Specific cognitive abilities and motor function were investigated at 5.5 years in 104 children with epileptic mothers and in 105 control children, all with normal general intelligence. The majority (89 per cent) of the children of epileptic mothers had been exposed to anti-epileptic drugs during pregnancy, most commonly phenytoin (69 per cent). Maternal seizures had occurred during pregnancy in 52 per cent. A significant difference, with poorer performance in the study group, was found in block design (WPPSI) and auditory closure (ITPA). Significantly more study than control children had some type of specific cognitive dysfunction. Within the study group, increased risk was associated with maternal partial seizures, with seizures occurring during pregnancy, and with low paternal education, but not with exposure to anti-epileptic drugs. Three possible mechanisms of this effect are suggested: subtle brain-damage associated with fetal asphyxia during the mothers' generalized convulsions; genetically transmitted brain abnormalities; and psychosocial disadvantage limiting partner choice.