A predominance of small low-density lipoprotein (LDL) particles, as determined by gradient gel electrophoresis (GGE), has been proposed to be a common genetic marker for risk of coronary heart disease. Previous analyses of the families in the Berkeley data set defined a dichotomous trait based on the LDL size distribution by GGE: subjects with a predominance of small LDL particles have LDL subclass "pattern B," while those with a predominance of large LDL particles have LDL subclass "pattern A." Using this definition, previous complex segregation analyses demonstrated a single major gene effect on pattern B with a dominant or additive mode of inheritance, and reduced penetrance in young males and premenopausal women. In the present analyses, a continuous variable denoted LDL peak particle diameter is used, reflecting the size (diameter in A) of the major LDL subclass, for comparison with the dichotomous LDL subclass pattern classification. After adjusting for age and gender, the results demonstrate a single major Mendelian gene effect, but the exact mode of inheritance could not be established. Thus, analysis of the continuous variable was not superior to that of the dichotomous classification in determining the mode of inheritance of the proposed single gene controlling LDL subclasses.