There is a belief that perinatal factors are a major cause of epilepsy. We studied a community-based sample of twins, a group with a marked excess of adverse perinatal events. The observed number of non-twin siblings with seizures did not differ from that predicted by the age-specific cumulative incidence rate of seizures (4.2% at age 10 years) in the twins. The types of epilepsies in the twins were largely benign and self-limited and not those associated with brain damage. Zygosity, birth order, and birth weight did not predict affected status. Within affected sibships, the frequency of seizures in co-twins of dizygotic probands (9%) was not different from the frequency in non-twin siblings (12%) but was much less than the frequency in co-twins of monozygotic probands (38%; p < 0.001), reflecting a major genetic component to certain epilepsies. These data show that twins do not have an increased risk of seizures and strongly suggest that perinatal factors have little bearing on the etiology of the common epilepsies in the community.