BY JAMES M. PETRIE. 839 



each was determined by Kjeldahl's process, using 10 cc. of 

 sulphuric acid, 0*2 gra. copper sulphate, and boiling till charring 

 ceased; then 10 gins, of potassium bisulphate were added, and the 

 boiling continued for three hours after the liquid cleared; the 

 ammonia was distilled off in the usual way into ~j acid. 



(1) 50 cc. for determination of total N. 



(2) 50 cc. were boiled with 20 cc. of Stutzer's reagent and lgm. 

 of alum, filtered and washed, and the N estimated. 



(3) 50 cc. were treated with 15 cc. of tannin-salt solution (i.e., 

 just in slight excess), centrifuged, washed with reagent, and the 

 N determined in the ppt. 



(4) 50 cc. were evaporated down to 10 cc, and poured into 

 90 cc. of alcohol 94 %, making the solution of the strength of 85 % 

 spirit, heated to boiling, allowed to stand 3 hours, and filtered. 

 From the filtrate the alcohol was then distilled off, and the N 

 determined in the residue. 



The above were done in duplicate, and the mean results are 

 given below. 



1. Total N ll*8cc. ^ acid = 0*0165 gm. N. 



2. N in copper ppt 4*2 0-0059 



3. N in tannin ppt 2 65 0-0037 



4. N in alcohol filtrate 8 '95 0-0125 



Therefore the non-protein nitrogen is represented by — 



2. N in copper filtrate 0*0106 gm. 



3. N in tannin filtrate 0*0128 



4. N in alcohol filtrate 0*0125 



We may conclude from these figures that, in a solution such as 

 the one used above, tannin and alcohol are practically equal in 

 precipitating power; and, further, that peptones which were known 

 to be present in the above extract are also precipitated by the 

 tannin-salt reagent, the filtrates being biuret-free. Copper 

 h}'droxi(le precipitated 13% more nitrogen than either of the 

 other two reagents. 



On looking over Schulze's results, a few of which are quoted at 

 the beginning of this paper, it is observed that the maximum 

 amount of nitrogen of non-protein compounds is about 1 1 % of 



