Dr. T. Thomson on Sillimanite. 537 



it quite free from zirconia, and identical in its composition 

 with bucholzite, which is a silicate of alumi^ia. How Dr. 

 Muir fell into the mistake I do not know. I looked over all 

 my specimens, and found in one small crystals which had the 

 aspect of zircons. It might be that the same crystals existed 

 in the specimen which lie analysed. That it was impure is 

 obvious from the quantity of iron which he obtained. 



The mistake of Dr. Muir had escaped my memory till it 

 was recalled by the notices in the new edition of Phillips's 

 Mineralogy by Mr. Alger. To set mineralogists right 1 re- 

 peated the analyses of sillimanite which I had made several 

 years ago. This I shall give at the end of this notice, after 

 drawing up a short description of the mineral. 



Sillimanite^ bucholzite and ^brolite are all synonyms of the 

 same mineral, the chemical constitution of which is Al S, or a 

 simple silicate of alumina. 



In all the specimens which I have seen the mineral is in 

 long crystals, consisting of oblique prisms, with angles of about 

 80° to 100^. 



The colour is gray or rather mottled, in some places nearly 

 white, in others gray, more or less dark. The texture is fo- 

 liated, or in some crystals fibrous. The lustre vitreous. The 

 hardness 5, or about that of phosphate of lime. It is opake, 

 or only slightly translucent when in thin plates. It is easily 

 frangible, especially in the direction perpendicular to the axis 

 of the prism. 



The specific gravity in very pure crystals I found to vary 

 from 3-200 to 3-357. 



Before the blowpipe it becomes white but does not fuse. 

 With soda it forms a bead, transparent and red while hot, but 

 becoming white and opake when cold. With microcosmic 

 salt it forms a colourless bead. 



Being subjected to analysis, I found it composed of — 



Silica 45-65 



Alumina 49-50 



Protoxide of iron ...<. 4*10 



99-25 

 This is exactly 



20*75 atoms of silica, 

 20 atoms of alumina, 

 0-827 atom of protoxide of iron. 

 From the colour of the mineral there could be no doubt 

 that the iron was in the state of protoxide ; while from its 

 mottled appearance it was evident that the iron was not com- 

 bined but mixed unequally with the pure sillimanite. If we 



