and Rocks tinder Pressure. 249 



become sufficiently large, the sand was in a position to 

 exert pressure on the interior walls, and in this way pre- 

 vented further disintegrating action. The eroded cavity 

 was very irregular in outline ; for the greater part of its 

 depth it was entirely unsymmetrical, being merely an 

 angular extension on one side of the original circular cav- 

 ity. This is as one would expect; an angle once formed 

 at any part of the surface would be relatively unstable, 

 and erosion would proceed more rapidly here. There was 

 no apparent relation betwen the axis of the eroded cavity 

 and the three systems of cleavage planes mentioned. 



The outside of the cylinder was entirely unaffected by 

 the erosion of the interior; there was no perceptible 

 change of dimensions, either diameter or length, and the 

 cracks were entirely confined to the interior. This is true 

 of many of the other specimens tried. A photograph of 

 this cylinder is shown in fig. 4. 



The second specimen of quartz was cut in two sym- 

 metrical pieces as in fig. 2. The two parts were always so 

 put together as to observe the original orientation in the 

 crystal. It was subjected to pressures of 4000, 6000, 

 8000, 10,000, and 12,000 for 10 minutes each, experience 

 with the first crystal having shown that applications of 

 pressure at 1000 kg. intervals was not necessary. The 

 results were very similar to those with the first specimen. 

 The same system of funnel-shaped cracks separated by 

 120° was observed, but they were not so prominent, indi- 

 cating that the fitting together was better, and the end 

 effects, therefore, not so important. The same flaking- 

 off of the inner surface was observed, first noticeable at 

 8000, and resulting in complete disintegration of the inte- 

 rior at 12,000. A photograph of one of these cylinders is 

 shown in fig. 5. There was no permanent change of out- 

 side dimensions, no evidence whatever for flow, and the 

 cracks did not reach to the outside. The density of the 

 sand which packed the cavity was 1-43 against 2-65 of 

 the original quartz; the interstital space was therefore 

 nearly 50%. 



The stress required to produce rupture of these crystals 

 is much higher than might be expected, and the manner of 

 rupture is apparently not like that contemplated in any 

 theory of rupture, nor is it such as would be suggested by 

 the elastic deformation before rupture. I have worked out 

 mathematically the nature of the stress-strain relation, 



