ALUMINA. 31 1 



me during the course of the summer of 1821, 

 and was repeated so frequently that I considered 

 myself as quite certain of its accuracy. But 

 after Mr. Phillips published his experiment on 

 sulphate of alumina, and his suspicion that 

 3*375 was the true weight of an atom of alu- 

 mina, I resolved to repeat the whole again. 

 This I accordingly did during the summer of 

 1823. The result which I obtained may be con- 

 sidered as absolutely identical, for the only dif- 

 ferences took place in the second decimal place ; 

 and the ratios of the constituents of alum de- 

 duced from this second analysis, did not de- 

 viate Toooth part from those of the preceding 

 analysis. 



If the weight of alumina were 3*375 instead 

 of 2*25, we should obtain from 60-875 grains of 

 alum no less than 10*125 grains of alumina* in- 

 stead of 6*75, which I obtained ; so that I must 

 have lost more than 30 per cent, of the alumina j 

 which, from the care with which the experiment 

 was made, 1 consider as impossible. 



But my analysis of alum is corroborated and 

 confirmed by the analysis of this salt made pre- 

 viously by Berzelius. When we correct his 

 statements by means of the more accurate data 

 given in this work, his analysis of alum is as 

 follows :* 



* Ann. de Clilm. LXXXII. 258. 

 U 4 



