Through the work of the committee for reference methods in NORDKEM and the Scandinavian Society for Clinical Chemistry, a hierarchy of methods was established in Denmark in 1982 for accuracy assurance of determination of total cholesterol in serum. The hierarchy makes it possible for any laboratory to relate their results of measurements on fresh samples from humans, to the results obtained on the same samples by reference methods, which have been adjusted to definitive methodology. Proficiency studies in 1985 and 1988 gave consensus values (mean of all participants data) for total cholesterol in frozen human serum almost identical to values obtained by reference methods. Interlaboratory coefficients of variations were 4-5%. It is essential for the clinical chemist as well as for the clinician to be aware of the relative quantitative importance of preanalytical, analytical and biological variation and to realize to what extent replicate measurements on the same sample and on several samples from the same person reduce the total intraperson variation of serum cholesterol concentrations.