1920] AGRICULTURAL CHEMISTRY — AGROTECHNT. 805 



tures of coffee and coffee substitutes. The error is thought to be due to the 

 impurities in the caffeln sublimate. In such cases the author recommends 

 either the determination of nitrogen in the crude caffein or a further purifica- 

 tion of the caffein. In the first process the nitrogen is determined by the 

 Kjeldahl method on the crude caffein as it is obtained after separation of the 

 fats in the method previously noted. In the final titration with n/10 Hi SO4, 

 Congo red is recommended as indicator. In the alternative method the crude 

 caffein is purified by repeated extraction with chloroform, followed by the addi- 

 tion of a small amount of sodium carbonate to unite with impurities on tlie 

 order of humic acid. The solution is then filtered and evaporated and the 

 caffoin dried at 100° C. and weighed. 



Samples of Kaffee Hag yielded 0.13 per cent of caffein by the author's pre- 

 vious method, from 0.04 to 0.05 per cent by weighing the purified caffein, and 

 from 0.05 to 0.06 per cent by the nitrogen determination method. 



Preservatives for fermentable beverage samples, J. S. McCune and A. N. 

 Thukston (Jour. Indus, and Engin. Chem., 12 {1920), No. 7, pp. 689, 690).— 

 Tests are reported of the efficiency of various preservatives for samples of un- 

 treated draft beer and of bottled beer treated with Fleischmann's yeast, which 

 are to be tested for alcohol content. 



Sodium benzoate, 0.2 per cent, proved unreliable, sodium bisulphite and boric 

 acid, 0.2 per cent, worthless, and mercuric chlorid and salicylic acid, 0.2 per 

 cent, satisfactory as preservatives. The use of 0.2 per cent of salicylic acid is 

 recommended for preserving fermentable samples when immediate delivery to 

 lie laboratory is not po.ssible. 



Report of the committee on cottonseed products, division of industrial 

 and engineering chemistry, American Chemical Society (Jour. Indus, and 

 Enffin. Chem., 12 (1920), No. 7, pp. '707-711). — This report includes methods 

 recommended by the committee for the sampling and examination of cotton 

 seed, cottonseed oils, cake, and meal, crude cottonseed oil, crude peanut oil, 

 crude coconut oil, crude soy-bean oil, refined oils, soap stock, and acidulated soap 

 stock, titer tests, and an appendix specifying oil grades, both crude and refined. 



A comparison of methods for estimating fat in butter, H. W. Gregory, 

 P. S. Lucas, and G. Spitzeb (Indiana Sta. Bui. 2U (1920), pp. 11, figs. S). — The 

 authors report parallel determinations of the fat in 20 samples of butter by the 

 Official and four unofficial methods, and likewise the moisture, salt, and curd 

 content by the Official and unofficial methods. Determination on 7 other sam- 

 ples by fewer methods are also reported. 



In the 20 samples the percentages of fat given by the Official method aver- 

 aged 0.21 higher than Kohman's gasoline extraction method (E. S. R., 40, p. 

 311), 0.08 lower than the Shaw method (E. S. R., 27, p. 614), 0.66 lower than 

 the results from the Wright butter fat test bottle, and 0.42 higher than the 

 so-called factory method. 



In the factory method the fat is calculated by difference on the assumption 

 that the curd forms 1 per cent of the butter, the moisture being estimated by 

 heating a sample over a flame and the salt, after a single extraction, being deter- 

 mined by the volumetric silver nitrate method with potassium chromate as 

 indicator. The factory method is considered the best .suited for use by prac- 

 tical butter makers. Actual determinations of the curd by the gasoline extrac- 

 tion method showed it to be lower than 1 per cent in 20 out of 24 samples and 

 lower than 0.8 per cent in 12 of them. 



Correction for volume of precipitate in the polarimetric determination 

 of lactose in milk, A. E. Perkins (Jour. Dairy l?ci., S (1920). No. 2, pp. 1S4- 

 142).— The author, at the Ohio Experiment Station, presents calculations from 

 19i)73°— 2U 2 



