Transmission models for high density lipoprotein (HDL) were evaluated in the Berkeley data set through segregation analyses using S.A.G.E. These preliminary analyses indicate that among the models fitted, an additive model with non-Mendelian transmission probabilities provides a good fit for HDL, suggesting the possibility of a significant environmental component in the transmission of HDL. After adjustment for triglyceride, however, Mendelian models for a major gene seem to provide as good a fit as non-Mendelian models but the results do not permit distinction between a dominant model and an additive model.