Prof. B. Silliman, Jr., on some American Minerals. 387 



quently analyzed by Dr. Thomas Muir. in the Laboratory of Dr. 

 Thomson, who found in it a large quantity of zirconia, an obser- 

 vation which all subsequent researches have failed to confirm. 

 Since that time it has been analyzed by various chemists, viz. : 

 by Connel, Norton, Staff, Hayes and Thomson. The most re- 

 cent of these analyses which has been published is that by 

 Thomson, who reports it to contain 45*65 per cent, of silica. 

 We have then the following discordant results in the amount of 

 silica found in Siilimanite by different chemists in the order of 

 their publication : 



1. 2. 3. 4. 5. 6. 1 



Bowen. Muir. Connel. Norton. Staff. Hayes. Thomson. 



Percent. 42-67 3867 3675 3740 37-36 4260 46 65 



The cause of this disagreement will undoubtedly be found in 

 the difficulty of effecting a complete decomposition of anhydrous 

 silicates of alumina, which contain a high percentage of alumina. 

 This decomposition can be completely effected only by the aid of 

 caustic potash applied to the mixture of carbonates and the min- 

 eral during the fusion, as first recommended by Berzelius, or by 

 fluo-hydric acid. 



Select crystals of this mineral were taken from the original lo- 

 cality at Chester, Conn., and their analysis afforded the following 

 results: Quantity taken, 775*5 grammes, found, 



Silica, 



Alumina, .... 



2 atoms silica 1154-62 



3 " alumina 1927-00 



•292= 

 •484 



-perct. 



37653 

 62-411 



•776 



100064 



Si 





Required. 



37 47 



Al 





62 53 



308162 10000 



• •• 



This result gives then exactly the formula of kyanite, viz 



Al 2 Si 3 . The analyses of Staff and Norton give also the same 



result* 

 We 



t 



Bucholzite is a name given by Brandes to a silicate of alumina 

 from Tyrol, which occurs in compact masses of a finely fibrous 



In Prof. Norton's analysis, which was made in Yale College Laboratory, the ex- 

 „ of 273 was owing undoubtedly to aluniinate of pota,h which remained with 

 the alumina after separating the peroxyd of iron by caustic potash. Subtracting this 

 8 um from the sum of alumina and peroxyd of iron, we have 6230 per cent, alumina 

 and peroxyd iron, which is almost exactly the quantity required by theory, and I 

 nave corrected the analysis accordingly with the consent of Prof. Norton. That 

 analysis was made on the siilimanite from Fairfield, New York. 



T In his translation of \tnlw vr.1 Hi 1 fU. 



