332 PEOF. W. MOEEIS DAVIS ON [Aug.. I909, 



rock-steps is most significant ; one would have thought that post- 

 Glacial time alone might have sufficed for a deeper entrenchment 

 than usually occurs. The waste-heaps of the Glanrafon slate- 

 quarries east of Llyn Cwellyn, in the face of the second rock-step 

 outside of the western (Clogwyn) cwm of Snowdon, would about fill 

 the little gorge that the stream (Afon Treweunydd) has cut in its 

 cascading course down the face of the step. Indeed, the enormous 

 waste-heaps by the quarries of Llanberis, Ffestiniog, and Nantlle 

 would, I fancy, without counting the great amount of good slate 

 that has been marketed, about equal the volume of rock eroded from 

 all the post- Glacial stream-gorges in the Snowdon district. 



Under the theory of ice-erosion, rock-steps in the valley-floors 

 are necessarily expectable features in a glaciated district where 

 the glaciation did not last long enough to allow the perfect 

 reduction of the ice-channels to a maturely even grade. It is not 

 necessary to make special suppositions regarding long halts of the 

 glacier-ends at particular places. It stands to reason that, when a 

 normally eroded pre-Glacial valley comes to be occupied by an 

 eroding glacier, the work of valley-deepening cannot be accomplished 

 evenly or all at once. The work demands time ; and, during its 

 advance, it is most reasonable that the deepening should be faster in 

 some parts than in others. It is a matter of course that channel- 

 floor hollows and channel-floor steps should be produced along the 

 course of the glacier, if it be regarded as a destructive agent. It is 

 equally expectable that, if time allowed, the steps developed in the 

 more youthful stages of glacial erosion should be worn away in the 

 more mature stages ; and hence that glaciated valleys (glacial 

 channels ) should, like cwms (glacial reservoirs), be described in terms 

 of the stage of advance reached in the long process of modifying a 

 normal pre-Glacial valley into a young, mature, or old glacial 

 channel. It is evident enough that the stage of maturity was not 

 fully reached in the Snowdon district. The valleys — or glacial 

 channels, as we should call them in the present consideration — are, 

 however, more nearly mature in their lower courses than in their 

 upper courses, and this is a not unnatural arrangement. 



It is difficult to determine whether rock-steps in valley- (channel) 

 floors are in any definite way related to differences of rock-structure. 

 So far as the geological maps go, no constant relation of steps to 

 structure can be found in the Snowdon district. In some cases a 

 rock-step lies close to a boundary-line between two formations ; in 

 other cases a step occurs within the area of a single formation : in 

 some instances the step-face is parallel to the strike of the for- 

 mation ; in others it is oblique or transverse. Even where the step 

 lies near the boundary between two formations, it does not clearly 

 appear that the upper level is of more resistant, and the lower level 

 of less resistant, rock ; for when the two formations concerned are 

 followed up to a ridge-crest, the height of the ridge shows no marked 

 change at the boundary between the two. In this connexion, 

 however, two points should be borne in mind. First, that all the 



