A revised scheme is described for evaluating automatic instruments used in clinical chemistry. Procedures are outlined for the assessment of mechanical and electrical features, and measurement of the accuracy and precision of individual units. Methods are given for the measurement of analytical precision, carryover, cross-contamination, accuracy, and linearity. The safety of equipment and methods of assessing costs are discussed, and the importance of subjective features is noted. The general principles of the evaluation scheme should be applicable to other types of equipment.

The renal responses to change of posture have been ascribed to changes in renal haemodynamics (Rosenbaum, Ferguson, Davis & Rossmeisl, 1952); but Thomas (1957) has produced evidence that, at least in prolonged experiments, there is a change of tubular function such as could be produced by an increased secretion of aldosterone. There is evidence (Muller, Manning & Riondel, 1958) that aldosterone output over periods of twelve hours is higher in ambulant subjects than in those who remain in bed. Changes in output in relation to posture have not, however, been demonstrated over shorter periods such as those here reported. It has often been suggested that secretion of aldosterone is regulated by 'volume receptors' at some unknown situation (e.g. Bartter, Liddle, Duncan, Barber & Delea, 1956); if posture can be shown to affect aldosterone output, it seems likely that the same volume receptor mechanism is involved, and postural studies may therefore contribute towards the location of these receptors.A preliminary account of this work has already appeared (Gowenlock, Mills & Thomas, 1958). METHODSObservations were made on healthy male subjects, 25 medical students and one of the authors (J.N.M.). For each experiment groups of 3-8 subjects were used; in order to investigate changes of aldosterone excretion over short periods, urine from the group of subjects was pooled, as only thus was it possible to obtain the 8-12 man-hours of urine required for accurate assay by the method used. Except in experiments 7 and 8 no two groups had the same composition; each subject was used in one to four experiments.After a light breakfast without tea or coffee, subjects voided and discarded their urine at about 10.30 and then lay down for 2 or 24 hr before voiding sample 1. They then either remained recumbent, for 5 hr in all, voiding sample 2 at the end; or they changed their posture to standing or sitting, voiding and discarding the urine after 1 hr and collecting the urine passed in the followng 2 hr; or they remained recumbent and received an intravenous injection of 0-25 mg DL-aldosterone-21-monoacetate (Ciba), in 95% ethanol, diluted with isotonic NaCl, repeated 1j hr later, and collected the urine passed during 3 hr from the first injection. They drank when thirsty. Small samples of urine from each subject were removed for separate analysis of Na and K * Present address: Department of Pathology, University of Manchester.

No abstract

Definition of termsIf a specimen is analysed repeatedly for a constituent using the same method throughout, the individual results will usually be distributed normally around a mean value which is the best estimate of the concentration of the constituent. This best estimate may be greater or less than the 'true' value and the term 'accuracy' is concerned with the agreement between these two figures; the closer the agreement, the greater the accuracy.The extent of the scatter of individual results around the mean is expressed by the standard deviation (S.D.) quoted in the same units as the mean. Alternatively the standard deviation may be expressed as a percentage of the mean value and is then referred to as the coefficient of variation (c.V). The smaller the S.D. (or C.V.) the greater the precision of the method.The term 'normal range' indicates that range of concentration of a particular constituent covered by two S.D.s on either side of the mean value observed in a group of 'normal' persons. In the case of a Gaussian distribution this range will include approximately 95 % of the 'normal', population.There are various 'normal' populations and the influence of such factors as age, sex, posture, time of day, diet and general health is the province of other papers in this symposium. Whatever is selected as 'normal' for determination of the normal range of a particular substance it is relevant to enquire whether the accuracy and precision of the methods and instruments used to carry out the determinations have any significant effect on the range determined.Influence of precision on the normal range for a particular substance

SUMMARY Two patients with primary biliary cirrhosis who were increasingly incapacitated by xanthomatous neuropathy are described. Treatment with a low fat diet and cholestyramine was unhelpful but repeated plasmaphoresis by simple venesection in one and plasma exchange using an IBM blood cell separator in the other over a period of several months completely relieved the symptoms of the neuropathy, caused skin xanthomata to recede, and lowered plasma lipid levels in both patients. There was no evidence that this procedure was associated with any deleterious effects on the liver. The size of the cholesterol pool in xanthomata in one patient was estimated to be approximately 35 g, and from the plasma cholesterol response to plasmaphoresis at varying frequency it was suggested that the excess of cholesterol synthesis over degradation was less than 0.3 g/day in one patient and less than 0.4 g/day in the other. On the basis of the response in these patients it is suggested that the turnover rates of lipid pools are relatively slow in biliary cirrhosis and that cholesterol accumulation is more likely to be due to a reduced catabolic rate than to an increased synthetic rate of cholesterol. Plasmaphoresis or plasma exchange are useful methods of treatment for the rare patient afflicted by this resistant and distressing complication of biliary cirrhosis.