340 PROF. W. MORRIS DAVIS ON [Aug. ^OQ* 



Valley. The valley of Afon Gwarfai receives other cascading 

 streams from two short hanging cwm-valleys on the south-west, 

 roughly shown in fig. 3 (p. 296) ; one is alongside of Llyn Cwellyn, 

 and the other is a mile farther down stream. Both of these are 

 fine examples : the second is the sharply defined Cwm Du in the 

 north side of Mynydd Mawr, already described. 



The Gwynant Valley has a number of small torrents on its sides, 

 cascading from small hanging valleys on the higher slopes; one 

 of the best of these is on the north-west side of the valley, about a 

 mile and a half above Beddgelert, and another enters the southern 

 end of Llyn Dinas from the south. The exceptional recession of 

 the hanging valley which enters the Gwynant Valley from the 

 southern cwm of Snowdon has already been mentioned. 



The north-western cwm-valley of Snowdon hangs 200 or 300 feet 

 over the Llanberis Valley ; the fine waterfall that is here cutting a 

 ravine in the lip of the hanging-valley mouth is a picturesque feature 

 in the ascent of Snowdon by the rack-railway, just after leaving 

 the town of Llanberis. Several small high-level hanging valleys 

 supply cascading streams on the north-east side of the Llanberis 

 Valley. 



It is difficult to imagine how these strikingly abnormal features 

 can be accounted for under the theory of glacial protection. In 

 every case, the main valley is not a narrow, vertical-sided chasm, 

 over which a hanging lateral valley is a normal feature, but a wide 

 open thoroughfare, with respect to which a hanging lateral valley 

 is absolutely abnormal. To suppose that the glaciers of the 

 hanging valleys held their ends in fixed positions during all the 

 time required for the deepening and widening of the main valleys 

 is unwarranted by anything that is known of existing glaciers. 

 To suppose that, if the glaciers really did hold their ends in fixed 

 positions so long, they could have selected stopping-places so 

 significantly related to irrelevant features — the main valleys to- 

 which the hanging valleys are tributary — is unreasonable, to say 

 the least. To explain these hanging lateral valleys by any 

 processes of normal erosion, aided by tilting, uplifts or anything 

 else, but independent of ice-protection or of ice-erosion, would 

 require the abandonment of the well-grounded principles of valley- 

 development as determined in non-glaciated regions. On the other 

 hand, the explanation of the hanging lateral valleys under the 

 theory of ice-erosion requires no special or peculiar suppositions. 

 The explanation is a corollary of the proposition that glaciers erode 

 their beds. It necessarily follows, either from the more rapid 

 channel-deepening by a main glacier than by a side-glacier in the 

 early stage of a cycle of glacial erosion ; or from the inability of a 

 smaller side-glacier (or a side water-stream) to erode as deep 

 a channel as that of the larger main glacier, however long the 

 glacial cycle continues. 



The depth of glacial erosion in a main valley is roughly indicated 

 by the discordant altitude of the hanging lateral valleys, as well as 



