THE NITROGEN-FREE EXTRACT OP FEEDING STUFFS. 653 



Eepeated combustions of the original paper and the product from 

 treating- it witli sulplinric acid and potassium liydroxid, carried on in 

 Ilenneberg's laboratory-, showed the composition invariably to be that 

 of cellulose, the carbon in the product ranging between 44.17 and 44.38 

 per cent, iind the hydrogen between G.17 and G.19 per cent. 



Kern found that the cellulose of feeding stuffs was similarly attacked. 

 Orchard grass gave by the Weende method 34 i)er cent of crude iiber, 

 and this product treated by F. Schulze's method (macerating with 

 potassium chlorate and nitric acid) showed 32.8 per cent of cellulose in 

 the hay; but when the hay was treated directly by Schulze's method 

 35 per cent of cellulose was found. In other words, 0.3 per cent of the 

 total amount of cellulose was dissolved by the AVeende method. 



It follows from the above experiments that the cellulose is always 

 somewhat attacked and dissolved by the reagents of the Weende 

 method, .and that consequently a part of the cellulose of feeding stuffs 

 is credited to the nitrogen-free extract. 



This diiliculty is not removed by the modifications that have been pro- 

 posed from time to time, including Wither's modification in which the 

 operation is reversed, tlie substance being boiled first with potassium 

 hydroxid and then with sulphuric a<;id. Furthermore, the experiments 

 by Huston and McBride have shown that with stronger sulphuric acid 

 and alkali the cellulose is atta(;ked to a greater extent than with 1.25 

 per cent solutions. 



MIXED CIIAIIACTER OF CUT'DK FIUER AND NITROGEN-FREE EXTRACT. 



From what has been said, it appears that the crude tiber and nitro- 

 gen-free extract, as determined by the Weende method, must be 

 mixtures whose constituents aie changeable and a matter of chance; 

 and that the assumption that the crude fiber is the cellulose of the 

 vegetable substance under examination is incorrect. 



The nitrogen-free extract obtained by the Weende method contains 

 three categories of substances: 



(1) True carbohydrates, including the pentosans. 



(2) A variety of substances which either do not belong to the carbo- 

 hydrates at all, as organic acids, ligniu substances, materials of the 

 aromatic groups, etc., or which, like mannit, are closely allied to the 

 carbohydrates. 



(3) The cellulose changed and dissolved by the 1.25 per cent acid 

 and alkali. Since several per cent of the cellulose is always dissolved 

 in the Weende method, somewhat less crude fiber is always found than 

 might be expected, and a correspondingly too high amount of nitrogen, 

 free extract. Furthermore, as the cn-nde fiber contains other substances 

 besides cellulose, more crude fiber than cellulose is Ibund. 



