Yol. 6^.~] CONSTANTS IN MINERALS AND IGNEOUS ROCKS. 149 



"When the final reading had been made, the top of the tube was 

 broken off, and the glass and beads were collected on a filter-paper. 

 They were then washed with carbon-tetrachloride and afterwards 

 with alcohol, the glass being preserved for further investigation. 



After crushing, the glass appeared to be almost entirely free from 

 bubbles ; if, however, any remained, the fragments enclosing them 

 rose somewhat above the general level of the zone, and could be 

 neglected. In the same way, any unfused crystalline particles 

 would tend to sink. 



Thus it can be seen that the more complete the fusion the better 

 defined will be the zone, and the very fact that it is possible to 

 ascertain the specific gravity of the zone ensures the rock having 

 been converted into a perfectly-homogeneous glass. 



The change in specific gravity having been thus ascertained, the 

 percentage-increase in volume of the rock on becoming glass can be 

 calculated from the formula 



V = ™ V = — = ^i • 

 1 d x ' 2 d 2 d 2 ' 



where d 1 = Density of rock, 



d 2 = Density of glass, 



Y l = 100, 



V 2 = Volume of glass. 



The following list comprises some of the results ascertained by 

 previous investigators. The numbers refer to percentage-increase 

 in volume : — 



Delesse. Bischqf. 

 Granite 9tollp.cent. | Granite 18*7 p. cent. 



Syenite 6 to 9 



Granitoid porphyry . 8 to 10 



Diorite 6 to 8 



Melaphyres 6 



Basalts 3 to 4 



Glassy lavas 1 to 3 



Basalt 7"4 



Trachyte 8*5 



Forbes. 

 Bowley-Rag dolerite... 6*36 



Carl Bams. 

 Diabase 11 



Nearly all the rocks experimented on come from well-known 

 localities, and have been analysed ; references to these analyses are 

 given at the end of this paper (p. 161). The melting of the powdered 

 rocks was watched through coloured glasses, and their behaviour 

 was found to vary only so far as their viscosity and coloration 

 were concerned. The acid rocks were hard to fuse, and crushing 

 and remelting had to be resorted to in several cases before a good 

 zone could be obtained in the diffusion-column. 



The colours of the glasses varied from a black almost metallic 

 appearance in the case of the gabbro, through stages of bottle- 

 green and pale-green, to the perfectly-transparent rhyolite-glass. 



Unfused crystalline particles could often be seen in the molten 

 glass, but these eventually disappeared after further heating. 



Bubbles seemed only to be formed at high temperatures ; these 



