SIXTH REPORT OF THE DIRECTOR I909 I93 



therefore coincide very closely with the period of most abundant 

 flow of chemically active water. The deeper and narrower por- 

 tions of crevices kept full by capillary attraction would lose no 

 appreciable amount of water by evaporation on account of the 

 small and protected exposed surface. Such waters while they 

 might approach saturation would rarely reach it because each fresh 

 filling of the space above the water surface maintained by capil- 

 larity would add mechanical pressure and force the older water 

 out at lower levels, thus refilling the crevice with fresh water. As 

 the expansion of the beds during the summer heating forced the 

 crevice walls more closely together, a portion of the water held by 

 capillary attraction would be forced to occupy new positions and 

 would take its dissolved load with it. Evaporation would rarely 

 be rapid enough to allow of deposit and the occasional advent of 

 summer rains would furnish abundant fresh water and help to 

 remove that more nearly approaching saturation. Fall cooling and 

 the more abundant rain of that season in this locality would be 

 very effective agents in removing material dissolved during the 

 summer. 



Exposed joint crevices, so situated as to receive surface flow, are 

 thus gradually widened and deepened. Other joint crevices con- 

 necting with these might not receive so much water directly from 

 the surface but they would help to carry the subterranean flow 

 and become also widened. As the system developed the waters 

 would move in certain channels with greater freedom. Dust and 

 disintegrated particles accumulating on the exposed rock surface 

 would be swept by rainfall into these openings, and the effects of 

 abrasion would be added to those of solution. Joint crevices thus 

 become joint channels. As these are widened they admit the cold 

 air of winter in greater volume and the circulation maintained at 

 this season serves to more effectively chill the rock mass and thus 

 open the crevices to greater depths. As these channels approach 

 local grade (determined by the level of a surrounding body of water 

 or by the water table of the mantle rock flanking these regions) 

 the deepening of the crevices ceases, but lateral cutting is continued. 

 Such basal widening of the system is an indication of maturity. 



Such work must be exceedingly slow, for the water supply is 

 only such as falls on the very limited area in question. The waters 

 doing the work would be free from humus acids on account of the 

 bare character of the rock surface. Expansion of freezing water 

 would be effective near the surface, but the depth at which it could 

 be considered as an important factor would be very limited. 



