150 MR. J. A. DOUGLAS ON CHANGES OF PHYSICAL [May I907, 



increased in size by coalition and rose to the surface, where they 

 were expelled. In the acid rocks, however, owing to their greater 

 viscosity, they were extremely difficult to drive off. 



The molten glasses, at temperatures well above their fusion- 

 points, more especially the basic forms, had the appearance of 

 boiling ; this, however, was not o£ long duration. 



The majority of the rocks examined were easily fused, giving 

 clear glasses of a bottle-green colour, almost free from bubbles. 



The following cases, however, require special mention : — 



1. Shap Granite. — A somewhat larger specimen of this rock was crushed 



up, owing to the presence of porphyritic orthoclase-crystals. It was 

 very hard to fuse, and had to be crushed and remelted twice before a 

 zone could be obtained. The glass was of a pale-green colour. 



2. Peterhead Granite. — Fused with difficulty; was crushed and re- 



melted. For some time small brown specks remained undissolved, 

 these being probably biotite. The glass was colourless, and gave one of 

 the best-defined zones. 



3. Eh yo lite. — Was very hard to fuse, containing many bubbles which 



were driven off with difficulty. Glass perfectly colourless and trans- 

 parent. 



4. Gabbro. — Very low fusion-point. Glass black and opaque, with an 



almost metallic lustre ; flooded the ribbon at high temperatures. 



5. Quartz-Enstatite-Diabase. — Fused with difficulty, probably owing 



to the presence of quartz. The glass was crushed and remelted. 



In no case did there appear to be any action on the platinum- 

 ribbon. I append (on p. 151) a tabulated list of the rocks that were 

 experimented on, showing the change of specific gravity and per- 

 centage-increase of volume on fusion. It will be seen that this 

 increase attains a maximum in the acid plutonic rocks, and decreases 

 with an increase of basicity. The presence of quartz in more basic 

 rocks seems to affect readily the increase of volume ; this is very 

 noticeable in the tonalite and quartz-enstatite-diabase, the latter 

 having an expansion nearly as great as that of Shap granite. The 

 three specimens of andesite resemble one another closely. 



These results, although slightly lower in many cases than those of 

 Delesse, do not differ from them to any great extent. They merely 

 serve to demonstrate, by an entirely- different method, the great 

 expansion which must eventuate when an igneous rock becomes 

 converted into the molten state. A considerable contraction takes 

 place when a molten magma solidifies as glass, and this must be 

 greatly increased if crystallization occurs. 



A comparison may, I think, be drawn between the conditions 

 under which the enormous lava-flows of Iceland, India, and North 

 America, known as ' fissure-eruptions,' were formed, with the con- 

 ditions which may be obtained in a small platinum-furnace, such 

 as the one used in the foregoing experiments. The behaviour of the 

 gabbro may be taken as typical of the other basic rocks. 



It was found that, on placing a small portion of the powdered 

 rock at the bottom of a U-shaped loop in the ribbon, and suddenly 

 raising the temperature to a point many degrees above its fusion- 

 point, an instantaneous melting took place, and so great was the 



