286 EARTHS PROPER. 



R. Philips has published some experiments on the 

 sulphate of alumina, from which he considers it 

 most probable that the atomic weight of alu- 

 mina is 3*375. This opinion, from a man of Mr. 

 Philips' precision and extensive chemical know- 

 ledge, induced me to repeat the whole of my 

 experiments, and to extend my researches as far 

 as possible, in order to throw as much light as I 

 could upon this very important, but difficult sub- 

 ject. I shall in this section lay the result of my 

 experiments before the reader. 



There are three methods, by means of which 

 ing the we may hope to come to an accurate conclusion 



atom of J , -i n i 



alumina, respecting the atomic weight 01 alumina. 1st. It 

 enters as a constituent into a great number of 

 crystallized minerals, which appear to be definite 

 combinations of acids and bases. By a rigid 

 analysis of these minerals in a state of purity, 

 and by our knowledge of the atomic weights 

 of all the other constituents, we may determine 

 the quantity of alumina which unites with a deter- 

 minate weight of the acid in the mineral, which 

 will give us the atomic weight of alumina, or 

 at least an approximation to it. 2d. Several arti- 

 ficial salts exist, containing alumina as one of the 

 constituents. A precise analysis of these salts, 

 in a state of purity, must lead to an accurate de- 

 termination of the saturating power of alumina, 

 and consequently give us its atomic weight. 3d. 

 Alumina enters into definite combinations with 



