OF ARTS AND SCIENCES. 55 



of Rose* and Fresenius,f their analyses show from 99.96 percent, 

 to 100.48 per cent, of the true amount of phosphoric oxide. The 

 » weight of phosphoric oxide used in each of their experiments was 

 not far from 0.1200 gr. The precipitate was washed until the reac- 

 tion for chlorine with argentic nitrate failed to appear in the filtrate 

 acidified with nitric acid. The method of precipitating with magnesic 

 chloride and ign'ting over the Bunsen lamp gave in their hands from 

 99.9 G to 100.48 per cent. ; the method of precipitating with magnesic 

 sulphate and igniting over the blast, from 100.4 to 100.7 per cent.; 

 and the method of precipitating with magnesic sulphate and igniting 

 over the Bunsen lamp from 101.4 to 103.1 per cent, of the actual 

 amount of phosphoric oxide. 



All agree that the precipitated ammonio-magnesic phosphate is 

 soluble to a very considerable extent in dilute ammonia, and to a 

 greater extent in ammonia containing ammonic chloride ; that the 

 preseuce of a magnesia salt tends to prevent the solution of the pre- 

 cipitate ; and that when magnesic sulphate is used in excess as a pre- 

 cipitant the precipitate includes, mechanically or otherwise, an 

 appreciable amount of magnesic sulphate, or magnesic hydrate, or 

 both. Kissel and Schumann aim to arrive at correct results by pre- 

 cipitating with as little excess of magnesic sulphate mixture as 

 possible, and compensating for foreign inclusions by dissolving a part 

 of the precipitate. Kubel and Heintz strive to reach the same end 

 by removing the excess of the precipitant, together with the included 

 impurity by filtering off, dissolving, and again throwing down the 

 precipitate. 



With this preface I proceed to the description of my own experi- 

 ments upon this subject. 



I have worked with measured portions of solutions of hydro-disodic 

 phosphate, or of microcosmic salt, the standards of which were, in 

 all cases but one, determined by evaporating a known volume to dry- 

 ness, igniting the residue and weighing the remaining sodic pyrophos- 

 phate or metaphosphate as the case might be. In the one exceptional 

 case a known weight of microcosmic salt, freshly crystallized and 

 dried, the volatile constituents of which had been determined by ignit- 

 ing separate portions, was dissolved in a known volume of water. 

 The experiments which fixed the standards of these solutions are 

 given below. 



* Rose's Analytische Chemie, VI. Aufl., 512, bearbeitet von Finkener. 

 t Anleit. zur Quant. Anal. 6 Aufl. 403. 



