Table 1 . Mean Chemical Composition and Calculated Total Volume

laboratory;the other analyses were performed intriplicate inour laboratory (Table 1). To convert the mass analyses from mass per 100 mL to mL/100 mL, the relative volume for proteins was taken as 0.75 and for lipids as 1.1; for the electrolytes, Na and K were added and calculated as 75% NaC1 and 25% NaHCO3, with a relative volume of 0.46. The standard error of the means for the total volumes was estimated by treating each of the triplicate analyses as separate series. Only specimens 4 and 5 had a total volume significantly less than 100 mL. Assuming that specimen 4 contains 2.85 mL of contrast medium per 100 mL and specimen 5 contains 1.08 mL per 100 mL, the measured data for these two specimens were recalculated per 100 mL of plasma; these calculated data are displayed in parentheses. The total lipids for these two samples have been recalculated by using the equations provided above. Use of the standard t-test to appraise the corrected data, and taking p <0.02 as significant gave the following results: for water, 2>6t5413*1; for protein, 1>4>3*.5>6>2; for total lipid, 14>5*.3t-6*.2; for Na, 2*.1i3t.6>5>4; and for K, 1>6>3>5>4>2. For aldosterone and renin, an F-test with the highest and lowest values consecutively excluded showed none of the results to differ significantly at the p <0.2 level. It is not surprising that blood from the renal veins or from the upper part of the inferior vena cava should differ significantly in composition from blood drawn at other sites. However, the numerous differences between the six samples cannot be rationalized by any simple hypothesis. Nonetheless, the data do show clearly that comparing a specific analysis on blood taken from two or more sites could be misleading unless attention is paid to the composition of the samples as a whole.