Chemical Constitution of SlUimanite* 235 



2.92 9G.G8 



I had not sufficient material to ascertain wliether the loss 

 on the analysis was due to the presence of any alkali or 

 other substance ; but as the other analyses do not indicate any 

 other constituents than those which have been under our 

 view, except indeed a half per cent, of water, I believe the 

 loss to have been merely due to causes incidental to the em- 

 ployment of so little material. 



On the whole, there seems every reason to regard Mr Hai- 

 dinger's view of the nature of the mineral to be correct. The 

 analysis of varieties of Disthene by Klaproth and Laugier do 

 not greatly differ from those of Mr Bowen and myself of Silli- 

 manite. They give, 



Klap. Laug. 



Silica, . . , 48.0 38.5 



Alumina, . . . 55.5 55.5 



Oxide of iron, . . 0.5 2.75 



Lime, , . . ... 0.5 



Water, . . . ... 0.75 



99.0 98.00 



I conceive, however, that the formula which will best express 

 the constitution of this species, including both Disthene and Sil- 

 limanite, when in a state of complete purity, is AL' S * 

 (A^ S ^), which Dr Thomson, founding on the analyses of Dis- 

 thene by Ardwedson, has suggested as the formula for that 

 mineral.* It gives, 



Silica, 4 . . . . 37.47 

 Alumina, 02.52 



99.90 



One of Ardwedson's analyses gave exactly these propor- 

 tions, and the others did not deviate very considerably. On 

 this view, the mineral is a subsesquisilicate of alumina. 



* Outlines i., 242. 



