and Rocks under Pressure. 265 



the sand was reduced after compression were entirely 

 irregular and of all sizes from a few hundredths to frac- 

 tions of a thousandth of a millimeter. It is therefore 

 obvious that any calculation of the density to be expected 

 from the average density of closely piled spheres is inap- 

 propriate. 



A single experiment on pure feldspar gave a density 

 of 247, on pure talc 2-616, and on a mixture of equal parts 

 by weight of quartz and feldspar sand, 2-520. The initial 

 densities of feldspar and talc were 2-57 and 2-76. The 

 average interstitial spaces were, therefore, 2-6, 3-9, and 

 5-27« in quartz, feldspar, and talc respectively. The high 

 value for talc was a surprise, but it receives possible 

 explanation from the remarks of the next paragraph. 



The powder after compression has been referred to as 

 in the form of coherent cakes ; this requires considerable 

 qualification. It was never possible to obtain a single 

 coherent button of the dimensions of the containing box, 

 but this button always broke up spontaneously into 

 laminae, slightly cupped, the faces of the laminae at right 

 angles to the direction of pressure. These laminae varied 

 from small scales a few tenths of a millimeter thick to 

 comparatively large plates 2 mm. thick and nearly 1-5 cm. 

 in diameter. Density determinations were made on the 

 largest coherent pieces. Spontaneous break-up of the 

 buttons into laminae was always a comparatively slow 

 affair, and might occupy 10 or 15 minutes, the button 

 slowly puffing up with a crawling simulation of life. The 

 force involved in this break-up was considerable, and was 

 frequently sufficient to pull apart the copper box. 



This phenomenon is not what one would at first expect, 

 elastic recovery from stress being immediate, and it 

 seems to me of considerable importance for an inter- 

 pretation of the other results. A plausible explanation 

 seems to me as follows : Under the intense stress, commi- 

 nution into minute fragments proceeds so far that there 

 are few actual voids left, but many spaces that would be 

 void under no stress are closed by the elastic deformation 

 of the walls. The majority of such spaces are probably 

 lens-shaped, like the erosion cavities in the crystals above, 

 and these are squeezed flat by the pressure. But on the 

 surface of each grain there is a film of adsorbed air, which 

 is squeezed extremely thin by the pressure, but neverthe- 

 less keeps the walls of the cavities from actual molecular 



