Donald D. Van Slyke 275 
that have undergone bacterial decomposition, either in the blad- 
der as the result of cystitis, or outside the bladder as the result 
of preservation with insufficient antisepsis. As long as the 
ammonia formation leaves the urine still acid, it does not much 
alter the NH;-+ acid figure, since CO,, of the ammonium car- 
bonate formed by bacterial action on urea, escapes, while the 
ammonia remains and neutralizes approximately an equivalent 
of acid. The net effect of an increase of ammonia is therefore an 
approximately equal decrease of titratable acid, with no signifi- 
cant influence on the resultant sum of the two. 
0.1 N NH; fae. 
OPM NGaCIg. tes 
be used as a fairly sensitive indicator of ammoniacal decomposi- 
tion. In the urines of Paper IV, which were all analyzed while 
perfectly fresh, the ratio varies from 0.3 to 4.8 averaging 1.6 in 
normal men and 2.8 in diabetics. In only one case was a value 
of 4.1 exceeded. It appears therefore that when the ammonia 
exceeds four or five times the titratable acid there is ground for 
suspecting the origin of a measurable portion of the ammonia in 
bacterial action. Although the 24 hour urines reported in Paper 
VI are believed to have been collected with at least ordinary 
care and were preserved with toluene, it will be seen from the re- 
sults that in these urines the ammonia:acid ratio averages higher 
(about 4.1) than in the quickly collected and analyzed specimens 
of Paper IV, and frequently exceeds the maximum of the short- 
time urines, at times rising as high as 8 or 9. In no case had 
decomposition gone so far as to neutralize all or nearly all of the 
titratable acid, so that the results, based on the sum of ammonia 
and titratable acid, could not have been significantly affected. 
The frequency of high ammonia:acid ratios in the 24 hour urines 
nevertheless indicates the readiness with which decomposition 
may occur in 24 hour specimens even when collected with routine 
precaution. 
For the above reasons we believe that when diabetic acidosis 
must be estimated from acid excretion, the most satisfactory 
formula at present available for expressing the results in terms of 
alkaline reserve is 
Our data indicate furthermore that the ratio 
0.1 N (acid + NH3) per 24 hours 
Pi CO, = 80 — 5 @— 
aes : \ kilos body weight 
THE JOURNAL OF BIOLOGICAL CHEMISTRY, VOL. XXXIII, NO. 2 
