54 PROCEEDINGS OF THE AMERICAN ACADEMY 



but which were otherwise treated in the same manner as the first, 

 gave figures for phosphoric oxide varying from 99.8 per cent, to 

 100.5 per cent, of the real amount. When the precipitate was twice 

 dissolved and twice reprecipitated, the amount of phosphoric oxide 

 indicated was 99.8 per cent., and when three times dissolved and re- 

 precipitated, 99.2 per cent, of the real amount. While noting that 

 the precipitate is not absolutely insoluble in ammonia, more especially 

 if it contain ammonic chloride, Kubel concludes, supposing that a 

 basic sulphate or hydrate of magnesium is thrown down with the first 

 precipitate, that this precipitate must be dissolved after washing and 

 again thrown down with ammonia. 



Kissel,* by using large amounts of ammonic chloride, washing 

 copiously in order to make the solubility of the precipitate compensate 

 for the inclusion of foreign matter, and taking care to use no great 

 excess of the magnesic sulphate mixture, obtains in one set of experi- 

 ments from 99.4 to 99.6 per cent., in another from 99.8 to 100 per 

 cent., of the correct amount of phosphoric oxide. Of three filtrates 

 from the precipitated ammonio-magnesic phosphate two gave weak 

 reactions for phosphoric acid with acid amnionic molybdate, and one no 

 reaction. The wash-water tested by the same method gave in every 

 case a plain reaction. Heintz f supports Kubel so far as to say that 

 in presence of any considerable excess of magnesic sulphate mixture 

 the precipitate must be dissolved after incomplete washing. Brunner t 

 points out that the solution must not be precipitated hot by magnesic 

 sulphate mixture on account of the danger of throwing down 

 magnesic hydrate. Schumann § supports Kissel in the view that 

 the solution of the first precipitate is unnecessary, but adds a 

 caution against the addition of ammonia after the magnesic sulphate 

 mixture. 



Finally, Abesser, Jani, and Marcher || substantiate Kubel's results, 

 and having proved the presence of a sulphate in the precipitate 

 obtain somewhat better figures by igniting over the blast. Adopting 

 the method of precipitating by magnesic chloride mixture previously 

 suggested by Brassier^ and Bunsen,** and recommended in the works 



* Zeitschrift fur Anal. Cheni., VIII. 169. 



t Ibid., IX. 16. 



J Ibid., XI. 30. 



§ Ibid, XI. 382. 



|| Ibid, XII, 239. 



T Ann. de Chim. et de Pliys. [4] VII. 335. 



** Zeitschrift fur Anal. Cheni. X. 405. 



