We examined the relation between epilepsy and birth order, using data on 1,950 probands with epilepsy and 4,636 of their full siblings without epilepsy from the Epilepsy Family Study of Columbia University. The proportion of first-born individuals appeared to be higher among probands with epilepsy than among their unaffected siblings, but this relation disappeared after we controlled for the confounding effect of sibship size. With sibship size controlled, the proportion of first-born individuals was similar to that in unaffected siblings for probands with idiopathic/cryptogenic epilepsy, generalized and partial onset seizures, and all ages at onset of epilepsy. Probands with remote symptomatic epilepsy had higher birth orders than their unaffected siblings, even after we controlled for sibship size.