Initiation, distribution, concomitants and follow-up of cholesterol testing were studied in Grampian. Data were examined for 4979 patients, representing all patients in one year from those general practices who made exclusive use of the clinical chemistry laboratory for cholesterol testing. A random sample of 215 patients was studied in further detail. Age and sex distribution, results of cholesterol tests and their follow-up, nature and results of associated biochemical tests, test initiation, testing rates by practice, and prescription rates of lipid lowering agents by practice were measured. Cholesterol testing was mostly in line with current knowledge of cardiovascular risk, and associated with further cardiovascular and biochemical assessment. There was a 90-fold range in practice cholesterol testing rates, and a similarly wide range in prescription rates of lipid lowering agents; there was a significant correlation between these. Rates and results of follow-up testing suggests a 'rule of halves 'for cholesterol testing.