Plasma cholesterol was measured in the fifth decade of the life of 249 pairs of monozygotic (MZ) and 262 pairs of dizygotic (DZ) World War II veteran twins and 70% of the same cohort 10 years later. There were no significant differences between the mean cholesterol values for MZ and DZ twins, and the within DZ pair mean squares were significantly larger than the within MZ pair mean squares for all of the cholesterol variables measured. However, the DZ twins were found to have greater total variance, positive skewness, and leptokurtosis than the MZ twins for total and high-density lipoprotein cholesterol, and the total/high-density ratio. Comparisons with published data revealed that the variance of DZ twins was similar to that of singletons while the MZ twins have smaller total variance, perhaps owing to a missing component of variation. Hypotheses for the source of the differences in the zygosity distributions are proposed including environmental influences (pre- or post-natal and within- or among-families), genetic differences, and selection at the time of induction into the armed services. Because of the differences in total variance of the two zygosities it is difficult to know which estimates of genetic variance or heritability have the least bias. However, these data provide clues that may lead to further understanding of sources of plasma cholesterol variation that could be important to the future understanding of risk for coronary heart disease.