Analytical and intra- and inter-individual components of variation were assessed over a 40-week period in 15 apparently healthy subjects, seven men and eight women, for serum creatinine, urine creatinine expressed in both concentration and output terms, and creatinine clearance, both uncorrected and corrected to standard surface area of 1.73 m2. Serum creatinine, even when considered separately for men and women, has marked individuality, and conventional population-based reference ranges are consequently of limited value. In contrast, urinary creatinine and creatinine clearance have less individuality, and reference ranges are more useful. Desirable analytical standards are not attained for serum creatinine assays, but are achieved for urine creatinine and clearance determinations. Creatinine clearance is, therefore, the favoured first-line test for initial assessment of patients. However, the small critical difference required for two serum creatinine results to be significantly different, and the comparatively large critical difference for clearance, make serial serum creatinine assays more useful for monitoring individuals.