We have previously reported segregation analysis of alcoholism in 35 multigenerational families, each ascertained through a pair of male alcoholics by using the mixed model implemented by POINTER. This analysis suggested that liability to alcoholism was, in part, controlled by a major effect with or without additional multifactorial effects. The hypothesis that the major effect was explained by a single genetic locus with strictly mendelian transmission was rejected. The purpose of the present analysis was to use the regressive model implemented by the REGD program from the Statistical Analysis for Genetic Epidemiology computer package (S.A.G.E.) to confirm by replication that a major effect was present in these 35 families. Evidence for the major effect found in Pointer was replicated in the present analysis by using S.A.G.E. Also, we found strong evidence for parental effects that were independent of the major locus transmission from ancestral relatives to children. Mendelian transmission of this major effect was rejected when models incorporated parental effects. When the major effect was calculated adjusting for parental phenotypes, the relative risk of affection for children was about twice as high with affected parents vs. unaffected parents.