1918.1 AGKICULTURAL CHEMISTRY AGROTECHNY. 615 



is more sharply marked on account of the small volume, and (5) there is little 

 or no return of any blue color at the end of the titration. 



By means of a special capillary tip delivering from 45 to 55 drops of urine 

 per cubic centimeter, the titration can be made on undiluted urine vv^ith an ordi- 

 nary burette. The burettes are preferably filled by suction and the titration 

 made by the drop system, starting -with 25 drops of urine added to a clear solu- 

 tion obtained by heating 4.5 gm. of the dry salt mixture with 5 cc. of the cop- 

 per sulphate solution. For accurate results the drops should be delivered not 

 faster than 1 drop per second, and the portion of the burette most used should 

 be calibrated. 



The authors recommend for convenience of manipulation and for work in- 

 volving very small amounts of material special 5 cc. burettes graduated in 0.02 

 cc. In the case of urines containing albumin, rather large test tubes should 

 be used on account of foaming. The albumin alters the appearance of the 

 cuprous precipitate, but does not change or obscure the end point of the titra- 

 tion. 



The determination of lactose in milk, O. Folin and W. Denis {Jour. Biol. 

 Chem., 33 (1918), No. 3, pp. 521-524) .—The titration method of Folin and Mc- 

 Ellroy, noted above, was found by the authors to be applicable to the determi- 

 nation of lactose in milk without preliminary i-emoval of the protein materials. 

 The method of procedure was practically the same as that described in the 

 preceding paper, except that it was found advisable to dilute the milk 1 : 4 for 

 cow's milk and 1 : 5 for human milk. 



A colorimetric picrate method simpler than the one of Dehn and Hartman^ 

 was also used. The method is described in detail and a table given of com- 

 parative results with human and cow's milk of both the titration and colori- 

 metric methods. Of the two methods, the authors believe the titration to be 

 the more accurate, although the colorimetric method has the advantage that 

 by means of it a large number of determinations can be made more or less 

 simultaneously. 



A new microscopic method of counting bacteria adaptable to all grades 

 of raw and pasteurized milk, P. W. Allen (Jour. Infect. Diseases, 22 {1918), 

 No. 3, pp. 245-251, fig. 1). — The method consists of adding to the milk to be 

 tested a water suspension of aluminum hydroxid which readily collects the 

 bacteria. By centrifuging, the precipitate is thrown down and can be easily 

 separated from the fat, casein, and water. It is dried to a thin microscopic 

 film on a glass slide and stained with methylene blue, for which the hydroxid 

 has slight affinity. A bacterial count is made, using an oil immersion lens. 



The method is described in detail and tables of its accuracy reported. From 

 the data the author concludes that about 95 per cent of the bacteria in the 

 average sample of milk appears in the hydroxid thrown down by centrifugali- 

 zation. 



Detection of peanut oil in oils and fats, D. J. De Jong {Pharm. Weekbl., 

 54 {1917), No. 47, pp. 1390-1398; abs. in Chem. Abs., 12 {1918), No. 5, pp. 536, 

 537). — From experimental data a comparison is made of the relative value of 

 three methods of detecting peanut oil in commercial oils: (1) That of Jean, 

 recommended in the Dutch Pharmacopoeia for testing olive and sesame oils, 

 in which the oil is saponified with alcoholic potassium hydroxid and kept at 

 18° C, a precipitate appearing within an hour indicating the pre.sence of 10 

 per cent or more of peanut oil; (2) that of Franz-Adler, previously noted 

 (E. S. R., 30, p. 14) ; and (3) the solidification point method. The Bellier and 

 Renard-Archbutt methods are also discussed. 



» Jour. Amer. Chem. Soc, 36 (1914), No. 2, pp. 403-409. 



