298 TROF. W. MORRIS DAVIS ON [Aug. I909,. 



were drained by a large river, and the cwm by a trickling rivulet ; 

 for a number of examples can be shown around Snowdon where 

 the main stream is less than ten times the volume of the cascading 

 side-stream. The main valley is therefore due to some excavating, 

 agency that deepened the valley-floor very effectually, while the 

 deepening of the cwm-floor by the side-stream was as effectually 

 suspended. The cwm-stream is now doing its best to cut down the 

 floor of the cwm from which it issues ; it may have already cut a 

 little cleft from 5 to 50 feet deep into the rock-face of the main 

 valley-side, and it will eventually bring about an accordant junction 

 of the main valley and that of the cwm. 



Some of the Snowdon cwms have a significant amount of drift on. 

 their floors ; for example, the front step of Cwm Dur-arddu seems- 

 to be rather heavily drift-covered ; and a drift-sheet covers the- 

 front step of the lower floor of Cwm Clogwyn near the slate- 

 quarries ; yet the bare ledges on either side of these steps are so- 

 disposed as to make it very improbable that the cwm-steps are- 

 largely built up of drift. Indeed, a remarkable feature of the 

 Snowdon district is the small quantity of drift that is left upon it. 

 No large moraines were seen in the valleys ; and in the cwms the 

 few moraines noticed seemed to be of less volume than the post- 

 Glacial talus. The best cwm-moraines are those of Cwm Dur- 

 arddu ; some of these, rocky and bare, like fallen stone walls, lie 

 close around the blue lake in the cwm-floor ; others of somewhat 

 greater age, already rounded and grassed over, lap over the southern 

 part of the front step of this fine cwm. The difference in the age 

 of these two sets of moraines, as indicated by the freshness of one 

 and the grass-cover of the other, suggests that they might be 

 referred to different epochs of the Glacial Period ; if so, the only 

 effect of the latest epoch w^ould have been the making of a small 

 glacier in this cwm, for nowhere else in the Snowdon group are 

 moraines of so very fresh an appearance to be seen. 



The discordant or hanging relation usually found in the Snowdon 

 district where a main valley and a side-valley join is only an ex- 

 tension of the discordant relation between a cwm-floor and that of 

 the valley to which it opens. The discordance of level between the- 

 two valleys appears to increase with the difference of their drainage- 

 areas. These hanging valleys are entirely unlike the hanging side- 

 valleys in a normal river-system, which (though rare) are of well- 

 known occurrence. Normal hanging lateral valleys are found only 

 in that very early stage of the cycle of erosion in which the main 

 stream has entrenched itself faster than its side-streams ; and in 

 all such cases the valley of the main stream is only a narrow cleft 

 or gorge, as in fig. 4 (p. 299). When the main river approaches and' 

 reaches grade, so that its own downward erosion is greatly retarded, 

 the side-streams soon overtake it, and thus establish accordant junc- 

 tions which are maintained through all the rest of the cycle, as in fig. 5. 

 Even in a valley so immature as the Colorado Cafion, where the main 

 river is large and of continuous flow, while the side-streams are- 



