Events which might have a casual significance for epilepsy were collected from the obstetrical case reports of 100 epileptic children who were born in four hospitals in a large town. 100 healthy children of the same age, who had been born in the same hospitals, served as a control group. The differences were worked out by means of the Chi square test and where small numbers were involved the Yates theory was used. It indicated that some possibilities of damage (e.g. higher age of the mother at birth, toxemia of pregnancy, premature birth, heavy weight at birth) are important either with respect to all patients or to the different types of attacks, whereas other factors (e.g. abnormal positions, instrumental delivery, coiling of the umbilical cord) whose roles are likewise never usually doubted, can be neglected. Exogenous reactions were found most frequently in the case of patients with grand mal and focal attacks, although, as was expected, they were missing in patients with absences, whose main genetic nature is known. It is pointed out that only with extreme caution may the various possibilities of damage be found responsible for epilepsy in children in general or for the individual types of attacks.