﻿Penfield and Harper — Composition of JRalstonite. 383 



Ratio. 

 •183 ) 



•093 )• '278 1-00 



Mg 



4-39 



Na 



4'27 



K 



•12 



Ca 



•03 



Al 



24-25 



F 



39-91 



OH 



16-21 



HO 



10-12 



■002 



[ 



1-101 ) 

 •957 j 



•882 3-17 



3-058 11-00 



•562 2-02 



99-36 

 It will be seen that the assumption that hydroxyl replaces 

 fluorine not only makes up for the deficiency in the analysis 

 but also leads to a very satisfactory ratio. This assumption is 

 also well supported by actual experiment. When the mineral 

 is cautiously heated in a closed glass tube at first neutral water, 

 by stronger ignition acid water, is driven off. The first that 

 comes off is undoubtedly water of crystallization, afterwards 

 the hydroxyl is decomposed and fluorine comes off in combi- 

 nation with the hydrogen. By drying the air-dry powder at 

 100° C. there is a loss of only 0'10 per cent; by heating in an 

 air bath to a temperature never exceeding 250° C. the mineral 

 lost 10*37 per cent; the water goes out very slowly; the 

 experiment was carried on for over a week, during the last 

 three days of which the weight remained very constant. If 

 this 10*37 per cent is regarded as water of crystallization, the 

 remaining 8*36 per cent would correspond to 15*78 per cent of 

 hydroxyl, which agree closely with the figures in our last 

 an Jysis. The ratio of Mg : Na is almost exactly 1:1, there 

 seems to be no simple ratio between F and OH. The excess 

 of the aluminum in the analysis may be owing to some slight 

 impurity. We have never seen perfectly transparent, glassy 

 crystals of ralstonite, and their turbidity may be owing to 

 some slight decomposition ; if this is the case the alkalies 

 would naturally be most readily removed, causing the aluminum 

 to be too high. If our fluorine determination should be as 

 much as one per cent too low, which is probably not the case, 

 our results would not be materially changed. Using the actual 

 water of crystallization and hydroxyl determinations and 

 determining the fluorine by difference, we would have for the 

 latter part of our analysis : 



Ratio. 

 F 40*79 2*147 [ . . 



OH 15*78 -928 f 3075 " 7 " 278 1106 



H a O 10*37 *576-f-278 2-07 



Probably this determination of fluorine by difference 40*79 per 

 cent represents the true amount of that element more closely 

 than the results of our actual determinations. 



