'Arnold K. Balls 



201 



that required for the equivalent amounts of A1P04.^ The data :n 

 Table 3 show that, if the quantities of aluminium oxid in such 

 precipitates are determined in the usual way^ (by subtracting the 

 P2O5 found from the total weight of AI2O3 plus P0O5), the results 

 for AI2O3 are low, indicating incomplete precipitation of aluminium 

 as phosphate. In view of the solubility of aluminium phosphate, as 

 shown by Carnot,^ such loss is quite possible, especially since the 

 precipitate, obtained by the Schmidt-Hoagland method, is washed 

 with an abundance of dist. water. 



Precipitates of aluminium phosphate, prepared by the method 

 under investigation, were ignited, in coarsely granulär condition, in 

 an electric muffle to constant weights at different high-temp. levels. 

 They were then very finely pulverized in an agate mortar, and the 

 powders heated for 5 min. to dull redness, to expel atmospheric 

 moisture, and weighed. The powders were then re-ignited at the 

 corresponding temp's used previously to obtain constancy of mass, 

 and invariably lost weight. Illustrative data are given in Table 4.^ 



TABLE 4 



Data pertaining to further losses in weight of ignited, granulär, precipitates of 

 AlPOi, after thorough pulverization, desiccation and re-ignition. 



5 See Gmelin-Kraut : Handbuch der anorg. Chem., 1909 (vol. ii, pt. 2), p. 

 639, for comment on the great variety of basic phosphates. 



6 Bull. 107, U. S. Dep't of Agric, Bur. of Chemistry, 1912, p. 177. 

 ''Carnot: Compt. rend., 1881, xciii, p. 155. 



8 These, and the previously mentioned, precipitates of AIPO, were prepared 

 from potassium alum (p. 196). AIPO4, made from AICI3 and^(NH0=HPO«, 

 washed free from Chlorides and dried to constant weight at 110° C, also very 

 readily lost weight during streng ignition. 



