462 MR T. J. JEHU ON 



occupy a long trough-like depression, with a somewhat flat floor and steep cliff-like 

 walls, giving in cross-section a U-shaped figure. The trough is deepest towards the 

 upper end, and shallows gradually as traced downwards. This is just what we should 

 expect, for the erosive power of the glacier would be most marked at the upper end, 

 and its amount would tend to diminish downwards. The aspect presented by the lower 

 reaches of the Pass suggest the possibility that the valley has here been over-deepened 

 considerably as compared with what it was in pre-glacial times, and that the trough in 

 which the lakes lie was dug out of the floor of this already over-deepened valley. 



Llyn Cwellyn. — Here, again, we have a similar U-shaped, trough-like depression, 

 which is deep along its whole extent. The valley, which is wide towards the upper end, 

 narrows below the foot at Nant-Mill, and this would form an obstacle to the flow of 

 the glacier, and so favour the erosion of a hollow just above. A barrier of rock 

 extends right across the valley to Nant-Mill, over which the stream from the lake 

 runs in a waterfall. Mr Marr, assuming the lake to be only 50 feet deep, 

 thought he could find traces of a buried gorge at the upper end of the lake, along 

 which the waters of Llyn Cwellyn might formerly have been carried away in the 

 opposite direction. No proof has been offered that such a gorge occurs here ; and 

 Mr Dakyns, in his paper on " Some Snowdon Tarns," has shown that its existence 

 is extremely improbable. And the further fact that the lake proves to be so much 

 deeper than had been supposed (122 feet, as against 50 feet), militates against that 

 hypothesis. The depth of the lake and the general form of the hollow are more iu 

 consonance with the view that we have here a rock-basin excavated in the floor of the 

 valley by glacial action. 



Llyn Gwynant and Llyn Dinas. — We have already seen that the Vale of G-wynant 

 has in all probability been broadened and deepened along a great part of its course. But 

 at two places the valley becomes constricted, owing to the mutual approach of its flanks. 

 Above the upper constriction lies Llyn Gwynant, and above the lower lies Llyn Dinas. 

 So that these lakes occupy just those positions in the valley where the glacier would be 

 retarded in its flow by these obstacles, and where, therefore, the conditions for erosion 

 were most favourable. The hollows in which they lie are rapidly becoming silted up, 

 and so the lakes are more shallow. In Gwynant, the streams coming in on the left side 

 have caused the lake to shallow more rapidly there than on the right side, as shown in 

 the cross-section E-F (Plate VIII. ). (The section C-D is probably more typical of the 

 original form of the hollow.) Dinas is still shallower, and shows a more advanced stage 

 in the process of silting up. The bottom of the valley between Gwynant and Dinas is 

 largely covered by drift. The exit from Dinas is almost certainly over rock ; but in order 

 to avoid the hypothesis of a rock-basin, Mr Marr pointed to the possibility of a drift- 

 filled depression, leaving the lake on the right side a little above the exit, and passing 

 round a rocky knob into the road. Though one cannot exclude the possibility of these 

 lakes being drift-dammed, the more probable and natural hypothesis is that they lie in 

 rock-basins which have to a great extent been filled up by the deposition of sediment. 



