Vol. 65.] GLACIAL EROSION IN NORTH WALES. 325 



surmounting part of the peak was worn down by superglacial 

 weathering to the blunted form of to-day, as indicated beneath the 

 peak. Oo the other hand, if glaciers are destructive, pre-Glacial 

 Snowdon must have been a large dome, which may be roughly 

 reconstructed by extending the arch of the present rounded top over 

 the undercut cwms, leaving normal valley-heads along the cwm- 

 axes, as in fig. 25. Glaciers then excavated their great reservoirs 

 in the valley -heads and sapped the surmounting slopes, leaving only 

 a remnant of the original dome-like form. The first of these 

 suppositions does not seem to be so consistent as the second with 

 the pre-Glacial history of the region, or with so much of the pre- 

 Glacial form of the Welsh mountains as has been preserved, except 

 for normal weathering during the Glacial Period. 



The cwms in the neighbouring mountain do not afford relief for 

 the difficulty that is found with the cwms around Snowdon. Cwm 

 Da, for example, a very sharply limited cwm in the northern slope 

 of the dome-like mass of Hynydd Mawr, roughly drawn in fig. 3 

 (p. 296), is distinctly unlike the normal valley-heads of the sub- 

 mature Cevennes or of the late mature mountains of North 

 Carolina. The cwm-cliffs are too steep to be referred to the 

 Tertiary cycle of erosion, which lasted so long that all cliffs must 

 have been destroyed when uplift introduced the post-Tertiary cycle ;;' 

 and the revived erosion of the post-Tertiary cycle cannot be held 

 responsible for the cliffs of the cwm, for revived erosion in non- 

 glaciated monadnocks does not produce such features. Indeed, 

 Cwm Du can be regarded as the result of pre-Glacial erosion only 

 if pre-Glacial erosion is regarded as an irresponsible and arbitrary 

 process, capable of producing, in mountains that were afterwards to 

 be glaciated, curious forms, the like of which is not to be found in 

 non-glaciated mountains. It is interesting to note, in passing, that 

 a pillow-like mass of drift seems to lap on the border of Cwm Du 

 from the west ; and, as this drift-mass seems to be continuous with 

 the sheet of drift that envelops the neighbouring Moel Trefan, I 

 have taken it to be the work of the great northern ice-sheet, and 

 not of the local glacier. A small moraine within the cwm may be 

 referred to a late stage of the local glacier. 



Under the theory that glaciers are destructive agents, the valley- 

 head cwms of the Snowdon district are regarded as normal pre- 

 Glacial valley-heads that have been significantly enlarged and 

 deepened by glacial erosion, with pronounced steepening of the side 

 and head walls, partly by direct glacial action, partly by super- 

 glacial weathering. So much of the dome-like mountain as is still 

 seen above or alongside of the cwms is regarded as preserving 

 in a general way the quality of subdued form that the mountain 

 possessed in pre-Glacial time, but as having been lowered by an 

 unknown amount by general weathering during Glacial time. The 

 difficulty, in this view of the case, is that no one yet knows just 

 how glaciers can erode retrogresbiTely, so as to enlarge as well as 

 to deepen their neve-reservoirs. It is supposed that they scour and 



Q. J. G. S. No. 259. z 



