396 S. L. Penfield — Anthqphyllite from Franklin, N. C. 



We see from the above that the optical axes are about at right 

 angles to one another. There is a marked dispersion about 

 the a axis, which is p^>v, the brachy axis being the acute bisec- 

 trix for green and yellow but obtuse for red. One of the pris- 

 matic crystals after polishing the "faces served as a prism for 

 determining two of the indices of refraction. The prismatic 

 angle after polishing measured 52° 55'. The minimum devia- 

 tion for yellow was 41° 0' for rays vibrating parallel to the 

 vertical axis, and 40° 14/ for rays vibrating parallel to the 

 macro-axis, from which we calculate y = 1*6404 and /? = 1*6301, 

 while /? above was found to be 1*6353. From the values of 

 2 V, y and /3 for yellow we calculate a— 1*6288. The optical 

 orientation is therefore a = a, h = b and b = c. The double 

 refraction is negative for green and yellow, a being the acute 

 bisectrix and positive for red, c being the acute bisectrix. 



Hardness about 6. Specific gravity, by floating in the heavy 

 solution, 3-093. 



The purest crystals were selected for analysis by hand- 

 picking which was easily accomplished as the mineral separated 

 readily from the matrix. 



The results of analysis on the air-dry powder are as follows t 



Eatio. 

 •966 

 •1441 

 •005 J 

 •717 }> -963 

 •004 | 

 •093 J 



69-99 



The ratio of Si0 2 : RO (H 2 being included) is *966 : *963 

 or almost exactly 1:1, giving the formula RSiO s , R = Mg, 

 Ye, H 3 and traces of Mn and Ca. The mineral when heated 

 in a closed tube over a Bunsen burner was found to be anhy- 

 drous and the analysis was made accordingly. After drying 

 the powder at 100° C , the mineral was heated to faint redness 

 in a crucible and lost only 0*19 per cent. On summing up the 

 analysis, which was very carefully made and everything tested, 

 a deficiency of 1*50 per cent was found and moreover the ratio 

 of Si0 2 : RO = -966 : *870. which indicated a large excess of 

 silica. In trying to account for this deficiency a second sample 

 of the mineral was heated over a blast lamp in a closed glass 

 tube when water was given off in perceptible quantities. The 

 H„0 in the analysis was determined by loss of weight on 

 igniting in a current of CO„ gas to prevent oxidation. A 



Si0 2 



57*98 



FeO 



10-39 



MnO 



•31 



M^O 



28-69 



CaO 



•20 



H 2 



1-67 



A1 2 3 - 



•63 



Loss at 100 



3 "12 



