SPHERULITES AND LITHOPHYS^ 275 



Another fact which bears on the present problem is the increase in 

 solubility of gas in a liquid with falling temperature. The effect of this, 

 if pronounced, would be, in the case of a simple bubble, a reduction in 

 its size with lowering temperature. Opposing such reduction, however, 

 is the hydrostatic tension which develops in the central portion of the 

 magma on cooling and which tends to enlarge the bubbles. The ultimate 

 effect which these opposing forces may have had on the bubbles in Ice- 

 land obsidian is not known, but the fact that the cavity was lined with 

 crystallized material would tend to retard the magmatic resorption of 

 the gas, and thus tend to produce larger cavities than othei-wase. 



To summarize the conclusion briefly: Gas escaping from the. magma 

 on crystallization was an active factor in the development of the lithu- 

 ])liysie in the ol)sidian from Hrafntinnuhryggur. It caused rocrystaliiza- 

 tion and aided to a large extent in enlarging the cavity as crystallizatiun 

 proceeded. The amount of energy required to effect the observed re- 

 crystallization in the cavities need not have been great, because the energy 

 necessary for the solution of the spherulite crystals first formed was prob- 

 ably largely oft'set by the energy liberated during the crystallization of 

 the lithophysal minerals. The shrinkage of the viscous lava on cooling 

 tended, of course, to reduce the uniform hydrostatic pressure; but the 

 chief effect of such reduction on the size of the bubble cavities was to 

 increase the :ate of evolution of gas from the magma (reduction oL' 

 solubility of gas in magma under reduced pressure). Shrinkage of the 

 magma alone without evolution of the inclosed gas would tend to produce 

 t;racks (Jomts) bearing some relation to the inclosing walls. That the 

 magma contained abundant gas, however, is proved Ijoth hy the presence 

 (if piniiiceous layers and ])\ the recrystallizing action of the gas on the 

 walls of the cavity. The conclusion seems, therefore. \\;tir;uite(l that in 

 i]\v II laruiiiiiiuhryggnr obsidian, and probably in most obsidians, the 

 pressure of tlie gas set frae from the magma as a result of ciTstallizatiou 

 and also of r(>duction of hydrostatic ])rcssure induceil h\' llie shi'inkinu 

 of the ceiili'al [)()i'tions of Ihe inagma on C(.)oling has heen an inipoi-tant 

 faeloi- in llie develo])n)enL of the lithophysa-'. To ascrihe the total effect 

 to Ihe nnifonn |iull or tension developed hy shi'inkage of the cooling 

 magma is not an ailci|iiali' hypothesis to account for iln' din'ei'cnt facts 

 anil iclations uliich haxc hccii ohserxcil. 



h!j;<'OM>M{y MIXFJlALfi A.\D ETVlliyd PHENOMENA PnODVCED fiV JTOT 



aih'CULATING SOLUTWNii 



AlLlioiigh not sti'ictly germane to the theme of this paper, it may he 

 of inlei'esl to desi-rihe Ihe cITects pi'otliiccd hv hoi solutions on the nioi'e 



