276 Studies of Acidosis. X 
and that neither the titratable acid nor body weight may be 
neglected without increasing the chance of error in the estimation. 
As, thanks to Barnett, the acid excretion formula is.simplified 
to the above by elimination of one unnecessary variable, and as 
Palmer and Van Slyke.in Paper IX (5) have published data which 
add the bicarbonate retention to the indirect acidosis tests that 
have been composed with the direct, it appears desirable to re- 
vise accordingly the summarizing table on page 412 of Paper 
VI. In the following table we have therefore substituted the 
simpler expression for acid excretion and have added the data 
for the bicarbonate retention test. We have also indicated the 
chief fallacies to which, according to the results published in the 
present series of papers, each indirect test is liable when applied 
to the detection of diabetic acidosis. 
It should be noted that the data obtained from kidney and lung 
excretion as measures of alkaline reserve in diabetic acidosis do 
not necessarily hold for other types of acidosis. In nephritis, for 
example, the two tests based on kidney excretion become falla- 
cious, while Peters (6) has recently shown that in cardiac dyspnea 
and in conditions involving great diminution of lung capacity 
the mechanics of respiration are so disturbed that the alveolar 
carbon dioxide ceases to be an approximate measure of blood 
bicarbonate. The indirect tests may be trusted as approximate 
indicators of alkaline reserve only in conditions where they have 
been previously tested by comparison with the blood bicarbonate. 
We have made this comparison in diabetes, but the results do not 
hold for other pathological conditions. 
SUMMARY. 
Acid excretion as a measure of diabetic acidosis is, according to 
present data, most significantly expressed in terms of ammonia 
plus titratable acid per unit of body weight. The average 
error involved in estimating alkaline reserve from acid excretion 
is, as shown by Barnett, appreciably reduced by simplifying 
our original empirical formula, plasma CO. = 80 — ‘7 1/C, to 
plasma CO, = 80 — 5 V7 The absolute difference in the re- 
