A knowledge of birth order can be contributory to the evaluation of hereditary and environmental factors in disease. There is, however, no agreement as to whether a characteristic birth order pattern exists in epilepsy. Thus Brain (1956) and Lennox (1960) (1950) and Orr and Risch (1953) by their use of sibs of the reference epileptics as controls but here bias was possible to the extent that parents who have had an epileptic child might limit family size afterwards, causing such an epileptic to hold a relatively late birth order position in the family.
METHOD OF ENQUIRYWith all these considerations in mind, the following method of enquiry was adopted. From the hospital records all patients diagnosed as epileptics were collected and follow-up enquiries on them conducted during 1957 and 1961. There were then selected for further investigation those who were either first born with a second born sib control, or second born with a first born sib control, and had suffered 10 or more 'convulsion days', a 'convulsion day' being defined as any day in which the patient had one or more convulsions. The first stipulation secured the advantages of Alstro.n's enquiry but placed bias on the epileptics of the enquiry series being second rather than first born, and the second stipulation excluded convulsions which were exclusively febrile and also slight, infrequent convulsions, so minimizing bias which might arise from the possible tendency of parents to be more worried by illnesses in first children, that is, the standard of severity adopted was such that the affected children were likely to have been referred for hospital investigation whatever their birth order. It remained to arrange that the sibs of each pair be observed for 59