258 AGRICULTURAL ANALYSIS 



Drown was employed. 18 While the separation of silica by this 

 method is satisfactory, it was found impossible to completely 

 precipitate the aluminum phosphate in the presence of a thiosul- 

 fate. 



The presence of more than 1.25 grams of sulfuric acid pre- 

 vents the complete precipitation of aluminum phosphate, while 

 2.75 grams give a decided error. 



The Effect of Fluorin. The presence of a fluorid in a solution 

 from which it is attempted to separate aluminum by this method, 

 is as disastrous to the results as is the presence of sulfates. 



In none of the current methods is the presence of fluorin men- 

 tioned as a disturbing factor. 



The work so far done shows that alumina can be quantita- 

 tively separated as phosphate from a hydrochloric acid solution 

 containing aluminum, iron, manganese, lime, magnesium, sodium 

 and potassium, when only small quantities of sulfate are pres- 

 ent ; ,that the presence of silica in the solution produces a plus 

 error too large to be neglected ; and that the presence of large 

 quantities of sulfates or the presence of fluorids prevents the 

 complete precipitation of aluminum phosphate. Therefore, to 

 obtain accurate results, silica and fluorin must be removed while 

 sulfates, not more than the equivalent of 1.25 grams of sulfuric 

 acid, may be present. 



Proposed Method. The following method for estimating alum- 

 ina in phosphates is based upon the results of these experiments : 

 Treat one gram of the substance in a platinum dish with from 

 five to 10 cubic centimeters of hydrofluoric acid, let stand in the 

 cold from two or three hours, heat on the water bath to com- 

 plete dryness, add two cubic centimeters of concentrated sulfuric 

 acid, running well around the sides, and heat at a low tempera- 

 ture until the substance no longer flows in the dish. By this 

 process fluorin is completely expelled. Cool and add from 10 

 to 20 cubic centimeters of concentrated hydrochloric acid, and 

 warm a few minutes to soften the mass ; transfer to a small 

 beaker, and boil until all aluminum compounds are surely dis- 

 solved (from 15 to 30 minutes) ; filter from undissolved residue, 



18 Transact'ons of the American Institute of Mining Engineers, 1878-79, 

 7 : 346. 



