452 Prof. B. SilUman oti some American Minerals, 



gonal ; folias easily separable, but inelastic and brittle. Co- 

 lour white ; transparent in thin foliae. Lustre, silvery, 

 vitreous and pearly. Hardness, 3-5. Gravity, 2995. B. B. 

 in forceps, exfoliates and emits a strong light; fuses on the 

 edges of thin laminae. In the close tube it yields water, and 

 gives very feeble traces of fluorine. It behaved with the fluxes 

 like the Turkish mineral. A qualitative analysis showed the 

 presence of silica, alumina, lime, magnesia, soda, a trace of 

 potash and iron, water and fluorine, the last in very feeble 

 quantity. 



The quantitative analysis of this species is still incomplete 

 as to its alkaline constituents, which are given by the difference, 

 and the amount of water is probably placed too high*. The 

 analysis was conducted, under my direction, by my pupil, 

 Mr. W. J. Cravve. Three analyses gave him as follows : — 



I. 11. irr. Oxygen. 



Silica .... 32-311 31-060 31*261 16-24: = 4 



Alumina . . • 4.9-243 51-199 51-603 23-74. 6 



Lime .... 10-663 9-239 10-1461 



Magnesia . . . 0-298 0-283 0-499 k 3-42 ^j,! jl 



Soda and potashf 2-215 2*969 1-221 J ^ 



Water .... 5-270 5*270 5-270 4.*72 1 



100-000 100-000 100-000 

 It is obvious that this is very nearly the true constitution of 

 the mineral. The following formula corresponds very closely 

 with the analytical results, viz. — 



SRO, 4Si03, 6A103 + 3HO = 3(CaO, NaO, MgO)Si03 

 + 3(2A103, Si03) + 3l-IO. 



4 atoms Silica . . 2309-24 30*51 per cent. 



6 ... Alumina . 3854-00 50-92 



3 ... Lime . . 1068-06 14-11 Ai','Oi8E 



3 ... Water . . 337-44 4-46 



7568-74 100-00 



This result leaves but little doubt that the mineral here 

 examined is the same as the Turkish emerylite. That the 

 American species will be found constant in containing water 

 I have no doubt. 



Great care was bestowed on the trials to detect zircon ia, 

 but none was found. 



Corundellite. '^^ 



The next mineral belonmng to this series Ihave called'* 



Corundellite. This species in external characters much re** 



• The mean of two determinations. f By the differenci^ ^^fj 



