﻿Penfield and Harper — Composition of Ralstonite. 385 



In these analyses the ratio (MgNa 2 ) : Al=l : 3 very nearly, 

 especially considering the small quantity of mineral which was 

 used in making the analyses and in Brandl's analysis, the only 

 complete one, the fluorine is just sufficient to satisfy the metals, 

 while the ratio (MgNa 2 ) : Al: F : H 2 = 097: 3 : 10*88 : 2*02, or 

 nearly 1:3:11:2, the same as required by the formula proposed 

 by us, only in this case the compound is a pure fluorine compound, 

 containing no hydroxy 1. Following Brandl's analysis we have 

 given the percentage composition calculated from the above 

 ratio in which the Mg : Na=l : 1. It will be noticed how closely 

 the figures agree with the analysis of Brandl. 



The great difference in the proportion in which the metals 

 are united, (MgNa 2 ): Al = l : 3 in ralstonite and Ca:Na:Al = 

 1:1:1 in thomsenolite, would account for the decided change in 

 the formula derived from an analysis of a mixture of ralstonite 

 with a little thomsenolite. Brandl's formula, 3(Na 2 MgCa)F 2 , 

 8A1F„ 6H 2 0, is therefore a little too low in A1F 3 . The differ- 

 ence in the deportment of the mineral when treated with strong 

 sulphuric acid may be owing to the fact that although the 

 fluorine compound is readily decomposed by that acid so that 

 Brandl was able to determine the fluorine by driving over SiF 4 , 

 the hydroxyl compound in our mineral in some way hinders 

 the decomposition of the fluorine compound, perhaps by being 

 in itself with difficulty decomposed and inclosing and thus pro- 

 tecting some of the molecules of the fluorine compound from 

 decomposition. 



In thin sections under the microscope all of the ralstonite 

 appears very transparent and free from visible inclusions and 

 decomposition products. Some of the crystals on the original 

 specimen were colored yellow and where one of these had been 

 cut through, the yellow substance was seen to consist of a very 

 thin film, probably of iron oxide, coating the crystal. The 

 larger crystals were zonal in structure, the zones lying parallel 

 to the faces of the octahedron. This zonal structure is scarcely 

 perceptible in ordinary light, being indicated by faint greyish 

 streaks running parallel to the contours of the cross section, 

 which could not be resolved by the use of high powers into 

 visible inclusions. In polarized light the zonal structure was 

 more perceptible ; all of the crystals show slight double refrac- 

 tion and a division of the cross sections into fields reminding 

 one of the double refraction of analcite. The slightly double 

 refracting ralstonite with its absence of cleavage is in marked 

 contrast to the strongly double refracting thomsenolite, showing 

 brilliant polarization colors, blue of the second order, distinct 

 cleavage and inclined extinction. 



In closing we wish to express our thanks to Professor Geo. 

 J. Brush for his liberality in providing us with the rare mate- 

 rial for carrying on this investigation. 



Mineralogical Laboratory, Sheffield Scientific School, Oct. 12th, 1886. 



