560 EXPERIMENT STATION RECORD. 



of alumina of known composition. The sum of alumina and iron shown 

 by this method was somewhat lower than that given by direct deter- 

 mination in the Glaser method. This is explained by the fact that in 

 the Glaser method no account is taken of the difference in molecular 

 weight of the iron and aluminum oxids in calculating the results of 

 analysis. 



Solubility of phosphates in citric acid and ammonium citrate, 

 O. FoERSTER {Glieni. Ztg., 20 [1890), No. 103, pp. 1020, 1021).— It has 

 been generally observed that calcium phosphates as well as aluminum 

 and iron.x)hosphates are more soluble in ammonia-free citric acid than 

 in ammonium citrate. This has also been observed in the case of slags, 

 and is ascribed especially to the presence of iron and alumina phos- 

 phates. These general conclusions were confirmed by the author in 

 examinations of tricalcium and tetracalcium phosphate, iron phosphate, 

 alumina phosphate, and 8 samples of Thomas slag, the results of which 

 are reported. In only two cases did the slag show a greater solubility 

 in ammonium citrate than in iiure citric acid. It is suggested that the 

 unusual results in these two cases may be explained by variations in 

 the conditions of solution and in composition of the dissolved matter. 

 It is proposed to make this the subject of further investigation. 



Natural and industrial phosphates : III. The insoluble residue, 

 H. Lasne {Ann. Ghim. Analyt., 1, p. 207; ab.s. in Analyst, 21 [1890), 

 Sept., p. 216). — The residue from the hydrochloric acid solution is cal- 

 cined and silica driven oft" with hydrotiuoric acid, the residue being 

 recalciued and weighed. The new residue is dissolved in sulphuric 

 acid and tested for alumina, alkalies, phosphoric acid, titanic acid, and 

 zirconium, the presence of the latter indicating the origin of the.iihos- 

 phate. Barium should also be looked for, its presence indicating adul- 

 teration with plaster of Paris. 



The determination of phosphoric acid in potable Avater, C. 

 Lepierre [Bui. Soc. Cliim. Paris, 15-16 {1896), No. 23, pp. 1213-1217, 

 Jigs. 2). — It is maintained that the method in which the phosi^horic 

 acid is determined by weighing the phosphomolybdate is unreliable on 

 account of the solubility of this compound. 



A method based on the color of a nitric acid solution of the phos- 

 j)homolybdate is proposed as follows : Evaporate 1 liter of the water to 

 dryness in a platinum dish, with the addition of nitric acid, and heat 

 to separate silica. Take up in dilute nitric acid and evaporate to dry- 

 ness again to eliminate the last traces of silica. Moisten with nitric 

 acid and filter the solution, washing the filter until the filtrate amounts 

 to 50 cc. Add 2 cc. of molybdic solution and compare the coloration, in 

 the cold, with that of standard solutions of phosphomolybdate, pre- 

 pared by adding molybdic solution to water solutions of phosphate 

 containing from 0.1 to 25 mg. of phosphoric acid. For exact work the 

 temperature should be in all cases about 30° C. 



The determination of albumen in cows' milk, L. L. Yan Slyke 

 [Neiv York State Sta. Bpt. 1894, pp. .522-526).— In the Annual Ileport of 



