ESTIMATION OF ALUMINA AND FERRIC OXID 253 



is present, by the modification of Grueber no alumina at all is 

 obtained sometimes, when it may be present to the extent of half 

 a per cent. The results of the investigation favor entirely the 

 adoption of the method of Lasne to the exclusion of the others. 9 



221. The Estimation of Alumina and Ferric Oxid in Natural 

 Phosphates. The search for an accurate and rapid method for 

 the determination of alumina and iron oxid in the presence of 

 phosphoric acid has occupied the attention of analysts for years, 

 and many methods have been proposed for this difficult opera- 

 tion. It may be said generally that even those methods that 

 have stood the tests of extended use have not escaped severe 

 criticism; they are only accurate within narrow and rigidly de- 

 fined limits or they are tedious and time-consuming. 10 



Aside from its interest from the scientific point of view, this 

 subject is of importance in its technical and commercial aspects. 

 The value of raw mineral phosphates is judged largely by their 

 content of alumina and iron. 



In phosphatic slags the estimation of these oxids is more 

 difficult, though possibly not so important. 



Sources of Error in the Older Methods. The Glaser alcohol, 

 the acetate with its various modifications, and the caustic alkali 

 methods as carried out by Lasne, Lichtschlag and Gladding, 

 have all been criticised and the sources of error pointed out. 



1. In the Glaser alcohol method the precipitation of manganese 

 with the iron and aluminum phosphates and the solubility of the 

 phosphates in the wash-water are important sources of error. 

 Probably the manganese can be eliminated by a second precipita- 

 tion in the presence of a large amount of ammonium chlorid. 

 Possibly the presence of a large amount of ammonium sulfate 

 may also affect the accuracy of this method, in those cases where 

 the excess of ammonia is completely removed by boiling. Alumi- 

 num phosphate is noticeably soluble in a strong sulfate solution, 

 which is neutral or faintly acid from SO 2 . 



2. In the acetate method and its variations there is an error 

 caused by the precipitation of the lime with the iron and alu- 



9 Chemiker-Zeitung, 1897,2! 1*414. 



10 Veitch, Journal of the American Chemical Society, 1900, 22 : 246. 



