Phenotypes that predict coronary heart disease (CHD) are the consequence of interactions between many genetic and environmental factors. Quantitative measures of plasma apolipoproteins, lipoproteins and lipids are examples of phenotypes that link genetic and environmental factors to the CHD end-point. Population studies in Hawaii, Michigan and elsewhere have established that a significant fraction of variability in these phenotypes is attributable to genetic differences among individuals. Recent advances in molecular biology provide measures of the gene loci that code for the apolipoproteins, the cellular receptors for lipoprotein particles and the catalysts and cofactors in lipoprotein metabolism. By measuring polymorphic protein variability and restriction site variability in small regions of DNA known to contain genes that code for the proteins involved in these functions, it is possible to assign polygenetic effects to specific alleles or haplotypes. This 'measured genotype' approach may be used to study the genetic architecture (number of loci involved, the frequencies and effects of their alleles, and the type of loci, i.e., structural or regulatory) of quantitative variation in the plasma apolipoproteins, lipoproteins and lipids. This paper reviews statistical models, sampling designs and results of studies designed to estimate the genetic architecture of selected apolipoproteins, lipoproteins and lipids. The usefulness of these studies for answering questions about the prediction of CHD in the population, the family and the individual are discussed and the directions that human quantitative genetic studies will take in the future are considered.