268 P. W. Bridgmun — Failure of Cavities i H (Jryslalti. 



Consistently with this view one would expect the char- 

 acter of the surface to have an important effect. If the 

 surface is perfectly smooth, as in the negative quartz 

 crystals, or if as in glass there is a protecting covering of 

 water 200 molecules deep like a jelly in which there may 

 be surface tension effects, one expects very slight disin- 

 tegrating tendency. Any artificial cavities, on the other 

 hand, must contain large irregularities and be favorable 

 to the effect. ■ 



This paper mentions the results of a new mathematical 

 analysis of the effect in crystals of hydrostatic pres- 

 sure applied as in these experiments. It appears that the 

 new phenomena introduced by crystalline structure are 

 not prominent enough to lead one to expect rupture 

 because of them, and that in most cases an approximate 

 solution may be obtained by treating the crystal as iso- 

 tropic with mean values of the elastic constants. 



The stresses which these brittle materials stand are 

 many times higher than would be predicted by ordinary 

 compression tests. If one neglects the flaking-off effect, 

 which is entirely uncontemplated in mathematical theory, 

 stresses at least 20 times higher than those of ordinary 

 compression tests may be reached without rupture. At 

 the same time the possible stresses are very appreciably 

 lower than those found by Adams. His results were 

 affected by the unknown action of shrunk-on steel jackets. 



Attempts to weld together finely powdered quartz, 

 feldspar, and talc failed up to 30,000 kg/cm 2 . There is, 

 however, no evidence that such welding would not take 

 place if the adherent film of air could be entirely 

 removed; this is a matter of extreme experimental dif- 

 ficulty. The amount of interstitial space in compressed 

 powders has been measured, but caution must be used in 

 inferring from these figures the density of a compressed 

 sand while actually under pressure. The results of 

 these collapsing tests makes it extremely probable, how 

 ever, that minute crevices, at least large enough for the 

 percolation of liquids, exist in the stronger rocks at 

 depths corresponding to 6000 or 7000 kg/cm 2 ., and pos- 

 sibly more. 



The Jefferson Physical Laboratory, 

 Harvard University, Cambridge, Mass. 



