212 Prof. Connell on (he Elie Fi/rope or Garnet. 



As there was thus some loss on the analysis, a search was 

 made by ordinary processes for any constituent which might 

 have been as yet unnoticed, such as moisture, phosphoric and 

 fluoric acids and alkalies, but no such constituents were found. 

 I have reason to believe that the loss fell chiefly on the 

 magnesia, no steps having been taken to separate any farther 

 traces of that constituent which might have escaped precipi- 

 tation by the carbonate of potash. 



Even if the oxide of iron in this mineral were held to be 

 protoxide, there would be quite as much difficulty in bringing 

 the result under the garnet formula as there is in bringing the 

 leading analyses of Bohemian pyrope under it. This circum- 

 stance, as well as the general conformity between the result 

 here given, and the analyses of pyrope refei-red to, comprising 

 those of Klaproth,Wachtmeister, and Von Kobell, particularly 

 as respects the considerable quantity of magnesia, and the 

 comparatively small quantity of oxide of iron, although with 

 some apparent discrepancy as to the state of oxidation of the 

 latter substance, tend to shew a close connection between the 

 Elie mineral and pyrope. The occurence, also, in the former, 

 of a very minute but decided trace of oxide of chrome, appears 

 still farther to establish this relation, and perhaps shews that 

 that substance can hardly be viewed in pyrope as an essential 

 constituent. 



The specific gravity leads to a similar view as to this con- 

 nection ; for, whilst that of precious garnet is upwards of 4, 

 that of pyrope is only 378, and that of the Elie mineral 3-661, 

 as already stated. 



If we should hold the iron to be in the state of peroxide, 

 as the reactions, so far as tried, indicated, we might repre- 

 sent the constitution of the Elie mineral by the formula 



C } »S' -f 3 =; > S ; and, on a similar view, this formula would 



Mn\ ^ ) 



not be much at variance with the leading analyses of Bohe- 

 mian pyrope ; but, as the iron in pyrope is usually reckoned 

 protoxide, such a view would, I fear, hardly be admitted. 

 The colour of the glass obtained on fusing garnets, seems, 

 however, a doubtful source of evidence on that point. Thus, 

 melanite gives a black glass with the blowpipe, although the 



