Sibling correlations and segregation analysis were used to examine the familial distribution of age-sex-adjusted measures of nuclear sclerosis in 1,247 individuals from 564 sibships in the Beaver Dam Eye Study. There are highly significant sibling correlations for all sibs, and separately for sister-sister, sister-brother, and brother-brother pairs. Two transformed normal distributions give the best fit to the data. The hypothesis of mendelian transmission of a major effect cannot be rejected, but the hypothesis of a random environmental major effect is rejected. The parameters of the tau AB free model showed close similarity to the values expected under a mendelian hypothesis. Our results suggest that a single major gene can account for 35% of the total variability of age-sex-adjusted measures of nuclear sclerosis.