L'tit; P. }]'. Bridgman — Failure of Cavities in Crystals 



contact, for otherwise there would be welding- of the 

 grains. When pressure is released, the lens-shaped cav- 

 ities tend elastically to recover their form, but are pre- 

 vented by the air films, which in such excessively thin 

 layers will act like an extremely sticky and viscous glue. 

 The slow viscous yield of these thin films of air accounts 

 for the visibly slow break-up into laminae. 



This explanation suggests that the figures given above 

 for the densities of compressed powders must be' con- 

 siderably less than the density when actually under pres- 

 sure, and that for example the interstitial space in quartz 

 sand under high pressure may possibly be considerably 

 less than 2-6%. It is certain that 2-6% represents an 

 upper limit. I made an attempt to find the actual volume 

 of the sand while under pressure from the dimensions of 

 the copper box, but unsuccessfully. 



"When these considerations are applied to geology, 

 there are further complicating elements. In a cavity sur- 

 rounded by solid walls the process of erosion by flaking- 

 off may well be stopped long before the sand has acquired 

 the mean pressure of the surrounding rock; this would 

 allow a larger interstitial space than might be computed 

 from the mean pressure. 



Discussion and Summary. 



Cavities in the materials dealt with in this paper, which 

 may be broadly characterized by the property of brittle- 

 ness, exhibit a method of failure under high compressive 

 stresses not shown by ductile materials like the metals. 

 This method consists in the shooting-off of minute frag- 

 ments with considerable violence from the walls of the 

 cavity. The frequency, and probably the velocity, of 

 projection varies with the pressure, the rapidity of 

 disintegration becoming greater at higher pressures. 

 This mode of disintegration is shown both by rocks 

 and by single crystals ; in rocks the splinters show no 

 relation to the boundaries between chemically homogen- 

 eous parts of the mixture, and in the crystals there is no 

 obvious connection with the crystalline symmetry. The 

 rate of change of speed of disintegration with pressure 

 may vary greatly from substance to substance, being 

 comparatively small for quartz and high for tourmaline 

 and andesite. 



