'Arnold K. Balls 197 



In the determinations referred to above, the precipitates were 

 ignited at high temp's, for several hours at a time, in an electric 

 muffle and cooled in the covered crucibles over sulfuric acid. It had 

 been observed that, in practically all cases, the ignited products were 

 hygroscopic. After exposure to the open air, for 10 min., a precipi- 

 tate weighing 86.3 mg. increased in weight to 89.1 mg. Re-ignition 

 reduced the weight of this mass to 86.1 mg. Aluminium phosphate 

 is not hygroscopic but aluminium oxid has that property. The 

 conditions of precipitation in the Schmidt-Hoagland method make 

 it extremely unlikely that oxid of aluminium is separated, at that 

 point, with the phosphate. It was inferred, therefore, that either 

 oxid, or a basic phosphate, was produced during the ignition. This 

 opinion was confirmed by the results of analysis of several of these 

 precipitates, which were found to contain less P2O5 than the 

 amounts required by the formula for the corresponding quantities 

 of ortho-phosphate. 



The precipitates of AIPO4, as obtained by the Schmidt-Hoag- 

 land process, contain large proportions of sulfur, resulting from the 

 decomposition of ammonium thiosulfate by the acid in the sol. 

 This fact suggested that, during the ignition of the precipitate, 

 P2O5 is replaced by oxides of sulfur, and that the resultant 

 AUC 304)3 is converted into AUOg, at least to some degree. The 

 ignited precipitate of AIPO4 referred to above, weighing 86.1 mg., 

 was mixed with about its own bulk of pure flowers-of-sulfur, and 

 re-ignited. The weight became constant in 6 hr., and the precipitate 

 was found to have lost 1.7 mg. 



Similar losses, indicating an influence of associated sulfur, in 

 accord with the foregoing opinion, were obtained in three analogous 

 tests. (See contradictory data on page 198.) 



After adding a small crystal of ammonium phosphate to the 

 ignited mass just referred to (84.4 mg.), moistening it with a drop 

 of water, and re-igniting, the precipitate increased in weight to 

 89.6 mg. A second ignited mass of AIPO4, weighing 87.6 mg., was 

 treated in a similar manner with ammonium phosphate. Its weight 

 increased to 90.8 mg., and the product was no longer hygroscopic. 

 Re-ignition with ammonium phosphate increased the weight of this 

 precipitate to 92.8 mg., in practical agreement with the standardiza- 



