Sodium and potassium were measured in sets of 102 to 107 patients sera, and in 31 commercially available control sera. The results from four routine analytical methods/systems (indirect potentiometry: two; direct potentiometry and enzymatic assay: one each) were compared with those from a flame photometry-based reference method. In the assay of patient sera, substantial agreement was observed in some comparisons, clinically relevant bias in others. The inter-assay changes observed for the control sera differed significantly from those shown by the patients sera (i.e. commercial control sera were non-commutable) in about 12% of the comparisons, as a whole. Recalculation of serum sample results with a single control serum as calibrator lowered or increased the bias originally present according to whether the serum itself was commutable or not. Moreover, the inter-method variability in the assay of commercial control sera was lower with commutable sera, higher with non-commutable sera. With the exception of liquid sera stabilized with ethylene glycol, there was no evident link between any specific characteristic of the commercial control sera (matrix and physical state) and their degree of commutability. The routine measurements were performed with the following four methods/instruments/systems. Method 1 Indirect potentiometry, E2A instrument, from Beckman, reagents and calibrators from the same source. Method 2 Indirect potentiometry, Hitachi 717 analyzer, from Boehringer Mannheim, reagents and calibrators from the same source. Method 3 Direct potentiometry, Cobas Mira S plus analyzer, from Roche, reagents and calibrators from the same source.