15 



much smaller than from the simpler amids, and varied much more with 

 the precise conditions of temperature, pressure, and strength of solu- 

 tion used. In one experiment serum albumen gave but 2.G8 per cent, 

 and in another but 2.92 per cent was obtained from haemoglobin. 



On the whole, the indications pointed to the simpler amids and 

 amido-acids being most easily decomposed by nitrous acid, the proteoses 

 and peptones perhaps next, and the proteids proper least. But no 

 differences were observed upon which any useful analytical process ot 

 separation or distinction could be based. Experiments made in this 

 way are, moreover, troublesome, and require strict observance of the 

 necessary precautions to avoid error from introduction of air into the 

 apparatus and its action on nitrogen dioxid. 



INTERACTION WITH POTASSIUM PERMANGANATE IN PRESENCE OF 

 FREE ACID OR ALKALI. 



The writer had not much hope of obtaining useful results by this 

 method, in view of former experience gained in connection with the 

 extended research upon the determination of organic matter in drink- 

 ing water carried out many years ago for the U. S. National Board of 

 Health, of which the results were published in the annual report of 

 that board for 1882. Still, as the work then done was upon extremely 

 dilute solutions, comparable in respect to the amount of organic mat- 

 ter present with natural potable waters, it seemed possible that results 

 not altogether of the same sort might be obtained with solutions of 

 greater concentration. Hence some thirty or forty experiments were 

 made with permanganate strongly acidified with sulphuric acid, and a 

 rather larger number with the same salt after potassium hydroxid 

 had been added in about the same proportion as is usual for the Wank- 

 lyn, Chapman, and Smith so-called albuminoid-ammonia process. In 

 the former set of experiments the amount of oxygen withdrawn from 

 the permanganate, and in the latter set the amount of ammonia pro- 

 duced, were determined. In both cases the reactions were carried out 

 at various temperatures up to the boiling points of the several liquids. 



Some difference was observable in the results obtained with the com- 

 paratively strong solutions used, such difference being more notable 

 for alkaline than for acid permanganate. But in the main these results 

 only confirmed the conclusions arrived at in the earlier research. The 

 effect of the reagent employed is imperfect, and varies much with 

 the nature of the individual organic materials tested; much, also, 

 with the conditions of the experiment and with the rate at which 

 the action proceeds. Ko valid evidence is obtained in support of 

 Wanklyn's view that simple and definite fractions of the total nitrogen 

 present are evolved as ammonia on treatment with alkaline perman- 

 ganate. 



Some special difficulties and sources of error already known were 

 reobserved; as, for instance, the continuous evolution of ammonia by 



