To evaluate the relationship between people with epilepsy and birth order. A case-control study of 336 epileptics, 15 years and above, and their 1961 full siblings. The data was analysed by birth order and then stratified by sibship size. The 95% confidence interval (CI) for each odds ratio (OR) was obtained. ORs were calculated in the 259 probands in whom the seizure and epileptic syndrome were classifiable against their corresponding 1313 siblings. The OR in birth order 1 is 2.08 (1.6-2.8) on comparing probands to their unaffected siblings. In birth order 3 the OR was 1.64 (1.2-2.2) and ORs declined as birth order increased. The chi-square test for the decline was significant P < 0.05. OR in birth >2 in probands against unaffected sibs was 0.42 (0.2-0.62) in partial seizures and 0.27 (0.17-0.43) in the cryptogenic category, 86% of whom had partial seizures. In spite of some limitations in the study it seemed that there is a significant association between low birth order and the risk of epilepsy when all cases were computed together. The cryptogenic type showed the clearest association between low birth order and the likelihood of epilepsy.