; 44 (36%) mothers had had generalized convulsions during pregnancy. One hundred fourteen mothers and 87 fathers of study group children and 101 mothers and 58 fathers of control children were also examined. A significant excess of minor anomalies considered characteristic of hydantoin syndrome was observed in children of epileptic mothers and in epileptic mothers, compared with the control group. There was no excess of other minor anomalies studied. Several minor anomalies previously regarded as typical of fetal hydantoin syndrome were shown to be genetically linked to epilepsy. Only hypertelorism and digital hypoplasia were associated with phenytoin exposure. The current concept of the syndrome seems to be incorrect; most of the "typical" characteristics are not caused by phenytoin. None of the phenytoin-exposed children had all of the main characteristics of hydantoin syndrome (typical acrofacial features, intellectual deficiency, growth retardation, and microcephaly). The risk of developmental disturbance associated with intrauterine phenytoin exposure seems to be much lower than the 7% to 11% risk of fetal hydantoin syndrome reported earlier.