LITHOPHTSES IN THE OBSIDIAN OF THE ROCCHE ROSSE. 441 



•which we have described, and thus to the original presence of vesicles 

 in the rock. 



Commonly, however, the process of infilling does not go so far as 

 this. On the ends of the delicate branching fibres of felspar, plates 

 of tridvmite are deposited ; and this seems to close the growth of 

 the 'stalk' or of the 'bridge' (PL XII. figs. 3 & 4). The angles and 

 edges of the tridymite crystals impart a serrated and toothed 

 character to the spheroids within the vesicles, as is seenwhen the 

 rock is viewed with an ordinary lens. 



That the lithophysal structure in the Lipari obsidians was 

 developed during the cooling of the mass, and not by subsequent 

 amygdaloidal infilling of vesicles, is clear from the formation of the 

 compact spherulitic zone in the glass itself, and from the rooting 

 of the clearer fibres also in the glass, not merely upon the walls of 

 the cavities. "We conceive that in highly siliceous lavas the liberation 

 of water, on relief from pressure, renders the mass rapidly more 

 viscid. The heated vapour is liable to be retained after it has 

 expanded, and even to become superheated as fluctuations of 

 temperature travel through the mass. That solidified portions of 

 the Rocche Rosse lava have thus become reheated has been shown 

 in a previous paper .^ While Mr. Iddings^ has suggested that 

 crystallization occurs in regions of greatest hydration of the glass, 

 the water being liberated and forming vesicles during the develop- 

 ment of the crystals, we should be inclined to regard the loss of 

 water on relief from pressure as in itself an important factor in 

 consolidation. W^here the water has escaped freely, we have the 

 pumice or the occasional vesicular layers. If the bubbles are very 

 numerous, they open into one another, the extreme result of such 

 confluent vesicles being the ' thread-lace scoria ' of Prof. Dana.^ 

 If the vapours escape thus completely from the lava, their action 

 upon the rapidly cooling and often filamentous glass around them 

 is very transient or absolutely inappreciable. 



W-'here, however, the vapours are confined, they may be kept at a 

 high temperature for a considerable time, the vesicles being farther 

 apart, and each becoming a sphere of hydrothermal action. If the 

 surrounding glass remains at a temperature little below its fusion- 

 point, crystallization will be promoted in it ; and at the same time 

 the hydrothermal attack of the vapour in the vesicle will produce 

 reactions upon its walls. 



Thus we conceive the zones formed around the vesicles to result 

 from the crystallization of the constituents of glass that has been 

 kept at a high temperature ; and much interstitial glass remains in 

 the outermost porcellanous region. The vapour, however, attacks 

 and etches the walls of the cavity, like steam in an ordinary glass 



^ Gr. A. J. Cole,' Devitrification of cracked and brecciated Obsidian,' Mineraloar. 

 Mag. vol. ix. (1891) p. 272. 



2 ' Spherulitic Crystallization,' Bull. Phil. Soc. Washington, vol. ix. p. 447 ; 

 •Nature and Origin of Lithopbysje,' Ainer. Journ. Sci. vol. xxxiii. (1887) 

 p. 43. Also Seventh Ann. Report U. S. G-eol. Survey, p. 285. 



^ ' Cbaractetistics of Volcanoes,' p. 164. 



