248 PKOF. S. J. SHAND ON [vol. lxxv, 



eventually reached the limestone on the north side of No. 1 fault, 

 while on the south side of this fault it was still in gypsum. Erosion 

 then proceeded more rapidly in the latter part of the stream-course 

 than in the former, with the result that the limestone escarpment 

 began to be laid bare and a waterfall formed where the stream 

 crossed it. But this growing difference in level brought about an 

 increased rate of flow and greater corrosive power above the fall, 

 while at the same time the stream discovered the minor faults in 

 the limestone, and, adapting its course partly to them, cut out in 

 time the present deep zigzag cafion. In the same way, when the 

 gypsum within the valley had been excavated sufficient!)', the 

 limestone was again encountered at No. 5 fault; the stream being, 

 however, now confined in a deep channel in the gypsum was not 

 deflected by the obstacle, but was able to saw its way down 

 through it, aided doubtless by the joints in the limestone. 



In brief, the river, although younger than the initiation of the 

 faulting, is older than the sculpturing of the inside of the fault- 

 trough, which was largely accomplished by itself and by subsequent 

 tributaries which developed along the main fault-lines of the rift. 

 As regards the limestone arch, the stream is clearly a superimposed 

 one. Nearly all the drainage of the valley falls into this stream, 

 but the eastern end is tapped by another and smaller stream with a 

 precisely similar development. 



Other interesting features appear where the canon opens out in 

 the escarpment of the northern fault. The water falls into a 

 beautiful saucer-shaped pool, which is bounded for two-thirds of its 

 circumference by overhanging Avails of limestone. A band of a soft 

 shaly facies of the limestone crops out at the level of the pool, and 

 it has been cut back by the water into a deep recess or gallery as 

 much as 20 feet wide, which encircles the pool on one side. Owing* 

 to this process of undercutting, the waterfall is receding in a 

 well-known manner. The water which escapes over the southern 

 lip of the saucer disappears down the fault-plane, and for 

 some distance below the stream-bed is dry. The same curious 

 feature is shown by a smaller wet- weather Avatercourse that cuts 

 No. 4 fault half-way up the valley. 



The excellent preservation of the fluted surfaces on the fault- 

 scarps suggests that the last downward movement took place at a 

 very recent date; and, as the Avhole of this country is much affected 

 by earthquakes, such a conclusion would not seem unreasonable. 

 The eA'idence of the canon is against it, hoAvever : for the Avaterfall 

 at its mouth does not now lie on the fault, but has receded some 

 60 feet from it, Avhile at the point where the stream-bed crosses the 

 fault-plane there is no fall at all. Recent movement along the 

 fault-plane Avould have produced a fall here, and the time necessary 

 for the obliteration of the fall would, I think, have accomplished 

 the destruction of the engraved face of the escarpment also. For 

 this reason I attribute the preseiwation of the latter feature simply 

 to the rapid remoA-al b} r solution of its covering of gypsum. 



