104 J. W. Mallet, Esq., on the Analysis of Enclose. 



niometer. The mean results of some angular measurements 

 over the obtuse lateral edges of four distinct vertical prisms 

 were, 115° 6', 127° 51', 140° 44', and 149° 32', all of which 

 agree nearly with numbers given by Phillips. The only 

 cleavage I observed was that parallel to the terminal plane, 

 replacing the acute lateral edge of the vertical prism, which 

 is mentioned in mineralogical systems as the only cleavage 

 easily obtained. 



The specific gravity of these fragments was 3-036. They 

 were reduced to fine powder, and fused with the mixed car- 

 bonates of potash and soda, and the analysis was then con- 

 ducted according to the usual routine for silicates. The alu- 

 mina and glucina were separated according to the old method 

 by carbonate of ammonia, as from previous experiments I 

 found the use of caustic potash, which has been more recently 

 proposed for this purpose, both difficult and uncertain. The 

 analysis gave the following constituents per cent. : — 



Atoms. 



Silica 



44-18 



, 



950 



Alumina 



31-87 



. 



620 



Glucina 



21-43 



t 



564 



Peroxide of iron 



1-31 



. 



016 



Peroxide of tin 



•35 







99-14 

 These numbers agree very fairly with those of Berzelius, 

 and dividing by the atomic weights of the several constituents, 

 give their equivalent proportions as in the second column. 

 These are very nearly in the ratio : — 



Si0 3 : A1 2 3 : G 2 3 = 3:2:2. 

 And hence we have the formula : — 



2(Al 2 3 Si0 3 ) +G 2 3 Si0 3 . 



Or if the two earths, alumina and glucina, be isomorphous : — 



4(Al 2 3 + G 2 3 )3Si0 3 . 



Scacchi, taking glucina as a protoxide, suggests an ana- 

 logy between euclase and epidote, but if the corrected atomic 

 weight of this earth be U3ed, the formulae of these two minerals 



