22 Studies of Acidosis. X 
bonate from acid excretion in practically every form of diabetic 
acidosis. It does not appear to us that a valid reason has been 
advanced for modifying the conclusion based on those data; 
viz., that the ammonia plus acid excretion is quantitatively re- 
lated to the bicarbonate deficit in the blood, and that the rela- 
tionship is sufficiently uniform to be useful in estimating this 
deficit when the limitations of accuracy, as shown by our data, 
are taken into consideration. 
Barnett’s claim that all the variables of the formula except the 
two most important ones, the blood bicarbonate and ammonia 
excretion, may be replaced by a constant without significantly 
diminishing the average accuracy of the formula, would likewise 
appear to us, even if entirely justified, to detract nothing from 
the correctness of the above conclusion. 
Any simplification in the calculation not resulting in loss of 
accuracy would be an improvement, however, and therefore it is 
desirable to examine somewhat closely the formula minus one, 
two, and three of its variables in order to decide how many of 
them may be deleted without lessening the reliability of results. 
Concerning the dropping of C, thereby simplifying the formula 
aS N 
Plasma CO: = 80 — NEO VC to, Plasma CO, = 80—5 Vir 
we agree with Barnett that it: makes the equation not only 
simpler, but also measurably more accurate and is therefore in 
every way desirable.'. In preliminary experiments in which we 
tested several formulas before one was chosen to apply system- 
atically, the value of C appeared to have an appreciable effect, 
although much less than that of D;2.e., other conditions being the 
same, a greater volume of urine appeared to carry out somewhat 
1 From the negligible average influence of C it appears that, at least as 
long as either normal or exceptionally high volumes of urine are excreted 
(which was the case with all our patients), variations in the volume are of 
no effect on the acid excretion. We are not, however, prepared to follow 
this conclusion as to the non-effect of C to its logical limit and state that 
in a patient with severe diabetic acidosis flushing with water is of no 
value. A liter of urine is not an abnormally small amount in a normal 
adult, but it could carry out only about 15 gm. of 6-hydroxybutyric acid, 
if, as indicated by the data of Magnus-Levy and others, this acid is never 
excreted in concentration greater than 1.5 per cent. To remove the 
