In a study of the childhood leukemia experience in a California twin birth cohort, there was an apparent deficit of leukemia in twins compared to the occurrence in single births. The rough estimate of the deficit of leukemia in monozygotic twins tends to confirm that found in 2 other studies. The concordance rate for monozygotic twins was estimated to be 25%, which is in agreement with that made by MacMahon and Levy. Relative to the comparison group, the leukemic twins and their unaffected co-twins had a significantly greater disparity in birth weights, and heavy leukemic females made the greatest contribution to this result. More first-born twins died of leukemia than second-born, but an independent twin birth order effect was not shown when birth weight and sex composition were controlled.