ELI8HA MITdlELL .SCIENTIFIC SOCIETY. 25 



The aU)unHn()i(l enjuivalent of tin's was found to vary from 4 

 per cent, to 5 })er eont. with one treatment with alkali. The 

 acid subsequent treatment did not remove any more, but subse- 

 quent treatments with alkali left no nitrogen. 



Chemicai- Lahokatouy, 



Cornell University. 



SOME MODIFICATIONS OF THE METHOD FOR 

 DETERMINING CRUDE FIBER. 



BY W. A. WITHERS. 



With the method now in use for determining crude fiber 

 (vide Bui. 19, 1888, U. S. Dept. of Ag.) there is little trouble 

 in the manipulation, except in cases of feeding-stuffs, in which 

 the percentage of albuminoids is rather large. In all such 

 cases, however, the albuminoids precipitated by the treatment 

 with HgSO^ seriously irupede the filtration, rendering it very 

 slow, and from this longer contact causes a part of the crude 

 fiber to be dissolved. 



This difficulty can be obviated by treating the substance with 

 the NagO solution before rather than after the treatment with 

 H2SO4 solution. 



To test this modification, I compared it with the ordinary 

 method, on samples of Timothy hay and cotton seed meal. In 

 both cases clear filtrates were secured by the modification in 

 from 15 to 30 minutes, while with the ordinary method the fil- 

 tration with the hay required about the same time; but the 

 cotton seed meal from 18 to 24 hours, and then the filtrate was 

 not clear. With Timothy hay the two methods gave practically 

 the same results. With cotton seed meal the ordinary method 

 gave 2.68 per cent, and 2.91 per cent., an average of 2.80 per 

 cent., while the modification gave, with five determinations, 

 3.63 per cent., 3.58 per cent., 3.69 per cent., 3.49 per cent.. 



