514 JOSEPH BARRELL 



between grains becomes proportionately less. For the prevention 

 of ready recrystallization and the maintenance of this pore space 

 the granulated rock, according to present theory, must be con- 

 ceived of as dry and the grains accordingly unsupported except 

 at the points of contact. The shear strains within each grain 

 become very great in proportion to the diminution of contact, and 

 increase in proportion to the regional pressure. If the points of 

 contact, for example, cover only one-fourth of the surface, the 

 compression on those points would be four times as great per unit 

 of surface as if there were continuous contact between grains. On 

 the intervening parts of the surface there would be no pressure 

 Internal shears would result in this way from the hydrostatic 

 pressure of dry rock due to depth and are not dependent upon 

 a pressure-difference in the rock as a whole. The internal strains 

 would tend to produce molecular changes of state as in the plastic 

 flow of metals. There would be melting to relieve the strain, and 

 refreezing by which the molecules would build out the crystals 

 into the pore spaces. By this means recrystallization can go on 

 without the aid of crystallizers, though presumably with more 

 difficulty, and the comminuted crystals come to fit compactly 

 as they are observed to do. This elimination of porosity pre- 

 sumably goes on approximately with the process of granulation, 

 though it may lag somewhat. It would go forward more effectively 

 with depth, irrespective of temperature, since there would be the 

 greater static load upon the rock and the greater differential 

 pressures within the mineral particles. It might be expected that 

 such reduction of pore space would go forward to a limited extent 

 only, leaving a residual porosity. Observation, however, shows 

 that the pore space has been almost completely eliminated. 

 Furthermore, the rocks now exposed at the surface acquired their 

 absence of pore space at depths of only a few miles from the surface. 

 At depths measured in tens of miles there seems then no expectation 

 that density would be notably decreased because of a development 

 of porosity. 



To sum up the modes of yielding within the lithosphere: at 

 the surface is seen to exist a thin outer crust intimately cracked 

 on the outside by closely spaced parallel joint systems. Local 



