489 



at 100° ('. in a cuneiit oCliyfhom'ii [)]e\ ioius to v\i iju'tioii. A table gives 

 the percentage ot extract witli the dilfereiit .solvents from each nuitc;- 

 lial. 



"It apj.eais from these results that petroleum-beuzine and cliloroform 

 dissolve chior()i)h.vl quite as readily as ether. In all eases but one ben- 

 zine extracted considerably less than ether, but it has yet to be shown 

 that the benzine-extract any more nearly represents true fat than the 

 ether-extract. 



"Petroleum-benzine is also especially unfit for use as a solvent in 

 quantitative work for the reason that it is an indetinite and variable 

 mixture of a number of solvents, which, as the results indicate, diller 

 ill their solvent action on the dry matter of feeding stuffs." 



The determination of rnospiiouic acid m feutilizeks by 

 THE "cituate method," S. W. Johnson, M. a., and T. B. Osbokne, 

 Ph. D. (pp. 254-207).— A recapitulation of the studies of H. L. Wells 

 and others on this subject, together with an investigation by the authors 

 leading to a modiiication of the method. 



In the method as modilied "the volume of the solution of the phos- 

 phate was in all cases 100 cubic centimeters, and this was mixed with 

 10 cubic centimeters of a 50 per cent solution of citric acid. After neu- 

 tralizing with ammonia, 50 cubic centimeters of 'United States mag- 

 nesia-mixture' (containing twice as much ammonium chloride, but 

 otherwise the same as the German mixture) was slowly added with 

 constant stirring, and when the precipitate had sei)arated 30 cubic cen- 

 timeters of concentrated ammonia water was poured in. After 2 hours 

 the precipitate was filtered on a Gooch crucible, ignited, and weighed. 

 If the sample analyzed had more than 10 per cent of phosphoric acid 

 1 gram of it was commonly used for the determination. If less than 

 10 i)er cent 2 grams were usually employed. In the case of ashes con- 

 taining soluble-silica this substance was iirst separated. 



" Of sixty-seven comparisons of this modified method with the molyb- 

 date method the results of but three differed from the molybdate method 

 by more than 0.3, and but four others by more than 0.2 i)er cent. The 

 greatest difference was 0,41, the average difference 0.09 per cent. In 

 thirty eases the citrate method gave 0.117 percent more, on the average, 

 than the molybdate; in thirty-three cases 0.079 per cent. * * * 



"The process thus found so satisfactory with a large number of ferti- 

 lizers gave trouble when ai)plied in the same manner to Thomas slag 

 and Keystone concentrated phosphate. * * * 



"To further examine the sources of error involved in this method, 

 several of the ignited and weighed precipitates from determinations by 

 the citrate method were subjected to partial analysis. These analyses 

 indicate that the ])recipitate is nearly constant in composition whether 

 salts of calcium or those of iron and aluminium preponderate in the solu- 

 tion from which it is thrown down. From the (irand Cayman's phos- 

 phate, rich in iron and aluminium, but little more of these metals enters 

 the precipitate than from Carolina rock or bone-black." 



