RE-EXAMINATION OP AMERICAN MINERALS. 127 



York. The asbestiform mineral, carefully freed from the 

 chlorite and other impurities, gave on two analyses 



1 2 



Silica 58.20 58.47 



Magnesia 28.96 29.71 



Protoxide of iron 8.46 9.06 



Soda 88 .88 



Potash trace trace 



Ignition 2.26 2.26 



Alumina trace trace 



98.76 100.38 



These correspond to the formula R 4 Si 3 or ft 3 .si 2 +&Si. 



Atoms. At. weight. Percent. Oxygen ratio. 



Silica 3 1732.00 61.16 3 



Magnesia 3£ 875 00 30.90 \ „ 



Protoxide of iron \ 225.00 7.94/ 1 * 



This is the formula given for the last mineral, and the com- 

 position is that of an asbestus or magnesian hornblende. 



22. MONROLITE, IDENTICAL WITH KYANITE. 



This mineral was described by Professor Silliman, jr.,* as a 

 hydrous silicate of alumina resembling woerthite. Professor 

 Silliman, however, observes that the water varied in several 

 specimens examined from 3.09 to 1.84 per cent.; subsequent 

 examinations f made by one of us showed that the water in 

 the pure mineral was not over 1 per cent. 



In the analysis recently made we find that the silica and 

 alumina are the same as in kyanite, and that the high silica 

 obtained by the analyst quoted was undoubtedly owing to the 

 impurity of the mineral, as a careful examination with the 

 magnifier shows plates of quartz interlaced with almost every 

 specimen. The results of analysis on specimens carefully 

 selected to avoid the quartz gave 



1 2 



Silica 37.20 37.03 



Alumina ■.. 59.02") «i on 



Peroxide of iron 2.08/ D1 " yu 



Ignition 1.03 .85 



99.33 99.78 



These correspond to the formula A^l 3 Si 2 . 



*Amer. Jour. Science and Arts, 2d series, viii, 385. 

 t Dana's Mineralogy, third edition, p. 317. 



