16 PROCEEDIN-GS OF THE NATIONAL MUSBUM vol.68 



upon the idea that the different modifications are all pi'esent in the 

 solution in equilibrium with each other. In the supersaturated solu- 

 tion it is possible for the metastable nucleus to be first formed. If the 

 rate of change to attain equilibrium is small or if cooling is suf- 

 ficiently rapid, or if both of these conditions maintain, it is easily 

 possible for an unstable form to separate out. To some extent, there- 

 fore, the phase to form is conditioned by the rate of cooling. From 

 Ostwald's principle of the succession of the phases of a metastable 

 system, we would expect to find a tendency for the silica to pass from 

 glass through the intermediate unstable phases cristobalite and tridy- 

 mite to reach the stable form quartz. In that portion of the flow 

 where the escape of the solvent was rapid, the least stable crystal- 

 line phase, cristobalite, could form, and, in the absence of prolonged 

 action, persist. Where the action of the solvent is continued for a 

 longer period, the somewhat more stable form, tridymite, could form, 

 and, finally, where the action is of sufficiently long duration, the 

 stable form quartz might be attained. This does not imply that 

 cristobalite is first formed to be later transformed into tridymite 

 and quartz, for this certainly is not the case. The various forms of 

 silica once produced, persist with great tenacity, and the use of 

 mineralizers is usually required to change one into the other. 



Whether the solution attacking the glass was liquid or gaseous can 

 not be determined with certainty. At the low pressure and com- 

 paratively high temperatures prevailing, they were probably gaseous. 

 The thickness of lava above the solutions was not sufficiently great to 

 cause pressure high enough to keep the solutions liquid at the tem- 

 perature of the flow, and the slight, concentration of the solute prob- 

 ably had only a small, perhaps negligible, effect upon the boiling 

 point of the solution. Just how competent gases are to dissolve 

 silicate glass at a temperature of about 500° C. can not be told pre- 

 cisely at this time, but undoubtedly some solution does take place. 

 Certain lines of evidence indicate that the solubility of silica in 

 gaseous water may be fairly great at comparatively low temperatures. 

 The solubility, however, need not be great nor the amount of solvent 

 large, for the solvent in a case such as this simply acts as a catalyzer. 

 Its action is continuous. As rapidly as the crystalline phase sepa- 

 rates from its supersaturated solution, the solvent is capable of dis- 

 solving more glass, this process continuing until the transformation 

 is complete or until the escape of the solvent stops the reaction. In 

 this manner the final effect of a small amount of solvent can well be 

 the transformation of a considerable bulk of glass into crystalline 

 matter. 



The feldspars have apparently separated from the solutions be- 

 fore the quartz, tridymite, cristobalite, or fayalite, for invariably 

 these last minerals are perched upon the feldspar crystals. They 



