360 P. W. Ilr'ah/maii— Failure of Cavities in Crystals 



Andesite was exposed to 1000, 2000, 3000, 4000, 6000, 

 and 7000 kg. for 30 minutes each, then to 7000 attain for 9 

 hours, and then to S000 for 5y 2 hours. At 6000 the 

 flaking-off of the interior was just perceptible. A most 

 interesting effect in connection with this flaking-off was 

 observed. The brass rod supporting the solder disc 

 inside the cavity assumed a curious whitish appearance, 

 suggesting chemical action with gaseous fumes. Exam- 

 ination with a high-power microscope showed, how r ever, 

 that the whitish coating was a covering of minute splin- 

 ters of the rock, which had been projected with such vio- 

 lence when they had flaked off as to penetrate some dis- 

 tance into the solid brass and stick themselves in position. 



Under the first application of 7000 the flaking-off was 

 greater than at 6000, but still only slight. The brass rod 

 was peppered with splinters as before. The second appli- 

 cation of 7000 produced still more flaking-off, but not 

 nearly as much as I had expected from the longer time 

 interval. At 8000 the disintegration of the interior was 

 complete. The eroded region was much larger than in 

 porphyry, and was in shape an irregular pointed ellipse. 

 There was no flow whatever of the outside, and no cracks 

 in the solid mass, even in the immediate neighborhood of 

 the hole. The average density of the sand was 1-14 

 against 2-69 of the original rock. The interstitial space 

 was therefore more than 50%. 



Microscopic examination of the sand showed that the 

 pieces were most irregular in shape and of great range 

 of size. Furthermore, coincidence in one piece of the 

 light and dark material of which the rock is composed is 

 extremely common, showing no tendency for each flake 

 to be all of the same composition. It is evident, there- 

 fore, that the flaking-off is something which has no par- 

 ticular connection with structure, at least when it is 

 on as fine a scale as in this rock. The same remarks apply 

 to porphyry. 



The compressive strength of basalt, which is closely 

 allied to this fine-grained andesite, is given in engineering 

 works as varying from 1000 to 3200 kg/cm-. The com- 

 pressive stress at the interior of this specimen when 

 failure took place was 16,000 kg. 



Granite and Limestone. — These materials formed the 

 subject of the preliminary experiments mentioned in the 

 introduction ; they were much rougher in character than 



