Although there is widespread acceptance that total cholesterol (TC) value reference ranges should be based on epidemiological rather than statistical considerations, the epidemiological action limits for low-density lipoprotein cholesterol (LDL-C) are still incomplete and only statistical reference ranges for apolipoprotein B (Apo-B) levels are available. The combined use of epidemiological reference ranges for TC and incomplete or statistical reference ranges for LDL-C and Apo-B is illogical, since these parameters may fall into discordant risk categories that will hamper and complicate the management of hypercholesterolaemia. Based on a study of lipograms obtained from +/- 3,000 inhabitants of two industrialised Transvaal towns, the age-related epidemiological reference ranges for LDL-C and Apo-B were established. A comparison with published observational studies of other populations, in which comparable lipid, lipoprotein and apolipoprotein methodologies were used, reflected the severity of these lipid-related abnormalities in white South Africans, especially after the age of 30 years. In addition, the serum TC values found in this survey were not significantly different from those obtained 10 years ago.