Dr. Johnston- Lavis — Litliophy&eB in Lipari Obsidian. 489 



tbe cones would tend to become more acute. Crystallization at the 

 centre of the spherulite would be fairly complete, but as growth 

 took place the supersaturation of the surrounding glass would tend 

 to cause much of such glass to remain between the fibres unindi- 

 vidualized ; the outer surface fibres would again jam tight on account 

 of the relief of solution-tension by vesiculation. Thus there would 

 be more interfibrillar glass in the middle zone and more contraction 

 in that region when the spherulite breaks up. The chief cause, 

 however, of the concave sides to many of the cones is that they are 

 sectionized obliquely, and assume the concave-sided look through, 

 the microscope and an appreciable thickness of the section. 



After vesiculation, growth continues normal to the vesicle wall 

 and therefore no longer parallel with the fibres of the conical wedge, 

 hence the mushroom or fan-like appearance. 



Sometimes this wedge or cone is pushed into the glass by the 

 expanding vapour behind it, as in your fig. 3. 



You will see that I differ from you in holding that not only large 

 part, but all of the felspar needles have developed from witMn 

 outwards. 



Of course I consider that tridimite, fayalite, magnetite, haematite, 

 are true sublimates. If spherulites assume large size they become 

 Hthophysal for a second, third, and so on, spherical shells of vapour 

 form around the new shell of spherulitic matter, and repeat a similar 

 process to what occurred in the nucleus. Your figures are very 

 good and characteristic. I interpret them as follows : — 



Fig. 1. — Rapid and intense vesiculation around very small 

 spherulites, which in the upper cavity have grown out a little, 

 having for the most part been reduced to fine dust spread over the 

 bubble walls. In the other vesicles the expansion and cooling was 

 so rapid that no further growth could take place from the minute 

 cones attached to the walls. 



Fig. 2. — Well shows the broken-up cones and the difference in 

 their circumference curve and that of the enlarged vesicle and the 

 independent direction of the new formed fibres. 



Fig. 3. — Shows one cone holding remains of nucleus of spherulite. 

 It has been pushed into the wall of the vesicle by the expansion of 

 the vapours on the opposite side. 



Fig. 4. — Shows the union of one vesicle around several spherulites. 



These phenomena occur only where expansion can take place from 

 slight pressure. 



The difference between my interpretation and yours seems to lie 

 essentially in this — that your paper explains the universal association, 

 in your specimens of cavities and spherulitic growths, as due to 

 great part at any rate of these growths in the Lipari Obsidians 

 having been initiated by the formation of steam vesicles ; while I 

 hold that such association may be attributed — (1) to the liberation 

 of steam during crystallization ; and (2) to the fact that the crystalline 

 aggregates (spherulites) when formed would afford surfaces favourable 

 to the disengagement of gas from the surrounding glass. 



It is not only towards the surface of the Kocche Eosse lava stream 



