44 JOURNAL OF THE 



TABLE IL 



Fusion method. Philadelphia method, 



1. 11.70 11.67 



2. 14.12 13.93 



3. 13.52 13.50 



4. II. 19 II. 10 



The painful precautions which Fresenius recommends to remove 

 cidorides from a solution of phosphates before adding molybdic so- 

 lution are, I believe, entirely necessary. But that the presence of 

 free HCl is objectionable is shown by the following results, obtained 

 from identical HCl solutions of phosphates, but in the one case pre- 

 cipitating in the. HCl solution direct, and in the other neutralizing 

 the HCl with NH.HO and adding a little HNO3. 



It will be noticed that the difference is considerable in all cases 

 except 5 and 8, and is on the other side in one case only (6), due, 

 perhaps, to some error. 



It seems to be a mooted question how far the presence of silica in 

 solution effects the results in a determination of phosphoric acid. 

 If it does form a precipitate and thereby makes the results too high, 

 then it is the only objection I know of to the fusion method, for the 

 NaoCOs necessarily fuses some of the sand present in a fertilizer 

 into soluble silicate. I resolved, therefore, to separate soluble silica 

 from four solutions by fusion and compare per cents of P2O5 with 

 per cents found before separation of silica. The silica was separated 

 by careful evaporation over ths water bath, transferring to hot air 

 bath with higher temperature until all acid (HNO3) was entirely 

 removed, moistening with strong HCl, drying and heating thor- 

 oughly again, and finally taking up in hot HoO with some HNO3 



