and Rocks under Pressure. 267 



The phenomenon of rupture by fiaking-off is independ- 

 ent of other phenomena accompanying high stress. Some 

 substances develop cracks at the same time that they 

 erode ; the number of cracks may be great as in calcite, 

 or small as in quartz. Or the erosion may be accompa- 

 nied by no cracks whatever, as in feldspar, porphyry and 

 andesite. The substance may show no viscous flow dur- 

 ing erosion, or it may flow like granite and baryte. The 

 formation of cracks was never in these tests the cause of 

 final rupture, except with glass. Cracks are probably 

 in many cases due to the attempt of the solid to slip 

 bodily into the cavity, but such slip can never go far 

 before it is stopped by the mutual supporting action of 

 the walls. Such slip may be prominent in a substance 

 with easy cleavage, or slight as in quartz. It is probable 

 that the cracks in quartz and calcite were essentially the 

 same in character, one being merely more prominently 

 developed than the other. 



Flaws in the original specimen are apparently so 

 tightly closed by pressure that they play no part in 

 fracture. 



Rupture by flaking-off is not even suggested by any 

 mathematical theory of rupture, and probably cannot be 

 in the nature of things. Mathematical theory treats the 

 material as mathematically homogeneous, whereas we 

 probably have to do here with a phenomenon of molec- 

 ular agitation. All the observations are consistent with 

 the view that a microscopic splinter flies off when its 

 kinetic energy of temperature agitation has by chance 

 become sufficiently higher than the average. The ten- 

 dency to fly off will evidently be higher when the stress 

 is high, and if the stress is high enough the process of 

 disintegration will become rapid enough to be appre- 

 ciable. Thermodynamics is familiar with something 

 similar when the vapor pressure of a liquid is increased 

 by an increase of pressure acting on the liquid phase. 

 One expects that there will be no sharp point at which 

 spontaneous disintegration suddenly begins, but that the 

 effect continues over a range of pressure. This is con- 

 firmed by the experiments above. On the other hand, as 

 with all such effects, there must be a pressure at which 

 the effect practically ceases, so that at low pressures the 

 chance of disintegration is of the same order, for exam- 

 ple, as the chance that a pail of water will freeze on a red 

 hot stove. 



