STUDIES OF ACIDOSIS. X. 
By DONALD D. VAN SLYKE. 
(From the Hospital of The Rockefeller Institute for Medical Research.) 
(Received for publication, December 31, 1917.) 
The present paper includes a reply to Barnett’s (1) critique 
in the preceding article and a summary showing the nature of 
the results obtained with the methods for estimating alkaline 
reserve from urine and alveolar air analyses. 
Fitz and the writer (2) succeeded recently in demonstrating a 
quantitative relationship between the alkaline reserve and the 
excretion of acids in excess of fixed base, as measured by am- 
monia plus titratable acid in the urine. The empirical formula 
utilized to demonstrate the relationship was, plasma CO, = 80 — 
y2 1/ C, D representing the cc. of 0.1 N titratable acid and 
ammonia excreted per 24 hour time unit, and C (concentration) 
the amount excreted per liter. 
Barnett states the belief that this formula does not express the 
relationship between plasma bicarbonate and acid excretion with 
sufficient accuracy to justify its use. The reason for this opinion 
is that the results show that the average error in calculating the 
plasma bicarbonate from the excretion exceeds that which would 
result from maximal analytical errors in all of the determina- 
tions involved. This seems to us insufficient ground for the 
criticism. The essential question, for our purpose, is not whether 
the plasma bicarbonate can be estimated from urine excretion 
with no error except that of the chemical determinations, but 
whether the total sum of the errors of analysis, individual varia- 
tion in kidney function, and fault in the empirical formula is 
within such limits that the excretion may be of any assistance in 
estimating the alkaline reserve when conditions prevent the direct 
determination of the latter in the blood. 
We have given in Papers IV (2) and VI (8) data which show 
the limits of accuracy encountered in estimating plasma bicar- 
271 
