Summary: In clinical chemistry today there are often several different methods and instruments in use for the analysis of a particular constituent. When these different methods and instruments are used for careful analysis of the same control specimen, as in assigned value determinations, the assigned values obtained may be significantly different, and these differences may be of clinical importance. In order to determine the reasons for such differ-ences, reference systems or reference points are needed. The reference systems for calibration (calibration materials, standard reference materials) and control (control materials), where the matrices of these different materials must have quite different characteristics, are referred to jointly as reference materials. In order to compare methods, reference methods are needed with known, high reliability. Because of the amount of specimen material needed, the time needed for analysis and the facilities required, reference methods are not suitable for routine analysis. Ideally, standard reference materials are developed first, followed by a definitive method, which is then used to evaluate a candidate reference method. This is currently the case for only a small number of the constituents analyzed in the clinical chemistry laboratory. In this paper the characteristics of the different kinds of reference materials and of reference methods are described. The differences in terminological usage found in the literature are discussed. Limitations of knowledge and technique may necessitate certain compromises, with respect to the ideal character-istics of reference materials and reference methods. Possible compromises are discussed, and the various sources

calcium lithium magnesium potassium, chloride phosphate sodium, chloride cholesterol creatinine glucose glycine urea uric acid

(1, 2) gradually increased. The in-crease exceeded the expected varia-tion reported for a reference method (3). Chohesterylacetatehas been re-ported to yield a greater absorptivity with the Liebermann-Burchard re-agent than cholesterol (4); and we suspected that the cholesterol had gradually esterified. To test this hy-pothesis, we prepared a fresh solution of 100 mg of thrice-recrystallized (from absolute ethanol) cholesterol in 100 ml of glacial acetic acid and allowed it to stand at room tempera-ture. Immediately after preparation and seven days, 24 days, and 53 days thereafter, 0.1-mi ahiquots of the solu-tion were removed and dried under decreased pressure at room tempera-ture. When all the dried samples had been collected, they were redissolved in 0.1 ml of isopropanol, and 10-id aliquots were spotted on a silica gel thin-layer chromatographic plate. At the same time, lO-zl ahiquots of fresh-ly prepared isopropanol solutions of cholesterol and chromatographicahly pure cholesteryl acetate were also spotted, as was a lO-zl aliquot of a solution of cholesterol in isopropanol that had been standing for 53 days at room temperature. The plate was de-veloped in chloroform and the spots were made visible by spraying with sulfuric acid:water (1:1 by vol) and heating briefly in an oven at 110 #{176}C. As Figure 1 shows, within seven days of preparation and progressively thereafter, cholesteryl acetate was in-creasingly present in the solution. Al-though exactly when esterification began cannot be ascertained from this experiment, another experiment, with use of larger aliquots of the cho-lesterol solution, suggested that with-in 72 h of preparation some cholester-yl acetate was qualitatively detecta-ble. No cholesteryl ester was present in the 53-day old cholesterol solution in isopropanol. Further supporting evi-dence was obtained by submitting al-iquots from these same samples to gas-liquid chromatography on a

The dynamic interplay between preparation, handling, containment, and analysis of ultrapure chemical standards and reagents for the clinical chemistry laboratory is described. Inorganic as well as organic standards are reviewed. The contamination problems associated with water, acids, and solvents in trace-element analysis are defined and resolved. In addition, less well-known sources of error in trace analysis are reviewed, as are the common contaminants in the clinical laboratory. It does little good to use a 99.99 % primary standard with water of unknown quality in a heavily contaminated laboratory. Fortunately, recognition, definition, and solution of this problem is possible and should introduce new levels of precision and accuracy in the clinical chemistry laboratory. Additional Keyph rases sources of contamination, positive and negative preparation of high-purity standards, organic ond inorganic #{149}trace analysis